Boston Studies in The Philosophy of Science

BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE
Editor
ROBERTS. COHEN, Boston University
EditorialAdvisoryBoard
mOMAS F. GLICK, Boston University

ADOLFGRUNBAUM, University ofPittsburgh
SYLVAN S. SCHWEBER, Brandeis University
JOHN J. STACHEL, Boston University
MARXW. WARTOFSKY, Baruch Collegeof
the City University ofNew York
VOLUME 169
REALISM AND ANTI-REALISM IN THE
PHILOSOPHY OF SCIENCE
Tscha Hung with C. G. Hempel. Vienna 1982.
REALISM AND ANTI-REALISM
IN THE PHILOSOPHY OF
SCIENCE
Beijing International Conference, 1992
Edited by
ROBERT S. COHEN
Boston University
RISTO HILPINEN
University of Turku and University of Miami
and
QIU RENZONG
Institute of Philosophy, Chinese Academy of Social Sciences
....
"
Springer-Science+Business Media, B.Y.
Library of Congress Cataloging-in-Publication Data
Realism and anti-realism in the philosophy of seienee Be t j Inq

International Conferenee. 1992 I edited by Robert S. Cohen . Risto
Hilpinen. Oiu Renzong .
p. em. -- (Boston studies in the philosophy of seienee ; v.
169)
"Be i j i ng International Conferenee on Philosophy of Seienee"--Pref .
p , 10.
Ine 1udes index .
1. Realism--Congresses. 2 . Seienee--Philosophy--Congresses.
1. Cuhan , R. S. (Robar t Sonne) I!. H t lp men , Risto . II1. Ch' tu ,
Jen-tsung. IV. Beijing International Conferenee on Philosophy of
Seienee (1992) V. Series.
0175. 32. R42R43 1995
501 - -de20 94-39752
ISBN 978-90-481-4493-8 ISBN 978-94-015-8638-2 (eBook)

DOI 10.1007/978-94-015-8638-2
Printed on acid-free paper
All Rights Reserved

© 1996 Springer Science+Business Media Dordrecht
Origin ally published by Kluwer Academic Publishers in 1996.
Softcover reprint of the hardcover 1st edition 1996
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
TABLE OF CONTENTS
PREFACE ~
ROBERT S. COHEN / Recollections of Tscha Hung xiii

FAN DAINIAN / Hong Qian (Tscha Hung) and the Vienna Circle xvii
PHILOSOPHICAL ARGUMENTS
RISTO HILPINEN / On Some Formulations of Realism, or How

Many Objects Are There in the World?
RUDOLPH KREJCI/Dissolution of the Realism/Antirealism
Problem 11
ALAN MUSGRAVE / Realism, Truth and Objectivity 19
ILKKA NIINILUOTO / Queries about Internal Realism 45
QIU RENZONG / How to Know What Rises Up Is the Moon?
- On the Concept of Realism and the Irrelevancy of Quantum
Mechanics to the Debate on Realism vs. Antirealism 55
DEMETRA SFENDONI-MENTZOU / The Reality of Thirdness
- A Potential-Pragmatic Account of Laws of Nature 75
ARCHIE J. BAHM / Tentative Realism 97
F. WALLNER AND M. F. PESCHLI Cognitive Science - An
Experiment in Constructive Realism; Constructive Realism
- An Experiment in Cognitive Science 103
REALISM AND GENERAL METHODOLOGY
JAMES ROBERT BROWN / Phenomena 117

ALLAN FRANKLIN / There Are No Antirealists in the Laboratory 131
KOSTAS GAVROGLU / Can Theories of Chemistry Provide an
Argument Against Realism? 149
MATTHIAS KAISER / Empirical Versus Theoretical Progress
in Science 171
VB
viii TABLE OF CONTENTS
CHIN-TAl KIM / Science as Ideology 195

MARX W. WARTOFSKY / Three Stages of Constitution:
Historical Changes in the Ontological Status of the Scientific
Object 207
JOHN WATKINS / Scientific Realism versus Common-Sense
Realism? 219
JOHN WRIGHT / Metaphysical Realism and the Explanation
of the Success of Science 227
WU QIONGBING / Is Natural Science Free from Morality? 245
YIN ZHENGKUN / Truth and Fiction in Scientific Theory 265
PHYSICAL SCIENCES
FAN DAINIAN / Niels Bohr and Realism 279

HENRY 1. FOLSE / The Bohr-Einstein Debate and the
Philosophers' Debate over Realism versus Anti-Realism 289
HE ZUOXIU / On the Einstein, Podolsky and Rosen Paradox and
the Relevant Philosophical Problems 299
HONG DINGGUO / On the Neutral Status of QM in the Dispute
of Realism vs. Anti-Realism 307
lIN WULUN / Double Structure of Tao Reality 317
LARS -GORAN JOHANSSON / Realism and Wave-Particle
~~~ 3~
LARS -GORAN JOHANSSON / Van Fraassen's Constructive
Empiricism - A Critique 339
LI LI /On Complementarity Reality 343
LUO lIACHANG AND HU XINHE / Relational Realism on
Reform of the View of Physical Reality and Its Logical
Manifestation 359
MAO SHIYING / A Realistic Interpretation of Quantum
Mechanics 381
MORTON L. SCHAGRIN / Rumford 's Experiments on the
Materiality of Light - More Heat Than Light 395
RICHARD H. SCHLAGEL / Resolving the Realist-Antirealist
Dilemma 405
ZHOU 11 HONG / The Mathematicized Practical Attitude 419
TABLE OF CONTENTS ix
SOCIAL AND BIOLOGICAL SCIENCES

USKALI MAKI / Scientific Realism and Some Peculiarities of
Economics 427
TIMOTHY SHANAHAN / Realism and Antirealism in
Evolutionary Biology 449
PAUL C. L. TANG AND RALPH W. BROWN III/Anti-Realism
and the Complementarity Model of Mind-Brain 467
Name Index 481

PREFACE
This book contains the main papers from the first Beijing International
Conference on Philosophy of Science, held in 1992 and dedicated to
issues of 'Realism and Anti-Realism in Science' . The Conference was
organized by the Chinese Society for Dialectics of Nature (Philosophy
of Nature, Science and Technology), and sponsored by the China
Association for Science and Technology (CAST), the China International
Conference Center for Science and Technology (ICCST), the Boston
University Center for Philosophy and History of Science, the International
Union of History and Philosophy of Science through its Divi sion of
Logic, Methodology and Philosophy of Science (IUHPSIDLMPS), and
by Mr. Joseph C. T. Lee of Hong Kong. The Organizing Committee
was chaired by Qiu Liang-hui, the Programme Committee by Robert
S. Cohen, but we are especially mindful of the responsible role of our
real 'Secretary-General' Qiu Renzong and his endlessly helpful colleague
Ju Zhang .
The Conference was blessed by the late Professor Tscha Hung who
had hoped to join with us, with Chinese and non-Ch inese alike. He had
for many years been Director of the Institute for Foreign Philosophy at
Beijing University, and indeed for decades of thinking and teaching Tscha
Hung was a link between the ' scientific philosophy' of his logical empiri-
cist Vienna and the new time of China after the Second World War.
Both the teacher and also the critic of his own mentors and colleagues,
he was a scholar who quietly carried the roles of being the Ayer and
Hempel of China.
Where might philosophy of science within China be located today?
As in the West, modern Chinese philosophers deal with issues in addition
to those of science, technology, medicine and the natural world , and
they do so within various frameworks, some from traditional Chinese
civilization, some from the currents of classical, modern and post-modern
Western sources, others within a Chinese Marxist outlook. Philosophy
of science however is especially needful of, and open to, international
collaboration and mutual learning . Like the sciences, like mathematics,
physics, biology, chemistry, computer science, modern logic, nuclear
Xl
R. S. Cohen . R. Hilpin en and Qiu Renzon g (eds.) , Reali sm and Anti-Reali sm in the
Philosophy of Science, xi-xii.
xii PREFACE
engineering, like Western medicine and surgery, so philosophy of the

sciences should be a field of international cooperation. This Conference
was a contribution to the extension of such cooperation in philoso-
phical clarification of the currently living concerns about science, among
scientists as well as philosophers.
The renaissance of this endeavor in the People's Republic was evident
at this conference, among younger and elder scholars from every part
of China. The restrictions imposed by dogmatic habits or doctrines, so
hurtful to the honest progress of science and certainly of its philoso-
phical understanding, are hopefully behind our times (however diffi-
cult it may be to be freed from habit and its rituals) . Even that contentious
phrase from old Engels, adapted from Hegel, the ' dialectics of Nature ',
may be seen to be a suggestive problematic for scientific theorists, or
perhaps as heuristic for philosophy of nature and of science-in -society.
And within science, 'materialism' may be realist, objectivist, but not
thereby opposed to theoretically conceived entities or forces or whatever
the scientific imagination may conjure forth . So, 'realism' , whether
another 'neo-materialism ' or not, is an epistemological issue, an onto-
logical puzzle, and even a target for the pragmatic 'realist' of a different
sort. Is there a new 'i dealism' within the social constructivist interpre-
tations of today, even within the cultural anti-science which has arisen ?
Our Conference dealt with such matters, perhaps without sufficiently
vigorous exposition of the case for anti-realism but the reader will
judge.
I am reminded of a contrast drawn by the British philosopher Winston
Barnes many years ago. Barnes said that the idealist holds that "There
is a Mind with a world in it", while the materialist holds that "there
is a world with minds in it". Now we have many-world theories, and
non-reductive emergent realisms, and non-ontological non-idealist anti-
realisms. The issues were before this gathering of speakers and partici-
pants, and revised for this book. What would our friend Tscha Hung have
thought?
Robert S. Cohen
(for the editors)
ROBERT S. COHEN
RECOLLECTIONS OF TSCHA HUNG
I first learned of Tscha Hung in 1955 during conversation with Marie

Neurath . Ten years earlier, on the last day of Otto Neurath's life, Tscha
Hung had come to Otto at his home in Oxford, to talk once more about
protocol sentences and the Schlick-Carnap-Neurath debates. Who was
this Chinese philosopher, I wondered, and why was he in Oxford, con-
cerned about clarifying logical empiricism, in 1945? But Marie had no
notes of Otto's conversations with Tscha Hung, only a pleasant memory
of the visitor's sweet and inquisitive intelligence, his modesty, and his
ease with both German and English.
During the first of my visits to China, in 1985, I was introduced to
Tscha Hung by Fan Dainian ; we had in fact exchanged letters before then,
mainly about Vienna Circle studies. We talked at some length about
developments in the philosophy of science since the Second World War,
and to my delight his interests were very far-reaching. We met again
in China in 1988, and also during the Schlick-Neurath centenary sym-
posium in Vienna. Our talks were not systematic and the topics tumbled
over each other. He knew of my attempt to bring Neurath into greater
appreciations, both as critical philosopher and as practical social thinker.
This led us not only to Neurath's confrontation with Marxist theory
and Communist practice but to Tscha Hung's too. Then we discussed
Neurath's optimistic feeling, in his last years, about the future influ-
ence of a rational and humane social liberalism in the British mode.
Quickly we turned to issues of science, and the roles of scientists in
society. There was no doubt of Tscha Hung's concern about misuse of
science and technology, but equally no doubt for him that scientific
reasoning is the guide for human life. This was, he agreed, the realistic
policy of political and economic forces in the modernization of feudal
China, first in the Three People's Principles of the democratic revolu-
tion led by Sun Yat-sen before the First World War, and then in the
developing Marxist socialisms after that war. Democracy and science
were joined in social/perspective, often as a slogan along with other
slogans for popular support.
To Tscha Hung, there was an evident task for the philosopher at this
XIII
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds .), Realism and Anti -Realism in the
Philosophy of Science, xiii-xv .
© 1996 Kluwer Academic Publishers.
xiv ROBERT S. COHEN
point , which was to offer clarity about science, and about what is not
scientific, and to investigate central concepts of the sciences. For himself
this task took two forms: his research, as in his dissertation on the concept
of cause in contemporary physics, and in his careful attention to prob-
ability; and his teaching, together with his writing of textbooks and
editing of translations from Western philosophers. He seemed pleased
to be compared in his Chinese university environment with AJ. Ayer
in Ayer 's own 'Return from Vienna' to write Language, Truth and
Logic.
Certainly his situation had been difficult, and his Stoic equanimity
striking when I saw him. We talked sometimes about Needham's
researches in the history of the sciences in ancient and classical China,
for which Tscha Hung had great respect, and this led to his questions
about Vienna Circle investigations of the social conditions of knowl-
edge in general, and natural sciences and mathematics in particular.
He had known Edgar Zilsel in Vienna days, and he was much intrigued
by my attempt to sketch Zilsel's works of the 1940s on the historical
sociology of science in the period of the scientific revolution and its
precursors. All the more, he was eager to read Needham 's use of Zilsel 's
historical theories for the Chinese cases.
Tscha Hung was much taken with all the work on external condi-
tions of science, on what Carnap had called the pragmatics, on the several
accounts of the 'historical tum' by Kuhn and others, by the analytic/syn-
thetic debate around Quine, and by both the older and newer apparent
relativisms typified by the image of 'Neurath's boat'. After his own
adherence to Schlick's teaching in early studies, Hung continued to worry
over the coherence/correspondence/formal conceptions of truth, and of
knowledge generally. The issue of realism vs. anti-realism, in our inter-
national conference, was I suppose the vital issue for him. To his pleasure,
this continued to be vital for others as well. With this went his happy
recognition that a new generation of philosophers of science were inves-
tigating and interpreting the history of the Vienna Circle. He was pleased
to accept membership on the Editorial Committee of the Vienna Circle
Collection.
What then, I asked, was his own relation to Chinese thought, and to
Chinese colleagues, whether scientists or philosophers? He had been a
young student of Liang Qi-chao but left China at age 18 for university
studies in Germany. His own philosophical formation was deeply within
the anti-metaphysical logical empiricism. So, in China, he would be
RECOLLECTIONS OF TSCHA HUNG xv
apt to try 'rational reconstruction' of traditional, or 'received', Chinese

science and philosophy, to establish criteria for meaning, to identify what
is testable , verifiable, probable, certain. But, like Schlick, he saw a role
for the metaphysical, the non-scientific, in a different linguistic type .
He liked my vague suggestion that perhaps metaphysical phrases and
stories, although not strictly cognitive , may convey 'how the world feels' .
He had spoken of metaphysics, at its best, as a ' conceptual poem'.
How, I wondered, did he see the other philosopher of his own time
in China, one who also had completed a western doctorate in philos-
ophy (Columbia University) well before the Second World War, and
one whom I had an opportunity to meet in China. This was, of course,
the innovative neo-Confucian Fung Yu-lan, also a surviver through the
stages of 20th century China, also at Beijing University. In the early years,
just before and after the founding of the People's Republic, they had
debated, and engaged in written controversy. Fung brought his own
new rational, even formal, reconstruction of the immense cumulative tra-
ditional thought, and cultural weight, of China. There was, for Tscha
Hung , little understanding of the world-significant achievement of
modern science in the ethical focus of Fung's work , and yet he urged
me to respect Fung's historical and critical expositions of Chinese class-
ical philosophers.
Moreover, when we talked about the dreams of Marxist humanists
for a society within which people might be happy, and about Neurath's
praise for Epicurus, Tscha Hung admitted, and lamented, the rel ative
weakness of the Vienna Circle , indeed of the logical empiricists at large,
in dealing with problems of ethics, of practical life. When I remonstrated,
speaking of ' practical' writ ings of his beloved Schlick, of his admired
Reichenbach, of his friend Feigl, of his hero Russell, he countered with
the great example of Kant including moral issues within theoretical phi-
losophy. Even Fung's neo-Confucian doctrine seemed to Hung in danger
of losing the intuitive feeling of practical life within a new formalism .
I wondered where Tscha Hung might have proceeded in emulating Kant's
problematic , but he shook his head, saying he had lost those three decades
in isolation, and it was too late now. He surely was no longer isolated
in his last years, neither from Western friend s nor from Chinese col-
leagues . He was, throughout his time, a teacher of integrity and a witness
to a stage of the history of philosophy.
FAN DAINIAN
HONG QIAN AND THE VIENNA CIRCLE
Hong Qian (Tscha Hung , 1909-1992), a member of the Program

Committee of this Conference, should have been here with us if the angel
of death had not taken life away from him three and half months ago.
Hong Qian was born on October 12, 1909 in Fujian province but
his native town was She county, Anhui province . During his youth, he
was a student of the great Chinese scholar Liang Qi-chao. In 1927,
encouraged by Liang, Hong went to Jena, Germany in order to study
spiritual philosophy under the supervision of Rudolf Eucken. Since
Eucken died before Hong 's arrival , he began to study physics, mathe-
matics and philosophy at the University of Jena. Deeply attracted by Hans
Reichenbach's Von Kopernikus his Einstein (1927) and Relativitiitstheorie
und Erkenntnis apr iori (1920), Hong moved to Berlin to listen to
Reichenbach's lectures. Knowing that Hong was interested in philosophy ,
Reichenbach suggested his going to Vienna to be the student of Moritz
Schlick.
Hong Qian studied at the University of Vienna after 1928. At Schlick's
suggestion, Hong attended Rudolf Carnap's course in mathematical
logic, Hans Hahn's course in mathematics, Felix Ehrenhaft's course in
physics, and Schlick's, Friedrich Waismann's and Victor Kraft 's courses
in philosophy. From 1930, at the invitation of Schlick, Hong Qian
attended the meeting s of the Schlick group held every Thursday evening,
and became the unique member of the Vienna Circle from the Far East.
He was acquainted with Otto Neurath, Philipp Frank, Herbert Feigl ,
Kurt Godel , Theodor Radakovic, Edgar Zilsel, Alfred Tarski , Walter
Hollitscher, etc. there . Under Schlick 's supervision, Hong Qian worked
on his doctorate. He was awarded his doctorate in 1934 for a thesis
entitled 'Die Frage der Kausalitiit in der neuen Physik'. After that, Hong
stayed at the University of Vienna to continue philosophical research. On
June 22nd, 1936, Hong 's beloved teacher Schlick was assassinated.
Thereafter, Hong went back to China early in 1937.
Hong Qian became lecturer of philosophy at Qinghua University when
he went back to Peiping in 1937. After the outbreak of the Sino-Japanese
War, Hong fled to the Southwest of China, and after passing through
xvii
R. S. Cohen . R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
Philosophy of Science, xvii-xxii.
xviii FAN DAINIAN
many different places became the professor of philosophy at National

South-west Associated University at Kunming from 1940 to 1945. In this
period, he lectured on the philosophy of the Vienna Circle, and Schlick's
in particular. He published many papers on this topic in Chinese philo-
sophical journals and compiled these papers into a small book entitled
The Philosophy of Vienna Circle which was published in 1945. In these
writings , Hong systematically introduced the philosophy and scientific
world-view of the Vienna Circle , criticized the traditional metaphysics,
Kant's Apriorism , Husserl's and Scheler's phenomenalism, Windelband's
and Rickert's view of 'spiritual' science, Mach's positivism and Feng
Youlan's new Confucianism, and made a significant contribution to the
dissemination of logical empiricism and analytical philosophy in China .
As to Mach's positivism, Hong Qian pointed out: Although logical
empiricism inherited Mach's positivist spirit against metaphysics, there
is a fundamental difference between them. According to Mach's posi-
tivism, the reality of material bodies is a complex of sens ations, but
logical empiricists think, "since there are possibilities that all scientific
abstract formalism can be tested by the given, their reality is not less than
that of tables and chairs which can be felt. Therefore, the members of
Vienna Circle said: 'the atom is not a logical construct', not a way of
' thinking ' , but it is absolutely real." I
Hong also pointed out:
The opposition to metaphy sics by the logical empiricism is slightly different from that
of tradit ional positivi sm. Logical empiricists negate the role of metaphysics in the theory
of knowledge , but do not deny its sign ificance in practical life.
Hong agreed with Schlick :

that metaphysical philosophemes are conceptual poem s: in the totality of culture they play,
in fact. a role simil ar to that of poetry ; they serve to enrich life, not knowledge.'
Metaphysics is meaningful to ethics and the philosophy of life , so it

should not be eliminated.
Hong 's book The Philo sophy of Vienna Circle is similar to Ayer 's
Language, Truth and Logic in some aspects. Both were intended to intro-
duce the philosophy of the Vienna Circle to readers of their native lands.
But China lacked the tradition of empirical and analytical philosophy,
while in contrast Britain was strong in these. Compared to Ayer who paid
more attention to explicate his own viewpoint, Hong's book was more
faithful and comprehensive. Ayer's book was republished and reprinted
HONG QIAN AND THE VIENNA CIRCLE xix
again and again, and translated into many different languages (including
Chinese in 1980) and thereby became a very popular, basic philosoph-
ical reading in the world. Hong's book, after the 1st edition of 1945,
owing to political and ideological reasons, could not be republished in
mainland China until 1989, and then it was again welcomed as a valuable
academic treatise. Due to the differences of cultural traditions and
political circumstances, how different was what happened to these two
books .
Hong went to Oxford in 1945 where he was elected a research fellow
at New College. But events in China impelled him to return home in
1947, and while he was Professor and Director at the Department of
Philosophy, Wuhan University, he published 'Moritz Schlick and Modern
Empiricism ' in Philosophy and Phenomenological Research, vol. 9,
no. 4 (1949). In this paper, Hong introduced the debate between Schlick
on one side and Neurath with Carnap on the other side on the problem
of the foundation of knowledge, and criticized Carnap's viewpoint of
reductionism and protocol sentences. It was two years before the pub-
lication of Quine 's paper 'The Two Dogmas of Empiricism' .
After the People 's Republic of China came into existence, Hong
Qian moved to Beijing in 1950 and became the Professor and Head
of the Philosophy Department, Yenching University in 1951 [Beijing
University]. From 1952 to 1965, he was the Professor and head of the
Seminar for History of Foreign Philosophy of the Philosophy Department,
Beijing University and then was the director of the Institute of Foreign
Philosophy of Beijing University until his retirement in 1987.
Since logical empiricism was condemned as a reactionary idealist
philosophy serving imperialism, Hong could not continue his teaching
and writing on logical empiricism. In the first half of 1957, encouraged
by the Double Hundred Policy, Hong wrote two papers to introduce
Mach 's philosophical thought and Kant's Allgemeine Naturgeschichte und
Theorie des Himmels. He suggested to leaders of the Chinese Communist
Party to implement the Double Hundred Policy fully, "needn't overem-
phasize the leadership of Marxism, needn 't be afraid of ideali sm", "pay
more attention to the study of Western philosophy". In the summer of
1957, the Anti-Rightist Struggle broke out, Hong was criticized together
with other famous philosophers such as Jin Yuelin, Feng Youlan and
He Lin. The atmosphere of free contending soon completely vanished.
The only thing Hong Qian could do was to keep his silence and do
some translation and editing work.
xx FAN DAINIAN
Then Hong was made editor-in-chief of a series of western philosophy

from the ancient Greeks to contemporary western philosophy (in 5 vols.)
which became the basic materials for students of Philosophy Departments
in mainland China.
But, Hong Qian insisted on reading the western philosophical journals
and literatures, kept communications with members of Vienna Circle such
as Carnap, Feigl etc. until the Cultural Revolution (1966), and followed
the recent advances of logical empiricism and analytical philosophy in
the Western world.
During the Cultural Revolution, like many other philosophers and
scientists, Hong was criticized as a 'reactionary bourgeois academic
authority'. His house was searched, rare copies of his early writings,
letters from members of Vienna Circle, and some books were confiscated
and destroyed.
Perhaps due to the fact that some famous foreign philosophers wrote
to the leaders of the Chinese government and expressed their concern
about the safety of Hong Qian, the revolutionary committee of Peking
University did not force him to join the May 7th Cadre School to do
physical labor, but asked him to attend political study routinely. In 1973,
in order to coordinate the study of Lenin's Materialism and Empirio-
Criticism, Hong Qian and two other professors were ordered by the
Philosophy Department to translate the related paragraphs of Mach's
Die Analyse der Empfindungen into Chinese. In 1975, the abridged trans-
lations of Mach's book accompanied by an official preface were published
by the Commercial Press. (In 1986, Hong Qian and his co-translators
translated Mach's whole book and published it without the official
preface.)
When looking back upon the three decades since 1949, Hong Qian
often said to us regretfully, "lacking the indispensable research materials,
without the academic circumstances for free discussion, and with no inter-
national academic communication, I did little research work in these three
decades. Much time vainly slipped by."
After the end of the 1970's, with the opening to the West, Hong Qian
was again allowed to visit the outside world. Though by now in his
seventies and relatively frail, Hong Qian became, once more, an inde-
fatigable traveller, charming his hosts with his philosophical acumen
and acerbic wit. Through the 1980's, he received visiting fellowships
at Queens College, Oxford (1980, 1984) and Trinity College , Cambridge
(1982), contributed to symposia on Wittgenstein (1980) and Schlick-
HONG QIAN AND THE VIENNA CIRCLE xxi
Neurath (1982) in Australia, and lectured in Tokyo (1986) and Hong

Kong (1988). In 1984, Vienna University awarded him a second, honorary
doctorate on the 50th anniversary of his first one. He also became
research professor and member of the academic committee of the
Philosophy Institute, Chinese Academy of Social Sciences since 1978,
honorary President of the Chinese Society of Contemporary Foreign
Philosophy since 1980. His great prestige was a crucial factor in estab-
lishing the Sino-British Summer School of Philosophy in Beijing in 1988
with Hong and Sir Alfred Ayer as honorary presidents . Sir Peter Strawson
became an Honorary President after Ayer's death in 1989.
In his last decade , Hong's academic contributions reached a second
peak. He edited and published collected papers on logical empiricism
(vol. 1 in 1982, vol. 2 in 1984), enlarged and reedited the Anthology of
Contemporary Western Philosophy. He wrote and published 17 papers
such as: 'Wittgenstein and Schlick' (Proceedings of 5th International
Wittgenstein Symposium, 1981), 'Moritz Schlick und der Logische
Empirismus ' (Grazer Philosophische Studien 16/17, 1982), 'Remarks
on Affirmations' (Synthese 64, 1985), "On Logical Empiricism" (Zhexue
Yicong, 1987), 'Some Problems about Logical Empiricism' (Journal of
Dialectics ofNature , 1989), 'Ayer and the Vienna Circle' (The Philosophy
of A. J. Ayer, 1992), ' Rudolf Carnap' (21th Century, 4, 1992), etc . In
these papers, Hong Qian discussed the historical sources of the Vienna
Circle, the conditions of its rising in Austria during the 1930's and its
dissemination in English speaking countries; commented on the philo-
sophical viewpoints of Schlick, Ayer, Carnap, Feigl, Kraft and others ;
reviewed the changes and developments of logical empiricism in recent
decades , and its present situation and new trends; expressed his skepti -
cism about foundationalism . His works are valuable resources for the
study of the philosophy and history of the Vienna Circle and highly
valued by the academic world.
In respect to the criticism of logical empiricism in the recent half-
century, Hong Qian paid much more attention to Quine's criticism
than to that of Popper, Kuhn and Feyerabend. Especially, Hong dis-
agreed with what Popper had said, that he had killed logical empiri-
cism. Instead, Hong believed that logical empiricism is still alive and that
there was a trend to its rejuvenation in recent decades. Nevertheless, Hong
acknowledged the weakness of logical empiricism. In his article
entitled 'Philosophical information from travel in Europe' (in 1980),
he wrote:
xxii FAN DAINIAN
No need for reticence, since the beginning of the Vienna Circle, most schools of con-
temporary analytical philosophy have not paid sufficient attention to ethics, have not placed
ethics in the proper position in philosophy. Just as someone pointed out: a complete
philosophical system, needs not only an integral part of theoretical philosophy, but also
an integral part of practical philosophy. For instance, in Kant's philosophy, there are
three great treatises of Critique; in Marxist philosophy, there are dialectical materialism
and historical materialism. In view of this, it is no wonder that Russell said emotion-
ally: Strictly speaking , in this kind of philosophy like logical positivism, actually there
is no philosophy, but only methodology.'
Hong hoped that analytical philosophers including logical empiricists will

pay more attention to practical philosophy and ethics.
Hong Qian was really very sad after the Beijing event in June 1989.
He deeply worried that it would block China's international communi-
cations in philosophy which had just opened. It was Hong's principal
advice that helped the Sino-British Summer School of Philosophy to
be revived last summer. This Beijing International Conference on
Philosophy of Science was also prepared under his energetic support.
So Hong Qian's death is really an irremediable loss to Chinese philo-
sophical circles and the philosophy of science in particular. We will
cherish the memory of him forever.
Zhongguancun,
Beijing 100080,
China.
NOTES
I Hong Qian (1989). The Philo sophy of Vienna Circle, Commercial Press , Beijing,
p.62.
2 Moritz Schlick (1979) . Philosophical Papers, vol. 2, D. Reidel, Dordrecht, p. II I.
3 Hong Qian (1990). Collected Papers on Logical Empiricism, Shanlian Shudian, Hong
Kong, pp. 247-248.
RISTO HILPINEN
ON SOME FORMULATIONS OF REALISM, OR

HOW MANY OBJECTS ARE THERE IN THE WORLD?
This paper has a very modest objective. The recent discussion of realism
contains expressions and metaphors which I find difficult to understand,
and I shall try to translate some of these expressions into a more com-
prehensible idiom. The arguments about realism - or about different
'realisms" - often involve claims about the dependence (or indepen-
dence) of the world on our theories and concepts. What is the nature
of this dependence or independence? Some philosophers make a dis-
tinction between our conceptualizations of the world and "the way of
the world really is" (italics mine), or at least attribute such a distinc-
tion to other philosophers.' I assume that this distinction is not meant
to be the same as the familiar distinction between a representation and
its object.' Do scientists not try to represent the world the way it really
is? (The representation of the world as something other than what it really
is would seem to be misrepresentation.) The metaphors of 'carving '
and 'cutting ' are common in this context: Hartry Field speaks about
the "carving up of the noumenal dough" by means of various "cookie
cutters"," whereas Hilary Putnam has argued that this metaphor is a
misleading characterization of scientific representation, that is, a mis-
leading summary of what he regards as the correct account, his "internal"
or "pragmatic" realism.' The target of Putnam's criticism, "external"
or "metaphysical" realism, includes, according to him, the thesis that "the
world consists of some fixed totality of mind-independent objects"." In
this paper I am mainly interested in this thesis and its negation. I shall
not discuss other aspects of what Putnam calls "external realism", viz.,
the assumptions of correspondence (i.e., the correspondence theory of
truth) and uniqueness (the uniqueness of a complete and correct descrip-
tion of the world) ."
R. S. Cohen. R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
Philosophy of Science, 1-10.
2 RISTO HILPINEN
II
In his Paul Carns Lectures Putnam says:"

(PI) How many objects there are in the world is relative to the
choice of a conceptual scheme.
According to Putnam, this statement may be regarded as an expression
of a "relativistic" view:"
How can one propound this sort of relativistic doctrine and still claim to believe that
there is anything to the idea of ' externality', anything to the idea that there is some -
thing ' out there' independent of language and mind?
Sentence (P I) can be paraphrased as follows :

(P2) The number of objects in the world is relative to the choice
of a conceptual scheme.
Putnam also speaks about the dependence and independence of objects
and facts on our concepts. This suggests that the expression "relative
to" in (P2) can be replaced by "dependent on", that is, that (P2) (and
hence (PI)) can be paraphrased as
(P3) The number of objects in the world depends on the choice
If (P3) is part of Putnam 's doctrine of 'internal' or 'pragmatic ' realism,
the doctrine he calls "metaphysical realism" presumably involves the
claim
(MI) The number of objects in the world is independent of the
choice of a conceptual scheme.
(M I) seems to follow from the thesis of metaphysical realism men-
tioned earlier,
(M2) The world consists of some fixed totality of mind-indepen-
dent objects.
The expressions 'mind-dependence' and 'dependence on a concep-
tual scheme ' are often used interchangeably in the discussion of realism,
and they will be used in this way also in the present discussion. In this
context, a 'mind' consists of a system of concepts by means of which
the reality is represented and understood . Sometimes philosophers have
ON SOME FORMULATIONS OF REALISM 3
confused this sense of ' mind' with its other meanings; for example, in
his discussion of realism Robert Nola characterizes after-images as mind-
dependent entities.'? After-images may be called 'psychological' objects,
but this does not make them necessarily "mind-dependent" in the present
sense of the word.
(P3) and (M I) appear to be propositions about a number, viz ., the
number of objects in the world: (P3) seems to entail, e.g., the following:
(P4) There is a number n such that n = the number of objects in
the world, but the number of objects in the world would not
have been n if our choice of a conceptual scheme had been
different from what it actually was.
Here 'n' is assumed to be a rigid number designator; (P4) says that the
expression 'the number of objects in the world' refers to different
numbers in different possible worlds, and this variation depends on
variation in our conceptual schemes.
If (P4) were a well-defined proposition, it would express a truly
relativistic doctrine. For example, the much discussed doctrine of
linguistic relativity can be expressed in a form analogous to (P4):
If a given speaker's language were (significantly) different
from what it actually is, his "picture of the world" (world
view) would differ from his actual view."
But (P4) is of course not what Putnam means by 'internal' or ' pragmatic'
realism. In Representation and Reality Putnam asks us to consider an
example in which a person is taken to a room and asked, "How many
objects are there in this room?"." He argues that the question has no clear
and unambiguous answer: there is no such thing as the number of objects
in a given room. This can be expressed by saying that the expression
'object' is not an individuating expression or a sortal expression." To
use Frege's words, the concept of an object "does not isolate in a definite
manner what falls under it".14 Putnam expresses this indirectly by saying
that the answer to the question about the number of objects in the room
"depends on what one means by 'object' ", and that "there are many ways
of using the notion of an object"." These characterizations are mis-
leading, because they suggest that the expression ' object' is somehow
ambiguous." To say that there are different "ways of using the notion
of an object" means in this context only that the expression 'object'
can be joined together with different individuating descriptions.
4 RISTO HILPINEN
The concept of an object, unlike the concept of a chair or the concept

of a horse, is not an individuating concept, and therefore the descrip-
tion 'the number of objects in the world' (or ' the number of objects in
this room') is not a proper definite description . Consequently the expres-
sion ' the number of objects in the world' does not denote anything (in
any possible world), and sentences (PI)-(P4) and (MI) cannot be under-
stood literally as statements about the number of objects in the world.
Literally interpreted, they should be regarded as false (if we accept
Russell's account of definite descriptions) or meaningless.
Charitably interpreted, (PI) may be regarded as an indirect or round-
about way of expressing the following 'grammatical' (semantical) fact
about the use of the expression 'object' :
(P5) The expression 'object' is not an individuating expression.
According to this interpretation of (P I) , the doctrine which Putnam
calls "metaphysical realism" involves a semantical mistake about
the world 'object' , and Putnam's "pragmatic" or "internal" realism , as
far as (PI) is concerned, merely reminds us that 'object' is not an indi-
viduating expression. Putnam's example about the number of objects
in a room supports only (P5), but no further "metaphysical" conclu-
sion . In Reason, Truth and History Putnam expresses this view as
follows : 17
(P6) What objects does the world consist of? is a question that it
only makes sense to ask within a theory or description .
According to the observations made earlier, this is true, but it does not
express an interesting metaphysical doctrine: (P6) should be regarded
as an uncontroversial observation about the semantics of the word
'object' . If there are "metaphysical realists" who seem to deny (P6),
they are either semantically confused or are using some technical or
special notion of Object; perhaps they have in mind some notion of
metaphysically basic or "ultimate" Object with its own individuating
characterization.
III
The word 'object' is not an individuating expression, but if F is any indi-

viduating expression, 'object which is F' or 'F-object' is also an
individuating expression : such an expression is synonymous with F. If
the expression 'object' in (P3) is supplemented by some individuating

expression F, we get
(P7) The number of F-objects in the world depends on the choice

For example, for alligators we obtain the follow ing thesis of "internal
realism"
(P8) The number of alligators in the world depends on the choice
of a conceptual scheme .
This claim seems false. It makes good sense to speak about the number
of alligators in the world, and for alligators the relativized thesis (M I),
the thesis of relativized metaphysical realism, seems to hold true. IS The
falsity of (P8) can be seen more clearly if it is rewritten as follows:
(P9) If our conceptual scheme were different from what it actually

is, the number of alligators would be different from the actual
number of alligators.
(P9) is clearly false . (P8) may be regarded as true only if it is not under-
stood as a statement about the number of allig ators, but rather as a
statement about (the extension of) the word (or concept) ' alligator ' .
Let Ext(G) be the extension of an expression (or concept) G, let
Extaclual(G) be the actual extension of G (the extension of G in the actual
world) , and let n(A) be the cardinality of a set A. Thesis (P9) must be
distinguished from the following proposition:
(PIO) If our conceptual scheme were different from what it
actually is, n(Ext(, Alligator')) would be different from
n(Extaclual' Alligator'j)."
(P I0) is not equivalent to (P9): (P I0) may be true, even though (P9) is
false. If " internal realism" involves the acceptance of (P9) (and not
merely that of (PIO)) , it amounts to the view that if a dog 's tail were
called a leg, dogs would have five legs."
According to the argument given above , theses (P3) and (M I) are
meaningful only when relativized to objects "under" some individu-
ating description, as shown in (P7). If the doctrine of realism depends
on the truth of such theses, it is necessarily realism with respect to F-
objects, where F is some individuating description. Thus the following
6 RISTO HILPINEN
thesis is not meaningless (or semantically confused) in the same way

as (M2):
(M3) The world contains a fixed totality of mind-independent F-

objects,
where F is some individuating description. By generalizing with respect

to individuating concepts, we obtain from (M3) the following doctrine
of "absolute metaphysical realism ":
(M4) For any individuating concept F , the world contains a fixed
totality of mind-independent F-objects.
By a similar generalization we get from (P7) the following thesis of

"internal" realism :
(PI I) For any individuating concept F, the number of F-objects
in the world depends on the choice of a conceptual scheme.
It is not quite clear what 'mind-independent' means here. If the unavoid-

able relativity of any "fixed totality" to an individuating concept is
regarded a form of "mind-dependence", (M4) is a self-contradictory thesis
and (P 11) becomes a conceptual truism . But this is not a plausible inter-
pretation of 'mind-dependence ' (and 'mind-independence'), since it is
possible to distinguish kinds of objects whose character is "constituted"
by people's beliefs, e.g., chess pieces and traffic signs, from other kinds
whose character is largely independent of what people believe about
them, for example, alligators ." Thus both (M4) and (PI!) are clearly
false . To ask questions about existence and identity, we need individu-
ating concepts or descriptions, and for some such descriptions the correct
(true) answers are objective in the sense that they are independent of
people's beliefs, but in other cases the answers are genuinely "mind-
dependent".
IV
When philosophers have discussed real ism relativized to individuat-

ing descriptions (sorts or kinds), condition (M3) has usually not been
regarded as a criterion of realism about F-objects. Sometimes realism
about F's has simply been regarded as the view that there exist F's.
For example, according to Mark Heller,22
to be a realist about a some particular thing or kind of thing is to believe that thing or
kind exists (or that kind has members).
Heller calls this form of realism "existential realism". This is not an inter-
esting sense of 'realism': according to Heller's definition, a person who
believes that there are swans and that all swans are white, is a realist
about swans, but anti-realist (or non-realist) about black swans. Moreover,
existential realism in Heller's sense is compatible with mind-dependence
in the sense suggested in the previous section: There are (i.e., there exist)
traffic signs and chess pieces, but they are not mind-independent; they
are constituted by people 's beliefs." If the requirement of "mind-inde-
pendence" is added to Heller's definition, we get the following thesis
of (existential) realism with respect to F-objects:
(RI) F-objects exist in the world in a mind-independent manner.
Michael Devitt has formulated an "absolutized" version of (R I) as
follows ;"
(R2) Tokens of most current common-sense , and scientific, physical
types objectively exist independently of the mental.
To make thesis (R 1) immune to the objection against Heller's defini-
tion of existential realism mentioned above , it should be reformulated
as
(R3) The existence of F-objects in the world does not depend on
the mental (or on the choice of a conceptual scheme).
In the same way, Devitt 's "absolutized" version of (Rl ) should be under -
stood as follows :
(R4) For most individuating concepts F, the existence of F-objects
does not depend on the mental (or on the choice of a con-
ceptual scheme).
Devitt 's version of realism (reformulated as (R4» is more cautious
than (M4), and therefore more likely to be true, since it allows the pos-
sibility that some objects exist in a mind-dependent way, or, to use the
terminology of conceptual schemes , it allows the possibility that there
are objects whose existence "depends on our conceptual scheme" . Such
objects (for example, chess pieces) satisfy the following modified form
of (P7):
8 RISTO HILPINEN
(PI2) The existence of F-objects in the world depends on the choice

But whether these objects or object kinds form a majority or a minority
among all object kinds is hardly of great conceptual and philosophical
interest. The interesting questions about realism concern the interpreta-
tion of the expression 'mind-dependence' used above and the truth of the
relativized thesis (R3) for various concepts (or object kinds) F. If
Putnam's internal realism makes everything mind-dependent on the
ground that the identification of anything as an object requires an iden-
tifying concept, he bypasses a significant distinction between objects
constituted by people's beliefs and those whose character and existence
are independent of what is believed about them.
Departments of Philosophy,
University of Turku, Finland;
University of Miami, USA.
NOTES
I Cf. Susan Haack (1987). " 'Realisms'''. Synthese 73, pp. 275-299.
2 Mark Heller (1988). 'Putnam, Reference. and Realism ', in Midwest Studies in
Philosophy XII: Realism and Antirealism, ed . by A. P. French et al., University of
Minnesota Press, Minneapolis, pp. 113-127; see p. 115.
3 The distinction between the world as it really is and as it is represented by our theories
resembles Charles Peirce's distinction between the immediate and the dynamical object
of a sign : The immediate object of a sign is the object " as the sign itself represents it",
whereas the dynamical object is the " reality which by some means contrives to deter-
mine the sign to its representation"; see Collected Papers of Charles Sanders Peirce,
vol. 4, ed. by C. Hartshorne and P. Weiss (1933). Harvard University Press, Cambridge,
Mass., paragraph 4.436 . According to Peirce, the immediate object of a sign is a repre-
sentation rather than the reality represented by a sign .
4 Hartry Field (1982). 'Realism and Relativism', The Journal of Philosophy 79, pp.
553-567; see p. 561.
5 Hilary Putnam (1987). The Many Faces of Realism, Open Court, La Salle , III. p. 19,
and Representation and Reality (1988). MIT Press, Cambridge, Mass . pp. 113-114.
According to Putnam , the cookie-cutter metaphor founders on the question, "What are
the parts of the dough?" I don 't think this is a convincing objection.
6 Hilary Putnam (\ 981). Reason, Truth and History, Cambridge University Press,
Cambridge, p. 49.
7 Ibid., p. 49.
8 Hilary Putnam, The Many Faces of Realism, p. 32.
9 lbid., p. 32.
(() See Robert Nola, ' Introduction: Some Issues Concerning Relativism and Realism in
Science', in Relativism and Realism in Science, ed. by R. Nola (1988). Kluwer Academic
Publishers, Dordrecht, pp. 1-35; see p. 5.
11 See John Carroll (ed .) (1956). Language , Thought, and Reality : Selected Writings
of Benjamin Lee Whorf, The MIT Press , Cambridge, Mass., p. 221.
12 Hilary Putnam . Representation and Reality, pp. 110--112.
13 See E. J. Lowe (1989) . Kinds of Being, Basil Blackwell, Oxford, pp, 9-10.
14 Gottlob Frege (1950) . The Foundations of Arithmetic, transI. by J.L. Austin, Basil
Blackwell, Oxford, p. 66. See also Nicholas Griffin (1977) . Relative Identity, Clarendon
Press, Oxford, p. 40, and Peter Geach (1968) . Reference and Generality, Emended Edition,
Cornell University Press, Ithaca, pp. 38-39.
15 Representation and Reality, pp. 113-114.
16 In his discussion of sortal predicates Geach discusses the problem of counting the
red things in a room , but he does not conclude from the impossibility of this task that
the expression ' red thing' is ambiguous or can be used in different ways; according to
Geach, the example shows only that 'red thing' is not what he calls a "substantival"
(that is, a sortal) expression. See Reference and Generality , p. 39.
17 Reason, Truth and History, p. 49.
18 This is even more obvious in the case of dinosaurs and (say) galaxies, since in such
cases we need not worry about the causal and indirect effects of our conceptual activi-
ties. (In this context "the F-objects in the world" should be taken to include all F-objects
which have existed or will exist in the world .)
19 According to the possible worlds semantics of conditionals, (P9) and (PIO) can be
formulated as follows :
(P9 *) In the nearby alternative worlds where our conceptual scheme differs from
that adopted in the actual world , the number of alligators differs from the
actual number of alligators.
(PlO*) In the nearby alternative worlds where our conceptual scheme differs from
that adopted in the actual world, n(Ext(' Alligator'» differs from
n(Ext aclual('Alligator'».
20 I am indebted to Professor Alan Musgrave for this observation.

21 C.S. Peirce's well-known definition of 'reality' is based on this distinction. According
to Peirce,
that is real which has such and such characters, whether anybody thinks it to have those
characters or not.
(Collected Papers of Charles Sanders Peirce (1934) . vol. 5, ed. by C. Hartshorne and
P. Weiss, Harvard University Press, Cambridge, Mass., paragraph 5.430.)
22 Mark Heller , 'Putnam, Reference, and Realism', pp. 113-114.
23 See also Robert Nola, op. cit., pp. 1-35; see pp. 4-5; the definitions given by Nola
also entail that one can be a realist about swans but an anti-realist about black swans .
10 RISTO HILPINEN
24 See Michael Devitt (1984). Realism and Truth, Princeton University Press, Princeton,
p. 22. Devitt's reference to phy sical (as opposed to, for example, psychological)
types here may be based on the confusion between the two senses of ' mind' and
' mental' mentioned in section II, viz., mind as a system of concepts and mind as the
psychological.
RUDOLPH KREJCI
DISSOLUTION OF THE
REALISM/ANTIREALISM PROBLEM
INTRODUCTION
This paper investigates the way to dissolve the adversary either/or

position represented by realism and antirealism by suggesting the com-
plementarity approach which was so successfully introduced by Niels
Bohr in his revolutionary way of thinking applied to modem physics.
Bohr himself hoped that one day the complementarity principle would
become the backbone of everyone's education and would be applied
in the entire spectrum of our society. However, Bohr's approach is
regarded as revolutionary only in the western philosophico-scientific
tradition. Outside of it there are similar philosophical , moral, and
spiritual perspectives to be found in Taoism, Buddhism, Upanishads, and
lainism whose followers regard complementarity of Bohr as more-or-less
commonplace.
The argument of this paper focuses on the priority of the question
and the underlying assumptions which are responsible for polarity
systems of answers in the form of the epistemological realism coupled
with ontological materialism on one hand and epistemological antirealism
and ontological idealism on the other hand. In this very adversary form
it has become a major philosophical, scientific , and ideological argument
of the western tradition. We seem to be at a crossroad; we can blindly
continue to ama ss other either/or positions in our realism/antirealism
conflict. If we decided to follow Niels Bohr 's sugge stion of comple-
mentar ity we will discover that any two opposing contradictory aspects
are always integral, inseparable parts of a whole . "One cannot bow
in front of somebody without showing one 's back to somebody else"
[l ].
Any consequent dissolution of the realism/antirealism problem can
be successfully achieved by adopting a new conceptual context going
radically beyond the Aristotelean tradition by exclusion of his Tertium
non datur principle and replacing it by the principle of complemen-
tarity which can be symbolized as excluding [p V q == -(p f\ q)] and
replacing it by [p V q == (p f\ q) f\ r 1\ s f\ 00].
11
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.), Realism and Anti -Reali sm in the
12 RUDOLPH KREJCI
A great insight is a statement whose denial is another

great insight.
Niels Bohr
Philosophical and scientific problems manifest their intrinsic intercon-

nectedness all through the history by articulations of peculiar mind-sets
driven by mankind's curiosity. Some of them became central to views
about the mystery of the universe and man's role in it. Their implica-
tions found, in tum, their manifestations in religion, arts , science, politics
and in due time penetrated the entire theory and praxis of mank ind. All
through history such a philosophical problem was and still is the problem
of realism based on assumptions that there is a world outside of the
perceiver, governed by laws, we can know and communicate to others.
Two forms of the western realism are notably represented in teachings
of Plato and his pupil Aristotle, and both of them are responsible for a
great part of the philosophical, scientific and religious cultural context
of the West. In Whitehead's comment that the entire history of the West
consists of a few footnotes to Plato we have probably the best defined
historical role of reali sm . As this type of realism in all its forms is
essentially dualistic, it will find its contradictory nemesis in antirealism
challenging the reali stic position by advocating a view that the world and
everything in it is nothing but a temporary useful fiction helping us to
organize our observations for our benefit, however, predestined to be
replaced by another useful fiction in the future. Both of these positions
represent a conflict of mind-sets within which corresponding self-
referential realities are being constructed.
Psychologically the realistic thesis provides an ultimate foundation for
attainment of certainty by getting away from oneself and finding the
anchor in the world of things in their totality. As one can see, the entire
cultural, religious, political and social striving was, and, to a great degree,
still is, carried on within the realistic tenets of mankind. On the other
hand, the historical position of the antirealism starting with Xenophanes'
anthropomorphism and method of projective hypostatization bases all
human striving in the man himself as an observer who, driven by his
curiosity, continues his search and re-search and never reaches the
ult imate finding! The search with its temporary finding becomes the
only uppermost reward. Following Xenophanes, the best of the philo-
sophical and scientific tradition became necessarily anthropomorphic,
DISSOL UTION OF THE REALISM /ANTIREALISM PROBLEM 13
hypothetical, open-ended. It found its corresponding world-view in the

concept of open universe, open-society, open-mind without ever being
able to reach certainty by either logical , mathematical, scientific, philo-
sophical, or mystical ways.
Not until the 20th Century was it possible to depart from the tradi-
tional dualistic categories dictating our thought pattern with implied
adversary relationships. The entire fabric of our western culture became
saturated with this dualistic categorization manifested in most of its
science, philosophy, religion, arts, politics, and social striving. When it
entered the global scene it assumed the role of messianism attempting
to disseminate its western mission all over the world . Given the view that
there exists an absolute truth and by conviction that it was in posses-
sion of certain selected groups we have had the greatest historical
examples of intolerance represented by the institution of the Inquisition
and the 20th Century National Socialist, Fascist, and Communist ide-
ologies. Realism always appealed to large collectives, while antirealism,
on the other hand, was the conviction of individuals working in isola-
tion from the masses.
II
When the province of physical theory was extended

to encompass microscopic phenomena, through the
cre at ion of quantum mech anics, the concept of con-
sciousness came to the fore again : it was not possible
to formulate the laws of quantum mechanics in a fully
consi stent way without reference to cons ciousness.
Eugene Wigner [21
It was a universal belief of the scientists of the 19th Century that the
20th Century would bring the completion of the Newtonian world-view
by providing all the missing details. Materialism became not only the
integral part of scientific method but the very condition under which
science could be conducted. There were only rare exceptions among
scientists and mathematicians who did not subscribe to this dogma
[3].
The predictive power of the mechanistic science celebrated successes
which silenced even the most vocal critics. Immanuel Kant himself
declared the findings in mathematics, logic, and physics, as completed
in principle, to which nothing substantial could be added. This very
14 RUDOLPH KREJCI
successful way of thinking in physical sciences inspired a speedy appli-

cation to all other sciences such as biology, psychology, sociology,
economics, anthropology, politics, and philosophy. The shift continued
and persists practically up to now. Its political, social, and economical
implications became responsible for much of the havoc of the 20th
Century.
One of the most significant realistic and materialistic stands against
the implications of the new advances in science at the end of the 19th
Century and the beginning of the 20th Century is Vladimir I. Lenin's
work Materialism and Empiriocriticism, regarded by his followers as
an epoch making work of genius [4]. There were two major philo-
sophical positions Lenin maintained in his desire to protect the dialectical
materialism of his days. The first was the theory of knowledge of
Marxism under attack from various streams of modern philosophy
including the Social Democrats and the Russian Marxists - the second
was the problem of the concept of matter which was under attack due
to the implications of the discovery of radioactivity, which in turn
released an avalanche of speculations suggesting "dematerialization of
the atom" and "disappearance of matter".
Lenin was the first one to recognize these new emerging developments
as mortally endangering the very realistic and materialistic foundation
of Marxism in both theory and praxis. Truly in Marx's way Lenin attacks,
under the name of 'physical idealism', scientists and philosophers, such
as Mach, Poincare, Avenarius, Duhem, Pearson , his Russian adversaries
Lopatin-and Bogdanev and a host of others . Lenin's defense of materi-
alism and realism with his epistemology of copy-theory laid down the
20th Century position of the Soviet School of Marxism (Leninism-
Stalinism) against the idealistic interpretations of Einstein's theory of
relativity with its equivalence of mass and energy and of quantum
mechanics of the Copenhagen School.
On the basis of Lenin 's restatement of the 19th Century realistic
and materialistic dogma it was rather consequential to reject the philo-
sophical concepts of complementarity and uncertainty-relations due to
their denial of causality and objectivity of micro-phenomena. The
Copenhagen 'idealistic interpretations ' was thus rejected because it
regarded the wave-function describing the state of the micro-particles not
as an objective state of affairs, but merely as an expression of the knowl-
edge of the observer where the wave is not the property of the objects
of the microworld but merely a wave of probability.
DISSOLUTION OF THE REALISM /ANTIREALISM PROBLEM 15
As we can see it in retrospect, Lenin's crusade against the spirit of

the modern way of thinking which was manifested in new scientific
and philosophical approaches, helped to create a closed system of
thinking, which in turn produced a closed society unable to compete with
the rest of the world. Its recent abdication from the stage of modern
history of mankind must be regarded as an overdue disappearance of
dysfunctional philosophy in both theory and praxis .
III
Physics does not deal with physics . Physics deals with

what we can say about physics. What we can say
depends in tum on what we can agree about; and that
depends on clicks of a counter, irreversible acts of
amplification, indelible records .
John Archibald Wheeler [5]
The above quotation is from Professor Wheeler 's 'working paper' series
of three lectures delivered by the author in Beijing, Hefei, and Shanghai
in October 1981. It was almost 50 years ago, when in 1932 Niels Bohr,
Professor Wheeler 's great teacher and mentor, returned from his visit
to China, inspired by the great cultural tradition of its past. The prin-
ciple of complementarity formulated by Bohr five years before his
pilgrimage to China acquired thus a new symbol represented by the
Chinese YIN-YANG accompanied by the Latin motto non contraria
sed complementa, and the hope that one day this very principle would
become the central core of our entire educational system in the West.
Professor Wheeler, one of the last surviving giants of Bohr's school could
not help but come and report to you the accomplishments which followed
Bohr's legacy during the remainder of the 20th Century . In light of this
new tradition we can observe a slow process of introducing comple-
mentarity as a way of thinking into all other aspects of our contemporary
modes of thought. Similarly, we see the implications of Heisenberg's prin-
ciple of uncertainty for the historical striving of mankind to escape
from itselt into some postulated certainty out there, and with it the
nonetheless stupendous implications of Godel's incompleteness theorem
which proved the openendedness of human thought as against the
axiomatic closed mind. In light of these innovative ways of thinking I
dare to argue for the dissolution of the realism-antirealism problem as
one of the important metaphysical underpinnings of the historical and
16 RUDOLPH KREJCI
cultural tradition of the West which is entirely foreign to the traditional

heritage of China and the great part of the Orient.
Professor Wheeler's 'Chinese' quotation provides a telling description
of the new view of the nature of the reality of the universe as no more
"sitting out there" independent of any act of observation. Rather, 'reality'
is directly the co-production and by-product of the observer's inten -
tional total tools, thus making observer's actions necessarily participatory
actions. (Esse est participare.) There are many implications of this tum
around, one of them is that we cannot have any information unless we
have acquired it by some specific method accompanied by detailed dis-
closure of both - what we report and how we arrived at it. The observer
is ourselves - thus we have to start and end with him (or her) - there
seems to be no escape any more . Again there are necessary implica-
tions. The most important ones are that the traditional dualistic division
of the subject and object, inner and outer world, body and soul , good
and evil, true and false, right and wrong, paradise and hell are no longer
adequate concepts for a more sophisticated description of our new
'realities' .
A long time before the impact of the above discussed ideas was felt
in our century , there was an important discovery of the German physi-
ologist Johannes Muller (1801-1858) who formulated the idea that the
world as we sense it is our own construction - known as ' MOller 's law
of specific nerve energies'. An important conclusion of this law is that
our sentences do not mirror the external stimuli but only the state of
the nerve being stimulated which provides us with the justification that
our system of cognition is self-referential and self-explicative [6].
MOller's law thus antic ipates very important epistemological conse-
quences of the new way of thinking in the 20th Century, expressed
through principles of complementarity and uncertainty relations and the
incompleteness theorem .
CONCLUSION
This paper attempted to expose the cultural, philosophical, and scien -

tific assumptions behind the historical conflict of realism and antirealism.
They were responsible for a state of petrification into dualistic categories
of adversary relationships of our traditional way of thinking applied to
the entire interpretation of reality in both theory and praxis. It was the
radical departure from this traditional way of thinking due to our new
DISSOLUTION OF THE REALISM /ANTIREALISM PROBLEM 17
scientific advances at the beginning of the 20th Century which forced

us to rebuild our world-view during the 20th Century. It has been proved
over and over again that deep-rooted prejudices of human thinking die
hard as it was the case with false hypotheses about ether, phlogiston,
or pure race, survival struggle, class struggle, Oedipus complex, geo-
centricity of the solar system, phrenology and so forth. The old world-
view was a result of questions we asked and extralogical strivings for
certainty - it helped us to collect facts we desired, and preferred inter-
pretations we were pleased with. The old world -view was self-referential
and self-explicative without being aware of it.
In view of the above insight many philosophical problems which
resulted in adversary positions will disappear because a new world-
view does not provide the soil in which they can flourish. Among these
will be notably our realism/antirealism problem and hosts of others
such as materialism and idealism, subjectivism and objectivism, mind-
body problem, individual-collective and any other adversary application
of categories of exclusion.
The transition from one to another world-view is the best testimony
of humanity triumphing over its own prejudices and limits and reaching
its higher level of self improvement on its way ad summum .
(p V g) == -(p /\ g)
is rather
p V q == P /\ q /\ r /\ s /\ 00
Department of Philosophy,
University of Alaska,
USA.
BIBLIOGRAPHY
I . Abbe of Galiana (1728-1787) quoted by John Wheeler, in his address at the ceremony
conferring the Atoms for Peace Award on Niels Bohr, 1957.
2. Wigner, Eugene (1967) . Symmetries and Refle ction s. Indiana University Press,
Bloomington, p. 189.
3. Isaac Newton himself hoped to explain with his mechani stic laws only "all corpo-
real things ", still maintaining an existence of non-corporeal ones. Mathematicians
and logicians due to their a prioristic approaches were outside of the realist-materi-
alistic orbits .
18 RUDOLPH KREJCI
4. Lenin, Vladimir 1. (1952). Materialism and Empiriocriticism, Critical comments on

a Reactionary Philosophy, 1909 (originally published), Moscow.
5. Wheeler, John Archibald (1982). 'Physics and Austerity - Laws without Laws',
working paper, Center for Theoretical Physics, University of Texas, Austin, Texas ,
p.32f.
6. MOiler, Johannes (1834-1840). Handbuch der Physiologie des Menschen. "Sensation
consists in the sensorium's receiving through the med ium of the nerves , and as a
result of the action of an external cause, a knowledge of certain qualities or condi-
tions, not of external bodies, but of nerves themselves ." Miiller's quotation is in E.
Boring (1950). A History of Experimental Psychology, 2nd ed., Appleton, New York,
p.82.
ALAN MUSGRAVE
REALISM, TRUTH AND OBJECTIVITY*
Bill Lycan once called me a mad-dog realist. Whereupon Deborah Mayo

and Dick Burian presented me with a large wooden sign bearing the
legend "BEWARE OF THE mad-DOG realist". I am very proud of this
sign and I keep it in my office back in New Zealand. I thought of bringing
it to this conference. But I decided not to. You see, the sign is misleading,
I am not a mad-dog realist at all, but rather a pretty tame one. More rabid
realists than me are common, as we will see.
I . REALISM AND TRUTH
I subscribe to the old-fashioned idea that scientific realism is first and

foremost a thesis about the aim of science. It says that the aim of a
scientific inquiry is to discover the truth about the matter inquired into.
This incorporates a semant ic thesis (inquiry results in true or false state-
ments about the world) and an axiological thesis (science aims for true
statements). Both are controversial: some antirealists deny the semantic
thesis, others the axiological thesis.
This cannot be the whole realist story, of course. To what extent has
the realist aim actually been achieved? It is now fashionable for realists
to give strong positive answers to this question. What Putnam (before
he abandoned realism) or Boyd or Leplin would like to say is that current
science is all true. But history intrudes and they hedge this in various
ways : they say that most of current science is true (supposing that the
notion of 'current science' is unproblematic, which it is not); or they
say that most of current science is true or nearly true (supposing that
the notion of 'near truth' is unproblematic, which it is not); or they say
that most of current mature science is true or nearly true (supposing
that the notion of 'maturity' is unproblematic, which it is not).
Furthermore, they make these strong claims definitional of scientific
realism : you cannot be a scientific realist unless you think that all (or
most) of current science (or current mature science) is true (or nearly
true) .
Similarly strong claims are made about the existence of the 'theo-
19
R. S. Cohen . R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
20 ALAN MUSGRAVE
retical entities' postulated by current science - or equivalently, about how

the 'theoretical terms ' of current science succeed in referring to real
things in the world. Such claims also become definitional of scientific
realism : you cannot be a scientific realist unless you think that all or most
current theoretical entities really exist, that all or most current theo-
retical terms really do refer.
Thus construed, scientific realism is not a thesis about the aim of
science at all. It is a substantive ontological or metaphysical thesis . It
is the thesis (I omit the hedgings for simplicity) that there really are
such entities as current science claims there to be and that what current
science tells us about such entities is true (see Leplin, 1982, p. 88 &
p. 90). This is mad-dog realism proper.
Some try to soften it by harping on reference but giving up on truth.
We are to believe in scientific entities (believe that theoretical terms
refer), without thinking true any theory about those entities. We are to
be ' entity-realists' without being 'theory-realists' . Or so say Nancy
Cartwright (1983) , Ian Hacking (1983), and Michael Devitt (1991) - if
I understand them.
This is incoherent. To believe in an entity, while believing nothing
further about that entity, is to believe nothing, I tell you that I believe
in hobgoblins (believe that the term 'hobgoblin' is a referring term).
So, you reply, you think there are little people who creep into houses
at night and do the housework. Oh no, say I, I do not believe that hob-
goblins do that. Actually , I have no beliefs at all about what hobgoblins
do Or what they are like . I just believe in them . (Compare contempo-
rary believers in God.)
Cartwright and Hacking, despite their entity-realist rhetoric, do not
actually adopt such a position. Their belief in some entity X always
involves as well belief in some X-theory, in Cartwright's case a theory
about what X's do, in Hacking's case a theory about what scientists
can do with X's. You cannot intervene in nature and do things with
electrons unless you believe (think true) all sorts of things about the prop-
erties and behaviour of electrons. Nor, despite what Devitt says, does
supposing merely that electrons exist enable us to "give good explana-
tions of the behaviour and characteristics of observed entities" (Devitt,
1991, p. 108). "Electrons exist", by itself, can explain nothing whatever
about the behaviour and characteristics of observed entities. To do any
explaining we need an electron-theory which says that electrons exist,
to be sure, but which says much more besides. And a successful electron-
REALISM, TRUTH AND OBJECTIVITY 21
theory will do more than get the existence of electrons right, will do more
than succeed in referring to electrons. For as Larry Laudan has cor-
rectly urged (1981, p. 25) , a theory may be referential yet false, and
because of that thoroughly unsuccessful. Reference is important for
realists because they think it is a necessary condition for success. But
it cannot be a sufficient condition - realists need truth as well .
It may be objected here that one can believe in X's and believe all
sorts of things about X's, without having a ' fully-fledged ' or 'global'
X-theory - and that it is only 'fully-fledged' or 'global ' X-theories that
we have no business believing. To which I reply that for me any col-
lection of beliefs about X's counts as an X-theory - though perhaps
not as a ' fully-fledged ' or ' global' X-theory, whatever that might be.
So you cannot soften mad-dog realism by harping on reference and
giving up on truth . What of mad-dog realism itself? What of the idea that
to be a scientific realist one must think current science true and current
theoretical entities real? (I omit the hedgings for simplicity.) This is
the latest manifestation of an old error: it erects current science into a
metaphysic and ties scientific realism too closely to that met aphysic.
Cartesians did it in the 17th century, Newtonians in the 18th century,
determinists in the 19th century. Theories which did not square with
the Cartesian, Newtonian, or deterministic metaphysic were deemed false ,
incomplete, or mere 'convenient fictions' . (Duhem rightly objected to the
practice, but wrongly thought that the only way to save science from
metaphysical interference was to adopt the view that convenient fictions
are all that science seeks.) And now our own mad-dog reali sts tell us that
you cannot be a realist unless you adopt the metaphysic of curved space-
time and quarks and gluons (or whatever). Realism thus defined might
be empirically refuted, as antirealists like Laudan are quick to see.
But scientific realism has survived the demise of Cartesian, Newtonian,
perhaps even deterministic metaphysics. And it could survive the demise
of some , most , even all, of current science. Realism would not be refuted
if curved space-time or quarks or gluons went the way of phlogiston
or the heavenly spheres. Or perhaps we should put it thus: epistemic
realism would survive, while substantive realism anno 1990 would not.
Do not misunderstand me. Realists think that science should inform
our metaphysical picture of the world, and I am no exception. Sensible
realists are so informed, and suppose that some scientific entities really
do exist and that some of what science tells us about them is true .
Again, I am no exception. You would have to be a pretty pessimistic
22 ALAN MUSGRAVE
realist to think that science has discovered no entities whatever and come
up with no truths whatever about them . Such pessimists may exist.
Popperians, gripped of all things by the pessimistic induction, sometimes
tell us that all scientific theories are false. This is a ludicrous proposi -
tion and it is backed by an argument which is as bad as an argument
can be: its premise ("All past theories are false ") is preposterous and
its conclusion ("All theories are false") does not follow.
Mad dog realists go to the other extreme: they are besotted with current
science and think that it has at last got to the bottom of things. But current
science includes the current frontiers of science, the places where the
current scientific action is, the places where scientists do not know and
are trying to find out. Scientists tend to be pretty cagey about entities and
theories at the frontier - realist philosophers of science would do well
to follow them in this. It is well-entrenched and well-tested past science
that should inform our metaphysic. We should be more confident about
atoms and molecules than we are about electrons, and more confident
about electrons than we are about quarks and gluons. Realism about
the entities and theories of current science should be rather guarded. And
whether guarded or not, it should not be seen as definitional of scien -
tific realism .
The maddest of the mad-dog realists go even further. They say that
only the entities of current science exist , or 'really exist' . When science
found out that tables and chairs were composed of atoms and mole-
cules, it found out that tables and chairs are unreal or do not really
exist. Remember Eddington's tale (1982, pp. xi-xiv) of the ' two tables',
the 'table of common sense ' and the 'table of physics', and his question
"Which table is the real table?" . Remember Wilfrid Sellars ' tale (1963,
p. 126) of the 'manifest' and the 'scientific ' images of the world and
his insistence that if the latter is true the objects of the former "do not
really exist - there are no such things". Eddington's mistake might be
called the 'explaining-is-explaining-away fallacy' . If science finds out
that atoms are composed of quarks and gluons, it will not have found
out that atoms do not really exist. If social science finds out that social
wholes can be explained individualistically, it will not have found out
that they do not exist.
Eddington's mistake is, curiously, the mirror-im age of the positivist
mistake. Positivists argued that since tables and chairs are real, atoms
and molecules (not to mention quarks and gluons) cannot be. Sensible
realists (lap-dog realists?) resist both mistakes. Tables and chairs are real,
REALISM , TRUTH AND OBJECTIVITY 23
and so (it is reasonable to suppose) are the atoms and molecules of which
they are made. Atoms and molecules are real, and so (it might be rea-
sonable to suppose) are the quarks and gluons of which they might be
made. Commonsense realism about tables and chairs neither precludes
(pace positivists) nor is precluded by (pace Eddington) scientific realism
about atoms and molecules.
This is not, of course , to say that science never conflicts with common
sense nor later science with earlier science. We think of a table as a
solid object. Perhaps (though I doubt it) this is to think that the space
occupied by a table is continuously filled with matter. If so, then the
atomic theory of matter conflicts with this commonsense belief, and if
we accept the former we must reject the latter. But this is to change a
belief about tables, not to start thinking that the 'table of common sense '
does not exist. (The 'table-of-common-sense' and the ' table-of-physics'
are ersatz entities dreamt-up by philosophers. Once you start taking them
seriously, nothing can stop you from sliding into conceptual idealism.
But I anticipate.)
Lap-dog realism differs from mad-dog realism in another respect that
I digress briefly to mention. Some lap-dog realists are suspicious of
platonic realism, realism about abstract entities. They would like to
find a way to drive a wedge between realism about some of the entities
posited by science and realism about the abstract entities posited by
mathematics and other disciplines. Again, some lap-dog realists have
sympathy with nonrealist accounts of morals. Can lap-dog realists be
discriminating in this way? In particular, does not a realist theory of truth
commit one willy-nilly to all kinds of realisms, some of them suspect
in the extreme? Which brings me to my next section .
2. TRUTH
I return to semantic realism (scientific theories are true or false state-

ments about the world) and axiological realism (science aims for true
theories). As I said, both are controversial. What theory of truth do realists
have in mind when they say that the aim of inquiry is truth?
It is obvious to me, and to others, that realism assumes some version
of the classical or objective or correspondence (or we might add, realist)
theory of truth. To say that the aim of inquiry is truth in any other sense
of the term 'truth' is not to advocate realism at all. For example, suppose
you go in for an 'empirical adequacy theory of truth'. Then to say that
24 ALAN MUSGRAVE
science aims for truth is to advocate, not realism, but Bas van Fraassen's
constructive empiricism. Or suppose you go in for some sort of prag-
matist or instrumentalist theory of truth (whatever that is). Then to say
that science aims for truth is to advocate, not realism, but some sort of
pragmatism or instrumentalism.
Arthur Fine sees the realism-antirealism dispute in a perverse way.
Both realists and antirealists think that truth is the aim. They disagree
only about the meaning to be attached to the term 'true': realists go
in for a realist conception of truth, antirealists for some antirealist
conception of truth . One can see the dispute this way - but it is not
the clearest way to see it. Nor do antirealists see themselves as giving
esoteric antirealist readings to the term 'true'. Van Fraassen would
not bar an 'empirical adequacy theory of truth ' . He is quite clear about
the difference between truth and empirical adequacy and his best anti-
realist arguments rely upon it. So let us set aside Fine's perversity and
say that realists assert and antirealists deny that truth is the aim.
The most fundamental objection to scientific reali sm concerns the
realist conception of truth. What is this conception and what are the
problems, or alleged problems, which beset it?
It is well to begin with Tarski, who thought that his semantic con-
ception of truth was a version of the classical correspondence theory.
Most philosophers think otherwise, as we will see. It is well to begin ,
in particular, with Tarski's famous requ irement that a materially adequate
definition of truth for a language L must entail, for each statement S
of L, a statement of the form
(T) The statement S is true in L if and only if p
where S names the statement of Land p translates that statement into

the language in which the truth-definition is formulated. This is the
famous T-scheme or equivalence thesis. It is well to begin with it, not
of course because it exhausts Tarski's semantic conception of truth, but
because it exhausts most of the philosophical worries about that con-
ception. Some philosophers think it trivial, others think it far from trivial.
I begin with the latter.
There are those who invest the T-scheme with magical powers. They
observe that it can be applied across the board, so to speak. Commonsense
realists apply it to commonsense statements. Scientific realists apply it
to scientific statements. But it can also be applied to statements about
numbers, moral statements, modal statements, or whatever. If the T-

scheme gives commonsense and scientific realists all they need from a
theory of truth, won't it give all sorts of suspect realists all they need
too? If commonsense instances of it yield realism about tables and chairs,
and scientific instances of it yield realism about genes and electrons, will
not arithmetic instances of it yield platonic realism about natural numbers,
and moral instances of it yield moral realism, and modal instances of
it yield modal realism?
This worry is quite groundless. You might as well worry that since the
T-scheme can be applied to "Margaret Thatcher gives me the creeps", we
must all be creeps-realists and believe in a mysterious entity (the creeps)
which Margaret Thatcher gives to me and simultaneously to others too.
The T-scheme by itself yields no reali st claim. We all avoid creeps-
realism by saying that although "Margaret Thatcher gives me the creeps"
is true (which it is), it is an idiom which is not to be taken at face-
value for logico-philosophical purposes. We replace it with an equivalent
non-idiom (say, "Margaret Thatcher makes me nervous") and avoid onto-
log ical commitment to the creeps. Similarly, one sceptical of moral
realism might refuse to take "Eating people is wrong" at face-value for
logico-philosophical purposes - which is just what emotivists, prescrip-
tivists , and the like do. Alternatively, one can go for an error-theory: take
moral statements at face-value and say that they are all false. This is what
John Mackie and my colleague Charles Pigden do to avoid moral realism.
Hartry Field does a similar thing to avoid arithmetical platonism: he takes
"Two plus two equals four" at face-value and says that it is false , not
because two and two add up to some other number than four, but because
there are no numbers for the numerals ' two ' and 'four' to be names
of. Similarly with scientific realism : some anti realists (the instrumen-
talists) refuse to take scientific statements at face-value, other antirealists
(the positivists) take them at face-value and say that they are all false
because there are no theoretical entities. Sim ilarly, finally, with com-
monsense realism : idealists either refuse to take commonsense statements
at face-value or take them at face-value and say they are false because
there are no tables and chairs.
In fact, antirealists usually go in for a mixture of these strategies. They
first tell you that a certain kind of statement is false if taken at face-value
or realistically. They then soften that conclusion by telling you what such
statements "really mean", that is, how they are to be taken for logico-
philosophical purposes. Think of phenomenalist 'translations', so-called,
26 ALAN MUSGRAVE
of statements about physical objects, or of emotivist 'translations' , so-

called, of moral statements.
So the idea that the T-scheme by itself has magical powers and begs
all kinds of metaphysical questions is quite mistaken. But if it does not
by itself yield realism about tables and chairs, electrons, numbers, moral
qualities, the creeps, or anything else, what is its importance for realism ?
The answer is that it makes realism about all these things possible. This
is its importance for realism, this is why it is the theory of truth that
realists need. Antirealist conceptions of truth make realism impossible,
they leave no room for it. Realists need the T-scheme - but they also
need much more. To be a realist about X's (whatever X's might be)
you must take statements about X's at face-value or realistically, apply
the T-scheme to them, and accept some of them (appropriate ones: "There
are no X's" will not do) as true.
So much for those who invest the T-scheme with magical powers. Most
philosophers have quite the opposite view - that it is trivial. Some think
it trivial because circular: the same statement appears on each side .
This is a misunderstanding: on the left is a statement about a statement,
on the right a statement about something else. The T-scheme is called
a 'disquotational' scheme and is said to encapsulate a 'disquotational'
or 'deflationary' or 'redundancy' view of truth . But an instance of T
will not be 'disquotational ' if something other than a quotation-mark
name of a statement appears on the left-hand side. And 'disquotatiori'
will result in nonsense if L is a different language from the one in
which the instance of T is formulated. (Consider: "The statement 'La
neige est blanche' is true in French if and only if la neige est blanche".)
Again it is said that for Tarski sentences construed as meaningless strings
of symbols are the truth -bearers. But as the adjective 'semantic ' and
the reference to translation make plain, the truth-bearers are meaningful
sentences or statements. To set up an instance of T you must know
what the referring terms in S refer to, and you must know what the
predicates applied to those terms mean in order to translate them into
the language you are speaking.
It is true that some instances of the T-scheme are 'disquotational' , like:
(I) The statement "Snow is white" is true in English if and only
if snow is white.
Confining himself to such instances, Alan Chalmers once worried that
the T-scheme is trivial in that it merely describes how one bit of language
relates to another bit of language, not how language relates to the world.
And he suggested that some notion of truth-as-correspondence must be
added to Tarski's theory to fill the gap. (Chalmers, 1976, p. 121. In the
second edition of his book Chalmers corrected this mistake.)
This is a queer worry. Instan ces of the T-scheme do relate language
to the world. True , if one is to speak about the way language relates to
the world, one must use language. But this is no deep truth - rather, it
is a pallid truism . Sweeney had it right:
I gotta use words when I talk to you

But if you understand or if you don 't
That's nothing to me and nothing to you
We gotta do what we gotta do.
(T. S. Eliot, 1963, p. 135).
The fact that "I gotta use words when I talk to you" does not mean
that I am trapped inside language and can only ever talk about words .
To think otherwise is to ignore the hard-won philosophical distinction
between using a word and mentioning it. (Wittgenstein ignored this dis-
tinction in his Tractatus and concluded that the way language relates
to the world cannot be 'said' but can only be ' shown '. Wittgenstein's
' Iogocentric predicament' is actually old psychologistic wine poured into
new linguistic bottles. The British empiricists thought that thinking
consists in having a stream of 'ideas', and concluded that all we ever
think about are our own ideas.)
What of Chalmer's idea that ' truth as correspondence' is needed to fill
some gap which Tarski leaves unfilled? Arthur Fine has a similar idea.
Tarski's semantic theory of truth is one thing and the correspondence
theory another. The correspondence theorist provides (or seeks to provide)
something which Tarski does not, a general account of the 'correspon-
dence relation ' between language and reality, a theory of the way the two
can 'match up ' so that truth results. Fine says that such a theory "would
explain what makes the truth true" (Fine 1984a, p. 97). Armed with an
account of the relation which all truths bear to reality, the correspondence
theorist will know what all truths have in common, their essence, what
makes them a natural kind. And here Fine gets sceptical: he doubts that
truths form a natural kind (Fine, 1984b, p. 56).
I share the scepticism. So, I believe, would Tarski. For Tarski it is
meaningful linguistic items that are true or false. Languages are con-
28 ALAN MUSGRAVE
ventional human devices, suited to different purposes. The truths they

contain are many and various, and so are the ways they correspond to
reality. Why suppose that they share an essence and form a natural
kind? It is surely better to suppose that Tarski captures 'the correspon-
dence relation' as well as it can be captured, that there is no more to
that relation than Tarksi gives us. As Michael Levin says:
Tar ski tells us that all true conjunctions have in common the truth of each conjunct,
that each true ex istential general isation is such that its matrix is satisfied by at least one
sequence, and so on. To be sure, at the level of the basis clauses, the definition goe s
strongly extensional, but that is the way it ought 10 go. The truth s " Ron Reagan is a
man" and "This tulip is red" are not shown 10 have much in common: each is a mailer
of a different object satisfying a different open sentence . BUIdo they have more in common
than being a man has in common with being red, which is 10 say, very lillie ? I cannot
see that they do, or at least more than Tarski gives them (Levin , 1984, p. 126).
On this view, one thing Tarski taught us is that the essence of truth is
a chimera, that an essentialist correspondence theory is out. Or if we must
talk of essences, we might talk thus: Tarski gives us the essence of the
correspondence theory without giving truth an essence.
But it does not matter if we talk otherwise. Suppose that a corre-
spondence theory worth its salt must give truth an essence, must provide
an account of the correspondence relation which all truths bear to reality.
Then Tarski's theory of truth is not a correspondence theory. Indeed, in
view of all the dust philosophers have raised down the ages with the word
'correspondence ', only to complain afterwards that they cannot see,
perhaps we had better drop that word altogether in this context.
Tarski's theory of truth, or rather Tarski 's T-scheme, has been accused
of being trivial in yet another sense. It does not tell us which state-
ments are true and which are false, nor does it say anything about how
we might determine this. It is a non-epistemic conception of truth. So
it is, and so it should be. The questions "What is truth?" and "What is
true?" are not the same question at all. The trouble is, many suppose, that
a non-epistemic conception of truth lets in the sceptic. Work out your
epistemology, your way of sorting the true from the false . If truth is
non-epistemic, the sceptic can still ask whether what your epistemology
throws up is guaranteed to be true . To beat the sceptic we need an
epistemic concept of truth, we need to define truth as whatever our
epistemology throws up. And so we have a rash of epistemic truth-
theories.
My stalking-horse will be the 'ideal limit theory of truth' recently
proposed by Carl Matheson. According to this theory, "the truth is to

be equated with the limit of ideal scientific practice : what is true is just
perfect science pursued to its conclusion" (Matheson, 1989, p. 253).
The claim is not that "All truths would as a matter of fact be part of
ideal science ", a claim which a mad-dog realist convinced of the omnipo-
tence of scientific methods might endorse. Rather the claim is that this
is a matter of (conceptual) necessity: "It is a conceptual truth, that all
truths would be part of our ideal science" (Matheson, 1989, p. 253, note).
Matheson does not tell us what ' ideal science' or ' ideal scientific
practice ' is, nor does he justify his presumption that it will reach some
sort of 'limit'. But never mind. One thing is clear: whatever account is
given, it cannot be couched in terms of truth and falsehood . For to be
told that truth is what ideal scientific practice reaches in the limit, and
then to be told that ideal scientific practice is that which in the limit
reaches truth , is to be told nothing. If we are to have a non-trivial epis-
temic conception of truth, then the epistemology must be defined in
austere truth-free terms . For then we will be told something non-trivial
when we are told "This is what ideal scientific practice is - and truth
is by definition what it comes up with in the limit".
This opens the theory up to Moore's 'Open Question Test '. However
ideal scientific practice is to be defined, it remains an open question
whether what it yields in the limit is true. The sceptic does not contra-
dict herself when she says that it is possible that ideal scientific practice
pursued to its limit might issue in falsehoods.
Besides, the anti-sceptical virtues of episternic conceptions of truth are
entirely spurious . To recognize the truth I am supposed to recognise what
ideal scientific practice is and recognise that on some issue it has reached
it 'limit'. What some 'ideally rational inquirer' (is this God, I wonder?)
will be coherently believing at the end of time is just as inaccessible
to you or me as is truth-as-correspondence. In fact, of course , it is much
more inaccessible.
Consider, finally, Matheson's assertion that "It is a conceptual truth ,
that all truths would be part of our ideal science". What is the meaning
of the word ' truth' (as opposed to the word ' truths' ) in this assertion?
Applying the ideal limit conception of truth to it yields "It is a conceptual
part of ideal science, that all truths would be part of our ideal science".
The contention is, then , that this particular statement (the statement of
the ideal limit theory of truth) has been or will be thrown up by ideal
science pursued to its limit. I defy Matheson or anyone else to locate
30 ALAN MUSGRAVE
this statement in the writings of any reputable scientist, or to give some

reason for thinking that it will occur in the writings of any reputable
scientist of the future. No, the ideal limit theory of truth is not science
but philosophy - and bad philosophy to boot.
So much for epistemic conceptions of truth, of which the ideal limit
theory is my exemplar. Returning to Tarski's T-scheme, it has been
accused of triviality in yet another sense. It is said (by Davidson, Devitt,
and others) that the T-scheme is trivial in the sense that it is common
to all conceptions of truth . This is not true. According to the T-scheme ,
what makes the statement "Snow is white" true in English is the fact
that snow is white. According to the ideal limit theory of truth, what
makes it true is the fact that ideal science will subscribe to it in the
limit. According to the empirical adequacy theory of truth, what makes
it true is the fact that it is empirically adequate. According to the prag-
matist theory of truth, what makes it true is the fact that it is "expedient
in the way of belief" (as William James put it). According to the coher-
ence theory of truth, what makes it true is the fact that it figures in
some coherent system of beliefs.
What happens if we combine the T-scheme with some epistemic con-
ception of truth? To fix our ideas, let us apply both to the statement
that electrons exist. The former gives us:
(I) The statement "Electrons exist"is true in English if and only if
electrons exist.
The latter gives us:
(2) The statement "Electrons exist" is true in English if and only if
the statement "Electrons exist" satisfies epistemic condition E.
It does not matter what epistemic condition E is, except that to avoid
circularity it must not involve "Electrons exist" being true in Engli sh.
Notice that (2) asserts the (necessary?) equivalence of two statements
about the statement "Electrons exist". It says nothing about the world,
nothing 'metaphysical'. Hence Putnam talks of 'internal realism ' as
opposed to 'metaphysical realism' .
Now (I) and (2) yield, by transitivity of the bi-conditional:
(3) Electrons exist if and only if the statement "Electrons exist " satis-
fies epistemic condition E.
Surely (3) is, or might be, false . Electrons existing (or equivalently
"Electrons exist" being true in English) is no part of the specification
of E. However our best methods for seeking truth are to be specified,
they are not guaranteed to reach the truth. There is nothing incoherent
or unintelligible about the idea that our best methods might lead us to
think that electrons exist when they do not. At least, this is so unless
we suppose that a statement's satisfying condition E (which specifies
those methods) somehow brings it about that electrons exist. This
' internal realism' is, of course, a form of idealism . It conflicts with
what science teaches about electrons, for example, that they existed
before any sentient creatures able to theorise about them had evolved.
You cannot combine an epistemic truth-theory with the T-scheme and
retain realism about the world and its contents being (largely) indepen-
dent of our epistemic activities.
Michael Devitt, following Stephen Leeds , thinks the T-scheme trivial
in that T-schemes for a language are two-a-penny. Given a bunch of words
and a bunch of things, we can set up a great many ' reference-schemes'
between the words and the things. One reference-scheme says "The
English word 'snow' refers to snow ", another says "The English word
' snow' refers to grass ", and so forth . Each of these reference-schemes
yields a Tarskian T-scheme for the language in question. And each T-
scheme sati sfies the equivalence thesi s (Devitt, 1984 , pp . 30-31). But
a T-scheme which assigned grass is green to the English statement "Snow
is white" (the example is Devitt's) would not satisfy the equivalence
thesis (and would not, by the way, be 'disquotational' either). "Grass
is green ' does not translate "Snow is white", because the English word
'snow' does not refer to grass. It might have so referred, of course, in
which case the Engli sh language would have been different, the refer-
ence-scheme for English would have been different and the resulting
T-scheme for English would have been different too . But as things are
with the English language, ' snow' refers to snow, not to gras s.
Is this naive? Can we say confidently that 'snow ' refers to snow and
not to grass? Can we speak confidently (as I just did) of the reference-
scheme for English or any other language? Hilary Putnam thinks not.
He has a 'model-theoretic argument against realism ' (Putnam, 1978,
1980). It is this argument which probably lies behind the Leeds-Devitt
point of view just considered. Let us consider the argument.
The confident assumptions just questioned make the realist theory
of reference and truth radically non-episternic. Those assumptions make
room for the follow ing possibility: there might be a theory that is empir-
ically adequate, that satisfies any other epistemic desiderata we care to
impose (is consistent, simple, elegant, or whatever), and yet is false . Such
an ' ideal theory ' might be false either because its theoretical terms fail
32 ALAN MUSGRAVE
to refer, or because it makes false claims about its theoretical entities.

Either way, its falsity is empirically undetectable, since by assumption
it is an empirically adequate theory. Putnam calls the view that an ' ideal
theory' might be false 'metaphysical realism' . And he argues that meta-
physical realism is mistaken.
The argument is this. Since an ideal theory will be consistent, it will
(by the Lowenheim-Skolem Theorem) have a model, an interpretation
in which it is true, whose domain is the natural numbers. We can map
the natural numbers onto objects in the world, and use this mapping to
define an interpretation in the world in which the theory is true. So the
theory is true of the world. So metaphysical realism is mistaken.
The obvious objection to this is that the interpretation which Putnam
has defined is not the intended interpretation of the ideal theory, that
the terms of that theory do not refer to whatever they are made to refer
to in Putnam's real world model. The arithmetical model, and the real
world model induced by the mapping, serve to establish the consis-
tency of the ideal theory but not its truth .
Putnam is unimpressed with this obvious objection . He finds the notion
of an ' intended interpretation' fishy, because he find s any appeal to the
'intended' or real referent of a term fishy . Any attempt to spell out
what the referents of our terms are is "just more theory". It can be
added to the original theory, in the form of constraints on the interpre-
tation of that theory, and the entire argument can be applied to the
augmented theory. This will yield an interpretation in the world in which
the original theory T and any constraints C on the interpretation of T
are both true . This will be the 'intended interpretation ' of the theory,
or as near to its 'intended interpretation' as we can ever get.
The crucial flaw in the argument is that this is not the case. As David
Lewis points out "The constraint is not that an intended interpretation
must somehow make our account of C come true. The constraint is
that an intended interpretation must conform to C itself" (Lewis , 1984,
p. 225). In other words, an interpretation of T which also makes C
come out true, which is all that Putnam's argument provides, is not an
interpretation of T which conforms to C. This disposes of Putnam's
paradox.
Lewis 's point seems subtle but it is really quite obvious, as a simple
little example will show. Let T be "Ravens are white" and let the actual
world W* contain just five objects, a white swam (i.), a black raven
(i2), a chair (i .), a table (i4)' and a cat (is). You might think that T is
false of W*, since w* contains a non-white raven i2• Enter Putnam. Being
consistent, T has a model in the natural numbers - actually, it has a model
in the first five natural numbers. Let interpretation I assign prime numbers
greater than two to the term ' raven' and odd numbers to the term 'white'.
Clearly, "Ravens are white " is true in this interpretation. Now define a
mapping between the first five natural numbers and the objects in W*,
which takes n to in' and through this mapping an interpretation 1* of T
in W*. Clearly , T is true of w* under interpretation 1* .
We protest that 1* is not the intended interpretation of T. We protest,
in particular, that the term 'raven' refers to i2• Putnam replies that this
constraint C is just more theory and invites us to add it to T, yielding
"Ravens are white and the term 'raven' refers to i,". This augmented
theory is consistent, so there is an interpretation in the natural numbers
(actually in the first five of them) in which it is true. Let this interpre-
tation be as before for the terms 'raven' and 'white' ; let it assign 1 to
the term "the term 'raven' ", 5 to the term 'i 2 ' , and let the pairs of numbers
1 and 3 and I and 5 stand in the relation 'refers to'. As before, use the
mapping between the first five numbers and the objects in w* to define
an interpretation in w * in which the augmented theory T&C is true.
Now the crucial point is that this interpretation does not conform to
constraint C. Constraint C requires that the term 'raven' refer to i2• In
the interpretation Putnam defines, the term 'raven' does not refer to i 2•
The constraint is true in that interpretation, but the interpretation does
not conform to the constraint.
So far we have had Putnam allowing us to formulate constraints on
the interpretation of theories and arguing (invalidly) that they do not help
the 'metaphysical realist'. But the position to which Putnam is led by this
argument does not allow us such constraints in the first place. He says
that statements like "The term t refers to the object 0" presuppose a
'magical ' theory of reference which sticks words onto things with some
sort of 'cosmic glue'. He says that we cannot 'single out ' any sort of cor-
respondence between words and things: to do so we would need
independent access to the things, which we do not have. For ''''objects
do not exist independently of our conceptual schemes. We cut the world
up into objects when we introduce one or another possible scheme of
description" (Putnam, 1981, p. 51). Putnam calls this 'internal realism'.
It is an odd doctrine. Consider the Putnamesque possibility of two
'ideal theories ' which give utterly different and mutually incompatible
accounts of the way the world is. (One might be a particle theory, the
34 ALAN MUSGRAVE
other a field theory - the details do not matter.) Having abandoned

metaphysical realism, Putnam thinks both theories are true of the world .
But utterly different and mutually incompatible theories can both be
true of the world only if the notion of truth is relativised to each of
them - which relativises the notion of 'the world' to theory also. There
is no real world independent of our theorising - or if there is we have
nothing to do with it. The world we theorise about is a world-as-theo-
rised-about-by-us - and there may be more than one of them. This is
not realism of any kind - it is idealism. It spells trouble not just for
esoteric doctrines like Putnam's metaphysical realism - it makes nonsense
of scientific realism and commonsense realism too.
Before I argue this, let me say what fun it is to see the antirealists
squabbling amongst themselves. Van Fraassen's chief argument against
realism is that realists, being truth-freaks, will want to know which of
two incompatible 'ideal theories' is true. Evidence cannot help them
(being 'ideal' the theories will be empirically equivalent and empiri-
cally adequate). So realists must choose on non-evidential or meta-
physical grounds. Realism lets in metaphysics, metaphysics is bad news,
so realism is bad news. I will not discuss this argument now (I have
done so elsewhere). I mention it only to observe that if Putnam's model-
theoretic argument is correct, van Fraassen's argument cannot get off
the ground. Any empirically adequate theory might be deemed 'ideal' .
Any 'ideal theory' is, by Putnam's argument, true. So the realist does
not have to choose between two ideal theories at all, but can accept
them both. But only at a very heavy antirealist price, as we will see in
the final section of my paper.
3. OBJECTIVITY
A Dutch sociologist (Gerard de Vries) once told me in all seriousness

that when the concept 'person with an IQ two standard deviations above
the mean' was invented new entities were brought into being. So it
turns out (I replied) that there are two ways of making babies, the way
we all know and love, namely love, and this new way, psychological
theorizing! I was told not to be silly, that the new entities were not babies.
I tried to find out what they were but I did not succeed.
My ex-friend the Dutch sociologist was a conceptual idealist. And
conceptual idealism, a view which I find quite absurd , is arguably the
dominant metaphysical position of our age among prominent philo so-
REALISM, TR UTH AND OBJECTIVITY 35
phers. Putnam's 'internal realism ' is a species of it. And there are many
other species of it. Remember all that guff about 'the social con-
struction of reality' . Recall Kuhn 's suggestion that scientists imbued
with different paradigms 'inhabit different worlds' . Think on Nelson
Goodman 's ' ways of world-making' and his extraordinary contention that
"We can have words without a world, but no world without words or
other symbols" . Think too on Richard Rorty's 'the world well lost' .
Consider Simon Blackburn's view that we can ' project' , "speak and think
as though there were a property of things" when there is no such property,
and yet make no mistake (Blackburn, 1984, p. 171). An extraordinary
state of affairs!
Conceptual idealism was ushered into the world by Kant , who dis-
tinguished between the phenomenal and noumenal 'worlds' , between
things-in-themselves and things-as-experienced-by-us. There is the Moon-
in-itself (call it Moon .), about which we know nothing, and there is the
Moon-as-experienced-by-us (call it Moone)' Moon; is not the same object
as Moone - if it were the distinction would have no point and we could
rest content just with the Moon , without subscript, unhyphenated. For
example, Moon; is not located in space and time, these being ' forms of
sensibility' in which only Moone is located. Nor does Moon ; cause or
help cause moon-experiences in humans down on earth, causality being
a ' category of the understanding' which applies only in the phenom-
enal world. Some of Kant's immediate followers realised that Moon; is
nowhere, at no time, and does nothing, concluded that it was an idle meta-
physical posit , and did away with it altogether to become fully-fledged
idealists. I do not blame them. My Kantian friends tell me that I mis-
understand, that Kant was an ' empirical realist', and not an idealist at
all but only a 'transcendental idealist' . And I am reminded of how fond
Bishop Berkeley was of presenting himself as a defender of common-
sense realism.
Kant, like Berkeley before him, thought that there was only one 'phe-
nomenal world' : he assumed that humans all have the same immutable
set of basic concepts with which they structure incoming stimuli.
Contemporary philosophical wisdom has outgrown that assumption.
Modern conceptual idealism is Kantian idealism relativised to concep-
tual or linguistic scheme . Concepts and languages vary and change. There
is not a unique 'phenomenal world ' or world-as-conceived-by-humans at
all. The world-as-conceived-by-the Aristotelian differs radically from the
world-as-conceived-by-the-Newtonian. The world -of-the-Eskimo is not
36 ALAN MUSGRAVE
at all the world-of-the-Kalahari-bushman. This gets really exciting if

we drop human chauvinism and consider non-human animals too: the
world-of-the-chimpanzee is not at all the world -of-Albert-Einstein, and
both are worlds apart from the world-of-the-honeybee. (Does Kant think
that non-human animals have experiences? If he does, does he think
that they possess all those categories of the understanding deployment
of which is a condition of the possibility of all experience? It is pretty
implausible to suppose that chimps and honeybees and flatworms struc-
ture incoming stimuli the same way humans do . It is even more
implausible to suppose that they have no experiences at all. The upshot
seems to be that chimps can have experiences without deploying the
Kantian categories. But if chimps can do without the Kantian cate-
gories, why can't we do without them too?)
Of course, all this idealist talk of different worlds -as-experienced (or
conce ived or talked about)-by-X 's need not be taken seriously. We can
see it just as a fancy way of drawing attention to the diversity of expe-
rience or concepts or talk of the world . On this view, hyphenated entities
like the Moon-as-experienced-by-humans are all ersatz entities. The
Moon-as-experienced-by-humans is just the Moon - and similarly for
all other hyphenated entities, including the Kantian Moon-in-itself. And
a profundity such as "The Moon-as-conceived-of-by-Aristotelians was
perfectly spherical" is just philosopher's gobbledy-gook for "Aristotelians
thought that the Moon was perfectly spherical".
Conceptual idealists do take seriously talk of different world s-as-expe-
rienced (or conceived or talked about)-by-X's. Given their metaphysic,
the realist idea that our experience or thought or talk might correspond
with a reality independent of experience or thought or talk is seen to
be thoroughly naive. The things we experience or think about or talk
about are partly of our own experiential or conceptual or linguistic
making. What arguments lead to this extraordinary metaphysic?
To begin with there is the idea that what I can see (or say) depends
partly on the concepts or words that I possess. Now it is trivially true
that a being lacking the typewriter-concept (the word ' typewriter ' ) cannot
say that there is a typewriter on the table. Assuming that a being which
cannot say that P cannot see that P either (an assumption I will not
here dispute), it follows that a being lacking the typewriter-concept (the
word ' typewriter ') cannot see that there is a typewriter on the table either.
But it does not follow that such a being cannot see the typewriter on
the table. The Kalahari bushman may see the typewriter perfectly well,
as evidenced by his response to the request (couched in Kalahari-

bushman-ese) to pass us that thing over there. The cat may see the
typewriter perfectly well too, as evidenced by her not bumping into it
when the mouse she is chasing hides under it. This argument wrongly
conflates seeing-that with seeing. (In between these there is also seeing-
as: the bushman or the cat may see the mouse as food , and the typewriter
as non-food , even though they lack the mouse-concept and the typewriter-
concept.)
[Digression: It is not, in fact, "trivially true that a being lacking the
word ' typewriter ' cannot say that there is a typewriter on the table". A
French-speaker who knows no English and who therefore lacks the
word ' typewriter ' can say this - in French. Such a being lacks the word
'typewriter' but possesses the typewriter-concept, since French has a
word for typewriters too. I ignore this complication, and the knotty
problems involved in it, as being tangential to my main concerns. One
can also dispute the assumption which for simplicity I granted just now,
the assumption that a being which cannot say that P cannot see that P
either.]
The conflation mentioned a paragraph back between seeing-that and
seeing leads to the view that beings possessed of different 'conceptual
schemes ' or languages or theories literally see different worlds. This is
nonsense. The Aristotelian and the Copernican, watching the sunrise, see
the same thing. What might be true is that the Aristotelian says of the
sunrise "I see that the sun is still orbiting the earth", while the Copernican
says "I see that the earth is still rotating on its axis". The profundity "The
limits of my language are the limits of my world" is false. What is true
is the triviality that my language limits what I can say of the world.
Or consider the anthropological profundity "The world of the Kalahari
bushman contains no typewriters". Is this the hidden platitude that,
Kalahari language or concepts being what they are, the bushman cannot
see or say that a typewriter is on a table? Or is it the gratuitous paradox
that typewriters fail to interact causally with the visual system of the
Kalahari bushman? Since when did reflected light get to be so discrim-
inating or the bushman 's eyesight so weak? Or is it the even sillier
view that the (assumed) absence of some concept or word makes certain
physical objects dis appear? What magical causal powers are being
invoked here?
Arthur Fine thinks that the realist 'metaphysical picture ' of an objec-
tive, independent, external world is misguided. It is misguided, he says,
38 ALAN MUSGRAVE
because of "the obscurity of the correspondence relation and the

inscrutability of realist-style reference". He elaborates:
The problem is one of access . The correspondence relation would map true statements
(let us say) to states of affairs (let us say). But if we want to compare a statement with
its corresponding state of affairs, how do we proceed? How do we get at a state of
affairs when that is to be understood, realist-style, as a feature of the World ? . . . A
similar question comes up if we move to reference .. . , for there again what the realist
needs by way of the referent for a term is some entity in the World The difficulty is
that whatever we observe, or, more generously, whatever we causally interact with, is
certa inly not independent of us. This is the problem of reciprocity. Moreover, whatever
information we retrieve from such interaction is . .. information about interacted-with
things. This is the problem of contamination How then, faced with reciproc ity and con-
tamination, can one get entities both independence and objective? Clearly, the realist
has no direct access to his World . . . (Fine , 1986, p. 151).
What exactly is the problem here? Suppose somebody says "The Moon
is full tonight" and I look up into the night-sky and ascertain that the
statement is true . (I use a humdrum commensical example rather than
an esoteric scientific one because if there is a problem here it is a quite
general one which will afflict the commonsense realist metaphysic just
as much as the scientific realist metaphysic.) I have access to both terms
of the so-called 'correspondence relation' : my linguistic competence
gives me access to what was said, my eyes give me access to the Moon
out there in the world . Fine does not think that we lack either linguistic
competence or sensory awareness. Nor is his worry the traditional scep-
tical one - that my perceptual belief that the Moon is full is fallible,
because I might have been the victim of some illusion or hallucination.
No, Fine's worry is that reciprocity and contamination mean that I have
no access to the Moon at all.
Reciprocity was supposed to show that the Moon is not independent
of us because we can see it or otherwise causally interact with it. But
implicit in this is a silly account of independence: an object will be
independent of us only if we cannot see it or otherwise causally interact
with it. The only independent entities in this sense will be platonic
entities, which do not exist in space and time and which have no rela-
tions causal or otherwise with beings like us which do exist in space
and time . No realist, commonsense or scientific, thinks that the only inde-
pendent entities are platonic entities and the only independent reality
the platonic realm. When a commonsense or scientific realist says that
the Moon is (largely) independent of us, she means that it is non -mental,
continues to exist when we are not looking at it, existed long before
we did, and so forth.
What of Fine's 'problem of contamination'? That was supposed
to show that when we see that the Moon is full, we gain information
about an interacted-with-Moon, not about an objective Moon out there
in the World. Is this interacted-with-Moon then subjective and inside
our heads, like the ' Moonish-sense-datum' that was once thought to be
the immediate object of perception? Realists are perceptual realists and
have rightly discarded that view: perceiving external objects may involve
us in having perceptual experiences, but it is the external objects that
we perceive nonetheless. (Hallucinators, who are commonly said to be
'seeing things ', do not see things at all, they just may mistakenly think
that they do!)
Fine may not be seeking to revive the theory of sense-data - but he
arrives at something pretty close to it. His interacted-with-moon, his
Moon-as-observed-by-humans, is not 'objective' in the sense that it is
somehow partly constituted by the Moon-concept which is our invention.
(The Moon-in-itself is, of course, no better off in this regard!) Like
Putnam, Fine thinks that objects do not exist independently of concep-
tual schemes, that we carve up the world into objects when we introduce
a conceptual scheme, and that having carved-up the world we cannot
partake of it. (Evidently, the world is not like the Sunday joint!) The
Moon which we observe, talk about, theorise about, is an object that
we create by observing or talking or theorising. And this is said, believe
it or not, to be part and parcel of a ' natural ontological attitude', which
accepts the results of science as true on a par with 'homely' common-
sense truths. Well, some pretty mundane and well-entrenched bits of
science tell us that the Moon (not some hyphenated Moon, not even
the Kantian Moon-in-itself, just the Moon) is objective and indepen-
dent of us: it existed long before we did, was not created by us , and
so forth . Someone with a natural ontological attitude, someone who
accepts these bits of science as true, will have precisely the realist 'meta-
physical picture ' which Fine rejects.
Although he does not know it, Fine is heir to a long and distin-
guished philosophical - that is, idealist - tradition. The giveaway is
the hyphenated entity, the ' interacted-with-thing'. Fine implicitly argues
as follows: "We cannot find out about things without interacting with
them. Therefore, we cannot find out about things-as-they-are-in-them-
selves". What can we find out about, then? Why, interacted-with-things,
40 ALAN MUSGRAVE
of course. The argument is invalid . It is one of the countless versions

of what David Stove (1991) calls 'The Gem', an argument which has
converted most philosophers to idealism from Berkeley to the present
day. The general form of The Gem is: "We can X things only if condi-
tion C, which is necessary for us to X things, is met. Therefore, we cannot
X things-as-they-are-in-themselves." For X you can substitute ' know' ,
' perceive' , 'think of ', 'talk of', 'refer to', and so forth . For the neces-
sary condition C you can substitute Berkeley 's 'have in mind' , Kant's
'bring under the categories of the understanding ', Fine's 'interact with',
and so forth . Gems, implicit and explicit, take many forms.
Conceptual idealism takes many forms, too. John Mackie had an 'error
theory' of moral judgements, according to which we 'project' moral prop-
erties onto the world. Simon Blackburn liked the idea of 'projection', but
did not like the idea that it involves us in error. So he decided to square
the circle, and say that we can 'project' things onto a world which lacks
those things and make no mistake . The project of projecting things has
recently been extended to science, and conceptual idealism is again the
result .
Richard Jennings does not like the scientific realist's view that "there
are in the world theoretical (i.e. unobservable) properties and entities
which are independent of our theories (and thus do not change when
our theories change)" (1989, p. 243). Instead, we are to have 'scien-
tific quasi-realism ' : "theoretical properties and entities are projected onto
the world, they are the offspring of our theories . . . they are things in
the projected world rather than in some inaccessible world with which
we are utterly unacquainted" (1989, pp. 235-6).
Once upon a time many people believed in witches and subscribed
to a body of theory (call it witch-theory) about them - they projected
the predicate "is a witch" onto the world. For a while witch-theorists were
quite successful: suspected witches were identified, tested for witch-hood
in accepted ways, and when they tested positive disposed of in accepted
ways too. Despite these 'successes' , realists think that witch-theory was
false and that there were not really any witches. But quasi-realists think
this metaphysically naive: given a consistent witch-theory, there is a 'pro-
jected world' of which that theory is true; so there were witches after
all, not inaccessible noumenal witches perhaps, but projected witches. Or
as Alan Chalmers' once put it (1976: p. 131): "In medieval Europe,
witches really inhabited the common-sense world while in modern times
they do not .. .".
RE ALISM , TRUTH AND OBJECTIVITY 41
A sixteenth-century witch hunter prone to bang the table and insist

that there really are witches, would not thank Jennings for telling him
that what he believes is true, but only of 'projected witches' in some ' pro-
jected world'. For the witch-hunter's project was to find the real witches
in the real world . Contrary to what Jennings says, no ordinary sense of
the term 'true' is involved in quasi-realism. for quasi-realist truth does
not seem to admit of an opposite. Suppose the witch-hunter armed with
his consistent witch-theory confronts an opponent armed with a consis-
tent no-witch-theory. Is there a truth of the matter, as plain folk think?
Jennings thinks not: both theories are true , but of different 'projected
worlds': ". .. theories with which we no longer agree ... are seen as
having projected different properties and entities onto the world , and
based on their projections, they spoke truly. We now, of course , project
differently . . ." (1989, p. 239) . The realist question "But are there any
real witches in the real world ?" is naive and the real world it asks about
is inaccessible to us. "There are witches" was true of the projected
world of the witch-hunter, "There are no witches" is true of the pro-
jected world of today. And that is all we can say.
Jennings claims that his "quasi-realist account of realist talk offers a
more reasonable account of scientific talk than the realist account" (1989:
p. 239). Stuff and nonsense. Science has discovered that there are in
the world no witches , heavenly spheres, phlogiston-rich substances, or
caloric fluids. (These negative discoveries stand no matter how scep-
tical we may be about whether science has any positive discoveries to
its credit.) An account of " scientific talk" which denies that science
has these negative discoveries to its credit is quite unreasonable.
As for the quasi-reali st metaphysic itself, think of the extraordinary
claim that " the theoretical things we talk about in our science are the
products of our talk" (1989, p. 238) . So our talk produce s electrons, black
holes , genes, or whatever. From which it follows (assuming that the word
' produce' means what it ordinarily means) that there were no electrons,
genes , black holes or whatever before people started talking of them. This
flies in the face of everything science teaches us about such entities.
Science teaches us, for example, that black holes (if there are any) are
located far out in space and cannot be produced by anyone's thought
or talk down here on earth. If the quasi-realist metaphysic is true, then
science is all false . Scientific entit ies are not ' projected entities', because
projected entities only "seem to belong to the outside world" (1989,
p. 245) but do not really belong there. Projected entities are mental or
42 ALAN MUSGRAVE
linguistic entities, because mere thought or talk can produce them. Quasi-
realism is just conceptual idealism, and conceptual idealism is just Bishop
Berkeley 's metaphysic dressed-up. We should be grateful to Jennings
for making this abundantly clear.
The ideal limit theory of truth, discussed in my previous section as
an exemplar of epistemic truth-theories, also has relativistic and ideal-
istic implications. Why suppose that ideal scientific practice will converge
upon a single limit? Might it not converge upon two different and
seemingly incompatible limits, sayan ideal particle theory and an ideal
field theory? Might not Martian science pursued to its ideal limit be
very different from our science pursued to its ideal limit? Matheson
canvasses both possibilities and embraces them: "Thus the ideal limit
theory might better be called the ideal limits theory. According to it, your
community's limit might not be my community 's limit; what is true for
you might not be true for me. In allowing for the possibility of relativism,
the ideal limit theory contradicts global realism" (1989, pp. 253-4).
Relativism and idealism go hand-in-hand. Seemingly incompatible ideal
limit theories are, according to the ideal limit theory of truth , both true.
Incompatible theories cannot both be true of the world. So each must
be true of its own world, each must be true of an ersatz world-as-it-is-
according-to-my-community's-ideal-theory. Matheson sees and embrace s
the relativistic consequence of his theory of truth - he does not see the
idealist consequence.
Brian Ellis sees neither consequence very clearly. He calls himself
in 'internal realist' and embraces something like the ideal limit theory
of truth - "truth is a kind of limit notion of reasonable belief" (1985:
p. 68). He keeps saying that an internal realist can believe that reality
is independent of what we think or say. But in the end the pressure
tells and we have ersatz or hyphenated entities again: "The way the world
is relative to the sorts of beings we are. That is one of the consequences
of internal realism . . . according to the internal realist, there is no way
that the world is absolutely, only ways in which it is relative to various
kinds of beings" (1985, p. 71). Here the hyphenated entities are the-
way-the-world-is (noumenal) and the-way-the-world-is-relative-to-us-
humans (phenomenal). Ellis finally gives up the realist ghost: "There
is not and cannot be any absolute truth, and therefore there cannot be any
way that the world is independently of how we, or some other kind
of creature, would evaluate its beliefs about it" (1985, p. 72). Goodness
me, science teaches (speculation about God and aliens aside) that the
world existed long before sentient creatures did . Is there no way that a
world without sentient creatures is? Did the world have to wait for
sentient creatures to evolve before it had any objects in it or any struc-
ture ? How, in an objectless and structureless world , did sentient creatures
evolve?
University of Otago ,
New Zealand.
NOTE
* An earlier version of this paper was read at a conference on Objectivity held at La

Trobe Univers ity, Melbourne , Australia, on 20-22 April 1990.
BIBLIOGRAPHY
Blackburn, S. (1984). Spreading the Word, Clarendon Press, Oxford.

Cartwright, N. (1983). How the Laws of Physics Lie, Oxford University Press.
Chalmers, A. F. (1976) . What is This Thing Called Science?, University of Queensland
Press.
Devitt, M. (1984). Realism and Truth. Basil Blackwell (Second edition 1992).
Eddington, Sir A. S. (1928) . The Nature of the Physical World, Cambridge University
Press.
Eliot, T. S. (1963). Collected Poems 1909-1962, Faber and Faber.
Ellis, B. (1985). ' What Science Aims To Do', in P. M. Churchland and C. Hooker (eds.),
Images of Science, University of Chicago Press, pp. 48-74.
Fine, A. (1984a). ' The Natural Ontological Attitude' , in J. Leplin (ed.), 1984, pp. 83-
107.
Fine, A. (1984b). ' And Not Anti-Realism Either' , Nous 18, pp. 51-65 ,
Fine, A. (1986). 'Unnatural Attitudes: Realist and Instrumentalist Attachments to Science ',
Mind 95, pp. 149-179.
Hacking, 1. (1983). Representing and Intervening , Cambridge University Press.
Jennings, R. (1989). 'Scientific Quasi-Realism' , Mind 98, pp. 225-245.
Laudan, L. (1981). 'A Confutation of Convergent Realism ', Philosophy of Science 48,
pp. 19-49 .
Leplin , J. (1982) . ' The Historical Objection to Scientific Realism ', PSA 1982 1,
pp.88-97.
Leplin, J. (ed.) (1984). Scientific Realism , University of California Press.
Levin, M. (1984). ' What kind of explan ation is truth?', in J. Leplin (ed.), 1984,
pp. 124-139 .
Lewis, D. (1984). ' Putnam's Paradox ' , Australasian Journal of Philosophy 62, pp. 221-
236.
44 ALAN MUSGRAVE
Matheson, K. (1989). ' Is the Naturalist Really Naturally a Realist?', Mind 98 , pp. 247-
258.
Putnam, H. (1978). Meaning and the Moral Sciences, Routledge & Kegan Paul.
Putnam, H. (1980) . ' Models and Reality', Journal of Symbolic Logic 45, pp. 464-482.
Putnam, H. (1981). Reason, Truth and History, Cambridge University Press.
Sellars, W. (1986). Science, Perception and Reality, Routledge and Kegan Paul.
Stove, D. C. (1991). The Plato Cult and Other Philosophical Follies, Basil Blackwell,
Oxford.
ILKKA NIINILUOTO
QUERIES ABOUT INTERNAL REALISM
I. IS KUHN AN INTERNAL REALIST ?
In his Presidential Address at the 1990 biennial meeting of the Philoso-

phy of Science Association (PSA) in Minneapolis, Thomas S. Kuhn
announced his project of writing a philosophical book on realism and
truth. The audience included also members of the Society for Social
Studies of Science (4S). To the disappointment of many sociologists of
science, who have taken Kuhn as an authority in their campaign for
relativist and anti-realist views, Kuhn now argued that the concept of
incommensurability is in fact needed "to defend notions like truth and
knowledge from , for example, the excesses of post-modernist movements
like the strong program".' A "strong concept of truth" is needed, he
argued , to understand the "normal rules of evidence" for accepting or
rejecting theories. '
Kuhn thus seems to be a kind of 'realist' after all - to the surprise
of some or many of his admirers. But what kind of realist? Kuhn himself
note s that his account of the relationship "between the lexicon - the
shared taxonomy of a speech community - and the world the members
of that community jointly inhabit" clearly cannot be the one Hil ary
Putnam calls "metaphysical reali sm".' As Kuhn also follows Putnam
in giving up the correspondence theory of truth," it seems tempting to
conclude that Kuhn 's "Post-Darwinian Kantianism" might be a version
of Putnam's " internal realism".
The sugge stion that Kuhn's and Putnam 's positions have important
similarities is not a new one . In his insightful book, Repre senting and
Intervening (1983), Ian Hacking characterizes both of them as "tran-
scendental nominalists ".' This description fits very well Kuhn's new
account: the world itself is not mind-dependent, or our construction,
but its structure or " taxonomy" is relative to a shared conceptual system
or "lexicon" which changes with the passage from one community to
another.
However, Kuhn is wisely cautious, and does not call himself an
"internal realist". We shall see that it is indeed difficult to classify Kuhn's
45
R. S. Cohen , R. Hilpinen and Qiu Ren zong (eds.), Realism and Anti-Realism in the
46 ILKKA NIINILUOTO
position in this philosophical taxonomy. What is more, I shall argue

that Putnam 's 1976 distinction between the two rival kinds of realism
introduces a misleading dichotomy: a reasonable realist should accept
neither metaphysical nor internal realism in Putnam's sense."
2. PUTNAM ON INTERNAL REALISM
Perhaps Putnam's best formulation of the metaphysical-internal distinc-

tion is given in Reason, Truth and History,' The "externalist perspective"
(or "God's Eye point of view") of metaphysicaL realism (MR) includes
three theses:
(MRl) The world consists of some fixed totality of mind-indepen-
dent objects .
(MR2) There is exactly one true and complete description of "the way
the world is".
(MR3) Truth involves some sort of correspondence relation between
words and external things or sets of things.
On the other hand, internaL realism (lR) accepts the negations of MR I ,

MR2 and MR3:
(Ik l) What objects does the world consist of? is a question that it
only makes sense to ask within a theory or description.
(IR2) There is more than one 'true' description of the world.
(lR3) Truth is some sort of idealized rational acceptability.
In 'Why there isn't a ready-made world' (Putnam, 1983), Putnam refor-
mulates or generalizes IR 1 to the form
(IR 1') The world is not ready-made,
which denies that the world has any built-in structure. In other words,
the world is carved or sliced into pieces (objects, essential properties,
facts, causal relations, etc.) only relative to descriptions, theories , or con-
ceptual schemes. In Raimo Tuomela's (1985) terms, IR I' denies the
ontological Myth of the Given.
While MR2 assumes that the whole truth about everything is uniquely
determined, IR2 asserts the plurality of truths.
QUERIES ABOUT INTERNAL REALISM 47
According to Putnam (1978), the existence of a fixed correspondence

between words and mind-independent things would imply that
(MR3') Truth is a radically non-epistemic notion.
Instead, IR3 claims that an epistemically "ideal" theory could not be
false ."
Putnam clearly thinks that the MR-theses and IR-theses go together
in natural "packages". It is of course possible to find, with some varia-
tions, more or less articulated formulations of MR (e.g., Wittgenstein's
Tractatus, where the world , as the totality of facts, is pictured by the
language) and IR (Putnam, 1981; Rescher, 1982; Tuomela, 1985).
However, it seems equally clear that theses MR 1, MR2 and MR3
(and thereby their negations IR1, IR2 and IR3) are logically indepen-
dent of each other. Therefore, they allow us to define, besides the 'pure'
doctrines MR and IR, six other 'mixed' positions.
For example, the version of "internal realism" defended by Brian
Ellis (1989) accepts an epistemic notion of truth (lR3), and the exis-
tence of various "epistemic perspectives" (lR2), but at least starts from
the assumption that the world has a categorial ontological structure
(MR1).
Before studying some of these mixed views, let us note that Putnam's
dichotomy allows a realist to advocate either the correspondence theory
of truth or the pragmatist theory of epistemic truth. It may be debated
whether the redundancy theory (or disquotational theory) is sufficient for
someone who wishes to be called a 'realist' ." But at least the nomi-
nalist Hartry Field (1982), who takes space-time points as the ultimate
individuals, suggests that a metaphysical realist could combine MR 1 with
the redundancy theory. Kuhn (1991) in turn combines the internalist
theses IR 1 and IR2 with the redundancy theory. When Kuhn states that
a conceptual scheme is not a "set of beliefs" - and that sharing such a
scheme involves a common structure, not its embodiment in the head
of an individual - I think he can be interpreted as denying that truth is
an epistemic concept in the sense of IR3. This means that Kuhn's position
cannot be classified as MR or IR.
3. THE WHOLE TRUTH ABOUT EVERYTHING
Field (1982) has argued that MR1 does not imply MR2 which "should
not be taken as a component of any sane version of realism". Putnam's
48 ILKKA NIINILUOTO
reply appeals to the following point: if by MRI the world consists of a

fixed set I or individuals and a set P of their properties and relations
(and nothing else), then there is an ideal language L which speaks of I
and P. Such a language L needs not to be denumerable. But the set Tr L
of true sentences of L, which is well-defined by MR3, is then the set
of all truths, as MR2 requires. 10 Another way of formulating this argument
is to say that, by MR I, the world is (or can be represented) by a set-
theoretical structure W = (I,P). If L is a language which gives a complete
description of W, then Tr L is the theory of W (in the model-theoretical
sense), i.e., Th(W) = {h in L I W 1= h}.
This debate seems to conflate different senses of 'exist' . Field is
right that we need not actually have a unique complete language L,
even if MRI holds . If the structure W = (I,P) is infinitely complex (e.g. ,
I contains all past, present, and future individuals, P contains n-order
predicates and n-place relations for all finite n), then L would be an infini-
tary language which need not to be finitely definable at all. Further,
the set of truths TrL in L may fail to be finitely or even recursively axiom-
atizable. To say with Putnam that such an ideal language "exists" is
indeed a very strong idealization.
Grim (1991) has recently argued, by using diagonalization methods of
set theory, that Tr L for a given language L does not constitute a set. This
is not quite conclusive, since the argument appeals to set-theoretical or
metalinguistic truths about TrL • Moreover, there are versions of set theory
which admit a universal set (i.e., the set of all sets). But at least Grim 's
claim adds a further obstacle to the derivation of MR2 from MRI.
But who then would be willing to accept the strong thesis MR2?
Somewhat surprisingly, one answer is Wilfrid Sellars (1968) whose
scientific realism starts from a criticism of the Myth of the Given (i.e.,
rejection of MRI) and concludes with the equation of truth as asserta-
bility relative to the ideal "Peirceish" conceptual system. In the Sellarsian
realism, the scientific community reaches in the asymptotic limit the
ultimate "God 's Eye perspective". This means that Sellars combines
the theses IR I , MR2 and IR3.
More generally, the combination of MR2 and IR3 is typical to most
versions of epistemic, pragmatist or coherence theories of truth: a
sentence is true if and only if it follows from, or coheres with, the set
of all true beliefs. II These theories presuppose the existence of an ideal
belief state which contains all truths about the world. If MR2 fails, then
these epistemic theories of truth collapse.
Putnam has recently clarified his earlier discussion of truth as ideal

acceptability (IR3) by emphasizing that the ideal state does not simul-
taneously concern any truth whatsoever - there is no single state of
justification that would entail all truths." This is important for Putnam,
for otherwise his IR3 would conflict with his IR2.
Peter Unger (1975) has claimed that our understanding of truth
assumes that a sentence is true if and only if it expresses the whole
truth about everything or else is a part of the whole truth. As the concept
of the whole truth about the world is inconsistent (i.e., Unger rejects
MR2), Unger concludes that truth is impossible.
Unger's argument fails, however, to shake such theories of truth which
do not rely on the fiction of the whole truth. Such accounts are given
by non-epistemic notions of truth, in particular by the correspondence
theory, which defines truth as a relation between a statement and the
world, not between a statement and the set of true beliefs.
In the next section, I defend the claim that the best version of realism
is neither "metaphysical" nor "internal" in Putnam's sense, but avoids
the problematic thesis MR2 by combining the correspondence theory
of truth (MR3) with the pluralist principle IR2 and with a refined version
of IRI which does not imply "conceptual idealism".
4. CORRESPONDENCE WITHOUT "METAPHYSICS "
A scientific realist accepts the minimal ontological assumption that

there is an actual world independent of human minds, concepts, beliefs
and interests. Let us call it THE WORLD. 13 If you wish, this assump-
tion may be called "metaphysical". But all the evidence of science assures
us that THE WORLD has existed long before the emergence of man
on earth. Man is an evolutionary product of THE WORLD, and he has
learned to interact with this reality and to transform it according to his
needs. Against Kant (and later "conceptual idealists")," we have reason
to believe that THE WORLD has a physical spatio-ternporal structure
and obeys natural causal laws." Thus, the mind-independent WORLD
is a lawlike flux of causal processes.
For a realist, there must also be a legitimate sense of saying that
electrons, planets, stones, trees, and dinosaurs existed before human
beings invented languages or conceptual schemes. The suspicion that this
is philosophically problematic arises easily from the following fallacy :
the equivalence
50 ILKKA NIINILUOTO
(1) x exists at t iff the sentence 'x exists' is true at t
implies that the left side is false at those moments of time t, when a
concept or a sentence referring to x does not yet exist. But the correct
form of (1) is
(2) x exists at t iff the sentence 'x exists at t' is true.
The meta-linguistic sentence on the right side is temporally indefinite,

and if it is true now it is true at any time. Even if the sentence 'Dinosaurs
exist' was not yet invented at 1:0 = 100 million B.C., it is nevertheless
counterfactually true that 'Dinosaurs exist' would have been a true
sentence at 1:0, if someone had formulated it at to'
It is another matter that dinosaurs were identified as dinosaurs, i.e.,
as special kinds of physical objects, through a human language. To
identify something, it has to be distinguished from other things; to
identify an object as a dinosaur, at least some of its characters or prop-
erties have to be described . These identifications involve human activities
(thinking , perceiving, using language, etc.) , as Rescher (1982) convinc-
ingly argues. Rescher further asserts that to be identifiable is mind-
involving as well, since the realm of possibility is mind-dependent, but
this is less convincing. If an object x has some mind-involving proper-
ties (e.g., can be identified, seen, thought , etc., by human beings), x itself
may nevertheless exist in a mind-independent way. The possibility of
the identification of a physic al thing, like a dinosaur or a chair, is indeed
based on its mind-independent properties (location in space and time ,
causal continuity, qualities)."
We may thus make a distinction between UFO 's ('unidentified flying
objects') and [Fa 's ('identified flying objects'). Thesis MRI is then true ,
and Ik l ' false , in the sense that THE WORLD contains UFO's, which
are not our constructions, or produced by us in any causal sense. But
these UFO's are not " self-identifying objects" in a bad metaphysical
sense:" they are potentially identifiable by us, as extended elements or
"chunks " of the world flux, by means of continuity and similarity." IFO 's ,
on the other hand, are in a sense man-made constructions, objects under
a description, and hence exist only relative to conceptual schemes.
Even if we reject the conceptual idealism ("all objects are IFO's"), our
minimal realism is compatible with a version of IRI: THE WORLD
does not contain self-identifying individuals, but can be categorized
into objects in several alternative, overlapping ways relative to concep-
tual schemes . For example , depending on the choice of a suitable con-

ceptual framework, THE WORLD can be 'sliced' or 'structured' to a
system of momentary events, mass points, physical systems , etc.
The distinction between UFO's and IFO's may seem to resemble
Kant's Dinge an sich and Dinge fur uns, but there is a crucial differ-
ence in interpretation: an IFO is not a veil which hides a UFO from
us, but rather a partial truth about a UFO . A UFO corresponding to an
IFO is not defined negatively, as a propertyless bare particular, i.e., as
nothing , but rather it is a complex composition of properties, and the IFO
as it were gives a partial description of it (i.e., of those of its proper-
ties expressible in the chosen language L). In this sense, the existence
and the properties of IFO's depend on the "reality", and knowledge about
IFO's gives us truthlike information about the UFO's.19
Thus every interpreted language or conceptual system L, whose terms
have a meaning through social conventions, "picks out" or determines
a structure W L' consisting of objects, properties , and relations, and
exhibiting THE WORLD as it appears relative to the expressive power
of L. 20 Such structures W L are fragments or "versions" of THE WORLD.
This view can be regarded as a formulation of the internalist principle
IR I, since it allows the reality to be structured in many ways. It also
denies the sort of metaphysical realism which assumes the existence of
an ideal "Peirceish" language L such that THE WORLD = W L'
It is important to point out that the concept of W L is not epistemic:
if a community shares the language L (to use Kuhn's phrase), W L is
not what they believe about the world , but rather "the way the world
is" relative to L (to use Goodman's phrase) . In other words, if a member
of this community has false beliefs (expressible in L), then his "life-
world" differs from WL '
Furthermore, this account does not lead to relativism, since all struc-
tures W L are fragments of the same WORLD - and therefore cannot
be incompatible with each other."
It might be objected that my suggestion leads back to unacceptable
metaphysics. Martin Kusch (1991) has argued that my WORLD is some-
thing like Schopenhauer's das Ding an sich, a blind will: how can THE
WORLD, which does not have any inherent categorical structure,
"choose" the structure WL?
The metaphorical talk about "choices" can be understood here as easily
as in decision theory and game-theoretical semantics: in the game of
exploring reality, the choice of the language L is my first move, and it
52 ILKKA NIINILUOTO
is followed by "Nature's choice" of the structure WL • The game continues

by my attempt to study the secrets of W L ' And all true information
about W L , viz. about a fragment of THE WORLD, also tells something
about THE WORLD.
As each W L is a structure for language L, we can directly apply
Tarski's model-theoretical definition of truth for sentences of L. For each
L, we can define the class of truths in L:
Tr L = Th(W L ) ·
This agrees with IR2, and does not imply the problematic MR2 , since
the union of all classes TrL over "all possible languages" L (whatever
that might mean) is not well-defined. This approach also allows us to
define an objective notion of truth about mind-independent reality: a
sentence h in L is true in THE WORLD if it is true in W L '
In this way we have combined IRI and IR2 with MR3: our concep-
tion of truth is clearly a version of the correspondence theory , but it is
not metaphysically suspectible in the sense feared by Kuhn and Putnam."
Hence, if Kuhn were willing to accept W L as an explication of his
notion of a world with a taxonomy, he could annul his rejection of the
correspondence theory of truth.
University of Helsinki,
Finland.
NOTES
1 See Kuhn (1991). p. 4. For my evaluation of the strong program, see Niiniluoto
(1991).
2 Ibid., p. 9.
J Ibid., p. 10.
4 Ibid., p. 6.
5 See Hacking (1983), p. 109. According to metaphysical nominalism the world consists
of a fixed class of individuals. while " transcendental nominalism" claims that ind ivid-
uals (and their classes) are "constituted" by human minds or languages.
6 My first statement to this effect was given in Niiniluoto (1980). See also Niiniluoto
(1984), pp. 177-178; (l987a), pp. 141-143.
7 See Putnam (1981), p. 49. In Putnam (1978), p. 123 internal realism was defined as
an empirical theory which explains the convergence and success of science.
S Cf. Putnam (1978) , p. 125; (1983) p.xvi i. To say that truth is non-epistemicdoes not
imply that truth is entirely inaccessible to us (cf. Niiniluoto, 1984, p. 178). Rather it means
that the concept of truth cannot be defined, or co-extensively characterized, by epis-
temic concepts . Tuomela's (1990) internal realism claims truth to be epistemic in a much
weaker sense: talk of truth presupposes knowledge about meanings . But Tuomela (1985)
has also argued that , by the scientia mensura principle, truth can be characterized in
terms of best explanatory theories.
9 See Horwich (1990) . Cf. Putnam (1983). p. xiv ; (1990), p. 31.
10 See Putnam (1990) , p. 31.
II See Walker (1989).
12 See Putnam (1990), pp. vii-ix. This shows that Hacking 's (1983) equation between
Peirce and Putnam is problematic.
13 In this paper, I am primarily interested in the Popperian World I (nature) rather
than World 2 and 3 (cf. Niiniluoto, 1984, Ch. 9). To deny the independent existence of
World I leads to ontological idealism .
14 See Rescher (1982). Rescher 's conceptual idealism argues that " nature for us" or
"reality-as-we-picture-it" is "mind-invoking" in a conceptual - rather than ontological
or historical - sense. But he also makes the more radical conclusion that an altogether
mind-independent reality is "an essentially empty idealization" (p. 154), and that " reality-
as-we-think-of-it (= our reality) is the only reality we can deal with" (p. 169).
15 Putnam (1983), p. 211, argues that the concept of total cause involves the non-physical
concept of explanation. This does not prove that causal processes are mind -involving.
16 According to nominalism, an object is green because we apply the predicate ' green'
to it. It is more plausible to accept the converse thesis : we can correctly apply the
predicate 'green' to objects that have the common quality of greenness (cf. Armstrong,
1978).
17 See Putnam (1981) , p. 53.
18 For a theory of individuation and identification of physical objects, see Hintikka
and Hintikka (1989) . In spite of his "semantical Kantianism", Hintikka emphasizes that
"constitution" or individuation does not "create" inhabitants of any possible world "
(Hintikka , 1975, p. 216).
19 See Niiniluoto (I 987a, b).
20 If the language L is not semantically determinate, i.e., some of the terms of L are
vague and do not have unique extensions, language L determines a class of structures -
instead of the unique W L ' See Niiniluoto (1987), p. 146.
21 Cf. Niiniluoto (1991) .
22 In particular, I don't need the assumption that " there is such a thing as the corre-
spondence (the One , metaphysically singled out correspondence) between words and
things" (Putnam, 1983, p. xi).
BIBLIOGRAPHY
Armstrong, D. M. (1978). Nominalism and Realism, Cambridge University Press,

Cambridge .
Ellis, B. (1988) . ' Internal Realism ' , Synthese 76, pp. 409-434.
Field, H. (1982) . 'Realism and Relativism', The Journal of Philosophy 79, pp. 553-567.
Grim, P. (1991). The Incomplete Universe: Totality, Knowledge and Truth. The MIT Press,
Cambridge, MA.
54 ILKKA NIINILUOTO
Hacking, 1. (1983). Representing and Intervening, Cambridge University Press, Cambridge.

Hintikka, J. (1975). The Intentions of Intentionality, D. Reidel, Dordrecht.
Hintikka , J. and M. B. (1989) . 'Toward a General Theory of Individuation and
Identification ', in The Logic of Epistemology and the Epistemology of Logic . Kluwer,
Dordrecht, pp. 73-95.
Horwich, P. (1990). Truth, Blackwell, Oxford.
Kuhn, T. S. (1990). 'The Road Since Structure', in A. Fine et al. (eds.), PSA 1990, vol.
2, Philosophy of Science Association, East Lansing, MI, 1991, pp. 3-13.
Kusch, M. (1991). 'Koko totuus totuudesta' , Tiede ja edistys 16:4, pp. 284-297.
Niiniluoto, 1. (1980). 'Scientific Progress', Synthese 45, pp. 427-464.
Niiniluoto, 1. (1984). Is Science Progressive ", D. Reidel, Dordrecht.
Niiniluoto.T, (l987a). Truthlikeness, D. Reidel, Dordrecht.
Niiniluoto, 1. (I 987b). 'The Varieties of Realism', in P. Lahti and P. Mittelstaedt (eds.),
Symposium on the Foundations of Modern Physics 1987, World Scientific, Singapore,
pp. 459-483.
Niiniluoto.L (1991). ' Realism, Relativism, and Constructivism', Synthese 89, pp. 135-162.
Putnam, H. (1978). Meaning and the Moral Sciences, Routledge and Kegan Paul, London.
Putnam, H. (1981). Reason, Truth and History, Cambridge University Press, Cambridge.
Putnam , H. (1983). Realism and Reason . Philosophical Papers vol. 3, Cambridge
University Press, Cambridge.
Putnam, H. (1990). Realism with Human Face, Harvard University Press, Cambridge, MA.
Rescher, N. (1982) . Conceptual Ideali sm , University Press of America, Washington,
D.C.
Sellars, W. (1968). Science and Metaphysics. Routledge and Kegan Paul, London.
Tuomela, R. (1985). Science, Action and Reality . D. Reidel, Dordrecht.
Tuomela, R. (1990). ' Causal Internal Realism ', in G. Pasternack (ed.), Philosophie und
Wissenschaft en, Peter Lang, Frankfurt am Main, pp. 165-179.
Unger, P. (1975). Ignorance: A Case for Scepticism, Clarendon Press, Oxford.
Walker, R. C. S. (1989). The Coherence Theory of Truth, Routledge, London.
QIU RENZONG
HOW TO KNOW WHAT RISES UP IS THE MOON?

- ON THE CONCEPT OF REALISM AND THE
IRRELEVANCY OF QUANTUM MECHANICS TO
THE DEBATE ON REALISM VS . ANTIREALISM
In recent years the debate on philosophical problems in science, par-

ticularly in modem physics, has been in the ascendent in China . I would
like to refer only to the literature such as [4,5 ,9, 10, 11, 12, 13, 14,
15, 16, 17, 18,25,26,27], but this list is far from complete. Much of
the debate involves the problem of realism vs. antirealism in science
which might be said to be an important issue in the frontier of contem -
porary philosophy of science . If this debate is unfolded properly, it is
expected that a fruitful result will be brought about with benefit to science
as well as to philosophy . However, both in China and abroad there are
some confusions concerning the concepts of 'real', 'reality ' and 'realism';
the different issues or different dimensions of one issue are too tangled
up to unravel in the debate on realism vs. antireal ism in science with
adverse impact upon the debate.
This paper consists of four sections. In the first section, I try to
clarify the concept of reality. In the second section I try to formulate
some criteria for philosophical argument and elaborate a case for realism,
and point out the metaphysical nature of realism . In the third section, I
try to argue that what constitutes the real world is the individual par-
ticulars that possess their own self-identity, and argue for the irrelevancy
of quantum mechanics to the debate on realism vs. antirealism. In the
fourth sect ion, I discuss the semantic and epistemological dimensions
of the debate, and argue for the logical and practical possibility of rational
theory-choice.
WHAT IS IT RISES UP ? : CONCEPTS OF REALITY
Although I do not think that philosophy can be reduced to the therapy

of linguistic diseases, as L. Wittgenstein claimed, yet his claim is
revealing in the case of the debate on realism vs. antirealism in science .
Many scientists as well as philosophers, including the great ones, used
55
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
Philosophy of Science, 55-73 .
56 QIU RENZONG
or use terms such as ' real' , ' reality' and 'realism' in a non-rigorous sense.
For instance, 'empirical realism' (I. Kant) or 'phenomenal realism' (N.
Bohr). It led to conceptual confusion.
What is meant by 'real' or 'reality' depends on the context of dis-
course or the context in which the question "Is it real?" is raised. One
meaning of real or reality is simply ' existing' or 'existence'. The question
is like "Does it exist?" To this kind of question we can give a positive
answer saying all exist even which are dependent upon our mind. For
instance, we always say something exists in our mind. All sensations,
perceptions, desires, beliefs, loves and hatreds etc . do exist, but exist
in our mind. It is this ' existence ' sense when people say 'empirical
realism' or 'phenomenal realism', that refers to the theory which claims
that the empirical experience or the empirical world, or the phenomena
or phenomenal world does exist , they are not merely illusion. However,
when philosophers and scientists argue on "Is the external world real?"
in the context of the realism vs. antirealism debate , the term 'real' or
'reality' is used in a sense more than in the sense of 'existence' . Precisely,
there is a sense of ' independent of human mind' apart from the sense
of 'existence' . 1
PI : "Does A exist?"
P2: "Is A real?"
These are two different problems. We can say that something exists but
is not real. PI only concerns its existence or non-existence, the questioner
does not necessarily want to know its relationship with human mind:
dependent upon the human mind or not. The asking of P2 started with
the time when one discovered the distinction between appearance and
reality. The appearance of a thing actually depends upon the human mind,
but its reality does not. When one saw the moon in the sky, the moon
in water, the moon in a mirror, the moon in a dream . . . then he/she
asked: "Is the moon real?" It means to ask whether the moon exists
only in the human mind or not. The answer given by Plato was: What
is real is not the moon we see but the idea of the moon: the ideas exist
independently of the human mind. When Berkeley put forward his famous
thesis, that is "Being is what is perceived", actually he was talking about,
or his thesis implies, the question "What is real?" but not the question
"What exists?" , or "Does the being depend upon human perception or
not?" In modern science the question was usually put in the form: "Is
HOW TO KNOW WHAT RISES UP IS THE MOON ? 57
a theoretical entity only the creation of our mind , or has it its own
existence independent of the human mind?"
I just said that even some great philosophers did not use the terms
'real' or ' reality' in a strict sense , and those philosophers include Karl
Popper [21]. His 'pluralistic realism' includes the 'World 2', that is the
mental or conscious states and events. In arguing for the 'reality' of
his 'World 2' he defines ' reality' as such a property that anything can
interact causally with ' World I'. So far as the mental or conscious
state and events interact causally with 'World 3' , 'World 2' which is
composed of these states and events is ' real'. However, his so-called
' World 2' cannot exist independently of the human mind. So strictly
speaking, his ' Three Worlds' theory is a doctrine about 'existence' , but
not about 'reality' . Similarly in discourses of ' empirical realism' or
'phenomenal realism' the terms 'real' or 'reality' are used in the sense
of 'existence' .
In his interesting article 'Five Moons Rise up' Zhang Huaxia [27] also
did not notice the difference between the concept of existence and the
concept of reality when he argued for his scientific realism in which phe-
nomenal world, classic world , quantum world, empirical world and
theoretical world exist. He said in one place of his article:
Clas sic reality is only one kind of obje ctive reality, and objective realit y is no more
than the existence independent of human consciousness (p. 45).
Here he pointed out that the concept of 'reality' contains two senses of
'existence ' and ' independence of human mind' . But he said in another
place:
Therefore, the secondary quality of a material object is also one kind of the propertie s
of mater ial world . Recognizing it, recognizing the phenomenal reality, is no more than
recognizing some objective reality of the world. It is absolute ly not ' a compromise with
idealism' . On the contrary, denying it amounts to denying the world which is full of sound
and colour , and can be seen and touched , the world is thus described as a dismal picture
in which only a swarm of atoms dance in the air (p. 44).
He is right, when he criticized Eddington and other physicists. They

like to describe the world as a picture in which only a swarm of atoms
(or other more elementary particles) dance in the air; I shall elaborate
it later. But what is meant by ' reality' here is actually 'existence' . There
is no problem that the secondary quality has its objective content;
however, it cannot 'exist independently of the human mind'. That is,
although electromagnetic waves of different wave length , or vibrations
58 QIU RENZONG
of the air, or a given distance between molecules of material objects,

etc. are independent of the human mind, yet if there had been no creature
which is equipped with sense organs and their centers, no creature
with the ability of seeing, hearing, touching, there would have been
no "world full of sound and color, being seen and touched" . The author
is faced with the same problem when he was talking about the
' phenomenal world' and the 'theoretical world' : How can these two
worlds exist independently of the human mind? He seems to feel that
there might be some problem, so he said, talking about the theoretical
world:
Scientific theory has the form and implication of its own relatively independent exi s-
tence (p. 50).
It is not clear what is meant by the form and implication of existence.

But here he was talking first about the 'existence' of scientific theory;
secondly, he used the qualification of 'relatively independent'. We can
accept the 'relative independence' of scientific theory. However, the
problem is: Can a theory exist independently of the human mind? A
theory is not a biological species which can evolve independently of
human mind. The evolution of scientific theory has to rely on human
ingenuity.
I maintain that a theory should refer to, and expla in, the external , objective world ; this
should be one of the aims of scientific theory (p. 50).
He is right when he contrasted theory with the external world. Theory

is the model of the world, not the world itself. Therefore, the concept
of 'theoretical world' is untenable.
Only if we distinguish the concepts of ' reality' and 'ex istence' , can
we know that what rises up is the moon itself or the shadow of the moon?
The so-called ' empirical world', ' phenomenal world' and 'theoretical
world' are actually the shadow of the moon - the experience of the moon,
phenomena of the moon, the theory of the moon, but not the moon
itself.
HOW TO KNOW WHAT RISE UP IS THE MOON? :

A CASE FOR REALISM
The debate on realism vs. antirealism is a philosophical issue. To resolve

a philosophical issue we need to employ philosophical argument. The
argument I shall employ here is to compare various philosophical expla-

nations ? to the following proposition (explanandum): I see that the moon
rises up. Its generalized form is that: S perceives that p.
At first, we should put forward some criteria according to which we
can say that this philosophical explanation is better than that one.
Following the example of the discussion of scientific explanation in
philosophy of science, we can work out the criteria for judging philo-
sophical explanations as follows :
I. A good philosophical explanation should be simple. The simplicity
is either ontological or structural. The ontological simplicity requires that
we do not presuppose any unnece ssary entity. The pursuit of this sim-
plicity has been an ideal of philosophers since ancient times. Both
Thales 's 'water ' and Heraclitus's ' fire ' and Chinese philosophers' 'Tao '
and 'Qi >3 embody the pursuit for the ideal. Structural simplicity requires
that there should be no omittable conjunction, disjunction, appendix,
or proviso .
2. A good philosophical explanation should possess greater explana-
tory power. That is, it can explain not only what other explanations
can, but also explain what others cannot.
3. A good philosophical explanation should possess greater predic-
tive power. That is, it can predict not only what other explanations can,
but also what others cannot.
4. A good philosophical explanation should possess more fertility. That
is. it can lead to more interesting problems and suggest new approaches
or new research lines.
Now let us to compare various available explanations (explanans)
for the proposition 'I see that the moon rises up' according to the criteria
mentioned above:
E I. "I see that the moon rises up, because there is an entity outside
me called 'the moon' from which the reflected light from the sun
stimulates my retina, and thereby the image of the moon is produced
on the retina." This is a realist causal explanation.
E2. "I see that the moon rises up, because there is a mechanism so
far unknown in my genetic structure which causes me to see the moon
rising up at a given time." This is an innatist causal explanation.
E3. "I see that the moon rises up, because God or other super-
natural beings use divine or magic power to make me see the moon rising
up." This is a theological or magical causal explanation.
E4. "Human life is a dream . That I see that the moon rises up is a
60 QIU RENZONG
part of the dream. And who knows whether it is I who see the 'Chang
E,4 or it is 'Chang E' who see me." This is Zhuangzhouist explanation.
E5. "We cannot go beyond our own experience. That I see that the
moon rises up is my own visual experience. It is impossible to provide
any explanation for the experience beyond it." This is a Humean expla-
nation, or radically empiricist explanation. Actually it declines to provide
any explanation.
E4 and E5 are the simplest explanation in ontology and structure,
but they explain nothing , they have no explanatory and predictive power
at all. Neither of them is a causal explanation. They do not explain why
we have this visual experience ('see that the moon rises up'). And they
do not predict anything either: We do not know what we would dream
or what we would see at the next moment. The only thing we can do
is to record this or that experience and observe the order with which
they emerge in my mind.
They are confronted with other troubles. If we explain 'I see that
the moon rises up' by the dream , then I ask whether the moon is my
dream is the same as the moon in others ' dreams or not? If they are
the same, how to explain it? If not, there are numerous moons in dream,
and this runs counter to people's intuition. And nothing explains why I
and other people dream. The radically empiricist explanation is faced
with the same troubles.
E2 and E3 are similar. Their difference lies in that: E3 ascribes the
cause of our experience to some external supernatural being, whereas
E2 ascribes it to some mystical internal natural factor. However, nobody
on earth understands the regularity of this supernatural being or natural
factor: Why it causes this kind of experience, and not another kind? What
will it cause at the next moment ? etc., and these remain to be explained.
So their explanatory power is very limited , and their predictive power
is near to zero.
EI can explain why what I see is the moon but nothing else . Because
the image emerged on my retina was caused by the light from the sun,
reflected by the surface of the entity called the moon. The astronom-
ical knowledge can be incorporated into El , and help it to form
counterfactual conditionals such as: "If we had met last time in the middle
of months of the lunar calender, and there had been no clouds in the
sky, we would have seen the full moon as we see it tonight." And we
can predict that we will see the full moon in the middle of months of
the lunar calender, see nothing at the beginning or end of months, and
HOW TO KNOW WHAT RISES UP IS THE MOON? 61
see the crescent in between. E1 can lead to many interesting problems

such as how about the surface of the moon, its other side, the temper-
ature of its surface, is there any air, is there any trace of life, can human
beings survive on it, how to climb up to it, etc.
It might be objected that we can make E2 and E3 possess the same
explanatory and predictive power as E1 does. For example, God created
the moon and made us see it, he arranged the configuration of the sun,
the earth and the moon so nicely that we can predict when there will
be full moon or crescent. Or some mystical factor in our genetic struc-
ture acts as God. Although E2 and E3 now possess all that El does,
the cost is that they become more complex than El both in ontology
and in structure , because they need an extra external supernatural being
or an internal mystical factor, and an additional explanation of why
they want us to see the moon.
It follows from the argument that E1 is better than any other avail-
able explanations. When I see that the moon rises up, the best explanation
of it is that: What rises up is the moon. The moon in the first sentence
is our visual experience, the moon in the second sentence is the entity
which exists independently of our experience.
The realist explanation for our experience entails an ontology or meta-
physics. In other words, realism is inescapably metaphysical: the physical
objects exist independently of our mind, and the external world is
composed of these objects. But when we say that it is the object of the
external world which caused our experience, it does not entail that our
experience and theory built on the experience must constitute true knowl-
edge of the object. Is there any external world outside experience?
Does/can experience provide true knowledge of the world? They are
not the same problem, and should have different solutions. F. Engels
[6] pointed out that "Is our thinking capable of cognition of the real
world? Are we able in our ideas and notions of the real world to produce
a correct reflection of reality?" are 'another side ' of 'the great basic
question of philosophy '. Our contemporary philosophers [I, 3] point
out it too.
When some philosophers talk about semantic realism or epistemo-
logical realism (for example [16]), they actually deal with another
problem. In other words, they are talking about the problem"Are human
beings able to obtain true knowledge of the external world?" or "Do
our thoughts (mental representations) or sentences (linguistic represen-
tations) correspond to the external world?", but not the problem "Does
62 QIU RENZONG
a world exist independent of human mind?" Of course, maintaining or

denying that human beings can reach true knowledge, maintaining or
denying that a theory can be true, may lead to realism or non/antirealism.
But they are different problems logically. As M. Denitt [3] point out,
realism is a doctrine which claims the existence of physical entities inde-
pendent of the mental, and this definition is entirely unsemantic, making
no mention of representation or truth. Realism is a philosophical doctrine
on the nature of the world, not a philosophical doctrine on the nature
of language or thought. He pointed out that people often refuted realism
with the following reasoning:
1. If the realist's independent reality exists, then our thoughts/theories
must mirror, picture, or represent that reality .
2. Our thoughts/theories cannot mirror, picture, or represent the
realist's independent reality,
3. So, the realist's independent reality does not exist [3].
But 1 is false . So 3 cannot be derived from 2. 1 is untenable because
it confuses the realist problem with the semantic/epistemological problem.
If it is clear that the crucial difference between realism and anti-realism
lies in "is it accepted that the external world exists independently of
human mind?", then a version of realism such as H. Putnam's 'internal
realism ' [22] is actualIy a hidden version of anti-realism, but does not
"continue to hold fast to a realist position" [8]. Because he claimed
that:
'Objects ' do not exist independently of conceptual schemes. We cut the world up into
objects when we introduce one or another scheme of description ." (p. 52).
As pointed out by C. Brown [I], both Kant and Putnam hold that the
world we know and talk about is empirically real, but hold also that it
is mind-dependent.
WHAT IS THE MOON ?: ON THE REAL WORLD OF INDIVIDUALS

WITH SELF-IDENTITY AND THE IRRELEVANCY OF QUANTUM
MECHANICS TO THE DEBATE ON REALISM VS . ANTIREALISM
It might be argued that the argument developed in the second section only
holds in the case of macro-objects, not in the case of micro-objects,
because the latter can not be perceived directly by our senses and our
mind, and some of them were discovered after scientists had predicted
them before. However, it will not spoil the argument in our second
section: We can transform the question "What is the better or best expla -
nation of 'I see that the moon rises up' " into the question "What is the
better or best explanation of all empirical data about electrons avail -
able to scientists?"
On the basis of the argument discussed above, the better or best expla-
nation of "I see that the moon rises up" is that there is an entity which
exists independently of me called the moon which caused the emer-
gence of my visual experience .' Analogically, there are other various
entities that exist independently of me which caused and will cause the
emergence of my other various experiences. The external world is
composed by these independently existing entities. These entities can
be grouped into different classes according to the similarity of their prop-
erties, but each element in the class is a single particular individual, from
elementary particle to galaxy, but the most delicate individual is the
person.
Each individual has its own self-identity which make s it discernible
by other individuals and enables us to identify it as distinguished from
other individuals. Self-identity means that an object is the same iden-
tical with itself (Hume) , or means that things are what they are (Russell).
Each object has its own ' life' or 'lifetime'. Throughout its lifetime an
individual object remains the same and maintains itself through the
passage of various outside influences. It is the self-identity that makes
it into an individual. The concept of 'self-identity' is closely related
with the concepts of ' existence' and ' reality' . A thing is said to exist,
that is to say it has self-identity. So it might be said that the existence
of a thing presupposes its self-identity. If a thing has no self-identity,
how we say it exists? There are two alternatives for a thing without
self-identity: It does not exist or it is another thing . If we say a thing
without self-identity exists , we cannot distinguish it from other things
or we identity it with another thing. A thing without self-identity cannot
be said to exist, and thereby cannot be said to be real.
It might be objected that the concept of self-identity does not hold
in the case of micro-particles, because they are indiscernible: All parti-
cles of the same class have the same physical properties. However,
self-identity cannot be reduced to physical properties. The physical prop-
erties of two electrons are the same, but their own histories (though
very short), and the context in which they exist during their lifetime,
are not the same. Just like two homozygote twins, whose genes are
identical, but the environment of their development and growth and
64 QIU RENZONG
their personal history cannot be identical. They still have their own
self-identity.
On the basis of the concept of individuals having self-identity, philoso-
phers should not, like some physicists, reduce all things in the universe
to the sum of some 'elementary' particles. A table has its own self-
identity, this self-identity is irreducible. This is the reason why I said that
Zhang [27] is right when he criticized Eddington's attempt to reduce a
table to the sum of atoms . Objects at different levels [of reality] have
their own irreducible self-identities, and the external world is composed
of multi-levelled individuals.
The Honourary President of the Sino-British Summer School of
Philosophy, Peter Strawson 's thesis on objective individualistic ontology
[23] may be helpful to our discussion:
We think of the world as containing particular thing s some of which are independent of
ourselves; we think of the world 's history as made up of part icular episodes in which
we mayor may not have a part ; and we think of these particular things and events as
included in the topic s of our common discourse, as things about which we can talk to
each other. These are remarks about the way we think of the world , about our concep-
tual scheme. A more recognizably philosophical, though not clearer, way of expre ssing
them would be to say that our ontology comprises objective particulars (Ibid. p. 15).
In [24] he reemphasized his individualistic ontology, and although he

accepted that it is reasonable for phy sicists to claim that the ultimate
material objects are atoms, electrons, and further, quarks and fields in
some basic sense, but not in the sense of philosophers' primary concern.
What mostly concerns philosophers is the conceptual scheme , or general
"f ramework of thinking about reality and the world, and even scientists
have to start their inquiry within a framework.
There are some elements of individualistic philosophy in Mao
Zedong's philosophy [9]. He emphasized that universals only exist in
particulars. It means that universals do not exist independently like Plato's
'ideal world' or Popper's 'World 3'. The world is full of particulars in
which universals exist. He pointed out:
These are two cogn itive processes: one is from particular to general, the other is from
general to particular [191.
That is to say, it is the particular which is the starting-point and end of

human cognition, but not the universal. Why? The ontological reason
is that the world is composed of numerous particular individuals, and
what we have to deal with are all particular individuals. Building this
conceptual scheme or thinking framework is very important to the inquiry

in all fields and particularly important to the fight against dogmatism and
conservatism in today's China.
An individual object cannot be reduced to its constituent parts, nor
to its properties or qualities either. The properties of an individual object
depend upon the context in which it is situated, and the perception of
these properties by the human mind depends at least partly upon the
condition or state of the perceiver. This holds for the secondary quali-
ties as well as the primary [15, 18]. To say "The table is red " or "The
table has the quality 'red colour'" needs two conditions: the light is
adequate and the speaker's system of sense organ s-brain as well as her/his
psychology is normal. However, an individual object such as a table
cannot be reduced to its property or quality such as ' red ' . A material
object has the property of extension, but its length depends upon the
choice of a reference frame : its length contracts in proportion to the
speed of motion of reference frame. Under these two conditions it cannot
follow that the object does not exist independently of our mind from
the fact that 'red' and ' length ' will be changed with a choice of refer-
ence frame.
From the view of individualistic realism, the problems facing quantum
mechanics are irrelevant to the debate on realism vs . antirealism, i.e .
irrelevant to the problem "Does the moon exist when nobody see it?"
The concept of relevancy has not been sufficiently discussed in philo-
sophical literature. L. J. Cohen? pointed out:
We say that a proposition R is relevant to the question Q if and only if R gives us a
reason to accept or reject a proposit ion A as an answer to the quest ion Q.
However, neither the indeterminacy, nor EPR correlation in quantum

mechanics itself gives us a reason to accept or reject the realist or
antirealist answer to the question "Does the external world exist indepen-
dently of our mind?". Only a conjunction of this property or relationship
with other propositions can give us a reason to do so.
I. Indetermina cy
That the momentum and position of a micro-object cannot be determi-

nately measured at the same time concerns the problem of measuring
some properties of micro-objects, but not the problem of the independent
existence of a micro-object itself. It is quite often that the state of an
66 QIU RENZONG
object is changed by observation in a psychological or sociological study.

Suppose that a subject S is originally in state p, and the psychologist
or sociologist E intends to study S in the state p. If the question asked
by E is not proper or is not properly asked, S can be shifted into state
q: thereby S gives an answer to the question different from the answer
if S is in the state p. In some sense we can say that the state q is
'created' by E or ' emerges' with E's suggestion. However, that the state
of a subject can be changed by the experimenter does not give us a reason
to reject the independent existence of the subject. In the same vein ,
that there is no indeterminacy in classical physics does not give us reason
to reject the phenomenalist or radically empiricist answer to the problem
of "Does the external world exist independently?" According to a phe-
nomenalist or radical empiricist we cannot go beyond our sense -data
or experience.
Now there seems to be a physicalist account for the indeterminacy:
If we want to measure the position of a particle, we have to operate
the apparatus to produce a light wave with short wave length and great
energy, and thereby change its speed, and as a result, we cannot accu-
rately measure its speed. If we want to measure its speed, we have to
produce a light wave with long wave length and it makes us unable to
accurately measure its position. Even if the interference of the appa-
ratus in the measurement is such that it changes the state of the measured
particle, or even 'creates' its new state or 'make' its 'emergence', it
concerns only the state of the measured particle, not the object itself.
In the same vein , if the indeterminacy is not ascribed to the interfer-
ence by the observation, but to the probabilistic property of the particle
.itself, it also does not give us a reason to accept the realist answer to
the problem "Does the external world exist independently?" because it
might be the case that all facts including the probabilistic property of
particles are facts only within our experience, and do not reach exis-
tence outside our experience.
2. EPR Correlation
EPR correlation may be the most bizarre phenomenon . There are optional
explanations to this bizarre phenomenon including :
(1) There is a distant action between particle A and particle B.
(2) There is an interference by an observer when measuring any of
these two particles.
(3) There is some unknown special interaction or correlation between

A and B.
Now some Chinese scholars [13, 27] hold that (l) is already precluded
by modern physics, (2) implies that the "realist presupposition would
be no longer correct", so the most plausible explanation is (3) which indi-
cates the non-separability and non-locality of quantum objects . However,
even if (2) is the most plausible, recognizing the fact that the property
or state of particles is interfered by observation or experiment does not
necessarily lead to denying the independent existence of quantum objects.
Now it has been confirmed by experiments that there is EPR correla-
tion between particles. However, no matter how bizarre the EPR
correlation is, it is a property of particles, and nothing here concerns their
very independent existence. In the same vein, even if one day (I) or
(3) turns out to be more plausible, it does not entail the realist answer
to the question of "Does the external world exist independently?" because
it might be the case that all these are only within the limit of our expe-
rience, that nothing can go beyond our experience.
The conclusion is: problems of quantum physics show how to more
deeply understand various bizarre properties (quite different from those
of macro-objects) and the mechanism of interaction between micro-
object, experimenter/experimental apparatus, and context in which they
encounter, but all these are irrelevant to the debate on realism vs anti-
realism. In other words, the problem raised in quantum physics is not
"Does the moon exist when nobody see it?" but "Is the moon when
somebody see it the same as the moon when nobody see it?" [10] If
the answer is "the same", it does not give us reason to reject anti-realism,
and if the answer is "not the same", it does not give us reason to reject
realism . The assertion that the puzzles in quantum physics are relevant
to the debate involves a misunderstanding of the concepts 'reality' and
'realism' .
THE MOON AND ITS MODELS :

RATIONALITY OF THEORY CHOICE
Our external world is composed of particular individuals like the moon,

these individuals interact each other, and are exerted upon by downward
and upward causation between different levels of individuals. On the
other hand, human cognition is limited in a given space-time, and cir-
cumscribed by our logic, and our linguistic and conceptual framework
68 QIU RENZONG
which is not infallible. So the theory of the moon , or its model, formed
in a certain historical period is always incomplete. Scientists and philoso-
phers once were over-optimistic, when they believed that if a theory
has been confirmed over a long period of time, it can be confirmed as
true. In view of the fact that Newtonian theory, which had been well-
confirmed for two hundred years, was refuted at last, and hypotheses
and theories in modern science were put forward and then refuted
successively with a dizzy speed, many are now inclined to over-
pessimism and scepticism: to obtain true knowledge is impossible,
or as rival theories can be confirmed by the same experimental data,
rational choice between two competing and incompatible theories is
impossible.
Both over-optimism and over-pessimism are untenable

In the long run we cannot deny that knowledge of the world (including
nature, society and human beings) themselves possessed by modern
people is much richer, and much more correct than the knowledge
possessed by our remote ancestors and by ancient people in the age of
Plato and Confucius. Modern science and technology put tremendous
power in the hands of modem people, and has brought much more welfare
to them. On the other hand, the power is formidable, and created many
serious problems facing human beings who are now working to deal
with them . This is an intuition shared by common sense. Any logical
or philosophical argument cannot reject or eliminate this intuition.
However, the question of whether a scientific theory can be said to be
true or when it can be said to be true is not a question that can be
answered instantly. The life of many theories is full of haps and mishap s.
Theory grows and develops with human ingenuity. It is very difficult
for us to answer the question "Is it true or false?" at the beginning of
its life. Just as when we see a movie, when a character appears on the
scene , kids ask their parents: "Is he a good man or bad man?" It is dif-
ficult for parents to answer the question, and have to say : "You will
find the answer at the end of the movie." Scientific theory is the same:
We don't know who is the winner and who is the loser until the game
ends . However, during the scientific game, we can judge which theory
is better according to some well-received epistemic values. That allows
a rational choice between competing theories. These epistemic values
include :
V1. Corroboration or support by empirical evidences and the degree

of corroboration or support. If a theory (T2) is corroborated by avail-
able empirical evidences better than another (T 1), or T2 's degree of
corroboration is higher than T I 's, then we have reason to say that T2
is a better theory than T 1.
V2. Explanatory power. If T2 can explain all explained by Tl , and
explain what cannot be explained by T I , then we have reason to say
that T2 is better than T 1.
V3. Predictive power. If T2 can predict all predicted by T 1, and predict
what cannot be predicted by T 1, then we have reason to say that T2 is
better than T 1. The predictive power of a theory can be measured by
the novelty of the prediction , and the novelty of a prediction can be
measured by the distance between the prediction and background knowl -
edge .
V4. Fertility. This means that a theory can lead to interesting scien -
tific problems and/or suggest new approaches and new research lines .
If T2 can lead to more interesting problems than T'l , or/and it suggests
a new approach and a new line of research, but TI cannot, then we
have reason to say that T2 is better than T I.
V5. Simplicity. If T2 is simpler than TI ontologically and structurally,
we have reason to say that T2 is better than Tl.
V6. Coherence within a theory. If T2 is more coherent within it than
TI, with fewer self-contradictory propositions or self-contradictory con-
sequences in it than TI, we have reason to say that T2 is better then
Tl.
V7. Compatibility with other well-received theories. If T2 is more
compatible with these theories than TI, then we have reason to say that
T2 is better than T 1.
V8. Compatibility with well-received ontology. If T2 is more com-
patible with the ontology than TI, then we have reason to say that T2
is better than Tl.
In the different disciplines or different fields, different weight can
be given to these different epistemic values . For example, V5 may be
given more weight in theoretical physics , but V I be given more weight
in applied sciences. And scientists with different value systems may
give different weights to different epistemic values.
Some philosophers, like Putnam [22], maintain that because two
incompatible theories can get support from the same empirical evidence,
they can be said to be equally good, or they have equal claims to truth
70 QIU RENZONG
(p. 73) So they both merit acceptance. The choice made between them
seems to be arbitrary. But even in the case of Newtonian physics cited
by him, the theory of action at a distance between particles and the theory
of action through fields are not equivalent. They cannot have same
value with the epistemic values listed above. Let it be that they are equiv-
alent with VI, but the former may have a higher value with V2,3,4.
whereas the latter may have higher value with V5. Perhaps, the choice
between these two theories depends upon the relative importance
scientists put upon the different epistemic values. Scientists may ratio-
nally have different opinion s on how much weight should be given to
these epistemic values, and make different choices that can be equally
justified. Be that as it may, we have no reason to say that any of them
have same epistemic values, or they are equally good, or they have
equal claims to truth [20].
That a scientific theory is given higher marks for these epistemic
values does not entail it must be true. Just as a Chinese proverb says:
"We cannot judge if a man is a hero or not according to his success or
failure" . The success of a scientific theory is not a sufficient condition
for its being true. However, the success of a theory does give scientists
reason to (sometimes involuntarily) believe or (voluntarily) accept what
the theory describes, and represents what is going on in the real world,
or what the theoretical entity refers to is the real thing in the external
world. Using the metaphor of the moon, we have reason to believe or
accept that the moon the theory describes represents the real moon in
the sky. Here I say : "give scientists reason to believe or accept" that
the theory is not only better than others, but also is true. Because after
all there is a logical gap between success and truth (or hero) which seems
still to be unbridgable in the near future . Highly probably, a very suc-
cessful theory turns out to be false . However, that scientists then believed
or accepted it as true was not irrational. Even if a theory is successful
and gets high marks on epistemic values, it cannot be perfect. Somebody
is very sensitive to its shortcomings, and gives it low marks . He must
have reason to do so, e.g. he may give different weight to epistemic values
than others . If a person gave a theory low marks with no reason, even
if the theory turned out to be false eventually, his judgement may not
be said to be rational.
On the other hand, the failure of a person cannot entail that he is
not a hero. I do not want to repeat many examples in the history of
science that have already been elaborated by some philosophers of

science to show how a failing theory recovered its vitality and staged
a comeback after many years failure. There is an unbridgeable logical
gap between failure and falsity too. However, a theory that is failing
for a long time cannot have a claim to truth.
If we want to know what rises up is the moon, and how many moons
rise up, it can be done only through comparison, competition, discus-
sion and debate between different, rival, or even incommensurable models
of a moon and their modification, change, and alternation. All models
of the moon, all scientific theories, are unavoidably embedded in a certain
logical, linguistic and conceptual framework, and none of these frame-
work are infallible, so none of these models or theories have a right to
pretend that it is the only truth reflecting objective reality.
Institute of Philosophy,
Chinese Academy of Social Science,
Beijing, China.
NOTES
I When I prepared the first draft of this paper for the National Symposium on Realism
vs. Anti-Realism in Science [26] in December 1991 I had not read the article written by
M. Denitt [3]. During my visit to the United States in July 1992, I had chance to read it
and found his view is similar with mine.
2 In August 1992 while I was chairing the Third Session of the Sino-British School of
Philosophy, Peter Carruthers presented me a copy of his new book where he developed
an argument similar to mine but more elaborated .
3 Chinese Taoists and Confuc ianists in ancient and medieval times held that the 'Tao'
or 'Qi ' are the first principle of all things in the universe .
4 In Chinese legend there is a beauty name 'Chang E' who lives in the moon. Sometimes
the moon is called ' Chang E'. The great ancient Chinese philosopher Zhuang Zhou once
argued that it is impossible to distingu ish real life from the dream, and he said when I
dream a butterfly it is impossible to know whether it I dream it or it dreams me.
S Of course, this is a simplified formulation. Actually the emergence of visual experi -
ence is a complex process which involves the integration of the neural system, especially
the brain 's , and interference by past experience or background knowledge . However, all
these actions cannot provide a reason to reject the cause independent of mind.
6 J. Cohen presented his definition of ' relevance' at a symposium of the Third Session
of Sino-British Summer School of Philosophy held in Tianjin, China.
72 QIU RENZONG
BIBIOGRAPHY
I. Brown, C. (1988) . 'Internal Realism? Transcendental Idealism?' in P. French et al.
(eds.), Realism and Antirealism , University of Minnesota Press, pp. 145-155.
2. Carruthers, P. (1992). Human Knowledge and Human Nature, Oxford University
Press.
3. Denitt, M. (1988). ' Rorty ' s Mirrorless World', in P. French et al . (eds.), op cit.,
pp. 157-177.
4. Dong, G-B. (1988) . The Puzzle of EPR Correlation , Shaanxi Science & Technology
Press, Xian.
5. Dong, G-B. (1989). 'Experiments of EPR and the Debate between Realism and
Positivism' , Studies in Dialectics of Nature, no. 4, pp. 31-37.
6. Engels, F. (1988). Ludwig Feuerback and the Outcome of Classical German
Philosophy, English translati on, 1947, Lawrence & Wishart, London, pp. 24-
28.
7. Fales, E. (1988) . 'How to Be a Metaphysical Realist' , in P. French et al. (eds.), op
cit., pp. 253-274.
8. Gu, S. (1990). ' Remarks on the View of Truth in Scientific Realism' , Journal of
Dialectics of Natur e, no. 3, pp. 9-16.
9. Gu, Z-X. (1990). 'On the Three Levels of Epistemic Subjectivity', Chinese Social
Sciences, no. 6, pp. 85-96.
10. Guan, H. (1991) . ' On the Problem of "Does the Moon Exists When It Is Not Seen?" '
Studies in Diale ctics of Nature , no. 3, pp. 39-45.
II . He, Z-X. (1990). ' Comments on "The Theory of Property and the Theory of Reality
in Quantum Mechanics"', Philosophical Research, no. 2, pp. 69-83.
12. He, Z-X. (1990) . 'Can Modern Physics Provide Scientific Foundation for the Theory
of Epistemic Subjectivity?', Chinese Social Sciences, no. 2.
13. He, Z-X. (1991) . ' EPR Paradox and Its Philosophical Problems' , Studies in Dialectics
of Nature' , no. 3, pp. 28-38.
14. Hu, X-H. (1991) . ' Comments on the Discussion of Modern Physics and Epistemic
Subjectivity', Trends in Philosophy , no. 6, pp. 8-11.
15. Hu, X-H. (1992). ' The Quantum Theor y and Realism ', Philo sophical Res ear ch,
no. 10, pp. 52-53.
16. Laudan, L. (1977). Progress and Its Problems - Towards a Theory of Scientific
Growth, University of California Press.
17. Liu, S-Z. (1991). ' Modern Physics and Epistemic Subjectivity ' , Chines e Social
Science s, no. 2.
18. Luo, J-c. (1987). 'Objective Realism', Chin ese Social Sciences, no. 2, pp. 21-
36.
19. Mao, Z-D. (1937) . 'On Contradiction', Selected Works ofMao Zedong, 1967, Peoples
Press, Beijing, pp. 274-312.
20. McMichael, A. (1988). ' Creative Ontology and Absolute Truth ' , in P. French et
al. (eds.), op cit., pp. 51-74.
21. Popper, K. (1985) . Popper Selections, ed. by D. Miller, Princeton University Press,
Princeton NC, pp. 58-77.
22. Putnam , H. (1981). Realism, Truth and History, Cambridge University Press,
Cambridge, MASS.
23. Strawson, P. (\ 957). Individuals : An Essay in Descriptive Metaphysics, Metheun,

London .
24. Strawson, P. (1992) 'The Theory of Property and the Theory of Reality in Quantum
Mechanics', Philosophical Research, no. II, pp. 62-67.
25. Wang , Y-B. (\ 989). 'The Theory of Property and the Theory of Reality in Quantum
Mechanics', Philosophical Research, no. II, pp. 62-67.
26. Xu, L. (1992). 'A Report on the National Symposium on Scientific Realism', Trends
in Philosophy , no. 5, pp. 1-4.
27. Zhang, H-X. (\ 991). ' Five Moons Rise up ' , Philosophical Research, no. 10, pp.
42-51 .
DEMETRA SFENDONI-MENTZOU
THE REALITY OF THIRDNESS -

A POTENTIAL-PRAGMATIC ACCOUNT
OF LAWS OF NATURE
INTRODUCTION
The dilemma between realism and anti-realism is focused on the char-

acter of reality : what is the real? Is it the singular or the universal, the
individual or the species? Anti-realism claims that reality consists of indi-
vidual, singular, actually existing things, while realism seeks the most
essential element of reality in the universal. Thus, when the question
arises as to the character of laws of nature, anti-realists can find no
place for a real connection between separate cases. If there is nothing
other than a long series of independent events, then law merely describes
the similarity with which things happened in the past. Realists on the
other hand claim that laws of nature express not only a uniformity of
behaviour, but an inner string passing through particulars and linking
them by a real necessity. It seems, therefore, that the issue, which
occupies the minds of many present day philosophers, is still the question
of the reality of that which traditionally has been called the fundamentum
universalitatis. Its history can be traced back to the medieval debate
between realism and nominalism and certainly further back to Plato
and Aristotle.
A revival of scholastic realism is Armstrong's account of laws of
nature. I What is interesting about his theory is that the traditional problem
of universals, in the sense of the nature of things, becomes that of laws
of nature.' The central idea propounded by Armstrong's alleged realism
is that laws of nature are relations between universals, which link prop-
erties with properties and things with things. However, my thesis in
this paper will be that a careful reading of Armstrong reveals that some
of his most basic assumptions cannot possibly live in harmony with the
traditional doctrine of universals. His theory, which starts as a vigorously
defended realism, eventually transforms into a disguised nominalism.'
My pointing out this tension in Armstrong's thought does not diminish
my esteem for his admirable work on universals. It is only that I take
75
76 DEMETRA SFENDONI-MENTZOU
it as a challenging invitation for a fruitful discussion of the issue in its

original scenery: that of Aristotelean-scholastic realism.
To respond to this invitation I will try to profit from the subtle
scholastic analyses of some key-concepts and problems connected with
realism so as to be able to bring to light Armstrong 's concealed leaning
toward nominalism. In doing so my hope is that the development of
my arguments will provide at the same time the necessary material for
my drawing an alternative picture of laws of nature . In my account, which
I will call a potential-pragmatic realism, because of its linkage both to
Aristotelean-scholastic philosophy and to Peircean thought," Armstrong 's
belief in the reality of universals will be retained. What will be criti-
cized is his overemphasis on the principle of instantiation on the one
hand, and his rejection of the idea of potentiality on the other. By contrast,
the burden of my analysis will be on potentiality as a sine qua non for
a realist theory of laws of nature .'
ARMSTRONG'S THEORY OF LAWS OF NATURE
The central idea propounded by Armstrong is this: Laws of nature are

relations between universals. The idea that laws of nature link proper-
ties with properties and things with things, N(F,G) holds in virtue of a
de re necessity linking the relations between universals and the unifor-
mity it produces" (1983, pp. 85 , 86). Furthermore, relations between
universals are treated by Armstrong as universals themselves (see ibid.,
pp. 88ff.), so that he can "assimilate the relation between law and positive
instantiation of the law to that of a universal to its instance" tibid.,
p. 89). His realism, therefore, which is grounded in a realism about
universals (see ibid., p. 7) makes him, as he claims, an opponent of the
regularity theory on the one hand, and a defender of the idea of neces-
sity involved in laws of nature on the other.
Armstrong's criticism is focused on the Humean idea that "the law
is exhausted by the fact that the observed Fs are Gs, and the unob-
served Fs are Gs" (ibid. , p. 55) . By contrast, he holds that "the law
involves an extra thing" the presence of which "serves first to explain
why all the observed Fs are Gs, and second, to entail that any unobserved
Fs there are will be Gs" (ibid., p. 55). The basis of such necessity can
"be found in what it is to be an F and what it is to be a G" tibid.,
p. 77), namely in F-ness and G-ness . If it is a law that all swans are white,
then the necessity is grounded in swanhood and whiteness. This is what
THE REALITY OF THIRDNESS 77
Armstrong means in claiming that "the necessitation involved is a law

of nature in a relation between universals" tibid., p. 78) .
However, a crucial point in his alleged realism is his insistence on
the principle of instantiation, which brings us back to the problem of indi-
viduation, much discussed in the Middle Ages : Laws "are abstractions,
that is, they cannot exist in independence of other things" (ibid ., p. 90).
His reason for choosing the principle of instantiation, as Armstrong
explains, is to uphold along with the relation between universals the
doctrine of naturalism (see ibid., p. 82) . And it is only through this
combination, he claims, that universals can be brought into the spatio-
temporal world. At the same time the idea of powers, dispositions,
potentialities 7 is excluded from his scheme.
Even when Armstrong makes some space in his theory for powers,
these have nothing to do with potentiality and are used by Armstrong
in a pure pragmatic (see 1978b, p. 45) and positivistic sense. Powers
are taken as equivalent to "the outcome which things having those
properties have the power to produce in variou s circumstances" tibid.,
p. 46). Moreover, it is claimed, " we have access to the properties and
relations of things only in so far as the things act upon us, in par-
ticular upon our sensory apparatus" (ibid .).8Therefore, any idea of poten-
tially real power is to be excluded from Armstrong's world-picture, which
"contains", as he repeatedly points out, "nothing but particulars and
their properties and relations, neither of which can exist in indepen-
dence of states of affairs" (1983, pp. 83-4). It is thus emphasized that
laws are states of affairs and also abstractions from particulars, which
instantiate them (see ibid., p. 91; cf. 1983, pp. 113, 121-3).
This is a synoptic picture of Armstrong 's theory of laws of nature,
which I take to be, in its general lines, a type of ' moderate realism'
(Armstrong himself calls it a posteriori realism)? as opposed both to
nominalism in its defense of the reality of universals, and to Platonic
realism, in its insistence on the principle of instantiation.
What I must make clear right from the beginning is that in my
criticism of Armstrong the basic assumptions of moderate realism will
not be questioned, because I share with him the Aristotelean-scholastic
belief that universals cannot be up there in heaven but must be brought
down to earth (see 1983, p. 85). In effect, I shall not be concerned with
the question whether there can be any uninstantiated universals; 10 I shall
rather concentrate on the question referring to the character of instanti-
ation and to the nature of the universal per se . What exactly does
Armstrong mean by the "existence" of the instantiated law? Is there

any space left in his theory for the reality of the universal apart from
its presence in its particular instantiations? My claim will be that a careful
consideration of Armstrong's principle of instantiation, in the context
of scholastic philosophy, will open the door to our tracing his leaning
towards nominalism. I I
Instantiation and "Moderate Realism"

My starting point will be the principle of instantiation which asserts
that for "each N-adic universal, D, there exist at least N particulars
such that they D". This idea, which is a dominant feature in Armstrong's
theory of laws of nature, is the descendent of classical anti-Platonic
realism . Indeed, all scholastic philosophers, following Aristotle" in his
opposition to Plato, believed that the only true substances are single, indi-
vidual (KU~ ' E:Kua'tov) concrete objects located in space and time .
Therefore, universals cannot have the character of Platonic forms, that
is, they cannot be individual substances, existing in a world separate from
their particular manifestations. 13 The universal must be found only in par-
ticular things." This is why scholastic realism , such as that of St. Thomas
Acquinas and Duns Scotus is called "moderate realism". We must though
notice that in this view the significance of the universal is not dimin-
ished. For what offers the grounds for the real unity and intelligibility
of individuals is still the universal, which is found both in rebus and
in intellectuP but never in a Platonic transcendental world.
Thus far, I believe, Armstrong is in agreement both with Aristotle
and the scholastics. His theory is built on the assumption of the objec-
tivity of universals, "which however cannot exist independently of
particulars" (l978a, p. XIV). "If universals could exist uninstantiated,
then", he claims, following the line of thought of Aristotle and the
scholastics, "they would be substances, and would require a relation (a
universal) to relate them to their particulars" (1983 , p. 165). A uni-
versal, therefore, "exists if there was, is or will be particulars having
that property or standing in that relation" (l978b, p. 11). It should be
noted, however, that this position represents only the first step taken
by moderate realism in opposition to Plato. The claim that universals
are not ante res but in rebus is not sufficient to solve the problem. For
a most fundamental question is immediately raised: the question
concerning the character of the universal per se. To appreciate the
significance of the issue, a clarification of the meaning of some key-

concepts will be requisite.
Some Vital Distinctions
a. Universal as unum de multis and unum in multis
It is worth noticing that in scholastic philosophy" there is a basic dis-

tinction in the use of the term universal: 17 On the one hand, the universal
means the predicate of many (unum de multis) and on the other, that
which is in many (unum in multis). However, the character of univer-
sality belongs only to the former. Thus, the term universal is properly
used only when one refers to the universal concept or idea in the mind,"
which has an objective foundation in things ifundamentum in rebus). The
latter corresponds to what Aristotle called oootc " and the scholastics,
natura, quidditas, essentia.
Accordingly, granted that Armstrong's concern is with the ontolog-
ical basis of laws of nature (1983 , p. 3), the use of the term universal
is misleading, even for Armstrong himself. As will be indicated, it makes
it easier for him to disengage universality from its most fundamental char-
acter, that of essence; consequently, it also contributes to his neglecting
some vital problems concerning the relation of essence to 'thisness', or
in Armstrong 's terminology, of universality to particularity.
b. Esse Essentiae and Esse Existentiae

An analysis of the above topics can lead to false conclusions, if we
lose sight of a fundamental distinction in Aristotelian-scholastic tradition
between ' being as such ' (esse essent iae) and 'existent being' (esse exis-
tentiae). In scholastic philosophy the notion of being (ens) enters into the
definition of everything," for nothing can be thought of except under this
category. In Scotus the ens communissime suptum" is a metaphysical
genus which is ontologically prior to the ten genera generalissima and
contains with itself the whole universe, both of thought and of physical
reality (esse ration is and esse naturaei." This is the "tlv ~tlv" of
Aristotelean philosophy, the "ens secundum suam eniitatem'i." studied
by philosophia prima. In this context, the particular, individual thing
represents a specifically determined grade of being ," since the reality
of being does not necessarily entail its existence." Nominalists, by
contrast, reject this distinction in their limiting reality in the actually

existent world.
This seems to be true also of Armstrong, who treats in several places
reality and existence as identical terms. The real particular, he claims,
is the existent particular, the real law is the existent law (see 1983,
p. 81),26 which is also characterized as "a thing existing independently
of the mind" (ibid. , p. 55; cf. pp. 59, 126). The consequences of
this seemingly rather innocent confusion will be shown, I hope, in my
discussion of the principle of instantiation.
c. "Essence" and "Existence"

Before we come to this, it is also worth noting another relevant and
equally important distinction, the distinction between "essence" as such
(quidditas) and "existence", which has been particularly emphasized
by St. Thomas," but is also shared, in its general lines by all moderate
realists. All created things are compounds of essence and existence, the
latter belonging properly to the individuals as such, because existence
does not appear to be included in essence." To exist, therefore, is to exist
as an individual, particular substance, and not as a universal. It is only
Platonic forms that have an independent existence; for moderate realism,
and in particular that of Scotus, existence is haecceitas, particularity, actu-
ality with determinations of space and time, while essence (Armstrong 's
universality) considered as such has only a mode of being which is real
but not actual.
We must remember that for the scholastics essence is to existence
as potentiality is to actuality." This is an analogy that Armstrong totally
excludes from his scheme, by rejecting the ideas both of potentiality
and essence. Although he follows the scholastics in assuming the involve-
ment of particularity and universality in all individual existent things,
he nevertheless rejects the Aristotelean idea of essence, which, as has
already been pointed out, in scholastic realism is identical with univer-
sality. Furthermore, he rejects the traditional distinction between essential
and accidental properties," as well as the category of substance," the
latter being an element which is excluded from Aristotle's idea of
essence.f Universals are, thus, characterized as properties and relations
(see 1978a, p. 87, 1983, p. 166). "The properties of a particular", claims
Armstrong, "form a single property. The nature of a particular" (l978a,
p. 44). Essence is nothing but a conjunction of properties. This is in
the line of the doctrines of particularism and naturalism which he is trying

unsuccessfully, as is my hope to show, to combine with his theory of
universals. According to the former, "there is nothing but particulars
having (universal) properties and relations" (1973a, p. 1) and according
to the latter, "nothing else exists except the single spatio-temporal world"
(1983, p. 82).33 This fact deprives Armstrong of the possibility of giving
a consistent realistic account, and more specifically it deprives him of
the possibility of properly investigating the relation between univer-
sality and particularity.
THE PRINCIPLE OF INST ANTIA TION
Keeping in mind the above remarks, we are now ready for a careful
reading of several important aspects of the principle of instantiation.
Its claim is that "for each property, P, there exists (not necessarily now)
a particular, x, such that Px" (1978b, p. 76) ,34 so that "particularity and
universality are both involved in all existence" tibid.; p. 1).35 We must
though notice that a mere assumption of this principle is not sufficient
for the establishment of moderate realism, for the following questions
are immediately raised: (1) what is the relation between essence (uni-
versality) and thisness (haecceitas or particularity)? (2) how does the
essence or common nature become "contracted", or how does it appear
as particular in the thing? (3) can common nature of itself be a this (de
se haec), or, if not, what kind of unity can it have?
a. The Unity of "Nature"

Two points deserve our attention here: the issue of the unity of nature
and that of its particularized character. (1) As we have already pointed
out, all individuals belonging to the same species share a common nature.
Socrates is a man in virtue of something real which is called natura,
namely in virtue of manness, and as Armstrong would put it, this swan
(Fa) is white (Ga) in virtue of swanness (F-ness) and whiteness (G-ness) .
But Socrates's nature as an individual, is numerically different from
Plato's nature, as his nature is a part of him." Can we then say that
Socrates's nature is as distinct from Plato's nature as it is from a stone's
nature? If this is not the case, and if we wish to maintain that Socrates
and Plato (or this swan, Fa, and that swan, Fb) have something real in
common (F-ness), then this real cannot be a de se haec (singular); it
82 DEMETRA SFENDONI -MENTZOU
cannot be a numerically one, or identical in Armstrong 's terminology (see

1978a, p. 87); it must therefore have a unity which is neither numer-
ical nor merely conceptual. This is a specific type of unity which Scotus
characterizes as "less than numerical" but none the less real."
In his discussion of the character of what in scholastic philosophy is
called essence, Armstrong rejects Scotus's idea of "lesser unity": "I
find Scotus's doctrine of ' lesser unity' difficult to understand. It seems
to imply that the identity of particulars is different from the identity of
essence" (1978a , p. 87). For Armstrong "a property is strictly identical,
strictly the same, in all its different instances" (ibid., p. 83). "Identity"
he claims, "is just identity " (1978b, p. 112).38 Thus, "the identical uni-
versal is present in each instantiation" (1983, p. 79) and therefore "the
law is fully present in each instance" (ibid., p. 102). If this is so, then
Armstrong is in trouble; he has to provide an answer to a serious point
raised by the scholastics and expressed by Scotus as follows : "If the
numerically one nature were fully present, in Dobbin, then he would
not be a horse but horseness itself"." I am afraid that this is a thorny
point which Armstrong would be unable to answer, unless he claimed
that the full presence of the universal (or law) turns out in a complete
particularization of the universal, if we remember that numerical unity
can only be attributed either to Platonic forms or to particular concrete
objects.
b. Particularization of "Nature"
But this is not the only difficulty that Armstrong is faced with; he must
also provide an answer to the question, how the common nature (F-ness) ,
which is not a de se haec (singular) becomes, in a sense, a part of the
individual (Fa); and consequently, how are laws of nature instantiated?
This is a crucial problem and calls for our special attention. At this
point Armstrong's answer is very close to that of Scotus . As the subtle
Doctor claims, common nature is "contracted" in the individual, by
haecceitas, so that it becomes, in some respect, identical with it." This
is one of the main reasons that Scotus has been accused by C. S. Peirce
for his nominalistic attitude: "Even Duns Scotus", claims Peirce, "is
too nominalistic, when he says that universals are contracted to the mode
of individuality in singulars, meaning, as he does, by singulars ordinary
existing things" (8.208).41
Indeed, in Scotus's idea of contraction there is an emphasis on the
particularization of the universal by haecceitas, which must be under-

stood as the ultima realitas" of the individual and which gives it its actual
existence. The particular exists in virtue of its haecceitas which makes
it this particular thing (Fa) and not something else (Fb) . As a result it
is contradictory for Socrates's nature to exist in Plato. Each particular
thing , he claims, is a compound of quiddity, which is the universal
element and haecceity, which is the "entitas singularis'i."
c. Universality and Particularity
The idea of the compound of essence and particularity is something

shared by all moderate realists, including Armstrong." So they all have
to face the problem of defining its character: How is it possible for the
individual thing to be a composite of two elements, and at the same
time be numerically one? Are particularity and universality completely
united , or can they retain their distinct character in this combination?
Scotus tried to answer this question by inventing a new term, his famous
"formal" distinction," which has given rise to criticisms about Scotus's
nominalistic tendencies." What is peculiar about it is the fact that it is
very close to logical , but not itself a logical distinction. It is a distinc-
tion a parte rei, i.e., having an objective ground in reality itself, which
however is not physical, since essence and thisness are not two separate
things but are "different moments or aspects" of the same particular.
Armstrong's criticism at this point is that although Scotus's solution
of "lesser unity" seems to be the most satisfactory one, still it is puzzling,
because he talks of "a mere .formal' distinction" . "Though distinct",
he claims, " their union is closer than relation" (l978b, p. 3). This is an
interesting point witnessing Armstrong's leaning towards nominalism.
He is not satisfied with formal distinction, which is already too nomi-
nalistic , but he seems to need an even closer relation between universality
and particularity. If this were the case , then we would have to assume
that the universal becomes particularized in the individual.
Armstrong is not quite clear at this point. He maintains that " partic-
ulars and universals do not stand in splendid isolation from each other "
(l978a, p. 80). However, he does not appeal to any of the traditional
terms to define their relation. The only characterization appearing in
his writings is that of a "state of affairs". "Particulars and universals",
he claims, "are found only in states of affairs" (ibid., p. 8), the defini-
tion being that of "a particular's having a property". What is universal,
F-ness or G-ness, cannot exist but as instantiated or particularized,

Fa, Fb, Fe .. . or Ga, Gb, Gc . . . But this cannot be the end of the
story, for the most fundamental question for any version of realism has
not been answered yet , the question as to the nature of the universal
per se. Is the universal exhausted in its instantiations, or shall we claim
that its reality is wider and much richer in content than that of its
manifestations?
d. The Character of "Nature Per Se"
If my reading of Armstrong is at all correct, we must assume that he is

rejecting the second case, although his theory is based on the assump-
tion of the reality of universals. My contention, thus, is that at this
point the scope of his realism is limited to a substantial degree . To be
more specific: as already indicated, Armstrong rejects both Scotus's
"lesser unity" and "formal" distinction. His alternative to the former is
" total" identity and to the latter a much closer relation than that of
the "formal" one . This opens the back door to nominalism, which
indeed finds a place in Armstrong's world-picture, where "nothing else
exists except the single, spatio-ternporal, world, the world studied
by physics, chemistry, cosmology and so on" (1983, p. 82; cf. p. 126).
Laws of nature are not simply relations between universals but "relations
between actually instantiated universals" (ibid. , p. 112),47 and the "law
is fully present in each instantiation" (1978a, p. 89),48 so that we
are forced by Armstrong's own analysis to draw the conclusion that
law becomes fully particularized, leaving no space in his scheme for
universality per se.
Furthermore, it is important to notice that the impression of
Armstrong's nominalistic inclination is strengthened by his repeatedly
characterizing laws of nature as "states of affairs". This is indeed
confusing," not only because this characterization cannot fit in the
traditional framework, but particularly because it has already been used
by Armstrong in his description of the relation pertaining between par-
ticularity and universality in the individual thing . In that case a "state
of affairs" was defined as "a particular's having a property".
Armstrong tries unsuccessfully, in my opinion, to overcome this dif-
ficulty by drawing a distinction between a " thin" and a "thick"
particular." A "thick" particular is defined as "a thing taken along with
all its properties", while a "thin" particular as "a thing taken in abstrac -
tion from all its properties". Furthermore, Armstrong makes clear that
in his notion of a "state of affairs" he is using the conception of a
particular in the latter sense. But remember, a "state of affairs " is defined
as "a particular's posse ssing a certain property, or two or more particu-
lars standing in a certain relation" (1978a, p. 80). If in this definition
we substitute the definition of the "thin" particular for the simple term
"particular" the result will be the following: A state of affairs is "a
particular in abstraction from all its properties, possessing a property,
or two or more particulars in abstraction from all their properties standing
in a certain relation". What becomes obvious here is the fact that the
relation expressed by "states of affairs" is still a relation between par-
ticularity and universality. This unavoidably leads to the conclusion
that in Armstrong there is no clear difference between what is taken to
be a relation between particularity and universality and what is taken
to be a relation between a universality and universality.
My claim is therefore that Armstrong's deep concern not to lose
sight of the fact that universals exist only in their instantiations blocks
his ability to see that part of the reality of universals, which is also
objective but not limited to the mode of actual existence in the spatio-
temporal world. This goes hand in hand with his exclusion of potentiality,
which, as I hope will be shown, is an inseparable element of univer-
sality per se.
e. The Metaphysical Mode of Univ ersal
For our purposes, it will be helpful to refer to an extremely significant

distinction between the physical and the metaphysical aspect of universal
made by Scotus" in the spirit of Aristotelean metaphysics: although
the unity of essence (genus, species) is less than numerical, claims Scotus,
and although essence cannot exist apart from the unity of the individual
substance, yet it is logically and metaphysically prior to it.52 Therefore,
the essence per se cannot be identical with the individualized nature.
In other words, although the nature exists objectively only along with
its individuating conditions, logically and metaphysically it is prior to
these conditions, as absolute quiddity is indifferent to singularity and
universality.53
We thus arrive at a crucial point concerning Armstrong's instan-
tiated laws. If it is admitted that the universal, which is combined
with the particular, ends up in the existence of its instantiation in the
spatio-temporal world, then in what respect would this universal be dif-

ferent from its particular instantiations? It must be made clear, therefore,
that the universal in rebus has not one but two modes of being: the
physical and the metaphysical, the latter being the most fundamental one.
A more careful consideration of this distinction is likely to throw further
light on our analysis of the nature of law.
We have already seen that the logical universal, which is also char-
acterized as unum de multis, has an objective basis in things. This is what
Scotus calls the physical universal, which is the common nature con-
tracted in the individual substance by haecceitas, or what Armstrong
means by an instantiated universal (Fa) , and respectively an actually
existent law. This is where Armstrong 's analysis ends, for he does not
seem to be interested in tracing any further either the reality of univer-
sals or the relation of "contracted" nature both to the universal in the
mind and the universal per se. Thus he fails to realize what is extremely
significant in the context of scholastic philosophy (which is also in line
with Aristotelean thought), that the "contracted" nature is only a possi-
bility of a universal, i.e., it is virtually or potentially a universal." in
the mind, for to be predicable as one thing of many (unum de multis),
it must be given a numerical unity and this is the work of the intel-
lect." Furthermore, Armstrong fails to see that the instanti ated universal
(Fa) can only have a limited mode of reality, while common nature,
considered per se, is unlimited and indeterminate.
We thus come to the basic features of the metaphysical mode of
essence or common nature as such . The first thing to remember is that
common nature per se is not a res but a realitas, which is neither par-
ticular not universal ;" it exists objectively only with its individualizing
conditions, and is conceived by the intellect as universal. It is the mind
itself which endows it with its character of universality, for of itself it
is indifferent," a mere natura: "Horseness is just horseness, neither of
itself one nor many, neither universal nor particular"." Accordingly, it
does not point to any particular individual case and this is exactly what
constitutes the indeterminacy of the ens reale r ? "it can be real without
being determined to exist in anyone thing"." We are thus led to the
fundamental character of essence or universality per se. It can be defined
as a mere possibility of existence, just as when it has become existent
it is the possibility of a universal.
TOWARDS A POTENTIAL-PRAGMATIC REALISM
If a development along these lines is acceptable, we can draw the con-

clusion that the involvement of potentiality in the very idea of universality
is such that a consistent account of properties of things in terms of uni-
versals cannot be given apart from this idea. Furthermore, if it is admitted
that the neces sity of laws of nature is grounded in the nature of things
(F-ness or G-ness), then the characteristics of the metaphysical mode
of universality, which have emerged in the preceding analysis, can also
apply to laws of nature. We are thus led to a world-picture which
combines the Aristotelean emphasis on the idea of potentiality" with
the Peircean account of laws grounded in the category of Thirdness."
In this picture reality is not restricted to the realm of actual, individual,
instantiated things; for there is much more to it, which is real though
not completely determined "since not presenting itself in actually objec-
tified form" (6.365), as a concrete existence in space and time . This is
the mode of potential being, implying a "capacity for realization", or
"some inherent tendency to actuality, which if not thwarted, leads to final
completeness of being" (6.365). As C. S. Peirce expresses it, " the will-
be's , the actually is's and the have beens are not the sum of the reals .
They only cover actuality. There are besides would-be's and can be 's
that are real" (8.216) . Hence, besides the nominalist's world of actually
existing things and properties, all of which come to us through our senses,
there is plenty of space for properties of things, interpreted as disposi-
tions, liabilities, tendencies, propensities, as well as for processes taking
place in the sub-atomic world, all of which could be taken as instances
of a transition from a state of potential to that of actual being. The
transformations of nuclei into one another, the mutual annihilation of
matter and anti-matter and their conversion into radiant energy, the trans-
formation, in radio-activity, of a neutron in the nucleus into a proton
without any external energy being supplied, and finally, the real ity of
quarks, which are never found as actually existing free particles, can
all be seen as manifestations of potentiality in the physical world .
However, as fascinating as the idea of potentiality in the realm of
elementary particles is, I will confine myself, for reasons of space restric-
tions," to a consideration of properties of things, such as hardness, and
of dispositions, such as brittleness, fragility, solubility. What is signifi-
cant about them is that they can be ascribed as real, even if they are
not actually tested in the case of properties or even if they are not actually
manifested per se, in the case of dispositions. This view has many
affinities to the theory of "causal powers" propounded by R. Harre and
E. H. Madden'? in those respects that the latter establishes a connec-
tion with some Aristotelean and scholastic concepts, such as the reality
of natural kinds and the role that the nature of things plays in natural
necessity. Finally, a commonly posed idea is the explanation of causality
through powerful particulars which leads to a model shifting away from
the Humean billiard ball paradigm.
I must say, though, that the very idea of "causal power", as some-
thing that can be experienced, is not what I am interested in, my concern
being with the categorically unobservable which can offer the right place
for the introduction of the idea of potentiality." That is why what is
extremely welcome from the theory of "causal powers" is the view that
the "concept of power catches what might be called the strong sense
of potentiality or potency", namely, "what would happen as a matter of
course, if interfering conditions were absent or taken away" (p. 12).
In this respect, the analysis of qualities, dispositions, tendencies can
be given in the form of subjunctive conditionals, or in Peircean terms
of ' would-be's?" of possible behaviour. For what a subjunctive condi-
tional is about is a possibility or potentiality grounded in the nature
of things and referring not to what is contrary to fact , as is the case
with Armstrong 's counterfactuals,"? but to the as yet unactualized.
In this respect, potentiality is to be understood as that state of capa-
bility, disposition, or liability to suffer or produce change in virtue of a
thing's nature. To say for instance that the diamond is hard is to assert
that it would resist pressure when scratched (8.208). For "hardness",
as Peirce claims, "is not invented by men , as the word is, but is really
and truly in the hard things and is one in them all, as a de scription
of habit, disposition, behaviour" (1.27 n. 1). This is exactly the crucial
point, I believe, that offers the real ground for a passage from a realism
about properties to a realism about laws of nature. " Seen from this
point laws can be said to embody in a stronger sense the charac-
teristics of indeterminacy and potentiality, which have already been
seen in relation to properties. A further analysis of these features will
serve in my relating them to laws of nature and also in draw ing my
concluding remarks.
Let me start with indeterminateness.F' As already indicated, nature
or essence is indeterminate with respect to particularity. Although it is
true that "there is no man unless there is some particular man", nonethe-
less "There is a real difference between man irrespective of what the other
determinations may be, and man with this or that particular series of
determinations"." The same can be said of laws of nature. The statement,
all swans are white , does not point to any particular swan . This is exactly
why, not withstanding the fact that there is no Fa unless there is F-
ness, law can be real without being determined to be instantiated in
any particular case .
Hence, law can be characterized as a "potential aggregate'?' which
is irreducible to any collection of actualities; it only "contains general
conditions which permit the determinations of individuals" (6.185) . Thus,
the content of the statement, "all men are mortal", cannot be exhausted
by any multitude of existent things (see, 5.103) , either in the past and
present or in the future , because it applies to an infinity of future men ,
including not only concrete but possible men, as well. In other words ,
"it involves the idea of every possible something"," or, as Aristotle would
put it, "the idea of that of which there is always something beyond"."
The proper way, therefore, to describe law is to assert that it has that
mode of being which is "in potency of determination", "in equal distance
from the physicist who considers it in its concrete determinations and
from the logician who considers it as determined to universality"."
CONCLUDING REMARKS
In light of the preceding analysis, it can be claimed that, instead of trying

to emphasize that aspect of law which appears in its instantiations, it is
far more important for realism to appeal to that mode of being of law,
which represents a possibility of existence, infinitely rich in content,
because of its virtually containing the inexhaustible concrete wealth of
its particular instantiations.
This is the core of the difference between Armstrong's instantiated,
i.e., existent and actually real law, and that type of law, which is essen-
tially "instantiatable", or potentially real as it points to its instantiations.
For the actual manifestations of law are the discrete units which burst
the flux of what is essentially a continuum, the very nature of which
involves the absence of distinct individuality. " And since there is no limit
as to the number of its instances, a law of nature can be described both
in an Aristotelean and a Peircean fashion as "the potential though not
the realized whole"," which embraces all phenomena as its possible
future manifestations. This view in no way involves a rejection of the
principle of instantiation as is understood in Aristotelean-scholastic

realism . It only involves a shifting of emphasis from Armstrong's "instan-
tiated" to "instantiatable" law.
What I hope to have shown is, at least, that a consistent attempt at a
realist account of laws of nature grounded in universals cannot be given,
if trapped in the narrow domain of actually existent law in space-time;
what is further needed is an appeal to the infinitely rich reality of that
aspect of law, which is in potency of determination, so that the relation
between law and its instantiations will be accounted as a relation of
quidditas to haecceitas, or potentiality to actuality.
Aristotle University of Thessaloniki ,
Greece.
NOTES
* In preparing the final version of this essay , commentaries were taken into account.
My thanks are due to the participants of the Beijing Philosophy of Science Conference
for their helpful comments and suggestions. I am especially grateful to James R. Brown
for valuable conversations on the issue and for critical comments on two earlier drafts .
I am also indebted to Susan Haack and Andrew D. Irvine for their incisive commentary
and useful suggestions on an earl ier draft , read at the Dubrovnik Philosophy of Science
Conference in April, 1992.
I D. H. Armstrong (1983).
2 The question concerning laws of nature is quite recent. Plato was concerned with
forms (!OEaL), Aristotle with essence or nature (EU>O~, ocotc, q),\XJl~) the scholastics
with universalia, in the sense of quidditas or natura. One might say that discussion on
laws of nature started with Hume. However, the first serious attempt at a realist account
of laws of nature, grounded in Aristotelean-scholastic realism, was made by C. S. Peirce .
See for example his, "What is a Law of Nature" in his article under the general title, "Hume
on Miracles and Laws of Nature" (see, 1958, pp. 289-321).
3 In her instruct ive paper (1992) on the relevance of C. S. Peirce ' s scholastic-realism
to contemporary discussions in Philosophy of Science, S. Haack suggests that Armstrong 's
theory of laws of nature be characterized as "nominalistic aristotelianism" (p. 36) . I
must say that, although I share the view that each of these characteristics considered
separately is true of Armstrong , I would not be happy to use them in conjunction, because
they are mutually exclusive.
4 I have elsewhere elaborated the close linkage of Peirce's thought to Aristotelean phi-
losophy . See my (1980) and (1991) .
5 In what follows I will be particularly concerned with Armstrong 's (1983) . However.
when necessary I shall also refer to his (1978a) and (1978b), in so far as his theory of
laws is grounded in his realism about universals.
6 The source of inspiration, as Armstrong reports, for the first formulation of this theory,
was Plato (Phaedo 102-170). However, Armstrong's view is essentially anti-Platonic,
as he himself explicitly claims in several places (see, e.g., 1983, pp. 86, 90) and as will
also emerge from our analysis .
1 See, e.g., (l978a) p. 85, (1978b) pp. 5, 157, (1983) pp. 120-121, 135-136, 168.
Acceptance of "unrealized possibilities", such as a solid lump of gold with a volume greater
than a cubic mile, or a race of white feathered ravens (see, 1983, pp. 17ff.) cannot be
equated with acceptance of potentiality, in its classical sense , as will be shown in what
follows .
x Cf., "The power is constituted the power it is by the sort of actualizations it gives
rise to in suitable sorts of circumstances" (1983, p. 123). As is obvious here , the idea
of power is connected with the immediate experienced results .
9 Armstrong's "a posteriori realism", is the view that "just what universals there are
in the world, is to be decided a posteriori on the basis of total science" (1983, p. 83;
cf. 1978a, p. xiii).
10 As Tooley (1977) has made a substantial and fruitful effort to show . However, an
analogous answer as to the potential character of laws will eventually be provided -
although through a very different path.
11 An incisive criticism of the inconsistencies involved in Armstrong's attempted
marriage of universals with naturalism, as it appears in his principle of instantiation, is
made by A. Irvine in his, " Universals in rebus or in contradictione'l" (unpublished paper,
sent to me by the author recently) . Irvine's claim is that Armstrong's view seems to fall
victim to a traditional difficulty of the "one over many" : How can the universal be both
completely and not completely present within a certain region of space-time? or how
can it have multiple location at all? Irvine, who seems to share with Armstrong the view
that universals are indeed located in space-time can see no solution to the problem of
instantiation. By contrast, it is my contention that a promising way to get out of the dif -
ficulty, while remaining faithful to the basic assumption of the universals in rebus, can
be found by an appeal to the Aristotelean-scholastic treatment of the universal per se.
12 See, e.g., Arist. Met. 1029a 27-28; Categ. 2a II.
IJ See, e.g., Arist. Met. 1086b 2ff.
14 See, e.g., ibid ., 1020a 33, 35b 2, b6, 1045a 23-35 .
15 Avicena, Logica, fol. 2v. b (ed. Venice, 1508): "Essentiae vero rerum aut sunt in
ipsis rebus, out sunt in intel/ectu, unde habent tres respectus".
16 The contribution of Harris (1959a, 1959b), Boler (1963), Copleston (1953) and Moore
(1950) to my understanding of some fundamental scholastic subtleties crucial to the
purposes of this paper has been invaluable. More specifically my debt to E. Moore
(1950, 1952, 1968) for his pioneering emphasis on the idea of potentiality in C. S. Peirce 's
scholastic-realism has been acknowledged earlier in my work.
11 Actually, there is a three-fold distinction of the universal in scholastic philosophy.
See, e.g., St. Thomas, Comm. in Sentt., Lib. II, dist. iii, q. 2, art. 2: "Est triplex univer-
sale: quoddam quod est in re, seu natura ipsa, quae est in particularibus .. . est etiam
quoddam universale quod est a re acceptum per abstractionem . . . Est etiam quoddam
universale ante rem, quod est prius re ipsa".
18 See, e.g., Scotus: Rep. Par. II. dist. ii, q. 5, n. 12: "Dico quod universale in actu
non est nisi in intellectu, quia non est actu universale, nisi sit unum in multis et de multis".
19 See, e.g., Arist. Met. 1029b 14-17 . It is important however to be indicated that in
Aristotle there is no specific word corresponding to the scholastic term 'e ssentia' , The
meaning of the word 'essence ' can be applied to a large number of Aristotelean terms,
one of which is ooot«.
20 Scotus, Quaest. de Anima, q. Ixvi: "sed ens, quod est universalissimum , intrat defi-
nitionem omnium . . . ."
21 Scotus, De Rerum Principio , q. xvi, n. II: " Et isto mode ens communissime sumptum
est genus metaphysi cum ad Creatorem et creaturam".
22 Scotus, Quaest. super Elench ., q. i: "Ens est duplex, scilicet. naturae et rationis ".
23 Ibid.: " Logica est de ente communi , sive considerat ens in communi . . . Logica est
de hujusmodi intentionibus, quae sunt applicabiles omnibus rebus" .
24 See Harris (I 959b), p. 32.
25 The above analysis of being is definitely Aristotelean. See, e.g., Met lOO3a 21-32.
26 Cf., 1978a, pp. xiii, 10; 1983, pp. 40, 53, 55, 59, 98,104,105,112,117,120,121,
126, 129, 167.
27 See, St. Thomas, Summa Theol . Ia pars. q. 39 art. 2.
2R "SO esse denotes simply the act of being, existence being denoted by ipsum esse" :
E. Gilson (1957), p. 38.
29 See, S. Thomas, Summa Theol . Ia pars, q. 39, art. 2. In this respect, Scotus shares
with St. Thomas the view that matter stands to form as potency to act. He also defines
form in Aristotele an fashion as the actuality of matter. However, matter for Scotus is
not only potentiality; it also has an actuality of its own apart from its mere potential
existence: "Forma communicat materiae suam actualitatem et suam actum essendi et suam
operationem" (Scotus, De Rerum Principio , q. ix, n. 53). See Harris (I 959b), pp. 3, 4,
91; Copleston (1953), pp. 31, 51-52.
30 See, for example, Arist. Met. 1030b 12ff.
31 See, ibid., 1028a 13ff.
32 See, Arist. Met. 1088a 22.
33 More of this below.
34 Cf., 1978b, p. 156; 1983, pp. 82, 91, 97, 165.
35 Cf., "particularity and universality combined in states of affairs" (1983, p. 122).
36 Scotus, Metaph ., vii, q. 13, n. 21: " humanitas, quae est in Socrate, non est human -
itas, quae est in Platone . . . ."
37 Scotus, Rep . Par. II, dist. ii , q. 5, n. II : "Dico ad quaestionem concedendo .. .
quod est unitas extra animam minor quam numeralis, ut specijica . . . ."
3R Cf., 1978a, pp. 75, 77, 108, 115; 1978b, pp. 66, 95, 112; 1983, p. 83.
39 See, Boler (1963), p. 49.
40 Scotus, Op. Ox. II , dist. xlii, q. 4, n. 7: " universale in sinqulari non est aliud quam
singulare". See, Boler, p. 63.
41 References to this form are to the C. S. Peirce (1939-58). 8.208, for instance, refers
to volume 8, paragraph 208. See also, Scotu s "inclined toward nominalism " (1.560);
defended a "halting realism" (6.175); was separated from nominalism "by a hair" (8.11).
42 Scotus, Op. Ox. II , dist, iii, q. 2, n. 15: " ergo ista entitas non est mater ia vel forma
nee compositum, inquantum quodlibet istorum est natura , sed est ultima realitas entis,
quod est materia et quod est forma vel quod est compositum",
43 Scotus, Op. Ox. II, dist. iii, q. 6, n. 15: " Sed haec est formaliter entitas singularis,
et illa est entitas natura e formaliter".
TH E REALI TY OF THIRDN ESS 93
44 See , for example Armstron g (1983), p. 122.

45 SCOlUS, Op. Ox. II, dist. iii, q. 6, n. 15: "Sed haec est form aliter entitas singularis,
et il/a est entitas naturae fo rmaliter".
46 See, Harr is (1959 a), p. 50.
47 cr., 1978a, p. 113; 1978b, pp. 130, 131; 1983, pp. 82, 112, 135, 172.
4~ Cf ., 1983, pp. 79, 102.
49 Armstrong seems to realize the diffi cult y in his mak ing a distinct ion, although not
quit e often , between "states of affairs ', i.e., tokens, and "sorts of states of affairs", i.e.,
types (see, 1983, p. 98; cf. p. 172).
50 See, 1978a, pp. 114ff.
51 See, SCOluS, In Metaphys. I, 7, q. 18, n. 6; Op. Ox. 2 d. 3, q. I , n. 7; Super Praedieam.
q. 2, n. 2. For a discu ssion of the metaphy sical mode of univer sality, see, Boler (1963) ,
pp. 56ff. See also Claudin Engel-Tiercelin (1992) , pp. 59ff.
52 SCOlu S, Op. Ox. II, dist. iii, q. I: "Licet nunquam sit realiter sine aliquo istorum,
non tamen est de se aliquod istorum , sed est prius naturaliter omnibus istis",
53 lbid., q. i, n. 7: "nee ex se una. nee plures, nee universalis nee particularis",
54 SCOluS, "est ergo natura in potent ia remota ad determinationem singularitas et ad
indeterminationem universalitas", In Metaph. I, 7, q. 18, n. 8; cf. Op. Ox. II, d. 3, q. I,
n. 9. See , E. Gilson (1952), pp. 536ff., Moore (1964) , p. 405, and Boler (1963), p. 60.
55 SCOluS, Quaest. in Lib. Periherm, Lib. I, q. viii: "Int ellectus enim in sua operatione
non dependent ab existentia rei. cum et defin itiones et demonstrationes fiant de non
existentibu s, multo magis non existentibus potest attribuere intentiones secundas, quae
omnino ab intel/eetu causantur".
56 SCOluS, Quaest. de Anima, q. xvii, n. 14: ".. . quae quantum est de se nee est uni-
versalis nee singularis, sed de se est indifferens, et tale est objec tum intellectus directum",
57 SCOluS, Op. Ox. II, dist, iii, q. 1: "Non solum ipsa natura est indifferens ad esse in
intellectu et in partieulari .. .."
5~ Avicenna, Logica, fol. 12 r. a: "in se autem hujus nee est universale nee est singu-
lare. Si enim esse t univesale, ita quod anima litas ex hoc. quod est animalitas, est
universalis, oporteret nullum animal esse singulare".
59 For an illuminating analysis of the idea of indeterm inacy see E. Moore (1964) , pp.
406ff., Fred Michael (1988), pp. 334-335 and Claudin Engel-Tiercelin (1992) especi ally
pp.57-70.
60 See, Boler (1963) , p. 50.
61 See , e.g., Arist ., Met. 1045b 25-1046a 3, 1049b 9-11 ; Phys. 20la 10-13, 257b 7-8,
225b 34-37, 261a 32-36.
62 For a fuller discu ssion of Peirce ' s three ontol ogical categories, see my (1980), pp.
80ff.
63 For a more extended treatment of the role of potentiality in Q. M. and Physics of
elementary particle s, see my ( 1993).
64 R. Harre and E. Madden (1975 ).
65 I have developed this in some detail in my (1985-88).
66 In Peirce ' s ontologi cal scheme the ' would-be ' is another expression for potentiality
and it is interrelated with gene rality and law of habit , which is expressed by a condi-
tional proposition. For a fuller discussion of the concepts of ' would-be ' and 'habit' see
my (1980 ), pp. 70ff., 107ff.
67 It is to be noted that all statements of counterfactua ls Armstrong refers to are state-
ments of uninstantiated law, which "a ssert only that if, contrary to fact certain sorts of
things existed, then these things would obey a certain law" (1983, p. 126; cf., p. 137).
"The statement that, it is a law that Fs are Gs supports the counterfactual that if a, which
is not in fact an F, were to be an F, then it would also be a G" (1983 , p. 46).
68 For a fuller discussion of this connection see my (1992) .
69 The three-fold relation of indeterminacy-potentiality-probability, and its plac e in
C. S. Peirce 's categories of Firstness and Th irdness has been elaborated in my 1980 and
1993.
70 C. S. Peirce (5.312); cf ., Arist. De Caelo 278 a 7-12: " if we supposed that
there were but one circle , none the less to be a circle and to be this circle would be dif-
ferent ; the one would be the form the other would be form in matter and would be a
particu lar".
71 See, C. S. Peirce (6.185-188).
72 For an illuminating discussion of the issue, see F. Michael (1988) , pp. 344-345, and
C. Engel-T iercelin (1992) .
13 Arist. Physics 207a 1-2: "ot ad n 'ESW \!on". This is how Aristotle characterizes
the idea of infinity (ll1t£lpOV), which is treated by him as a potential reality : "OtlVd /lEl
ElvUl to ll1t£lpOV" (Physics 206a 19). The same is true of the Aristotelean infinite time,
which is real by continuously coming into being. Therefore, time, as a whole, can never
be actually, but only potentially real (see, ibid., 206a 8ff.). This is an extremely enlight-
ening idea that could shed light on the issue of laws of nature and in particu lar of
" instantiatable" laws. However this is a topic that deserves special treatment.
74 This is how E. Gilson (1952 , p. 113), describes the metaphysical mode of universality.
75 See, C. S. Peirce (4.172): ..It is only actuality, the force of existence, which bursts
the fluidity of the general and produces a discrete unit . . . the possible is general,
and continuity and generality are two names for the same absen ce of d istinction of
individuals".
76 Aristotle , Physics 207a 21-3: "to OtlVd/lEl l:lAOV, evt€A€X€~ o'ou".
BIBLIOGRAPHY
Aristotle (1933-35). Metaphysics, 2 vols. Loeb ed. Translated by H. Tredennich. Harvard

University Press.
Aristotle (1929-34). Physics, 2 vols. Loeb ed. Translated by P. H. Wicksted and F. M.
Conford . Harvard University Press.
Armstrong, David M. (l978a). Nominalism and Realism (Univer sals and Scientific
Realism , vol. I). Cambridge University Press .
Armstrong, David M. (I 978b). A Theory ofUniversals (Universals and Scientific Realism,
vol. 2). Cambridge University Press.
Armstrong , David M. (1983) . What is a Law of Nature ? Cambridge University Press.
Boler, J. F. (1963) . Charles Peirce and Scholasti c Realism. University of Washington
Press.
Copleston, F. S. J. (1953). A History of Philosophy (Medieval Philosophy , vol. 2). Image
Books Edition (1962) .
Duns Scotus, J. (1851-95). Joannis Duns Scoti Opera Omnia 26 vols. Vives, Paris.
Duns Scotus, J. (1908). Duns Scoti Capitalia Opera. R. P. D. Marie (ed.). Le Havre .
THE REALITY OF THIRDNESS 9S
Engel-Tiercelin Claudin (1992) . ' Vagueness and the Unity of C. S. Peirce 's Realism',
Transactions of the e. S. Peirce Society 28, pp. 51-82.
Gilson, Etienne (1952). Jean Duns Scot. Vrin.
Gilson, Etienne (1957). The Christian Philosophy of St. Thomas Aquinas . Translated by
L. K. Shook. Victor Gollancz.
Haack, Susan (1992). ' '' Extreme Scholastic Realism": Its Relevance to Philosophy of
Science Today' , Transactions 28, pp. 19-50.
Harre, R. and Madden, E. (1975). Causal Powers. Blackwell.
Harris, C. R. S. (1927a). Duns Scotus (The Place of Duns Scotus in Medieval Thought,
vol. 1). The Humanities Press (2nd ed. 1959).
Harris, C. R. S. (1927b) . Duns Scotus (The Philosoph ical Doctr ines of Duns Scotus,
vol. 2). The Humanities Press (2nd ed. 1959).
Irvine, A. D. (1992). ' Universals in Rebus or in Contradictione?'. Unpublished.
Michael, Fred (1988) . ' Two Forms of Scholastic Realism in Peirce's Philosophy',
Transactions 24, pp. 317-348.
Moore, E. C. (1950). Metaphysics and Pragmatism in the Philosophy of C. S. Peirce (Doct.
Thesis). University of Michigan Press.
Moore, E. C. (1952). ' The Scholastic Reali sm of C. S. Peirce' , Philosophy and
Phenomenological Research 12, pp. 406--417.
Moore, E. C. (1964) . ' The Influence of Duns Scotus on Peirce ', in E. Moore and R.
Robin (eds.), Studies in the Philosophy of C. S. Peirce. The University of Massachusetts
Press.
Moore, E. C. (1968). ' On the World as General' , Transactions 4, pp. 90--99.
Peirce, Charles S. (1931-58). Collected Papers of Charles Sanders Peirce (vols. 1-8) .
C. Hartshorne, P. Weiss and A. Burks (eds.). Belknapp Press of the Harvard University
Press.
Peirce, Charles S. (1958). Selected Writings (Values in a Universe of Chance). Philip P.
Wiener (ed.), Dover 1966.
Sfendoni-Mentzou, Demetra (1980). Probability and Chance in e. S. Peirce's Philosophy
(Doct, Thes. in Greek). University of Thessaloniki.
Sfendoni-Mentzou , Demetra (1985-88). ' A Realist Explanation of Causality Through
"Causal Powers" " Scien. Ann. School of Phil. of Thess., pp. 245-255 .
Sfendoni-Mentzou, Demetra (1991). ' Towards a Potential-Pragmatic Account of C. S.
Peirce's Theory of Truth' , Transactions of the e. S. Peirce Society 27, pp, 27-77.
Sfendoni-Mentzou, Demetra (1992). 'Is there a Logic of Scientific Discovery? A Pragmatic
Realist Account of Rationality in Physical Theory', in M. Assimakopoulos et al. (eds.),
Historical Types ofRationality. National Technical University of Athens, pp. 239-250.
Sfendoni-Mentzou , Demetra (1993). 'The Role of Potentiality in C. S. Peirce's Tychism
and in Contemporary Discussions in Q.M. and Micro-Physics ' , in E. Moore (ed.),
Charles S. Peirce and the Philosophy of Science: Papers from the 1989 Harvard
Conference. The University of Alabama Press.
Tooley, M. (1977). 'The Nature of Laws ', Canadian Jour. of Phil. 7.
ARCHIE 1. BAHM
TENTATIVE REALISM
'Realism' is an epistemological theory claiming that some objects of

knowledge exist whether or not they are apparent. In everyday practice,
every person is a realist. [Little observation is needed to conclude that
all other anim als are realists also. The biological evolution of humans
was possible only because their physiological organisms were adapted
to sensing and responding to real environmental beings and processes.)
Hum an beings are naturally naive realists. Doubts about accepting
apparently real objects as "really real" occur when apparent error appears .
Error appears when something appears to be so and not so at the same
time. A popular example is a glass of water with a pencil inserted.
Visually the pencil appears as broken at the water surface but unbroken
to a descending finger. Apparent contradiction often results in uncertainty
about whether to trust what is apparently real. Repeated experience
with apparent errors tends to generate some distrust about what appears.
Some persons, and theorists, have jumped to conclusions that no
apparent reality should be trusted . Some consequent anti-realistic theories
include Subjective Idealism, Phenomenalism, Romanticism, Existen-
tialism, Scepticism and Agnosticism. Some persons, and theori sts, have
ignored error as a subjective deficiency and have continued to be realists,
some naively uncritical and some formulating theories about the nature
of the real. Speculations about the nature of the real have generated
several varieties of metaphysics, some advocating existence as constantly
changing and some as unchangeable, some as consisting of uncuttable
particles and some as a cosm ic whole , some as completely determined
(even predetermined) and some as resulting from pure chance, some as
structure by eternal forms (Platonism) and some as evolving evolu-
tionarily (Spencer), some as inspired by an elan vital and some as
designed by an external creator. The history of the physical sciences
has produced several different conceptions, including preferring an infi -
nitely powerful Big Bang to an Omnipotent Creator. None of these has
provided a final solution to originating doubts .
I propose that we can, as we must, trust what appears as real as long
as it continues to appear real. People survive by adapting to their envi-
97
R. S. Cohen . R. Hilpinen and Qiu Ren zong (eds.), Realism and Anti -Reali sm in the
© 1996 Kluwer Academic Publisher s.
98 ARCHIE 1. BAHM
ronments by developing both habits of trusting regularly recurring real-

istic appearances and habits of being ready to change and adjust to
whatever appears as really new. [Intelligence is ability to adapt both to
changing apparent realities and to unchanging apparent realities.] When
something new appears, common sense response formulates some idea
about it and some action to test its nature. If the idea and test seem
satisfactory, then, when such novelty recurs , the same idea and test will
be reused. As long as such idea and test continue to satisfy, the idea
will remain as acceptable knowledge.
But if such idea later becomes unsatisfactory and involves some other
error, habitual confidence in it ceases and some new idea and test are
undertaken. Pragmatists have summarized the theory that those ideas
which work are true, and survive, and those ideas which fail to work
are false, and perish . Involved in this view is that an attitude of tenta-
tivity not only is desirable but should be recognized as a normal condition
of reliable knowing. The word 'tentativity' means "holding on", both
"holding on tightly as long as an idea works" and "holding on loosely
enough to be willing to let go whenever the idea fails to work".
I propose that 'Tentative Realism' is the best solution to the problem
of the nature of reliable knowledge and as a basic assumption of scien-
tific methods. Some will prefer the term 'Pragmatic Realism' or simply
'Pragmatism ', although the meaning of this word has been misused so
often that misunderstandings may prevent its accept ance .
Is knowledge realistic or subjectivistic? It is both . All knowledge is
mental. No non-mental knowledge exists. All objects of knowing are
mental because they appear in awareness as objects. Some apparent
objects are merely apparent, as those in dreams, day dreams, and ide-
alized objects of desire . Some apparent objects appear as real, i.e., as
existing when not apparent. Although, since any actually existing real
thing cannot enter into a mind and its appearance, as long as what appears
appears as real , and as long as a person trusts one's appearances, the
idea that the apparently real object is actually real 'works ', and thus
realism is thereby assumed . Problems that are involved in such assump-
tion are many.
Is knowledge of apparently real objects certain or uncertain? As naive
realists, we are uncritical of appearances and thus remain without feeling
uncertain about them. But when appearances contradict, rational doubt
begets distrust of appearance. Persons, and scientists, then try to design
tests for deciding whether one or the other or both of the appearances
TENTATIVE REALISM 99
seem trustworthy. Tentative Realism accepts the fact that knowledge,

including scientific knowledge, normally involves both some certainty
(or lack of uncertainty) and some uncertainty. Hence its inclusion of
the term " tentative" in its title.
Is knowledge in each mind private or is it communicable? An hy-
pothesis or conclusion is usually regarded as scientific only if it is
communicable (to one or some specified number of other persons). The
requirement assumes that all minds (or at least the minds of the scien-
tists involved) are alike. But actually, minds are both alike and different,
both in biological inheritance, physiological development, and experi-
ential conditioning after birth . Not only personal experiences but cultural
influences condition minds in somewhat different ways that make
the assumption dubious. Again , although this question may be settled
by some thinkers in terms of previously-accepted theories (biological,
physiological, psychological, sociological, and metaphysical), the fact
that minds are both alike in some ways and different in others needs to
be recognized in any acceptable theory. Tentative realism recognizes
this fact and advocates that the kind and amount of sameness involved
in communication of ideas is itself subject to experimental testing, and
that an attitude of tentativity be sustained regarding this kind of realism
(the existence of other minds as real is assumed) also. Recurrence of
agreement about realistic ideas tends to be regarded as some assurance
that the knowledge is reliable.
Is existence unchanging or changing? That is, does apparent regular
recurrence of an apparently real object as unchanging imply that it is
actually unchanging? Although each person, and each scientist, must rely
on one's own experiences, evidences from other sources demonstrate that
at least some apparently real objects do change in unapparent ways. Such
evidence, and resulting general theories that the existing universe is
entirely unchanging or constantly changing or both changing and
unchanging, has implications for theory of knowledge. Tentative Realism
advocates that the existing universe is both unchanging in some ways
and is constantly changing in other ways . This theory implies that the
principle of tentativity should be employed both with regard to theories
about the existing universe and to implications of such tentativity for
each more specific, and , when possible, for each particular, inference
regarding the reliability of accepting an apparently real object as actually
real.
Do apparent numbers exist only in minds or do numbers exist inde-
100 ARCHIE J . BAHM
pendently of minds? Many scientists claim that "if it can 't be measured,
it doesn't exist". Such scientists seem to assert that numbers required
for measurement of supposedly real things are also real, and some, at
least, that they are more real than the things numbered. Logical and
mathematical realisms have their origins in Platonism. The forms of
things, and numbers that measure such forms, are eternal in nature, and
thus their use implies, and assures us, that, except for any incapacities
of minds for knowing them, our knowledge of real things can be certain.
Tentative realism regards numbers as apparent objects existing only in
minds, and that any inference regarding their existence as real is mistaken.
The fact that apparently real things are enumerable does not imply that
numbers are somehow embodied in them. Whether and to what extent
minds are alike regarding abilities to understand mathematics is itself
something that can be tested experimentally, as is often done by teachers
of mathematics. But the capacity for observing that two apparently real
fingers and two more apparently real fingers always appear as four
apparently real fingers seems to be normally common. And generaliza-
tions and abstractions, all occurring in minds, normally generate the same
inferences regarding numbers. Any logical realist must also assume that
minds are alike in capacity to apprehend numbers as apparently real
and to generate the same inferences regarding numbers. That numbers
are apparent objects is intuited as self-evident. That numbers are appar-
ently real objects involves inference . When, and to the extent that
inferences that numbers are actually real ' work' , and serve other sub-
jective purposes, must be granted by Tentative Realists. But again, all
such 'working' is regarded as tentative.
Is a scientific hypothesis verified by remembered data or by future
data or by both? Empiricist theory claims that the truth of an idea is
appropriately verified by tracing its origins in previous sensations. But
pragmatists claim that such previous sensations no longer exist and so
cannot serve to verify, and that verification of an hypothesis can occur
only by testing it in future experiences in which the data can be intuited
(i.e ., as apparently real corroboration). Since some past experiences of
apparently real objects can be remembered (and the reliabil ity of memory
habits can also be tested pragmatically), and used for purposes of cor-
roboration , Tentative Realism advocates use of both past and future
(and present, for any actual assent to verification occurs in some present)
apparently real data available for verification. When scientists have
TENTATIVE REALISM 101
agreed upon some universal principles as reliable, then they too may
be employed in inferential verifications.
Does science exist only as concern about true theories or also in the
results of applying theories in practice? Some philosophies of science
claim that science is concern for understanding. Any applications and
their consequences are beyond science. Other philosophies of science
claim that the purpose of science is to solve practical problems and to
improve human welfare and that thus scientific theories are not fully
verified until human benefits become obvious. Tentative Realism
observes that some scientific efforts are primarily theoretical and that
some are primarily practical in intent. Practice , as observing apparently
real uniformities, often is a source of hypotheses and their verification.
But also, when conclusions become regarded as reliable, they tend to
generate applications in practice. Both seem normal and necessary to
an adequate theory of science.
If we review the history of science, we can observe that very many
of the hypotheses agreed upon as verified by the scientists of earlier
periods in history later have been proved false . Such review provides
evidence that an attitude of tentativity is needed and that the Tentative
Realistic claim that such an attitude is essential to the nature of science
adequately conceived is verified. [If we include the whole history of
science in our conception of the nature of science, we are reminded
that it consists of more false than true hypotheses.]
University of New Mexico,
U.S.A.
F . WALLNER AND M. F . PESCHL
COGNITIVE SCIENCE - AN EXPERIMENT IN

CONSTRUCTIVE REALISM ; CONSTRUCTIVE REALISM
- AN EXPERIMENT IN COGNITIVE SCIENCE
1. DEVELOPMENT AND EVOLUTION OF

CONSTRUCTIVE REALISM
There are various origins for the development of the ideas of Constructive
Realism - they are coming from philosophical as well as (natural) sci-
entific disciplines. CR , thus, does not stand in a certain philosophical
tradition. Its uniqueness and strength are founded in the fact that its
theses are the result of a process of establishing relations between very
different sources and information. Of course there a multitude of points
of reference can be found; they acted , however, rather as a stimulating
instance for the development of Constrictive Realism. The view of CR
is influenced by:
(a) the ideas of Ludwig Wittgenstein which played an important role
in the development of the function of Constructive Realism in respect
to the sciences;
(b) the second point was the experience with interdisciplinary cooper-
ation and research in the (natural) sciences - by discussing and
investigating the process of science we learned about the implicit and
in most cases unspoken needs of the scientists and sciences of our
time.
Another experience, which was very important for the development
of Constructive Realism, was the debacle of philosophy of science in
respect to its normative claims. On the one hand it seems to be clear
that the reflections being made in philosophy of science are deficient,
if they are only describing the processes of science; on the other hand
the last decades of discussion in philosophy of science have shown
that it is getting more and more unclear, how normative claims can be
legitimized in the context of the sciences. This uncertainty of philos-
ophy of science led to the following development: scientists not only
felt themselves not understood and not in charge of the philosophy of
science, but also they turned to metaphysical philosophy or irrational
ideas.
103
R. S. Cohen , R. Hilpinen and Qiu Renzong (eds.), Realism and Anti -Realism in the
104 F . WALLNER AND M. F . PESCHL
On the one hand we have to see this in the context of the develop-
ment of natural sciences in recent decade s, and on the other hand modern
epistemology did not react in an adequate manner to this development.
Modern technologies, especially the use of modern computer technology
in the process of research, accelerated the trend to instrumentalism (which
already can be found in the basic structure of (natural) sciences); i.e. ,
without having given up the claim of (definite and objective) knowl-
edge this claim has been increasingly circumvented and has been replaced
by contexts of functioning relations. This implies a lot of needs for expla-
nations which remain, however, in most cases unspoken, because they
do not have a direct influence on the development of research. Because
the natural sciences minimize the claim of recognizing the world , the area
of irrational explorations is continuously increasing.
In many cases this has the consequence that irrational strategies of
explanations (i.e., strategic methods are not rationally legitimized) make
use of results or elements of natural sciences . This is completely harmless
if these explanations are only for entertainment. It can have unpleasant
or even dangerous consequences if such explanations are used as nor-
mative claims. In the view of such a dichotomy - between contexts which
can be dominated and irrational explanations - the claims of occidental
science to lead to scientific knowledge are given up . It could be , of
course , that this already has been in the structure of science for a long
time - Heidegger pointed out that occidental science would be in prin-
ciple wrong . In contrast to that the Constructive Realism tries to save this
kind of knowledge by a new interpretation of its procedure, i.e., scien-
tific knowledge is understood as a specific form of results of actions.
At first we want to discuss alternative approaches in modern episte-
mology. An increasing alienation can be found between epistemology
and philosophy of science on the one hand and natural sciences on the
other hand . Naturalistic epistemology took into account this process;
the most famous and most successful is represented by the evolutionary
epistemology. Naturalistic epistemologies take into consideration the
knowledge of modern natural sciences. Hence, they are explicitly doing
something which already can be implicitly found in Kant's philosophy.
In most cases, however, it remains unclear in modern naturalistic epis-
temologies which argumentative role disciplinary results are playing in
their epistemological concepts.
We are assuming cognitive science to be a continuation of natural-
istic epistemologies. Compared to naturalistic epistemologies cogn itive
COGNITIVE SCIENCE 105
science has an interdisciplinary structure - the problem is, however,

that there does not exist a theory yet making clear the methods of coop-
eration of different disciplines. Thus it is not possible to rationally execute
interdisciplinary cooperation. As an implication of this lack of theories,
cognitive science is in a strange ambivalence: the use of computers offers
to a very large extent the possibility of hiding the fact that there are
methodological difficulties. Using computers often hides the possibility
of controlling the relevance and plausibility of data, whereas in the
manual manipulation of data the inadequacy of the used method can
be noticed (which gets clear in the discussion with other disciplines). The
application of computers makes complex theories explicit and thus
ignores or at least is not primarily interested in the critique of the applied
method. The orthodox AI approach, for instance, represents such an
example where the question of functioning has a higher priority than
the question of understanding and basic knowledge.
Having this in mind, cognitive science is often in danger of being
dominated by an instrumentalistic approach - the higher priority is
making perfect the processes of cognition, the basic problems of cog-
nition are completely omitted. Another aim of Constructive Realism is
the development of a methodology of cognitive science as an interdis-
ciplinary cooperation between various disciplines taking part in the
discu ssion of very basic problems concerning, for instance, the questions
of what is knowledge, what is language, what is knowledge represen-
tation, etc. - an answer for these questions is very important, because,
if we want to investigate the process of science, we have to have at
least an idea of what we understand, for instance , by knowledge or how
this knowledge is represented in the disciplines, etc. We will look for
new paradigms in cognitive science as well as in philosophy of science
- the Constructive Realism approach represents the basis for the inves-
tigations and considerations in this dynamic development.
2 . STRUCTURE AND METHODS OF CONSTRUCTIVE REALISM
Looking at the structure of CR we can see that it is standing in a two-

fold position: on the one hand it is positioned between methodology
and ontology; on the other hand it can be found between naturalistic
epistemology and theory of action. What does this mean in detail? We
are investigating methodological rules as instructions for acting and
thus criticize methodologies of single disciplines in the context of theories
106 F . WALLNER AND M . F . PESCHL
of acting. CR develops ontological concepts from the result of natural-

istic epistemologies. It also reveals the implicit ontologies of results from
single disciplines which arise as placebo effect of certain methodolog-
ical strategies.
2.1. The Concept of "Strangification"

Compared to classical metaphysical theories or to theories of philos-
ophy of science of our century, CR refers neither implicitly nor explicitly
to a meta level. The meta level has been replaced by another method-
ological action - it could be compared to a methodological trick: we
are referring to this process as "strangification". This does not mean only
one strategy, but rather a set of strategies having one thing in common:
they are transferring one (logical) system of propositions from their
original context into another context - this system is judged then out
of this context. This process can not be planned in advance, because there
is no metatheoretical standardization - it rather represents a game with
different contexts. Changing the context many times enables us to get
new insights, perspectives and views in the structure of the system of
propositions. The important point is that, if we are looking at these
contexts where the systems of propositions get to be absurd , we are
noticing the implicit assumptions and considerations of this system - i.e. ,
we are investigating its tacit knowledge. J
One could think, for instance, of orthodox Artificial Intelligence which
is strangificated in the context of the humanities; in this moment and
in this context the insufficiency of the (in most cases unreflectedly)
applied methods of computer science (i.e., mathematical logic, etc .) for
modeling or even better for understanding cognitive processes gets clear.
Think, for instance, of the models in cognitive psychology which make
in most cases use of the symbolic processing paradigm of Artificial
Intelligence (e.g., J. R. Anderson).' The consequences which can be
drawn from such a strangification are two-fold: in one case the scien-
tists ignore the results of this strangification (because they pretend that
these results are "not relevant or adequate") or they are consequently
trying to look for an alternative approach which is , of course, more
uncomfortable than the first case.
We are differentiating three types of strangification: linguistic, onto-
logical and pragmatical strangification, excluding contexts in which a
system of statements gets absurd represents the first type of strangifi-
cation. Thus, this method enables the revealing of (implicit) assump-

tions as well as showing up the domain of application of a system of
statements without having to fall back upon metatheoretical standard-
ization instances. Talking in meta levels is not prohibited; it can be
seen rather as an example for the strategies of strangification.
Applying a system (a set) of methods of one discipline to a very dif-
ferent discipline represents the ontological type of strangification. We
are thinking, for instance, of the application of hermeneutical proce-
dures in the natural sciences. Gestalt perception in biology, for instance,
forms the boundary of the application of quantitative methods.
We are also thinking of the introduction of quantitative methods in the
humanities. The aim is not, however - as normally assumed - to make
the humanities more exact or to give them the methods to "correctly"
quantify these phenomena that can be quantified; it is quite the reverse:
the application of quantitative methods in the humanities makes clear the
phenomena which can not be quantified - it enables a better under-
standing for structures which are unquantifiable. As an example we could
think of psychology (if we think of it as a discipline of the humanities)
and its (absurd) methods and trials to quantify intelligent behavior.
What can be shown is the naivete of natural scientists and their methods
- these methods are too simple and too naive to quantify or even estimate
such a complex phenomenon as human intelligence; it is, however, not
only the method, but also the scientists and their assumptions which have
to be analysed. That is the third kind of strangification in which the social
and organizational context of the scientists is observed - we are calling
it pragmatical strangification.
Thus, CR has a strict interdisciplinary structure - it is, however, not
ontological, but rather methodologically motivated. We have to give up
the claim of traditional European epistemologies of complete insight; the
methodologies of CR have another aim : on the one hand they repre -
sent the basis for the construction of knowledge, on the other hand for
the indirect insights into these constructions. As an implication of this
we do not need a metatheoretical legitimation for the methodologies of
CR.
2.2. Two Types of Reality
In German there exist two words for 'reality' : we will refer to Realitiit
with 'reality' and to Wirklichkeit with environment(w). Stimulated by
108 F . WALLNER AND M. F . PESCHL
H. Maturana we are distinguishing (in contrast to evolutionary episte-

mology) between two types of reality: we understand by "Wirklichkeit"
(environmenuwj) the world we are living with, i.e., the world which is
presupposed to our perceptions and to our processes of life. "Realitat"
(reality) means our cognitive world being the result of a process of
construction. This differentiation can be derived from different contexts.'
We could compare it to, as well as differentiate it from, Kant's differ-
entiation between "Ding-an-sich" (thing-in-it-self) and the things'
appearances. The dilemma of Kant's approach, which claims the unifi-
cation of the "D ing-an-sich" and transcendental ego , was the reason for
introducing this differentiation. It avoids the differentiation (being pre-
supposed implicitly as well as explicitly by Kant) between absolute and
human knowledge, as such a differentiation represents a principal mis-
understanding of the concept of knowledge according to CR. It rather
introduces the differentiation between the world we are living with as
environment(w) and the constructed world of our cognition as reality.
By making use of the strategies of strangification CR offers to get a
better understanding of reality; the environment(w) can not be under-
stood. We can gain control , however, over environment(w) by our
constructions of reality. In respect to our knowledge we can understand
only what we have constructed.
The differentiation between reality and environment(w) does not serve
the relativization of our knowledge; it rather avoids the fact that we
are surrendering ourselves to the success of our constructs in the envi-
ronment(w). As an implication we are suggesting a proposal for the
organization of the process of research: we should differentiate between
the instrumentalistic aspect in science and the aspect of construction of
reality. Instrumentalism gives itself up to the criterion of " working well",
the latter claims to impart knowledge.
It was one of the biggest mistakes in European history to admit to
an empirical success of constructed knowledge a judicial instance over
our knowledge. The alternative for such a hypostatization of empiri-
cism can be found not only in ideological and arbitrary assumptions.
We rather must not lost sight of the constructive character of empirical
instances. This implies that empirical control of theoretical constructs
represents nothing but the comparison of two (different) constructs (of
our cognitive reality). Thereby we do not give up principally the relation
to environment(w); the relation to environment(w) is rather indirect. If
we are comparing two different constructs (of reality), we are "leaving"
reality; we are testing to what extent reality endures beyond the corre-
lations of our constructions. Stated another way we could formulate the
following: By making use of empirical control we are applying a strangi-
fication "transforming reality to environment(w)". Having in mind this
context of indirect control, empirical testing seems to be very useful.
3. KNOWLEDGE REPRESENTATION AND

CONSTRUCTIVE REALISM
This differentiation into two domains of reality has very interesting and
important implications for a central problem in Constructive Realism
as well as in cognitive science: the problem of knowledge representation.
We are understanding knowledge as the result of cognitive processes
in natural (or artificial) cognitive systems - hence, knowledge can also
be seen as the result of complex information processing. As we will
see in the course of this section, we have to use this metaphor of cog -
nition being information processing in a rather careful manner for the
following reasons.
3.1. Symbolic Knowledge Representation
In most cases information processing and knowledge representation is

reduced to linguistic knowledge, i.e., at least three steps of mapping
are assumed:
(a) there is a mapping from environment(w) to the nervous system,
i.e., stimuli coming from the environment are transformed into
patterns of neural activations.
(b) these neural activities are mapped to explicit linguistic structures.
This domain is assumed to be what we have called "cognitive reality"
above - this implies that this domain is restricted to explicit natural
language.
(c) these natural language symbols are mapped to a symbol system of
a formal language; this system offers the possibility of processing
"knowledge" on a syntactic level. This implies that knowledge pro-
cessing is reduced to the manipulation of syntactic structures
representing only the very "surface" of our language.
Hence, the semantic as well as the pragmatic aspect of knowledge
are reduced to the level of syntactics. Such a reduction is well known
(and was doomed to failure) in philosophy of science : early R. Camap
110 F. WALLNER AND M. F . PESCHL
is one of the most prominent exponents of this approach." In this book

Logische Syntax der Sprache, Carnap tries to reduce the semantic aspect
of scientific language to syntactics. He is not even aware of the prag-
matic aspect at this time; a few years later Carnap himself recognized
the shortcomings of such an unjustified reduction. We can learn from this
development in philosophy of science that if we are interested in the
investigation of structures of knowledge (or of reality) we have to go
beyond this syntactic border and seek for alternative mechanisms of con-
structing knowledge. Here we are going the long way of giving up
classical AI.
If we are looking at the orthodox approach of Artificial Intelligence
or traditional cognitive science we can observe this unjustified dominance
of language in their representation mechanisms: (natural) linguistic struc-
tures are transformed into very complex systems of rules , frames,
schemata, semantic networks, etc. - these symbol systems obtain their
meaning only in the observer 's or interpreter's cognitive domain, i.e., a
symbol is represented in a meaningless bit pattern having a meaning only
if it is interpreted by the user. These computer symbols are lacking the
direct relation to the environment(w) as they are the abstract result of
processes taking place in the (cognitive) reality. This approach of manip-
ulating (meaningless) symbols is - from the perspective of computer
science - very convenient; if we are interested in very basic problems
concerning basic questions of what is knowledge, what is language, what
is cognition, etc ., such an approach is not adequate and we have to
look out for alternative forms of present ing knowledge in a way which
is not as restricted as in this symbolic approach (for a more detailed
discussion, see M. Peschl).'
3.2. Neural Computing - an Alternative Form of

Understanding Knowledge
Cognition can not be reduced to a process of manipulating symbols -

we rather have to see its neural foundation in a cognitive system's
nervous system . This implies that we have to understand the metaphor
of information processing in a more general way: cognitive informa-
tion processing also includes basic processes such as physically
interacting with an environment or physical interactions within the
nervous system - these processes are the foundation for language, formal
manipulations, etc . Knowledge is not understood in such a restricted
manner of linguistically explicit units any more. It is the result of a

process of self-organization, adaptation and construction. Hence, we
prefer the notion of knowledge construction to the notion of knowledge
representation. The Parallel Distributed Processing paradigm, which is
often called connectionism, neural computing, or PDP, offers such mech-
anisms for presenting knowledge in an alternative way and makes it
possible to achieve a deeper understanding of the process of construc-
tion of knowledge as Constructive Realism outlined. Rumelhart and
McCIelland 6 represent the traditional exponents of the newer form of con-
nectionism which is a dynamically emerging field in computer science
as well as in cognitive science . In short, the aim is to simulate neurons
which are assumed to be the functional "atoms" of our brain and cog-
nition . These neurons, which are called units, represent a very abstract
model of natural neurons. They are forming a huge network which is
highly interconnected; each of these units is working in parallel and
computes its own activation by summing up the product of the activa-
tions and weights of the other units. The output acts as an input for all
the other units being connected to this unit.
What about knowledge representation in such networks? We cannot
find a linguistic representation in the form of explicit symbols any more .
We rather find a set of weights being responsible for the behavior of
the whole network. These weights are changing by applying a learning
rule, i.e., by a locally slight change of the value of the weights the
global behavior of the whole network is changing as these changes are
taking place in each of these weights. Simply spoken, PDP networks
are "learning" to associate patterns of activities in a subsymbolic way,
i.e., in a way which is not restricted to symbols any more. The PDP
paradigm is comparable as well as different from the structure of knowl-
edge construction. This special relation enables us to improve the insights
in the structure of knowledge construction by strangification. The domain
of acting is strangificated into the domain of connectionist structures.
3.3. The Constructivist Approach to Knowledge
The point in CR is not to assume a given body of knowledge which is

mapped to the computational structures (for instance, symbols or PDP
networks); we are trying to make transparent the fundamental processes
of constructing knowledge. From a more abstract point of view we have
to look at the Parallel Distributed Processing paradigm as a very complex
form of pattern processing. It very much depends on the computational

environment in which such a PDP network is applied: if the input and
output are encoded symbolically (for instance in Sejnowski 's NETtalk)7
the same problems will arise as in the approach of orthodox symbol
manipulation; the only advantage is the way in which these symbols
are processed: by propagating patterns of activations and by interpreting
these patterns as a code for symbols. Such an approach does not help
us in solving the problems of reduction of knowledge representation to
linguistically explicit structures.
The alternative approach being suggested by M. Peschl " tries to model
these processes in a bottom up manner. This means that the suggested
cognitive model is linked through sensors and effectors to its environ-
ment(w). By learning in the highly recurrent connectionist network an
internal cognitive reality is built up. The structure of this way of under-
standing knowledge of the cognitive system's environment depends on
the following items:
(a) the cognitive system 's current state of activities, which itself is deter-
mined by
(b) the ontogenesis of the cognitive system (i.e ., by its very private
history)
(c) by the structure (and architecture) of the nervous system
(d) by the structure of the sensors and effectors.
These four factors determine the construction of knowledge, i.e., the
process of how knowledge is constructed, by integrating the stimuli from
environment(w) and the internal stimuli to the system 's cognitive reality.
Hence, a closed feedback loop is established between the cognitive
system, the cognitive reality and environment(w). The relation between
these can be studied by simulating such a cognitive system having this
epistemological foundation of Constructive Realism (i.e ., differentia-
tion between cognit ive reality and environment(w), etc.) ."
4. COGNITIVE SCIENCE AND CONSTRUCTIVE REALISM -

A MUTUALLY INTERDEPENDENT RELATION
In this section we are going to discuss the relations between cognitive

science and Constructive Realism in detail. If we are looking at tradi-
tional approaches in cognitive science we can see that the role of
philosophy and epistemology is quite unclear. This uncertainty is recti-
fied by Constructive Realism . CR, however, does not understand itself
as the philosophy of cognitive science; rather it understands itself as
an instance of services which are offered to the (natural) sciences (this

is standing in relation to L. Wittgenstein's concepts of philosophy). There
is, however, a difference between Wittgensteiri 's concept and the concepts
of CR: L. Wittgenstein 's philosophy "leaves everything as it is", whereas
CR claims to support the self reflection of sciences and thus is capable
of changing them. CR instructs the sciences to newly rethink central
aspects of method and the claims which are made for objects - by
applying this procedure many self-misunderstandings of science could
be revealed.
In short , CR offers the cognitive sciences a new ontology by redefining
the role of ontology itself. It designs concepts of the world for the proces-
sual descriptions of cognition and knowledge in cognitive science - the
most general concept is the differentiation between reality and envi-
ronment(w) as it has been presented and discussed above. Similar things
could be said about the constitutive role of the interdisciplinarity of
cognitive science. It is the point of CR to show that interdisciplinarity
is possible only if we give up the traditional ontological concepts and
ideas (Wallner 1989).1 0
Why is cognitive science interrelated with Constructive Realism? To
answer this question we have to present briefly our approach to cogni-
tive science: we are understanding cognitive science as the continuation
of traditional epistemology in the context of the modern knowledge (of
natural science) and in the context of the new technologies of our century.
This means that it is our aim to integrate traditional approaches of
epistemology with results and approaches from computer science and
neuroscience. As has been mentioned above, this requires an interdis-
ciplinary approach to the phenomena being investigated in cognitive
science.
If we are looking at traditional approaches, however, we can see that
the investigations are rather multidisciplinary than interdisciplinary (cf.,
Stillings 1987, Adams 1985, Posner 1989, Osherson 1990, etc.)," i.e. ,
there are results coming from different disciplines (and their methods and
paradigms). These results are presented to the other participating disci-
plines and (in the best case) compared to each other. There are , however,
no consequences drawn from this process of comparing. Only the con-
sequent integration of results enables fruitful interdisciplinarity. The
concept of strangification in Constructive Realism could possibly offer
such a strategy for consequent integration. Results from neuroscience,
for instance, are strangificated into the context of epistemology or
neurocomputing. One thing (among many others) we can learn from
this strangification is that relations can be found between epistemology,

computer science (especially Parallel Distributed Processing, McClelland
1986)12 and neuroscience. Such strangification is only possible because
of the similar structures of these disciplines.
As has been mentioned, the differentiation between (cognitive) reality
and environment(w) plays an important role because we are under-
standing (all kinds of) knowledge as the product of a process of
construction. This construction is triggered by environment(w) and is the
result of a learning process in the (cognitive) reality. On this basis we
are investigating the basic questions of what is knowledge, what is knowl-
edge representation, cognition, language, etc., in the interdisciplinary
context of the disciplines of epistemology, computer science and neu-
roscience." Each of them plays a specific role:
(a) neuroscience represents the empirical part, i.e., the empirical results
represent a constraint for epistemology and computer science. On the
other hand we have to be aware of the problems of methodology
and "empirical input" - they could be investigated by epistemology;
(b) epistemology has a threefold function:
(1) to structure the outline of the research programs
(2) to make interdisciplinary interventions by strangific ations
(3) to show insights into the structure of cognitive reality .
(c) computer science 's aim is to combine these two disciplines by
making use of simulation techniques - as mentioned, the Parallel
Distributed Processing approach (PDP, connectionism) plays an
important role in our understanding of the role of computer science
in cognit ive science. This means that computer science tries to find
models coming up to the claims of neuroscience as well as of
epistemology - the PDP approach is capable of meeting these
requirements and represents the foundation of our research and con-
siderations. Computer science generates (from an observer 's point of
view) cognitive behavior by applying the simulation methods of
the Parallel Distributed Processing paradigm. By learning in a PDP
network we are building up an artificial (cognitive) reality being trig-
gered by the (artificial) environment(w). Thus we are capable of
studying structures of knowledge, how they evolve , and how they
are constructed; we could even learn something about the con-
structive relation between reality and environment(w).
Thus we can find a twofold relation between cognitive science and
Constructive Reali sm; on the one hand CR provides the methodology
of interdisciplinary cooperation between the disciplines - as has been

mentioned, the strategy of strangification and the differentiation between
reality and environment(w) are the most important concepts being applied
in our approach to cognitive science. On the other hand cognitive science
represents a very interesting source for CR for two reasons:
(i) generally speaking cognitive science investigates the phenomenon of
knowledge - thus it provides the Constructive Realism with infor-
mation being of great interest, because it is one of the aims of CR
to get deeper insights into the structures of knowledge (i.e., in the
structure of reality);
(ii) the methods being suggested by CR can be tested in the interdisci-
plinary cooperation of our approach to cognitive science. As a
consequence of this testing process the methods of CR could be
improved, adapted, etc ., because we are trying to develop the
concepts of CR not only by speculative ideas , but also by consid-
ering the problems of the actual process of research and thus need
feedback from the sciences.
5. CONCLUDING REMARKS
CR is a concept of science and research necessary in order to rescue

the idea of our occidental science in the face of a completely changed
domain of social structures . While the aim of occidental science according
to the original ideas was the democratization of human conditions, the
history of our universities often shows that strong tendencies counter-
acted this structure . If we do not establish the already discussed difference
between environment(w) and reality, the structure of science tempts us
to accept an undemocratic hierarchy, because the structure of the world
must be mirrored in the organization and structure of science. If we
assume research groups constructing alternative constructions of reality
are not only possible, but rather desired, it is because they enrich our
insights into reality. Therefore reorganization of the university cannot
solve the rather fundamental problems of university structures unless
we are willing to rethink and reflect the ontological presuppositions of
science.
Institut fUr Wissenschaftstheorie

und Wissenschaftsforschung,
Universitdt Wien, Austria.
116 F . WALLNER AND M . F. PESCHL
NOTES
I Polanyi , M. (1966) . The Tacit Dimension, Doubleday & Company, Inc., Garden City,
New York.
2 Anderson, J. R. (1985) . Cognitive Psychology and its Implications, New York.
3 E.g., Wallner, F., Thesen zu Maturana. Eine Aufforderung zum Widerspruch , Research
Report , Dept. of Epistemology & Cognit ive Science , Univ. of Vienna, IWTF-ECS-89-2.
4 Carnap, R. (1936) . Logische Syntax der Sprache, Wien .
S Peschl, M. F. (1989). 'An Alternative Approach to Modelling Cognition' , in Proceedings
of Man & Machine Conference published by John von Neumann Society for Computing
Science s, Hungary , pp. 46-56 and also ibid. (1990) . ' Cognition and Neural Computing
- an Interdisciplinary Appro ach ', in Proceedings of Intern ational Joint Conference on
Neuronal Network s, Washington (IJCNN '90), Lawrence Erlbaum Associates, Publishers,
pp.IIG-113.
6 Rumelhart, D. E. and McClelland, J. R. (1986). Parallel Distributed Processing ,
Explo rations in the Microstructure of Cognit ion, Volume I: Foundations, MIT Press,
Cambridge Massachusetts. Ibid . (1989). ' The Architecture of Mind : A Connectionist
Approach ' , in Posner M. I. (ed .), The Foundation s of Cognitive Science, MIT Pres s,
Cambridge, Massachusetts , pp. 133-159. McClelland, J. L. and Rumelhart, D. E. (1988) .
Parallel Distributed Processing, Explorations in the Microstructure of Cognition, Volume
II: Psychological and Biological Models, MIT Press, Cambridge, Massachu setts.
7 Sejnowski, R. J. (1981) . ' Skeleton Filters in The Brian ', in Hinton and Anderson (eds.),
Parallel Models of Associative Memory, pp. 189-212.
K Peschl, M. C. (1990). ' Cognition and Neural Computing - an Interdisciplinary
Approach ', in Proceedings of Intern ational Joint Conference on Neural Networks,
Washington (IJCNN '90), Lawrence Erlbaum Associate s, Publishers, pp. 11G-113. Ibid.
(1990). ' A Cognitive Model Coming up to Epistemological Claims - Constructivist Aspects
to Model ing Cognition' , in Proceedings of International Jo int Conference on Neural
Networks 1990 (IJCNN '90) , San Diego, CA, IEEE, Vol. III, pp. III-657-662.
9 Wallner, F. (1990). 8 Vorle sungen zum Konstruktiven Realismus, Wiener Univer-
sitatsverlag.
10 Wallner, F. (1989). Maturanas moglicher Beitrag zur Epistemologie, Wissenschaftliche
Nachrichten, Janner .
II Stillings, N. A., Feinstein , M. H. and Garfield , J. L. et al. (1987) . Cognitive Science,
An Introduction , A Bradford Book, the MIT Press, Cambridge, Massachusetts. Adams,
R. G. (1985). ' Cognitive Science: A Tripartite Approach ' , in: Cognit ive Systems 1-1,
pp. 17-29. Posner, M. J. (ed.) (1989). Foundations of Cognitive Science, MIT Press,
Cambridge, Massachusetts . Oshersons, Daniel N. (ed.) (1990). An Invitation to Cognitive
Science, MIT Press, Cambridge , Massachusetts.
12 McClelland, J. L., Rumelhart, D. E and Hinton , G. E. (1986). ' The Appeal of Parallel
Distributed Processing', in: Rumelh art, D. E., Parallel Distributing, Vol. I, MIT Press,
Cambridge, Massachusetts.
13 The reduction to these three disciplines has been suggested by M. F. Pesch I (1990) .
Auf dem Weg zu einem neuen Verstandnis der Cognitive Science , Informatik Forum, Juni.
JAMES ROBERT BROWN
PHENOMENA
PROOFS AND PICTURES
My interest is phenomena, but I'll begin with something apparently quite

different - mathematical diagrams. Consider first, the following diagram
which accompanies the proof of the Pythagorean theorem in Euclid's
Elements. We can stare at it for days and still not see that the theorem
is true; we need a proof. The common view of diagrams in mathematics
is this: they provide a heuristic aide, a help to the imagination when
following a proof. Diagrams cannot justify and are not to be confused
with real proofs, which are formulated in words and symbols; at most
they playa psychological role.
Fig. 1. Diagram that accompanies the proof of the Pythagorean theorem.
117
© 1996 Kluwer Academic Publishers .
118 JAMES ROBERT BROWN
For examples such as this one, the standard account seems right. But
there are a few rare and remarkable examples where something quite dif-
ferent is going on. The following theorem is from number theory ; and
it has a standard proof (by mathematical induction) which uses no
diagrams at all. But it can actually be proven with a diagram . (Take a
moment to study the proof, to see how it works.)
Theorem: 1 + 2 + 3 + . .. + n = n 2/2 + nJ2
Proof:
Fig. 2. A diagram as a proof.
Of course, there is lots of interpreting going on to make this a proof.

For example, we must consider the individual unit squares as numbers
and we must bring some geometric background information to the effect
that a square with sides of length n has area n 2• But these sorts of inter-
pretive assumptions are not less innocuous than those made in a typical
verbal/symbolic proof.
The moral I think we should draw from examples like this is simple
but profound: we can prove things with pictures. In spite of the fact
that the diagram seems to be a special case (n = 5), still we can see all
generality in it. And the proof does not work by suggesting the 'real'
proof, since in the diagram there is nothing which corresponds to the
passage from n to n + 1 which is the key step in any proof by mathe-
matical induction.
Those who hesitate to accept the picture as a proof might think that
the picture merely indicates the existence of a 'real' proof, a standard
PHENOMENA 119
proof by mathematical induction. Perhaps they even wish to appeal to

the well known distinction between discovery and justification: the
picture is part of the discovery process while true justification comes only
with the verbal/symbolic proof. But consider: would a picture of an equi-
lateral triangle make us think there is a proof that all triangles are
equilateral? No. Yet the above picture makes us believe - rationally
believe - that there is an inductive proof of the theorem . The picture is
evidence for the existence of a 'real' proof and the 'real' proof is evidence
for the theorem. But we have transitivity here; so the picture is evidence
for the theorem, after all.
Let us put this connection between theories and pictures in the back-
ground for now and turn to our main concern, phenomena.
DATA YS . PHENOMENA
My point of departure is a notable recent analysis of phenomena by James

Bogen and James Woodward who make a "distinction between phe-
nomena and data" (1988, p. 305). The former are constructed I out of
the latter.
Data, which play the role of evidence for the existence of phenomena, for the most part
can be straightforwardly observed. However, data typically cannot be predicted or
systematically explained by theory. By contrast, well -developed scientific theories do
predict and explain facts about phenomena. Phenomena are detected through the use of
data, but in most cases are not observable in any interesting sense of that term (1988,
p. 305) .
Data are .. . idiosyncratic to particular experimental contexts, and typically cannot
occur outside of those contexts . . . Phenomena, by contrast, are not idiosyncratic to specific
experimental contexts. We expect phenomena to have stable, repeatable characteristics
which will be detectable by means of a variety of different procedures, which may yield
quite different kinds of data (1988 , p. 317).
There are several important features and consequences of this view.

Among the more important are these: explanation is not a relation
between theories and observable facts; nor is prediction; theories are
not tested by comparing them with experience; and observation - whether
theory -laden or not - is "much less central to understanding science
than many have supposed" (1988, p. 305).
Typical of Bogen-Woodward phenomena are weak neutral currents.
The associated data are bubble chamber photographs. The relevant theory
which is supported by all of this is the Weinberg-Salam theory of weak
interactions. It is supported, not by the data, but by the phenomena of

weak neutral currents. The existence of the phenomena is in turn sup-
ported by the data, the photographs. According to Bogen and Woodward,
the data are far too messy to serve as evidence for any theory; the phe-
nomena play a crucial and irreducible intermediate role in the process
of scientific inference.'
A cursory glance at the social sciences suggests that the phe-
nomena/data distinction is important here, too, perhaps even more so.
Social scientists seem to do (at least) two quite distinct things. One is
to establish phenomena: e.g., that there is widespread child abuse, that
x% of the population is homosexual, that suicide rates in some cultures
are higher than in others, that x% of women are physically abused by
their male companions, that there is an x% unemployment rate, etc. These
are often extremely difficult to ascertain as we might imagine, espe-
cially when quest ions of sexuality are involved. The second job is to
explain these phenomena. And it is indeed phenomena that social theory
attempts to explain. Economists try to tell us why we have high unem-
ployment (a downturn in the economy), not why Joe Blow is out of
work (perhaps he was an incompetent worker) ; and Durkheim told us
why Protestant societies have higher suicide rates than Catholic ones
(they are socially less cohesive); he doesn't tell us why Joe Blow killed
himself (perhaps he was depressed after losing his job).
A number of examples, even if only briefly described, should help
to explain and reinforce the data/phenomena distinction.
EXAMPLES
High energy physics abounds with examples. We are all quite used to
having information from this field presented to us twice over, first as data
in the form of a photograph, then as phenomena in the form of an
artist's drawing. Here is a typical example (Figure 3).
What high energy physics does is explain and predict the drawing
on the right; not the photo on the left. The chicken scratches on the
left are far too variable, idiosyncratic, and downright messy for any theory
to deal with. Theories in high energy physics only try to cope with the
phenomena as represented in the artist's drawing.
The so-called mechanical equivalent of heat was established by James
Joule in the middle of the 19th Century. That is, "That the quantity of
heat produced by friction of bodies, whether solid or liquid, is always
PHENOMENA 121
,\
\
\
\
\
\
\
II ;".,.- - - - - - - --
.....
, '"
1/'
I '" ....
r
I
Fig. 3. The reaction K'+p-7A+1t°. The photo is a datum ; the drawing is a phenomenon.
proportional to the quantity of force expended". In a large number of

repeated experiments involving a paddle wheel contraption that heated
a quantity of water when the paddles were driven by falling weights,
Joule established
That the quantity of heat capable of increasing the temperature of a pound of water
(weighed in vacuo , and taken at between 55° and 60°) by 1°F, requires for its evolution
the expenditure of a mech anical force represented by the fall of 772 lbs through the
space of one foot (Joule 1850, p. 82).
This was not the result of any simple observation, but the culmination
and processing of an enormous amount of data .
The method of experimenting was simply as follows : The temperature of the frictional
apparatu s having been ascertained and the weight s wound up . . . the roller was refixed
to the axis. The precise height of the weights above the ground having then been deter-
mined by means of the graduated slips of wood . . . the roller was set at liberty and allowed
to revolve until the weights reached the flagged floor of the laboratory, after accomplishing
a fall of about 63 inches. The roller was then removed to the stand, the weights wound
up again , and the friction renewed . After this had been repeated twenty times, the experi-
ment was concluded with another observation of the temperature of the apparatus. The
mean temperature of the laboratory was determined by observations made at the com-
mencement, middle and termination of each experiment (Joule, 1850, p. 66) .
122 JAM ES ROBERT BROWN
It was out of the enormous amount of data that he had that the phenomena
of the mechanical equivalent of heat was brilliantly and painstakingly
constructed by Joule .
The Periodic Table of the chemical elements provides another illus-
tration. The Periodic Table is a classification scheme of the elements
in accordance with these properties. It is a paradigm of the construc-
tion of phenomena out of data. The phenomena are the entries in the table
- the chemical elements and their properties: atomic weights, atomic
numbers, chemical similarities, etc.
There is no algorithm for making phenomena out of data - it is a
fallible process. Dimitri Mendeleev ordered the elements according to
their increasing atomic weights. But he noticed that atoms with similar
chemical properties recurred periodically at fairly regular intervals. By
lumping together those which are chemically similar he created a clas -
sification of the elements known as the Periodic Table.
Though brilliantly conceived, Mendeleev 's taxonomy was somewhat
problematic. In the case of a few elements, ordering them by increasing
weight was at odds with ordering them in accord with their chemical
properties. And the discovery of isotopes (which have different weights
but are chemically identical) made matters even worse . This was the
background for Henry Mosely 's work, begun in 1913.
The characteristic frequencies associated with each of the elements
is due, according to Bohr's theory of the atom, to electrons in orbit around
the nucleus falling to lower orbits. When they fall from one energy
level, or shell, to a lower one they emit a photon of the appropriate
energy, or frequency.
Mosely fired cathode rays at several of the heavier elements and
recorded the x-ray frequencies produced. He focused on a particular
series known as the Ka-lines in a large number of elements. What he
discovered is that as the atomic number increases by I (i.e., as Z ~
Z + 1), the quantity (4/3 x v(Z) x R)' 12 also increase s by I. This led to
the following formula for the frequencies of Ka-series for the element
with atomic number Z: v(Z) = (Z - 1)2 X (1/1 2 - 1/22) X R. (Where R
is the Rydberg constant, known independently, and 1/1 2 - 1/22 is asso-
ciated with the first and second energy levels .)
Mosely's classification and Mendeleev 's coincide except in a few
cases. For example, potassium preceded argon in Mendeleev's table,
but Mosely reversed them. This resulted in Mosely 's Periodic Table being
in full agreement with both the recurring regularities of the chemical
PHENOMENA 123
properties and with the increasing atomic numbers . There is no internal

tension as there was in Mendeleev's taxonomy.
The history of the Periodic Table illustrates all sorts of interesting
things about phenomena. In it we see, of course, the construction of
precise elements out of the hodge-podge of chemical data. And we also
see that the Table is not explanatory; it is just a taxonomy. But there is
one more feature of the Table which strongly supports the data/phe-
nomena distinction. When the table was being constructed there were
'gaps' in it; that is, nothing had ever been observed which corresponded
to certain places in the table (e.g., germanium, Ge). Any theory (such
as Bohr's) that attempted to explain the features of the Table would be
required to account for every place in the Table , including the gaps.
(Or explain why the gaps had to exist, as quantum mechanics does in
the case of the very heavy elements - they are unstable .)
I've been saying 'gaps in the table ' . Strictly, this is wrong : the table
is complete; the gaps are in the data . This means that the entries in the
Periodic Table cannot be identified with what is actually observed, with
data - since there is none (or was none at the time) - but must instead
be thought of as phenomena.
PHENOMENA AND NATURAL KINDS
The world is full of data, but there are relatively few phenomena.
My suggestion is rather simple: phenomena are abstract entities which
correspond to visualizable natural kinds. When scientists construct the
phenomena out of a great mass of data what they are doing is singling
out what they take to be genuine natural kinds. In Plato's gruesome
metaphor, they are trying to cut nature at its joints. To this I would
only add: at nature's visualizable joints.
The shift in the ordering structure of the Periodic Table, from atomic
weights to atomic numbers, shows the complexity and ingenuity that is
sometimes involved in constructing phenomena out of data. But it also
shows the importance of natural kinds and their essential properties in
scientific thinking. Mosely expressed it well when he summed up his
experimental work :
We have here a proof that there is in the atom a fundamental quantity, which increases
by regular steps as we pass from one element to the next. This quantity can only be the
charge on the central nucleus , of the existence of which we already have definite proof
(Quoted in Trigg 1975, p. 32£).
Notice that Mosely is not claiming to have discovered that the nucleus
has an electronic charge, any more than he is denying that the elements
have an atomic weight. His claim is about which of these existing prop-
erties is 'fundamental', or essential (chemically), and which is not.
THOUGHT EXPERIMENTS
Thought experiments deal with phenomena.' Obviously, they don't deal

with actual experimental data - this much is true by definition. But the
fact that they involve picturable processes suggests that we need to
keep something observation-like centrally involved.
In Einstein 's elevator, to cite one important example, the observer
inside cannot tell whether she is in a gravitational field or accelerating.
A beam of light passing through would bend downward if the elevator
were accelerating, so, by the principle of equivalence, it would also
.......
Fig. 4. Light bending in Einstein 's elevator.

PHENOMENA 125
bend downward in a gravitational field. The conditions required to make

such an observation are so extreme that any actual observer would have
to be a puddle on the floor of the elevator. The observation in this thought
experiment is of phenomena, not data.
Newton's bucket thought experiment provides an instructive example
in a different way. The thought experiment asks us to imagine the dif-
ferent stages of a bucket partly filled with water as it is released and
allowed to 'rotate'. The water and bucket would be initially at rest with
respect to one another, and the water surface would be flat. Next they
would be in relative motion. In the third stage they would again be at
rest with respect to one another, but this time the surface of the water
would be concave.
Why the difference between stages one and three? Newton's expla-
nation is simply this : In the first stage the water and bucket are at rest
with respect to absolute space, and in the third they are rotating with
respect to absolute space .
After Leibniz, Newton's most forceful critics were Berkeley and Mach .
IT
ill
Fig. 5. Three stages in Newton 's bucket thought experiment.
Did they deny that absolute space was the best explanation for the
observed difference? Not really; instead, they denied the observed dif-
ference itself. They denied that in a universe without distant masses
(the fixed stars) the water would climb the walls of the bucket. Clearly,
Berkeley's and Mach's fight with Newton is not a dispute over empir-
ical data; it's not even a fight over rival explanations of what is given
in the thought experiment - it is a fight over the phenomena.
What we can see from these two thought experiments is that phe-
nomena must be play ing a role in scientific inference, a role which is
distinct from data. Though phenomena are picturable, they exist at a high
level of abstraction.
In passing, a word about the theory-ladenness of observation. No
one nowadays believes in raw data; observations are always conceptu-
alized. (This is undoubtedly one of the great results of modem philosophy
of science, due to Hanson , Kuhn, Feyerabend, Sellars, and many others .)
Isn't 'phenomena' just another name for this, that is, for theory-laden
data?" In many cases the distinction between phenomena and concep-
tualized or theory-laden data will seem artificial. (High energy physicists
like to say that they can just see the subnuclear process right in the bubble
chamber photo.) But there are clear cases which cannot be treated as
theory-l aden observ ations . The elements of the Periodic Table are phe-
nomena, and no doubt some of them, e.g., Fe (iron), might successfully
be treated as observable in some theory -laden way. But there are others,
e.g., Ge (germanium), which (at the time of Mendeleev) were simply
not seen at all. Similarly, the phenomena of a thought-experiment, e.g.,
the light bending in Einstein's elevator are not actually seen at all either.
So these examples of phenomena cannot be reduced to actual observ-
able data, theory-laden or not.
PHENOMENA AND INFERENCE
How is it possible that a great and grand theory can seem to be justi -
fied by only a tiny bit of sketchy visualized information? Recall the
mathematical example from the very beginning of this paper. A lesson
about inference can be learned from that rare form of mathematical proof.
The claim made there was that the diagram is a perfectly good proof.
One can see complete generality in the picture, even though it only
illustrates the theorem for n = 5. The diagram does not implicitly
'suggest' a 'rigorous ' verbal or symbolic proof. The regular proof of
this theorem is by mathematical induction, but the diagram does not
PHENOMENA 127
correspond to an inductive proof at all (where the key element is the

passage from n to n + I).
One of the morals to be drawn from this example is of great impor-
tance for the philosophy of mathematics, especially concerning the nature
of proof. But the moral I want to draw here is just this: We can in
special cases correctly infer theories from pictures, that is, from visu-
alizable situations.
The great inductive leap is really from data to phenomena; once we
have the phenomena, the further inference to theory often can be actually
rather small. This is because of the following feature of natural kinds.
Any natural kind has an essential property (or set of prop-
erties) that makes it the kind that it is. If any member of a
kind has essential property <I> then every member of the kind
has <1>.
If any sample of water, e.g. , has chemical composition H 20, then they
all do. Notice our reluctance to make a similar inference about, say,
the colour of ravens . We balk at: If any raven is black, then they all
are . We hesitate because of our belief that colour is not an essential
property of ravens. We do believe that all ravens are black, of course,
but this belief is based on the observation of an enormous number of
ravens. By contrast, our beliefs about, say, the mass of intermediate vector
bosons are based on only a handful of scratchy bubble chamber pho-
tographs. So, either our physics colleagues have absurdly low standards
when compared to bird watchers or something remarkably different is
going on in each case. Clearly, it is the latter. There is a profound dif-
ference between the two cases and it has to do, I suggest, with phenomena
as natural kinds. In particular, if any vector boson ha s mass m then
they have mass m. Natural kind inference is quite different than enu -
merative induction, the principle used in inferring the colour of all
ravens.'
While I have invoked natural kinds and their essential properties to
account for some of the aspects of phenomena, my commitment to natural
kinds is not too deep. Perhaps patterns would be a better notion." Instead
of seeing phenomena as constructed out of data, I should take patterns
to be so constructed. First, patterns would avoid the controversial meta-
physics of essences, etc.; second, patterns are obviously abstract and
hence clearly different from observable data; and third, inferences from
patterns are quite unproblematic and so lend themselves to quick con -
clusions of the sort we see in the mathematics and thought experiment
128 JAM ES ROBERT BROWN
examples given above. For now, at any rate , I prefer to remain agnostic
and leave this an open question, and for the sake of consistency I'll
stick with talk of natural kinds .
Of course , the question arises whether we really have a natural kind
(or an essential property of a natural kind) on our hands or not. Is mass
really an essential property, and colour not ? It seems like an a priori
assumption, and to some extent I dare say that it is. But the view that
the colour of ravens is not an essential property while the micro-struc-
ture of water is, is at least in part based on very broad experience and
the past success of various classes of theories that we hold. Theories
based on micro-structure have been enormously successful while those
based on colours have not. So the construction of phenomena out of
data is based on more than the data itself. It is theory-laden, but it needn 't
be laden with the theory that it will subsequently be used to test.
This raises an interesting an important point that I can only mention
here: the construction of pseudo-phenomena. Many scientific works are
replete with drawings. E. O. Wilson's Sociobiology, for example, has
almost no photographs, but has several beautiful drawings of animals
in various activitie s. One of these shows two dinosaurs fighting. Needless
to say, this was seen by no paleontologist. It is not a datum, but a phe-
nomenon. But is it a real phenomenon? I will leave to others the
assessment of sociobiology. But I will point out that the theory - like any
other - is in the business of explaining phenomena, not data. Perhaps
it even does this brilliantly, which is why many find it persuasive. But
there is a lot of room to ask pointed questions about the construction
of such phenomena (are they pseudo-phenomena?) and the role that
values may have played .
Let me now, by way of conclusion, summarize the main points.
Phenomena are to be distinguished from data, the stuff of observation
and experience. They are relatively abstract, but have a strongly visual
character. They are constructed out of data, but not just any construc-
tion will do. Phenomena are natural kinds that we can picture. They
show up in thought experiments and they play an indispensable role in
scientific inference mediating between data and theory. So let 's attend
to them.'
University of Toronto,
Canada.
PHENOMENA 129
NOTES
1 This term is unfortunately loaded . Sociologists of science often use 'construction' to

mean ' soc ial construction ', the very opposite of an independently existing object or fact,
as for instance, Pickering does when he speaks of ' constructing quarks.' I'm using the
term in a more innocuous sense perfectly compatible with describing the objective truth.
For example, when a mathematician 'constructs' a function, she is not creating it anew,
but merely (though perhaps very cleverly) characterizing it in terms of other already
given mathematical objects.
1 For further discussions, rich with examples, see Woodward 1989, Kaiser 1991, and
Kaiser this volume.
J For more on thought experiments see my The Laboratory of the Mind : Thought
Experiments in the Natural Sciences.
4 This challenge came from Simon Blackburn during the conference.
5 For more on natural kind reasoning see Harper (1989) . Generally, this is unexplored
territory and deserves a great deal more attention.

6 Thanks to Mary Tiles for making this point. I'm grateful for helpful discussions on a
number of other topics in this paper.
7 The themes of this paper will be treated in greater depth in my forthcoming book, Smoke
and Mirrors: How Science Reflects Reality .
BIBLIOGRAPHY
Bogen, J. and Woodward J. (1988) . 'To Save the Phenomena', Philosophical Review .
Brown, J. R. (1989) . The Rational and the Social, Routledge, London and New York.
Brown, J. R. (1991) . The Laboratory of the Mind: Thought Experiments in the Natural
Sciences, Routledge, London and New York .
Harper, W. (1989). ' Consilience and Natural Kind Reasoning ', in J. Brown and J.
Mittelstras (eds.), An Intimate Relation: Studies in the History and Philosophy of
Science [Boston Studies in the Philosophy of Science, v. 116), Kluwer, Dordrecht.
Joule, J. (1850) . 'On the Mechanical Equivalent of Heat ', Philosophical Transactions of
the Royal Society.
Kaiser, M. (1991). 'From Rocks to Graphs - The Shaping of Phenomena ', Synthese,
111-133.
Kaiser, M. (this volume) . ' Empirical versus Theoretical Progress in Science'.
Trig , G. (1975) . Landmark Experiments in 20th Century Physics, Crane , Russell & Co.,
New York.
Wilson, E. O. (1975). Sociobiology: The New Synthesis.
Woodward, J. (1989). 'Data and Phenomena', Synthese, 393-472 .
ALLAN FRANKLIN
THERE ARE NO ANTIREALISTS

IN THE LABORATORY
One of the continuing polemics in the philosophy of science has been

the battle between the real ists and the antirealists. Although there are
probably as many variants of these positions as there are adherents of
them, I shall adopt, and defend, Bas van Fraassen's characterization of
realism, "Science aims to give us, in its theories, a literally true story
of what the world is like; and acceptance of a scientific theory involves
the belief that it is true" (1980, p. 8). I shall argue that we have good
reasons to believe in both the truth of scientific laws and theories and
in the reality of the entities involved in those theories.
The battle has flared up since van Fraassen's seductive account (1980)
of an antirealist position, that of constructive empiricism.' In his own
words , "Science aims to give us theories which are empirically adequate;
and acceptance of a theory involves as belief only that it is empirically
adequate" (1980, p. 12), where empirical adequacy mean s that what the
theory "says about the observable things and events in the world, is
true - exactly if it ' saves the phenomena' " (1980, p. 12).2 I shall discuss
this view in some detail later.
My own position, which one might rea sonably call 'conjectural'
realism, includes both Sellars's view that "to have good reason for
holding a theory is ipso facto to have good reason for holding that the
entities postulated by the theory exist") (1962 , p. 97), and the 'entity
reali sm' proposed by Cartwright (1983) and by Hacking (1983) . Both
Hacking and Cartwright emphasize the manipulability of an entity as a
criterion for belief in its existence. "We are completely convinced of
the reality of electrons when we regularly set out to build - and often
enough succeed in building - new kinds of device that use various well-
understood causal properties of electrons to interfere in other more
hypothetical parts of nature" (Hacking, 1983, p. 265) .4 Cartwright also
stresses causal reasoning as part of her belief in entities. In her discus-
sion of the operation of a cloud chamber she states, "... if there are
no electrons in the cloud chamber, I do not know why the tracks are
there" (Cartwright, 1983, p. 99) . In other words, if such entities don't
exist then we have no plausible causal story to tell. Both Hacking and
131
R. S. Cohen, R. Hilpin en and Qiu Renzong (eds.) , Realism and Anti-Realism in the
Philosophy of Science. 131-148.
132 ALLAN FRANKLIN
Cartwright grant existence to entities such as electrons, but do not grant

'real' status to either laws or theories, which may postulate or apply to
such entities.
In contrast to both Cartwright and Hacking, I suggest that we can have
good reasons for belief in the laws and theories governing the behavior
of the entities, and that several of their illustrations implicitly involve
such laws.' I shall present an illustration of my own later. I agree with
them, however, that we can go beyond Sellars and have good reasons
for belief in entities even without such laws .
I. ARE THERE REALLY K MESONS ?
Before discussing some of the philosophical issues, I would like to

present an example from contemporary physics which illustrates my view.
I will argue that this experiment provides good reasons to believe in
both the existence of K mesons, as well as in the truth of several laws
involved in the operation of the experimental apparatus.
I do not, however, wish to imply by this that experiment always
provides good reasons for belief in entities and laws ." I do wish to argue
that it can, and does, do something I believe Van Fraassen would deny.
One must also be careful about which entities the experiment argues
for and which are assumed on the basis of prior knowledge. Thus, in
the example that follows , the experiment provides reasons for belief in
K mesons, but the existence of electrons, protons, and pions and their
properties is regarded as unproblematic. One should also note that the
status of an entity may change during the course of an experiment. In
the case of the experiments designed to demonstrate the existence of weak
neutral currents, the existence of the currents was shown only after all
the data was taken and an argument constructed that the events seen could
not be due to neutron background or to the failure of the apparatus to
detect muons 7 (See Galison, 1987 for details). As David Cline, one of
the experimenters, remarked toward the end of one of the experiments,
"At present I don't see how to make these effects [the neutral currents]
go away" (Galison, 1987, p. 235). I note also that in this experiment
the existence of neutrinos, neutrons, pions, muons, and electrons was
taken as given.
The evidence for the reality of entities may also change over time.
Thus, when Gell-Mann gave a seminar on the quark model in its early
stages, before any experiments were conducted, he suggested that one
THERE ARE NO ANTIREALISTS IN THE LABORATORY 133
might regard them either as useful mathematical devices or as real

entities. After experiments on deep inelastic electron scattering were com-
pleted there was more reason to believe that quarks were real. Subsequent
experiments strengthened that belief.
The experiment I wish to discuss was designed to measure the K~z
branching ratio , the fraction of all K+ mesons that decay into a positron
and a neutrino (Bowen et al. , 1967). The first order of business was to
obtain a supply of stopped K+ mesons. The experimental apparatus is
shown in Fig. 1. The group obtained an unseparated positive beam of
momentum 530 MeV/c from the Princeton-Pennsylvania Accelerator
(PPA). This beam included pions and protons, in addition to the K+
mesons (kaons) needed. The kaons were identified by their range in
matter and by time of flight. The beam telescope consisted of our scin-
tillation counters, C 1, Cz, C3 , and C 4, with 6.7 ern of copper placed before
the stopping region, which was counter C 3 . A stopped particle was indi-
cated bLa coincidence between C 1, C z, and C 3 with no pulse in C 4
(C 1CZC3C 4) . The copper eliminated virtually all of the protons, which
have a shorter range in matter than kaons, before they reached C 3 • In
addition, protons do not decay into positrons." Pions were a more serious
problem. There were about 1000 times as many pions as kaons in the
beam . Most of the pions, which have a longer range than kaons, passed
through the stopping region and counted in C 4 , and were eliminated,
reducing the ratio of pions to kaons to about 100 to 1. Time of flight
provided additional discrimination. Particles of the same momentum (this
was a momentum selected beam) but different masses have different
velocities and therefore different times of flight. The internal proton beam
at the PPA consisted of bunches of protons separated in time by 34 ns.
Thus, particles were produced every 34 ns. A signal from the RF (radio-
frequency) system of the accelerator signalled the production of particles
and could therefore be used to time the beam particles. For the beam
transport system used in this experiment the difference in time of flight
betweeny ions and kaons was 8 ns, so a narrow coincidence of 3 ns
(C 1CZC3C 4 + RF) was used to separate kaons from pions. The background
of unwanted pions was reduced to approximately 5%.9
Particles from decays at approximately 90° to the incident beam were
detected by two scintillation counters , C, and C6 , and then passed through
a gas Cerenkov counter, which was set to detect high energy positrons.
The time between the K+ stop and the decay particle was recorded for
each event.
134 ALLAN FRANKLIN
,
~ BEAM
k r .
. MAGNET
COUNTE R No . I COP PE R AB SORB ER

COPPER ABSORBER SIX 4 GAP SPARK
COUNTER No .2 CHAMBER MODULE S
S TOPP ING COUNTER No.3
VETO COU NT ER NO.4
fro VETO COUNTERS -I ~~ .! ~_ CERE NKO V

COUN TER
PHOTO TU BE
EL LIP SOIDAL
MIRROR
RANGE CHAMBER
Fig . 1. Plan view of the experimental apparatus used to measure the K: 2 branching
ratio. From Bowen et al. (1976) .
A K+ decay wa~ identified by a coincidence between a stopping

K+ signal (C,C 2C3C 4 + RE) and a decay particle pulse (Cj C6 ) . If the
events were really due to K+ decays the time distribution between the
K+ stop and the decay pulse should match the known K+ lifetime. This
experimental check was performed, the lifetime measured, and the
posit ive results found are shown in Fig. 2. An electronic gate was used
to eliminate pion background, as shown in the figure.
Thus, the experimenters determined that the particles had a definite
charge , mass, and lifetime" which, in addition, agreed with the known
properties of the kaon. II If there are no K mesons, then we have no
plausible explanation of what was being observed. It would seem odd,
in such circumstances, to refer to the kaons as merely useful fictions,
even if the particles could not be directly observed in van Fraassen 's
sense of unaided human sense perception. I note that the three proper-
ties were sufficient to identify the particle as a kaon ."
This procedure also seems to me to be quite similar to identifying a
THERE ARE NO ANTI REALISTS IN THE LABORATORY 135
4 r,- - - - - - - - - - - - - - - - - - - - - .
Fig . 2. The decay time spectrum obtained by Bowen et al. (1967). The K+ lifetime
is shown.
person by noting that measurements on them of all the properties listed

on a driver's license give definite values . Suppose we had an entity
who had a definite height, weight, gender, hair color, eye color, date of
birth , and home address, which were, in addition , exactly those listed
on the driver's license of Bas van Fraassen. Would we not be justified
in concluding not only that there is a real Bas van Fraassen, but that
136 ALLAN FRANKLIN
the entity we are observing is one and the same person . Yes, the skep-
tical reader, or an antirealist, might reply, but we can observe the entity
directly. Suppose, however, that we had determined all of these quanti-
ties without such direct viewing, as well we might have." Would we
not still be justified in believing in the real existence of the philoso-
pher? This seems to me to be by far the best explanation of the
observations. If there is no Bas van Fraassen then we are faced with a
remarkable and bizarre set of coincidences - similarly for K mesons."
It might be argued that there is a difference between the arguments for
the existence of K mesons and the existence of Bas van Fraassen. Van
Fraassen is an individual, whereas K mesons are a type. 15 I do not believe
that this is a valid objection. I could, for example, name my K mesons.
Lest this be regarded as totally facetious, let me point out that scien-
tists are, in fact, already referring to individual elementary particles by
name .
' Here .. . in the center of our Penning trap resides positron (or anti-electron) Priscilla,
who has been giving spontaneous and command performances of her quantum jump ballets
for the last 3 months'. There can be little double about the identity of Priscilla dur ing
this per iod, since in ultrahigh vacuum she never had the chance to trade places with a
passing antimatter twin . The well-defined identity of this elementary particle is some-
thing fundamentally new, and deserves to be recognized by being given a name, just as
pets are given names of persons (Dehmelt, 1990, p. 539).
A further example of the existence of, at least, a small number of

entities, namely xenon atoms is shown in Fig. 3. A group of scientists
used a scanning, tunneling electron microscope to manipulate a group
of xenon atoms to spell out 'IBM ' (The company the scientists worked
for is clear). This is an example in which the manipulability of the objects
themselves argues for their existence. As Hacking might say, "If you
can spell IBM with them, they are real".
In establishing the observed properties of the kaons we have also made
use of, and made a commitment to, several physical laws. Thus, the
momentum of the particles was fixed by requiring them to travel in a
circle of fixed radius in a known magnetic field and then determining the
momentum by using F = rna, a = v'lr, and F = qvB where F is the force
on the particles, q, m, v, and a are its charge, relativistic mass, velocity,
and acceleration, r is the radius of curvature, and B is the known magnetic
field. From this we find that momentum = mv = qBr. We have also
made use of the dependence of the range of charged particles on their
charge, mass, and velocity. Our use of these laws to establish the prop-
THERE ARE NO ANTI RE ALISTS IN THE LABORATORY
Fig. 3. A magnified picture of xenon atoms spelling out IBM.

137
138 ALLAN FRANKLIN
erties of kaons seems to me to give them the same epistemic status as

the particles and their properties. The successful performance of the
measurements gives us reason to believe not only in the kaons, but
also in the laws . If the laws weren't valid it is hard to imagine that the
measurements would be possible." In contrast to Sellars's view, these
laws do not involve K mesons per se, but are laws obeyed by all charged
particles. Belief in these particular laws does not give us good reasons
for belief in K mesons.
An example where the support for a law does give us good reason
to believe in an entity is the following. The discovery of the n- particle
supported the eightfold way, a particle classification scheme, which
both predicted and accounted for the existence of certain elementary
particles, including the n-. This discovery also gave us good reason
for belief in kaons , which are part of the scheme. This will be discussed
further below. Similar support would be provided for the existence of
elements by the discovery of new elements predicted by Mendeleev's
Periodic Table."
2. IS CONSTRUCTIVE EMPIRICISM EMPIRICALLY ADEQUATE ?
a. Is It Empirical ?
How might van Fraassen react to the argument given above for the
existence of K mesons? " His discussion in The Scientific Image (1980,
pp. 75-77) states that measurement of a particle 's properties, such as
the charge on the electron, does not imply that the particle exists. In
his view a theory leaves blanks for experiment to fill in and claims that
this type of experiment, "shows how that blank is to be filled in if the
theory is to be empirically adequate (emphasis in original)" (1980,
p. 75). I shall discuss in detail later why I believe that this analy sis, based
on Millikan's experiments, is historically inadequate. I also do not believe
that this is a sufficient answer to the argument in favor of the existence
of kaons . First, there is no theory of kaons for which the blanks have
to be filled in, unless it is the statement that every particle has a mass,
a charge, and a lifetime ." Second, the experimenters measured three
properties of the particle simultaneously, rather than just one , providing
more evidence for its existence. It is hard to imagine filling in three
blanks simultaneously without believing that there is actually some-
thing that has the properties.
THERE ARE NO ANTIREALISTS IN THE LABORATOR Y 139
Van Fraassen would also regard the K meson or its properties as unob-
servable because the story I have recounted involves instrumental
detection and not unaided human perception. "A calculation of the mass
of a particle from the deflection of its trajectory in a known force field
is not an observation of that mass" (1980, p. 15). There is a real question
as to why van Fraassen privileges unaided sense perception in arguing
for the existence of entities and for the validity of measurements.i"
Human perception is notoriously unreliable. It can be influenced by
weather conditions (mirages), the state of the body (alcohol, drugs,
etc.), stress, and so on. Eyewitness identification in trials has been shown
to be far from infallible, and optical illusions do occur. It seems to me
that the arguments one gives for the validity of human observations
are, in fact, the same arguments one gives for the validity of instrumental
observations and measurements and that neither is privileged over the
other. (For details of the strategies used to establish the validity of exper-
imental results see Franklin (1986, Chapter 6) and Franklin and Howson
(1988). A similar point on instrumental detection has been made by
Bogen and Woodward (1988) .)
Despite this argument, van Fraassen might very well still deny the
existence of the kaon on the grounds that we have only measured or
calculated its properties and not observed it directly. It is not clear to
me, however, how he would explain the measurements if not by the
existence of a particle. As Cartwright might say, "If there are no kaons
then we have no explanation of our measurements"."
Van Fraassen might respond that, although there are indeed K mesons,
constructive empiricism deals only with the attitude one should take
toward theories that involve kaons. He might further state that the
evidence in support of such theories only gives us reason to believe
the theory is empirically adequate and does not give us reason to believe
in the existence of kaons, whereas the measurements I have cited do
support such existence. I don 't believe that such an evidential distinc-
tion can be maintained, that between evidence in support of the existence
of a particle such as the kaon and evidence that supports a theory of
kaons, and therefore their existence.
Consider the eightfold way mentioned above. One found evidence
in favor of the existence of the n- particles, i.e., its mass and charge. The
existence of such a particle was predicted by the eightfold way and there-
fore its observation supports, at the very least, the empirical adequacy
of the theory. But, the content of the theory is precisely the existence
140 ALLAN FRANKLIN
of certain particles and their properties. Therefore, I believe that van

Fraassen would have to accept that detection of the Q- provided support
for the existence of K mesons . I don't believe that having accepted this
argument that van Fraassen is then justified in excluding evidence that
supports a theory, but does not bear directly on the entities involved in
the theory. For example, observation of atomic parity violation supports
the Weinberg-Salam unified theory of electroweak interactions. It also
supports the existence of the intermediate vector bosons that the theory
uses to explain this effect." It would seem then that he would have to
agree with Sellars and myself that any evidence that supports a theory
gives ipso facto good reasons for believing in the entities postulated
by that theory. Once van Fraassen has accepted the arguments given in
favor of the existence of K mesons, as I believe he should, then he has
lost the battle.
A second point concerns the question of what van Fraassen means
when he says that acceptance of a theory involves only the claim that
what it says about "observable phenomena" (1980, p. 57) is true. He states
earlier that "the term' observable' classifies putative entities (entities that
mayor may not exist" (1980, p. 15). It seems clear that he cannot
regard only entities as observables, for then theories would make very
few statements concerning observables. Theories also make statements
about measurable quantities and certainly acceptance of a theory must
involve, at the very least, the belief that what it says about such mea-
surable quantities is true. If a constructive empiricist does not subscribe
to at least this much then the science they believe in is so impover-
ished as to scarcely merit the name.
To make this clearer, let us consider a kinetic theory which consi sts
of a hard sphere model of gas molecules and Newton 's laws. One con-
sequence of this theory is that PV = constant, at constant temperature,
where P is the pressure of the gas, and V is its volume. We would cer-
tainly regard V as measurable, if not observable. For a reasonably sized
container, which has the shape of a rectangular solid , we can determ ine
the volume using only a ruler and the naked eye. What then of pressure?
We can certainly detect pressure with our unaided senses . All one needs
to do is to dive beneath the surface in a swimming pool. This would seem
to make pressure an observable. In order to see whether or not what
the theory says about these observables is true, we must not only be
able to detect pressure, we must be able to measure it. To do this we
will need measuring devices or instruments which not only will go
THERE ARE NO ANTIREALISTS IN THE LABORATOR Y 141
beyond the unaided senses" but will also, no doubt, involve a theory
of the apparatus." I don't see how a constructive empiricist can avoid
allowing quantities like pressure and volume to be observables. Otherwise
theories will have virtually no empirical content. Once, however, it is
granted that pre ssure is an observable then other quantities detected
only with instruments also become observable, such as the mass, lifetime,
and charge, etc., of particles. That being granted, we can then say that
we have good reasons, based on observation, for the existence of
entities.
b. Is It Adequate?
One of van Fraassen 's arguments in favor of constructive empiricism

is that it provides an accurate description of scientific practice. " However,
there is also a positive argument for constructive empiricism - it makes
better sense of science, and scientific activity, than realism does and does
so without inflationary metaphysics" (1980, p. 73). I wish to dispute
this claim.
If scientists are interested only in empirical adequacy then we cannot
explain why so much effort is devoted to resolving a contradiction
between two theories , each of which is empirically adequate within a
certain range of phenomena. Why should scientists have worried about
the contradiction between Maxwell's electromagnetic theory and Bohr's
atomic theory? Both were, after all, empirically adequate within their
own ranges of phenomena. It was the fact that Maxwell's theory pre-
dicted that the Bohr atom would be unstable (the Bohr atom should dec ay
in 10-9 seconds). Why not accept the empirical adequacy of the theories
for their separate phenomena, and let it go at that? This is not what
scientists do , or what they did in this case. They work to resolve or
eliminate the contradiction. The scientific realist has an explanation of
this . The reason scientists act as they do is that when they accept a theory
they believe in its truth, and truth cannot be contradictory.
Van Fraassen might respond that Bohr 's theory shows that Maxwell 's
is empirically inadequate, or vice versa. But this will not work. It is
true that the two theories are incompatible, but there is nothing in the
available data on atomic spectra that contradicts Maxwell's theory. The
theory didn 't say anything about atomic spectra. It is true that there
was no classical electromagnetic explanation of the phenomena, but
that might just provide a reason for looking for one. Still, there already
142 ALLAN FRANKLIN
existed an empirically adequate theory of atomic spectra , namely Bohr 's

theory. Why search for another one if your aim is only saving the
phenomena?
An antirealist might respond that the removal of the contradiction is
necessary in order to construct a unified theory of the phenomena, and
that unified theories have, in the past , turned out to be more empiri-
cally adequate, i.e., they fit more phenomena than the separate theories
they replace. I agree that the search for unified theories is part of
scientific practice, and that it has sometimes, although certainly not
always, been successful in the past. Recall Einstein's or Heisenberg's
futile searches for a unified field theory late in their careers." I do not
believe the antirealist position offers any advantage here over the realist
view in the explanation of scientific practice. At best, it seems equally
good. This judgment of equality depends, of course, on believing that the
search for a unified theory is usually, if not always, successful, and that
such unification results in theories that are more empirically adequate.
I am not convinced that this is always the case . The unifications we
remember in the history of science do satisfy this criterion, but we tend
to forget those that don't.
My suggestion of equality does not hold for van Fraassen's explana-
tion of the episode of Millikan's measurement of the charge on the
electron. As mentioned earlier, he views Millikan 's measurement of the
charge on the electron as filling in a blank in atomic theory. Although
he admits that the experiment was "a test of the theory that there exists
this elementary electrical charge, it was not surprising at this time that
such tests should bear out theory" (1980 , p. 75). He concludes that, "In
a case such as this one, experimentation is the continuation of theory
by other means" [emphasis in original] (1980, p. 77).
While I agree with him that at the time most, but certainly not all,
physicists believed that there was an elementary electrical charge , the
question was an open one. We can look at the introductions to Millikan's
own papers during this period to see what he regarded as the important
questions and answers (I note that at this time the question of whether
or not there was a fundamental unit of electrical charge and what the
value of that unit was were closely related) .
In 1911 Millikan reported that it had been found possible
To present direct and tangible demonstrat ion through the study of the behavior in elec-
trical and gravitational fields of this oil drop , carrying its captured ions, of the correctnes s
of the view advanced many years ago and supported by evidence from many source s
that all electrical charges, howe ver produced , are exact multiples of one definite
elementary, electrical charge , or in other words , that an electrical charge instead of being
spread uniformly over the charged surface has a definite granular structure, consisting,
in fact, of an exact number of specks , or atoms of electricity, all precisely alike , peppered
over the surface of the charged body .
To make an exact determination of the value of the elementary electrical charge
which is free from all que stionable theoretical assumptions and is limited in accuracy
only by that attainable in the measurement of the coefficient of viscosity of air (Millikan,
1911, p. 350).
The ordering of these statements seems to indicate Millikan's view of

their relative importance.
By 1913 Millikan regarded the question of charge quantization as
settled on the basis of his own previously published (1911) work. "The
total number of changes which we have observed would be between
one and two thousand, and in not one single instance has there been
any change which did not represent the advent upon the drop of one
definite invariable quantity of electricity, or a very small multiple of
that quantity" [emphasis in original] (Millikan, 1911, p. 360). His 1913
paper begins, "The experiments herewith reported were undertaken with
the view of introducing certain improvements into the oil drop method
of determining e and N and thus obtaining a higher accuracy than had
been in the evaluation of these most fundamental constants" (Millikan,
1913, p. 109). He is, in fact , filling in the blank , but only because he
regarded the more important question of charge quantization as having
already been answered.
During the course of these measurements Millikan was engaged in a
controversy with Felix Ehrenhaft concerning the quantization of charge.
(For details see Holton [1978] and Franklin [1986, Ch. 5].) During 1912
and 1913 a lull occurred in the controversy and opinion, as well as
other experimental results, was generally favorable to Millikan. Ehrenhaft
and two of his students, Zerner and Konstantinowsky, returned to
the attack in 1914 and 1915 (Ehrenhaft, 1914; Zerner, 1915; and
Konstantinowsky, 1915). Millikan answered the criticism, I believe
successfully, in a paper entitled "The Existence of a Subelectron?"
(Millikan, 1916).
The vast preponderance of evidence and the opinion of the physics
community strongly favored Millikan and he was awarded the Nobel
Prize in Physics in 1924. The presentation speech by Gullstrand further
emphas izes the importance of experimentally establishing charge quan-
tization .
144 ALLAN FRANKLIN
Mill ikan 's aim was to prove that electricity really has the atomic structure, which, on
the basis of theoretical evidence, it was supposed to have . . . By a brilliant method of
investigation and by extraordinarily exact experimental technique Millikan reached his
goal . . . Even leaving out of con sideration the fact that Millikan has proved by these
researches that e lectricity consists of equal units, his exact evaluation of the unit has
done physic s an inestimable service, as it enables us to calculate with a higher degree
of exactitude a large number of the most important physical constants (Gullstrand, 1965).
The filling in of the blank of the value of e, as van Fraassen has it,
while of great importance, is secondary to the issue of charge quanti-
zation. Even three years after Millikan had received the Nobel Prize,
O. D. Chwolson, a respected physicist, wrote
It (the Millikan -Ehrenhaft dispute) has already lasted 17 years and up to now it cannot
be claimed that it has finally been decided in favor of one side or the other, i.e., that all
researchers have adopted one or the other of the two possible solutions of this problem .
The state of affairs is rather strange (Chwolson, 1927).
Although by that time most physicists would have disagreed with this
assessment, it indicates that the issue had not been closed by 1911 as
van Fraassen states.
I believe that the realist account that Millikan 'discovered' or gave
strong evidence for the existence of a fundamental unit of electricity, and
then made a precise measurement of that unit, is a far better descrip-
tion of this episode than van Fraassen's view that Millikan was merely
filling in a blank left by theory .
3. CONCLUSION
I believe that the discussion of kaons has shown that we can have good
reasons to believe in both the existence of entities and in laws , and that
our belief in the laws is a belief that it is true. If my arguments are correct
then constructive empiricism is not philosophically justified. I also
believe that I have cast doubt on the empirical adequacy of construc-
tive empiricism. It has a vague notion of 'observable', which I believe
is far too strict to apply to the practice of science. If one extends the
notion as I have suggested then it supports a realist position. Constructive
empiricism is, at best, no better than the realist account in discussing
the elimination of contradictions between theories or the unification of
theories. It also fails to give an adequate account of Millikan's experi -
ment, which is better explained by a realist position.
Nevertheless, showing that constructive empiricism is not satisfac-
tory provides only a small amount of support for the realist view. The
elimination of one of many alternative explanations and there are many
of them) only slightly strengthens our belief in the remaining alterna-
rives." I have, however, also given positive arguments for the particular
version of realism that I support. This includes the discussion of the exis-
tence of kaons, the elimination of contradictions, and the historical
account of Millikan's experiments.
Supporting a realist position does not, however, mean that I believe
in either the absolute truth of the laws or in the 'real' existence of the
entities. It means only that I think we have good reasons for believing
in the truth of the laws and in the existence of the entities.
Department of Physics,
University of Colorado,
U.S.A.
NOTES
I See, for example , the recent book on realism edited by Leplin (1984) and the volume
devoted to discussion and criticism of van Fraassen 's view edited by Churchland and
Hooker (1985) .
2 Van Fraassen admits that 'observable' is a vague predicate but argues that one can
still make use of it. I believe , as will be discussed later, that his characterization of observ-
able is, in fact, too strict.
3 John Stachel has raised the interesting question as to whether or not my view commits
me to arguing that 'virtual' particles such as those referred to in the empirically very
successful quantum electrodynamics actually exist or whether I regard them as useful visu-
alizations or interpretations of the mathematical formalism . I tend to lean toward the
latter view, but I must admit I have not thought about the issue sufficiently. Although I
do regard Feynman diagrams, which include virtual particles, as very useful pictures,
each line or vertex in such diagrams can be replaced by a mathematical expression, but
how one should regard such lines is not clear to me.
4 Morrison (1990) has argued that manipulability is not sufficient to establish belief in
an entity. She discusses particle physics experiments in which particle beams were viewed
not only as particles, but also as beams of quarks, the particle constituents, even though
the physicists involved had no belief in the existence of quarks . Although I believe that
Morrison's argument is correct in this particular case, I do think that manipulability can,
and often does, give us good reason to believe in an entity. See, for example, the dis-
cussion of the microscope in Hacking (1983). More recently scientists have used the
scanning tunneling microscope to spell out IBM with xenon atoms . This seems to me to
be very good evidence for such atoms .
5 In Cartwright's discussion of the electron track in the cloud chamber, for example,
she can identify the track as an electron track rather than as a proton track only because
146 ALLAN FRANKLIN
she has made an implicit commitment to the law of ionization for charged particles, and
its dependence on the mass and momentum of the particles. The momentum is measured
by curvature in a known magnetic field in a way similar to that discussed below for K
mesons. A similar point has been made by Morrison (1989). See also my review of
Cartweight in Foundations of Physics (1984).
6 I am grateful to Bob Ackermann for raising this point .
7 I note that it is the argument that is constructed, not the weak neutral currents, as
some have stated (see Pickering, 1984).
K Although modem theory does predict proton decay into a positron, both the measured
proton lifetime and the predicted rate give a decay rate that is too small to be of any
significance in this experiment.
9 The reader may note that the pion also decays into an electron and a neutrino, with a
branching ratio approximately seven times that of the kaon. These pion decays could easily
be separated from kaon decays by their momentum. The momentum of an electron from
pion decay is 69.8 MeV/c, while that for an electron from kaon decay is 246 MeV/c.
The experimental apparatus had a momentum resolution of 1.9%, which could easily
distinguish between the two decays.
10 In fact, the experimenters had assumed that the charge on the particle was e. They did,
however, measure the range, the radius of curvature in a known magnetic field, and the
time of flight of the particles. From these three measurements, the three unknown quan-
tities, the charge, mass, and velocity of the beam particles could have been calculated.
II The reader may object that using agreement with known properties of the kaon already
assumes that the kaon exists . I believe, however, that establishing that a particle has a
definite mass, charge, and lifetime is sufficient to establish its existence. Recall J. J.
Thomson's 'discovery' of the electron by showing that cathode rays had a definite charge
to mass ratio.
12 There are of the order of 100 elementary particles, each with different properties (in
particular each particle has a unique mass) so that specifying three properties serves to
identify the particle .
13 One could , for example, measure the weight of the entity by using a scale which
was located behind an opaque screen, but which had a remote readout visible to the
observer, and so on.
14 I note that this story agrees with Cartweight's causal view, but differs slightly from
Hacking 's because it is the K mesons themselves that are under investigation.
IS I am grateful to Zeno Switinjk for pointing this out and also for providing the
quotation from Dehmelt, which argues against his own point.
16 I have used the term 'valid' to describe the laws. I don't think the laws have to be
true in order to have a successful measurement, and I don't know how to make sense
of the term "approximately true". What I mean by valid is that to within the required
experimental accuracy the laws give correct results . This does not mean that I do not
feel that we are justified in our belief that the laws are true . Observation of evidence
entailed by a theory should, and indeed does, strengthen our belief that the theory is
true. It also strengthens our belief that the theory is empirically adequate, but that is beside
the point.
17 For a discussion of this historical episode and a discussion of the evidential value
of prediction and accommodation see Howson and Franklin (1991) .
IK For the sake of economy I am attributing the arguments that might hypothetically
be offered by a defender of constructive empiricism to its foremost proponent, Bas van
Fraassen. He might not, of course, agree that he would offer any of these arguments.
19 This is not strictly true. The current theory of strongly interacting particles, quantum
chromodynamics (QCD), does, in principle, predict the mass of the kaon and other
elementary particles, given the masses of the up, down, and strange quarks. Because
there are more particles than quarks, one can use the observed masses of some of the
particles to calculate the quark masses, and then proceed to calculate the masses of
the other particles. These are very difficult and complex calculations and, at present,
are accurate to approximately 20%. This theory was not available in 1967, when the
experiment was performed.
20 Van Fraassen actually privileges vision over the other senses . No one would regard
hearing something as evidence for the existence of an entity. Recall the old television
commercial "Is it live or is it Mernorex?" in which a listener cannot distinguish between
a recording and a live person. The sense of touch also does not provide reliable evidence
for the existence of an entity. Remember the story of the five wise men each touching a
different part of an elephant and reaching different conclusions as to its nature .
21 Cartwright (1983), pp. 87-99, argues that when we have a causal explanation of a
measurement or an observation then we are justified in making an inference to the most
probable cause, i.e., an entity . I agree . Van Fraassen might ask why we need an expla-
nation at all.
22 I note that more direct evidence for the existence of these particles was found
later.
23 It may very well be true that any detection device, such as a mercury barometer,
will have an output detectable by unaided human senses, but that is not the point here .
24 For a discussion of how one comes to believe in an experimental result see Franklin
(1986), Chapter 6, and Franklin and Howson (1988). For further discussion of the theory-
ladenness of measurement see Franklin (1989).
25 The search for unification also applies to noncontradictory theories . Take, for example,
the V-A theory of weak interactions and quantum electrodynamics. These were not con -
tradictory. Each was empirically adequate for its own range of phenomena. The search
for unification was successful, leading to the Weinberg-Salam unified theory of elec-
troweak interactions. This unified theory was widely hailed as a major achievement and
led to the prediction and observation of new phenomena, weak neutral currents, atomic
parity violation, and the observation of intermediate vector bosons, the Wand ZO parti-
cles . Similarly, Maxwell 's theory unified the separate theories of electricity and magnetism
and led to the observation of electromagnetic radiation .
26 For a discussion of this see Franklin and Howson (1988) .
BIBLIOGRAPHY
Bogen, J. and Woodward, J. (1988). ' Saving the Phenomena' , The Philosophical Review
47, pp. 303-352.
Bowen, D. R. et 01. (1967). ' Measurement of the K: 2 Branching Ratio ', Physical Review
154, pp. 1314-1322.
148 ALLAN FRANKLIN
Cartwright, N. (1983) . How the Laws of Physics Lie, Oxford University Press, Oxford.
Churchland, P. and Hooker, C. (eds.) (1985). Images of Sciences, University of Chicago
Press, Chicago .
Chwolson, O. D. (1927) . Die Physik, 1914-1926, G. Kluge (trans.), F. Viewig and Sohn,
Braunschweig.
Dehmelt, H. (1990) . 'Experiments on the Structure of an Individual Elementary Particle',
Science 247, pp. 539-545.
Ehrenhaft, F. (1914). 'Die Quanten der Elektrizitat', Annalen der Physik 44, pp. 657-700.
Fraassen van, B. (1980) . The Scientific Image, Clarendon Press, Oxford .
Franklin, A. (1986). The Neglect ofExperiment, Cambridge University Press, Cambridge.
Franklin, A. and Howson, C. (1988). 'It Probably is a Valid Experimental Result: A
Bayesian Approach to the Epistemology of Experiment', Studies in History and
Philosophy of Science 19, pp. 419-427.
Franklin, A. et al. (1989). ' Can a Theory -Laden Observation Test the Theory?' British
Journal for the Philosophy of Science 40, pp. 229-231.
Galison, P. (1987) . How Experiments End, University of Chicago Press, Chicago.
Gullstrand, A. (1965) . Nobel Lectures in Physics, 1922-41, Amsterdam, pp. 51-53.
Hacking, I. (1983) . Representing and Intervening, Cambridge University Press.
Holton, G. (1978). 'Subelectrons, Presuppositions, and the Millikan -Ehrenhaft Dispute',
Historical Studies in the Physical Sciences 9, pp. 161-224.
Howson, C. and Franklin, A. (1991) . 'Maher, Mendeleev , and Bayesianism', Philosophy
of Science 58, 574-585.
Konstantinowsky, D. (1915). 'Elektrische Ladungen und Brownsche Bewegung sehr
kleiner Metallteilchen im Gase ', Annalen der Physik 46, pp. 261-297.
Leplin , J. (ed.) (1984) . Scientific Realism, University of California Press , Berkeley .
Millikan, R. A. (1911). 'The Isolation of an Ion, A Precision Measurement of Its Charge,
and the Correction of Stoke's Law', Physical Review 32, pp. 349-397.
Millikan, R. A. (1913). 'On the Elementary Electrical Charge and the Avogadro Constant',
Physical Review 2, pp, 109-143.
Millikan, R. A. (1916). 'The Existence of a Subelectron?' Physical Review 8, pp. 595-625.
Morrison, M. (1990) . 'Theory, Intervention, and Realism', Synthese 82, pp. 1-22.
Sellars, w. (1962). Science , Perception, and Reality, Humanities Press, New York,
p.97.
Zerner, F. (1915). 'Zur Kritik des Elementarquantums for Elektrizitat', Physikallische
Zeitschrift 16, pp. 10-13.
KOSTAS GAVROGLU
CAN THEORIES OF CHEMISTRY PROVIDE

AN ARGUMENT AGAINST REALISM?
Right after the formulation of quantum mechanics by Heisenberg, Born,

Jordan and Schrodinger, most chemists became aware of the amazing
explanatory power of the new quantum mechanics, yet it was difficult
to see how this newly developing explanatory framework would be
assimilated into the chemists' culture. Many chemists were apprehen-
sive that such an assimilation might bring lasting, and not altogether
welcome, changes to their culture. But for some, it was a risk worth
taking. Neville Sidgwick in his influential book The Electronic Theory
of Valency written just after the dramatic developments of 1926, would
have no inhibitions about letting the new quantum mechanics invade
the realm of chemistry. He expressed an unreserved enthusiasm about
the new quantum mechanics. Faced with the full development of the
new mechanics by Heisenberg and Schrodinger, but not, yet, with an
application of the theory to a chemical problem, Sidgwick in the very
first lines of the preface to his book attempted to clarify the method-
ological stumbling block that he sensed would be in the way of his fellow
chemists
In developing a theory of valency there are two courses open to the chemist. He may
use symbols with no definite physical connotation to express the reactivity of the atom s
in a molecule , and may leave it to the subsequent progress of science to discover what
realities these symbols represent: or he may adopt the concepts of atomic physics - elec-
trons, nuclei , and orbits - and try to explain the chemical facts in terms of these. But if
he takes the latter course, as is done in this book, he must accept the physical conclu-
sions in full, and must not assign to these entities properties which the physicists have
found them not to possess. He must not use the terminology of physics unless he is
prepared to recognize its laws. I have endeavoured to conform to this principle, and not
to lay myself open to the reproach of an eminent physicist, that 'when chemists talk
about electrons they use a different language from the physicists ' . 1 have been careful
to avoid as far as possible the introduction of any physical hypotheses which are not already
sanctioned by those who are best qualified to judge of them. 1
Exactly fifty years later, at a Conference in 1977 commemorating

half a century of Valence theory, Charles Alfred Coulson gave the closing
talk. He was a mathematician by training and a writer of what became
the standard text-book on valence . Here is a passage from his talk.
149
ISO KOSTAS GAVROGLU
Valence Theory is about bonds: what they are and how many there are from each atom
. .. So fifty years of valence theory really means fifty years of changing ideas about a
chemical bond. The first third of the period . . . was necessarily concerned with identi-
fying the electronic nature of the bond, and in escaping from the thought forms of the
physicist . . . Has the chemical bond now done its job ? Have we grown to that degree
of knowledge and that power of calculation that we do not need it? . .. This is a tanta-
lizing question. Chemi stry is concerned to explain, to give us insight and a sense of
understanding. Its concepts operate at an appropriate depth and are designed for the
kind of explanation required and given. If the level of enquiry deepens, as a result of
our better understanding, then some of our older concepts no longer keep their rele-
vance .. . Whether "something bigger" that should replace the chemical bond, will come
to us or not is a subject for a Symposium bearing for its title: the changing role of chemical
theory.'
Coulson could not have expressed any better the often misunder-
stood relationsip between chemistry and physics. Borrowing and using
concepts first proposed by physicists, and, subsequently, "escaping from
the thought forms of the physicists" had really been the dominant trend,
not only in the first years of quantum chemistry, but also in much of
the history of physical and structural chemistry in the latter part of the
19th century as well.
And though chemistry was striving to free itself from the "thought
forms of the physicist", the theoretical issues raised by this escape have
been, on the whole, ignored by philosophers of science. I am not, of
course, saying that there have been no philosophical discussions of the
problems of chemistry, since questions about the phlogiston and the
Daltonian atoms have been very systematically discussed. But the neglect
of the philosophical problems arising out of the particularities of chem-
istry after the development of microphysics since the middle of the 19th
century is also a fact.
In all the discussions about scientific realism, there is at least one
undeclared assumption. If pressed, the various arguments, claims and
conclusions can be illustrated by examples in physics . And though it is
acknowledged that biology and the social sciences pose problems of their
own, there is a rather peculiar silence about chemistry. Nothing is said
about chemistry because there is the implicit assumption that chemistry
is reducible to physics and, hence, physics is the paradigmatic science,
the science that really tells us about Nature. It is claimed, not unjusti-
fiably, that the specific role of mathematics in physics renders the
problems of scientific realism to be clearly delineated. Furthermore, it
is said that physics deals with the fundamental entities of the world
AN ARGUMENT AGAINST REALISM 151
and there are no intrinsic limitations as to how deeply it might probe .

Whether it studies the planets, billiard balls, atoms , nuclei, electrons,
quarks or superstrings, it is still physics and the change of scale does
not oblige the change, as it were, of the discipline itself - as it would
be the case in biology and chemistry. Although, on the whole , I agree
with this view, I have two reservations. The first, which I mention without
further discussion, is that it is not too clear to me that when we are
discussing quarks or superstrings or the big-bang we are doing theo-
retical, rather than mathematical physics . My second reservation is that
the view which confines the study of realism predominantly to the
problems of physics is just a convenient mode of discourse and it neglects
the theoretical particularity of chemistry. And what is much more impor -
tant , it supposes an absolute reductionism of chemistry to physics. If
nothing else, these undeclared assumptions deprive philosophy of science
of a vast area where issues about the ontological status of theoretical
entities and the criteria for empirical adequacy for the acceptance of a
theory have been passionately debated. Ever since the end of the 19th
century chemists have been debating whether their science may not be
the "science of bodies which do not exist". Or whether the unsettling dis-
covery of radium implied that "in relation to the ponderable, we seem
to be creating a chemistry of phantoms." The history of chemistry is
also a history of the attempts of the chemists to establish its autonomy
with respect to physics. Otherwise the chemist would be at a loss about
the underlying ontology, and would never be sure whether chemistry
should be doing the describing and physics the explaining, The chemists
have passionately debated these issues, and the myth of the reflective
physicist and the more pragmatic chem ist is, if anything, historically
untenable.
Since most of the problems debated in realism refer to quantum
mechanical entities or, to be more precise, to entities predominantly
used in quantum mechanics, I would like to present an account of the
historical becoming of quantum chemistry in such a way as to high-
light some of its philosophical problems.
Quantum chemistry has been invariably regarded as another instance
of the success story of quantum mechanics. Textbooks, leaflets, ency-
clopedias and the vast majority of scholarly articles discuss the successful
application of quantum mechanics to chemical problems. Its birth and
early development in the late twenties and early thirties was during the
period when physics was reasserted as the unquestionable paradigm for
152 KOSTAS GAVROGLU
a "good and proper" scientific discipline. Certainly the community of

physicists was under the spell of promises to be delivered by Dirac's
reductionist program, expressed as a theoretically correct, but practi-
cally meaningless dictum
The general theory of quantum mechanics in now almost complete ... The underlying
physical laws necessary for the mathematical theory of a large part of physics and the
whole of chemistry are thus completely known , and the difficulty is only [that] the exact
application of these laws leads to equations much too complicated to be soluble .'
There was, thus, no more fun in chemistry, its laws had been deciphered
and what was left was a matter of devising shrewd methods for carrying
out the calculations.
The point I want to argue is that the beginnings and the establish-
ment of quantum chemistry involved a series of issues which transcend
the question of the application of quantum mechanics to chemical
problems. Quantum chemistry developed an autonomous language with
respect to physics and what appeared to be disputes over methods were,
in fact, discussions concerning the collective decision of the chemical
community about methodological priorities and ontological commitments.
The outstanding issue to be settled in the community turned out to be
the character of theory for chemistry and, therefore, a reappraisal of
the praxis of the chemists. Is it the case that , especially after the advent
of quantum mechanics, all the different theoretical and cultural factors
precipitated in such a way as to determine and impose a common way
of doing chemistry? I am inclined to think that, as a rule, disputes and
dis agreements were as much about getting the correct solution to a
problem as they were part of a rhetoric about how to go about solving
similar kinds of problems. In many instances the scientific papers had
a strong rhetoric 'propagandizing' various changes in the chemists'
culture. In other words . during the 1930s the discussions and disputes
among chemists were to a large extent about the new legitimizing pro-
cedures and consensual activities to be incorporated within the chemists'
culture.
Drawing up a program to examine the nature of the chemical bond
presupposed a particular attitude on how to construct a theory in
chemistry, on how much one 'borrows' from physics and what the
methodological status of empirical observations for theory building is.
There were basically two different research traditions. Heitler and London
insisted on an approach which, while not as reductionist as Dirac's pro-
nouncement of 1929, followed this path of orthodoxy. Pauling and

Mulliken had a strong inkling to semi-empirical methods whose only
criterion for acceptability was their practical success. To suppose that the
question of a stronger command over the mathematical details is the
sole differentiating criterion between the two styles is quite misleading.
The difference could only be understood in terms of two different cultures
for doing (quantum) chemistry. It is a matter of explicating the theo-
retical and methodological coherence of the proposed schemata, and
realizing that they constitute two diverging programs . At times it appeared
to be a confluence of different styles of research and, at the same time
there was an uneasy feeling that not all differences could be reconciled.
THE MYSTERIOUS BOND
Among the first successful applications of quantum mechanics was the

calculation of the energy levels of the hydrogen atom and of the hydrogen
molecular ion. By the beginning of 1927 many physicists and chemists
talked about the possibilities of quantum mechanics to deal with actual
chemical problems. The simplest, but deeply intriguing, chemical problem
was the formation of the hydrogen molecule. The 'mechanism' respon -
sible for a such a formation - the homopolar bond - was quite puzzling,
since it joined two electrically neutral atoms to form a molecule. Earlier
on, there had been quite a few suggestions to explain the homopolar bond,
but it was only since 1916 and within the framework of the old quantum
theory, that some remarkably insightful proposals by Gilbert Newton
Lewis provided a rather simple rule to deal with such a puzzling bond.
Lewis had proposed that chem ical bonding - both the ionic type as well
as the homopolar type - could be explained in terms of shared electron
pairs . It was a semi -empirical ' theory' . He had started from the obser-
vation that almost all the molecules had an even number of electrons.
By requiring the molecule to have eight electrons in the outer orbits of
each atom he argued that two neutral atoms could only be joined together
by the sharing of pairs of electrons between them.
In 1923, after the publication of Lewis 's book Valence, the Faraday
Society organised a meeting at the University of Cambridge entitled
the "Electronic theory of valency". J. J. Thomson was the Chairman of
the meeting and in his opening statement he referred to the tasks awaiting
the chemists. He felt that the problem of the chemical bond dominated
chemistry and that the problem itself found its most suggestive mode
154 KOSTAS GAVROGLU
of expression in terms of electrons." Lewis gave the introductory talk and

though he thought that the aim of the meeting was to secure a better
mutual understanding of divergent points of view, the expressions he used
in his presentation failed to convince the audience of his tolerance about
the kind of pluralism he advocated. He asserted that the "cardinal phe-
nomenon of all chemistry" was the formation of electron pairs. For Lewis
this was an actual pairing and not a convenient mode of explanation.
He could not offer an explanation of the mechanism itself, but he had
no doubt that sooner or later quantum theory would provide an expla-
nation of electron pairing.
HEITLER AND LONDON IN ZURICH
During their short stay in Zurich, Fritz London together with Walter
Heitler managed to solve one of the outstanding problems of chemistry
by using the wave mechanical methods of Schrodinger. They showed that
the mysterious chemical binding of two neutral hydrogen atoms to form
a hydrogen molecule could only be undertood in terms of the princi-
ples of the new quantum mechanic s.
Undoubtedly the simultaneous presence of both Heitler and London
in Zurich was one of those unplanned happy coincidences. Walter Heitler
was born in Karlsruhe in 1904. His father was a profe ssor of engineering.
His interest in physical chemistry grew while he attended lectures on
the subject at the Technische Hochschule and through these lectures in
physic al chemistry he came into contact with quantum theory. Wishing
to work in theoretical physics, he first went to Berlin, but found the
atmosphere not too hospitable especially since a student was left to
himself to choose a problem and write a thesis. Only after its comple-
tion would the ' great men' examine it. After a year in Berlin he went
to Munich and completed his doctoral thesis with Karl Herzberg on
concentrated solutions. After completing his thesis, Sommerfeld helped
him to secure funding from the International Education Board, and he
went to Copenhagen to work with J. Bjerrum on a problem about ions
in solutions. He was not particularly happy in Copenhagen. Determined
to work in quantum mechanics, he convinced Bjerrum, the Education
Board and Schrodinger to spend the second half of the period for which
he received funding in Zurich.
Fritz London was born in 1900 and his father was a Professor of
Mathematics at the University of Bonn . To graduate from the University
of Munich in 1921 he wrote a thesis in philosophy under the supervi-

sion of one of the best known phenomenologists, Alexander Pfander.
In his thesis he discussed questions of deductive systems. After teaching
at the Gyrnansium, he went to Sommerfeld's institute at the University
of Munich and started working in physics. His first paper in physics,
in 1925, was on the broadening of the spectral lines. In 1926 he was
offered an assistantship from E. E. Ewald at the Technical College
in Stuttgart. There he started work in quantum mechanics publishing
papers in transformation theory and on the quantum mechanical inter-
pretation of Weyl's unifying electromagnetism and gravitation. London,
too, went to Zurich on a Fellowship from the International Education
Board.
When they met in Zurich, Heitler and London decided to calculate
the 'van der Waals' forces between two hydrogen atoms considering
the problem to be "just a small 'by the way' problem". Nothing indi-
cates that London and Heitler were either given the problem of the
hydrogen molecule by Schrodinger or that they had detailed discus-
sions with the latter while they were proceeding with their calculations.
The acknowledgements in their paper were quite diplomatic: They thank
Schrodinger for his hospitality and "for the kind and supportive interest
with which he has guided our work". Linus Pauling, who was also in
Zurich during the same time as HeitIer and London, notes that neither
he nor HeitIer and London discu ssed their work with Schrodinger, who,
nonetheless, did know what they were all working on as witnessed by
Robert Mulliken's reminiscences of his visit to Zurich in 1927.
Schrodinger had told Mulliken that there were two persons working in
his Institute and who had some results "which he thought would interest
me very much; he then introduced me to HeitIer and London whose paper
on the chemical bond in hydrogen was published not long after".' It
may be worth noting that Ewald remembered that the question of the
homopolar bond was in London's mind before going to Zurich, and
Pauling recalled discussions with Heitler about bonding when they were
both in Munich in 1926.
BINDING FORCES
HeitIer and London's initial aim was to calculate the interaction of the
charges of two atoms . They were not particularly encouraged by their
first results, since the attraction due to the 'Coulomb integral' was too
156 KOSTAS GAVROGLU
small to account for the homopolar bond between two hydrogen atoms.
But they were puzzled by the presence of the 'exchange integral' whose
physical significance was not evident at all. Heisenberg's work on the
quantum mechanical resonance phenomenon which had already been pub-
lished was not of particular help to Heitler and London, since the
exchange was part of the resonance of two electrons one of which was
in the ground state while the other was excited and both were in the same
atom. Heitler remembered that they were stuck and "we did not know
what it meant and did not know what to do with it".
Then one day was a very disagreeable day in Zurich; [there was the] Fohn . It's a very
hot south wind, and it takes people different ways. Some are very cross . .. and some
people just fall asleep .. . . I had slept till very late in the morning, found I couldn 't do
any work at all ... went to sleep again in the afternoon. When I woke up at five o'clock
I had clearly - I still remember it as if it were yesterday - the picture before me of the
two wave functions of two hydrogen molecules joined together with a plus and minus
and with the exchange in it. So I was very excited, and I got up and thought it out. As
soon as I was clear that the exchange did playa role, I called London up; and he came
as quickly as possible . Meanwhile I had already started developing a sort of perturba-
tion theory. We worked together until rather late at night, and then by that time most of
the paper was clear. . .. Well, I am not quite sure if we knew it in the same evening,
but at least it was not later than the following day that we knew we had the formation
of the hydrogen molecule in our hands . And we also knew that there was a second mode
of interaction which meant repul sion between two hydrogen atoms - also new at the
time - new to the chemists too . Well the rest was then rather quick work and very
easy , except, of course , that we had to struggle with the proper formulation of the Pauli
principle, which was not at that time available, and also the connection with spin . . .
There was a great deal of discussion about the Pauli principle and how it could be
interpreted."
The paper was sent for publication on June 30, 1927. In the manu-
script, the introduction and the discussion of the results are in London's
handwriting. The calculations are in Heitler's handwriting. Heitler wrote
the part on the Pauli principle and molecular forces as well and London
the conclusion. Heitler and London started their calculations by con-
sidering the two hydrogen atoms coming slowly close to each other.
Electron 1 belonged to atom a and electron 2 to atom b or electron 2
belonged to atom a and 1 to b. Because the electrons were identical,
the total wave function of the system was the linear combination of the
wave functions of the two cases.
'Jf= c,'Jfa(l)'Jfb(2) + c z'Jf.(2)'Jfb(l)

The problem now was to calculate the coefficients c I and c 2• This they
did by minimizing the energy
E
f 'I'*H'I' d.
= ---:---
f '1'2 d.
They found two values for the energy
C+A C-A
E I = 2E o + S; E 2 = 2E o + S
I + 12 I - 12
The integrals C (Coulomb integral) and A (exchange integral) had
negative values, but A was larger than C. E I implied c lfc 2 = 1 and E 2
implied c/c 2 = -I. Hence the wave function of the system could now
be written as
'1', = 'l'a(1)'I'b(2) + 'l'a(2)'I'b(1)
'I'll = 'l'a(l )'I'b(2) - 'l'a(2)'I'b(l)
Up to this point, the spin of the electrons was not taken into consider-
ation . The symmetry properties required by the Pauli exclusion principle,
were satisfied only by '1'1. It was the case when the electrons had
anti-parallel spins. But '1', corresponded to E I. E I was less than 2E o,
the sum of the energies of the two separate hydrogen atoms, and ,
hence, it signified attraction. 'I'll' which, when spin was taken into con-
sideration was a symmetric combination, corresponded to E 2 • But E 2
was greater than 2E o, and it implied repulsion. The "mechanism" respon-
sible for the bonding between the two neutral hydrogen atoms was the
pairing of the electrons which became possible only when the relative
orientations of the spins of the electrons were antiparalleJ. To form an
electron pair it did not suffice to have only energetically available elec-
trons, but the electrons had to have the right spin orientations. The
homopolar bonding turned out to be a pure quantum effect, since its
explanation depended wholly on the electron spin which had no
classical analogue. As Heitler and London noted in their paper, such a
result could only be described very artificially in classical terms." They
found the bond energy to be 72.3 kcals and the internuclear distance 0.86
Angstroms to be compared with the experimental values of 109.4 and
0.74 respectively.
They soon realised that the proposed exchange mechanism obliged
them to be confronted with a fundamentally new phenomenon. They
158 KOSTAS GAVROGLU
had to answer questions posed by experimental physicists and chemists,

like "What is really exchanged? Are the two electrons really exchanged?
Is there any sense in asking what the frequency of exchange is?"
It became gradually clear to me that it has to be taken as a fundamentally new phenom-
enon that has no proper analogy in older physics. But I think the only honest answer
today is that the exchange is something typical for quantum mechanics, and should not
be interpreted - or [one] should not try to interpret it - in terms of classical physics. X
Both London and Heitler in all their early writings repeatedly stressed
this 'non-visualizability' of the exchange mechanism. It is one aspect
of their work which in the name of didactic expediency has been con-
sistently misrepresented.
Though it appeared that the treatment of the homopolar bond of the
hydrogen molecule was an 'extension' of the methods successfully used
for the hydrogen molecular ion, there was a difference between the two
cases that led to quite radical implications. It was the role of the elusive
Pauli principle. In the case of the hydrogen molecule ion, its solution was
a successful application of the Schrodinger equation where the only forces
determining the potential are electromagnetic. A similar approach to
the problem of the hydrogen molecule leads to a mathematically well
defined, but physically meaningless solution where the attractive forces
could not be accounted for. There was a need for an additional constraint,
so that the solution would become physically meaningful. At least part
of the theoretical significance of the original work of Heitler and London
was that this additional constraint was not in the form of any further
assumptions about the forces involved . Invoking the Pauli exclusion prin-
ciple as a further constraint led to a quite amazing metamorphosis of
the physical content of the mathematical solutions. These solutions
became physically meaningful and their interpretation in terms of the
Pauli principle brought about the new possibilities provided by the elec-
tromagnetic interaction.
London, in his subsequent publications, proceeded to a formulation
of the Pauli principle for cases with more than two electrons and which
was to become more convenient for his later work in group theory: the
wave function can at most contain arguments symmetric in pairs; those
electron pairs on which the wave function depends symmetrically have
antiparallel spin. He considered spin to be the constitutive character-
istic of quantum chemistry. And since two electrons with antiparallel spin
are not identical, the Pauli principle did not apply to them and one
could , thus, legitimately choose the symmetric solution, With the Pauli
principle it became possible to comprehend 'valence' saturation , and
as it will be argued in the future work of both Heitler and London, spin
would become one of the most significant indicators of valence behav-
iour and, in the words of Van Vleck, would forever be "at the heart of
chemistry"."
CHEMISTS AS PHYSICISTS ?
Among the meetings where questions related to chemical bonding and

valence were exhaustively discussed there was the "Symposium on
atomic structure and valence" organized by the American Chemical
Society in 1928 at St. Louis.
G. L. Clark's opening remarks in the 1928 meeting of the American
Chemical Society noted some of the difficulties associated with atomic
physics, but ascribed them to the failure of the chemists to test "their
well-founded conceptions with the facts of physical experimentation, and
that far too few physicists inquired critically into the facts of chemical
combination". He thought that physicists and chemists were "firmly
entrenched, each in his own domain, a certain long-range firing of static
cubical atoms against infinitesimal solar atoms has ensued, with few casu-
alties and few peace conferences. 10
Clark was not alone in attempting to specify the newly acquired con-
sciousness about this strange relationship betwe en the physicists and
the chemists. Worth Rodebush, one of the first to receive a doctorate
in 1917 from the newly established Department of Chemistry at Berkeley
under the chairmanship of Lewis , went a step further than Clark. He
asserted that the divergent paths of physicists and chemi sts were being
drawn together after the advent of quantum theory and especially after
Bohr's original papers. But in this process the physicist seemed to have
yielded more ground than the chemist and the former appeared to have
learned more from the latter than was the case with the chemists.
Rodebush gracefully remarked that it was to the credit of the physicist
that he could now calculate the energy of formation of the hydrogen
molecule by using the Schrodinger equation. But the outstanding tasks
for a theory of valence was to predict the existence and absence of various
compounds, and the nature of valence which can be expressed by a series
160 KOSTAS GAVROGLU
of small whole numbers leading to the law of multiple proportions. The

"brilliant theories" of Lewis accounted for the features of valence "in
a remarkably satisfactory manner, at least from the chemist's point of
view"." London 's group theoretical treatment of valence was consid-
ered as an important piece of work even though it did not provide answers
to all the queries of the chemist as, for example, the differences in degree
of stability between chemical compounds.
Perhaps the most cogent manifestation of what would become the
characteristic approach of the American chemists was Harry Fry's con-
tribution in this Symposium. He attempted to articulate what he called
the pragmatic outlook. He started by posing a single question. What
would be the kind of modifications to the structural formulas so as to
conform with the current concepts of electronic valency? Such a question ,
he suggested, should by no means lead to a confusion of the fundamental
purpose of a structural formula which is to present the number, the kind
and the arrangement of atoms in a molecule as well as to correlate the
manifold chemical reactions displayed by the molecule.
The opinion is now growing that the structural formula of the organic chemist is not
the canvas on which the cubist artist should impose his drawings which he alone can inter-
pret , . , On the grounds that pract ical results are the sole test of truth , such simple
system of electronic valence notat ion may be termed ' pragmatic' , 12
'Chemical pragm atism' resisted the attempts to embody in the struc-

tural formulas what Fry considered to be metaphysical hypotheses:
Questions related to the constitution of the atom and the disposition of
its valence electrons. It was the actual chemical behaviour of molecules
that was the primary concern of the pragmatic chemist, rather than the
imposition of an electronic system of notation on these formulas which
was further complicated by the metaphysical speculations involving the
unsolved problems about the constitution of the atom. Fry had to admit
the obvious fact that as the chemists will know more about the consti -
tution of the atom, they would be able to explain more fully the chemical
properties. He warned, though, that premises lying outside the territory
of experiences of sensations are bound to lead to contradictory conclu-
sions quoting Kant and becoming, surely, the only chemist to use Kant 's
ideas in order to convince other chemists at a Conference about an issue
in chemistry!
CONCEPTIONS RELATED TO THEORY BUILDING IN CHEMISTRY
In April 1926, Linus Pauling, supported by a Guggenheim Fellowship,

arrived in Munich where he planned to work at the Institute of Theoretical
Physics. He was twenty five years old and had received his doctorate
from the California Institute of Technology working with Roscoe
Dickinson on the structure of molybdenite. When he met Sommerfeld,
the latter suggested that Pauling work on the electron spin . Pauling did
not follow this advice, since his main interest was in chemical problems,
especially of atoms with many electrons. He had a solid background in
the old quantum mechanics and had exceptionally good training in the-
oretical physics and mathematics for a person whose main interests
were in chemistry. His first paper based on quantum mechanics had
appeared in 1926 and it contained a calculation of the values of the
Sommerfeld fine -structure screening constants.
While in Zurich he spent most of his time attempting to treat the inter-
action of two helium atoms, trying to get good approximate values to
some integrals that come in the Heitler-London analysis of simple mol-
ecular systems. Despite discussions with both Heitler and London, they
did not invite him to participate in their work. And, as he said in his inter-
view , the group theoretical methods were not particularly appealing to
him .
Almost everything in the series of Pauling's papers starting in 1931
and titled The Nature of the Chemical Bond are included in his book
of the same title . There are, however, some details of significance. In
the opening paragraph of the first paper in the series, Pauling stated
his assessment of the situation concerning work on the chemical bond
as well as the method he would follow.
During the last four years the problem of the nature of the chemical bond has been attacked
by theoretical physicists, especially Heitler and London, by the appli cat ion of quantum
mechanics. Thi s work has led to an approximate theoretical calculation of the energy of
formation and of other properties of simple molecu les .. . and has also provided a formal
just ification of the rules set up in 1916 by G. N. Lewis for his electron bond . In [this]
paper it will be shown that many more results of chemical significance can be obtained
from the quantum mechanical equations, permitting the formulation of an extensive and
powerful set of rules for the ele ctron-pair bond supplementing those of Lewis."
Texts of this sort are , in a way, pace setting texts; they are rhetor-
ical texts contributing to the formation of the chemists' culture, to the
162 KOSTAS GAVROGLU
way chemists view others and themselves. It was the theoretical physi-
cists who applied quantum mechanics to a chemical problem, but at the
same time Pauling considered his own work as an extension of their
program. He declared his applications will provide "many more" results
which can be obtained in the form of rules supplementing other rules -
of Lewis , in fact, who had formulated them much earlier than the advent
of wave mechanics! But since Lewis' cardinal rule - that of electron
pairing - had been given formal justification, one can formulate new rules
supplementing Lewis's rules! Interestingly, concerning the question of
the relationship of the various alternatives to the Lewis schema, Heitler
and London thought that their work replaced that of Lewis, whereas
Pauling and Mulliken considered theirs as supplementing it.
The rules which are formulated later in the paper were provided with
a kind of quantum mechanical justification and they were by no means
rules derived from first principles. Pauling's papers were mathemati-
cally sophisticated and from the calculations he had published it was
evident that he was at home with the details of quantum theory. And,
furthermore, to be able to present, in his book, such a coherent and
convincing argumentation about the nature of the chemical bond with
little mathematics is not a tribute to Pauling's ability to polarize his work,
but is indicative of his method. In this manner Pauling was able to inau-
gurate the language of quantum chemistry which could be used by
chemists in a practical manner.
To account for chemical bonding, Pauling's schema made use of the
attraction expressed in the quantum mechanical resonance between two
'more basic' structures. But the ontological status of these 'more basic'
structures was rather problematic. In 1944 George Willard Wheland, who
was a student of Pauling's and one of the strongest propagandists
of the theory of resonance , published his The Theory of Resonance
and its Application to Organic Chemistry. Appropriately, the book was
dedicated to Pauling. Wheland's view was that
resonance is a man-made concept in a more fundamental sense than most other physical
theories. It does not correspond to any intrinsic property if the molecule itself, but instead
it is only a mathematical device, deliberately invented by the physicist or chemist for
his own convenience."
At the time, Pauling did not seem to disagree with such an assessment.
But when a later edition of Wheland 's book appeared in 1955, a lively
correspondence ensued between the two about the actual character of
resonance theory. Wheland thought that resonance was not an intrinsic

property of a molecule, but instead
something deliberately added by the chemist or physicist who is talking about the molecule.
In anthropomorphic terms, I might say that the molecule does not know about the reso-
nance in the same sense in which it known s about its weight , energy, shape and other
properties that have what I would call real physical significance."
Pauling disagreed. "I feel that in your book you have done an injustice
to resonance theory by overemphasizing its man-made character.?" Their
correspondence continued and neither appeared to be convinced by the
other. What Pauling greatly emphasized was not the 'arbitrariness of the
concept of resonance, but its immense usefulness and convenience which
"make the disadvantage of the element of arbitrariness of little signifi-
cance"." This according to Pauling became the constitutive criterion
for theory building in chemistry. It was the way, as he had noted, to
particularize Bridgman's operationalism in chemistry . In fact, Pauling felt
more at ease with the Schrodinger approach than with matrix mechanics
and did not worry about questions of interpretation of quantum
mechanics. "I tend not to be interested in the more abstruse aspects of
quantum mechanics. I take a sort of Bridgmanian attitude toward them"."
In his analysis of resonance, Pauling expressed in the most explicit
manner his views about theory building in chemistry. He asserted that
the theory of resonance was a chemical theory, and, in this respect, it had
very little in common with the valence-bond method of making approx-
imate quantum mechanical calculations of molecular wave functions
and properties. Such a theory was "obtained largely by induction from
the results of chemical experiments"." The development of the theory
of molecular structure and the nature of the chemical bond, Pauling
asserted in his Nobel speech in 1954, "is in considerable part empirical
- based upon the facts of chemistry - but with the interpretation of
these facts greatly influenced by quantum mechanical principles and
concepts"."
Both the discussions with Wheland and a vicious attack against his
theory by chemists in the Soviet Union" prompted Pauling to include
a discussion of the character of theory in chemistry in the third edition
of his book in 1960. The theory of resonance was not simply a theory
embodying exact quantum mechanical calculations. Its great extension
has been "almost entirely empirical, with only the valuable and effec-
tive guidance of fundamental quantum mechanical principles". Pauling
164 KOSTAS GAVROGLU
emphasized that the theory of resonance in chemistry was an essen-

tially qualitative theory, which, "like the classical structure theory,
depends for its successful application largely upon a chemical feeling
that is developed through practice"." Pauling himself has repeatedly
stressed the rather empirical character of his theory of resonance.
My work on the nature of the chemical bond and its application to the stru cture of mol-
ecules and cryst als has been largely empirical, but for the most part guided by quantum
mechanical principles. I might even contend that there are four ways of discussing the
nature of the chem ical bond : the Hund -Mulliken way, the Heitler-London way, the Slater-
Pauling way , and the Pauling semi -empirical way.23
Though the method of molecular orbitals was first introduced by Hund,

it was Mulliken who provided both the most thorough treatment of the
different kinds of molecules, and the theoretical and methodological
justifications for legitimizing the molecular-orbital approach. Mulliken
was born in 1897 and received his doctorate from the University of
Chicago in 1921 working with D. W. Harkins on isotope separation, espe-
cially of mercury. He had worked at the University of Chicago and at
Harvard as a National Research Fellow and by 1926 he was an assis-
tant professor at New York University. When he did his foundational
work on the method of molecular orbitals he had become associate pro-
fessor at the University of Chicago and had spent some months travelling
in Europe, before his extended stay there in 1930 as a Guggenheim
Fellow.24
In a remarkable paper titled 'On the method of molecular orbitals '
published in 1935, Mulliken expressed his views on what he consid-
ered to be the most characteristic and differentiating aspects of his theory.
The Heitler-London method
follows the ideology of chemistry and treats every molecule, so far as pos sible, as
composed of definite atoms . . . It has had the notable succe ss as a qualitative concep-
tual scheme for interpreting and explaining empirical rules of valence and in semi-
quantitative, mostly semiempirical calculations of energies of formation."
The method of molecular orbitals departs from "chemical ideology ...

and treats each molecule, so far as possible, as a unit". This seemingly
terminological - or shall we say procedural? - difference highlights the
more theoretical issues involved in the study of molecular physics .
It is the writer 's belief that, of the various possible methods, the present one may be
the best adapted to the construction of an exploratory conceptual scheme within whose
framework may be filled both chemical data and data on electron levels form electron
spectra. A procedure adapted to a broad survey and interpretation of observed relation

is here aimed at, rather that (at first) one for quantitative calculation, which logically would
follow later. 26
Mulliken had realized that one of the reasons for the poor quantita-
tive agreement using the molecular orbital approach was because of the
inability of this theory to include the details of the interactions between
the electrons. But even though their quantitative inclusion would make
a theoretical calculation from first principles an impossible job, "their
qualitative inclusion has always formed a vital part of the method of
molecular orbitals used as a conceptual scheme for the interpretation
of empirical data on electronic states of molecules"." Such considera-
tions, in fact, led to the qualitative explanation of the paramagnetism
of oxygen - one of the main weaknesses of the valence bond approach.
The approach where theoretical developments had their foundations
in the experimental results and where they were more or less general-
izations of these results, was the standpoint from which Mulliken built
up his own approach to the question of chemical bonding. For him
"quantum mechanics, following experiment, demands the existence of
stationary states of energy, for molecules as well as for atoms" (my
emphasis)." Coming to his critique of Heitler and London he remarked
that their "very valuable method .. . should be carefully distinguished
from [their] valence theory" (emphasis in the original)." The complicated
calculations were indeed a reason for not preferring the Heitler-London
method, but more important is the fact that this methods failed to provide
a detailed insight into the nature of the changes which took place in
the electron orbits when atoms came together. It was quite surprising that
a large part of the Mulliken's critique was based on a wrong interpre-
tation of the "exchange energy " which he thought was "connected with
the exchange or jumping back and forth of the electron between the
two nuclei "." Both Heitler and London in their subsequent papers
strongly emphasized the impossibility to give a visualizable interpreta-
tion of the exchange energy.
London and Heitler's theory is enticingly simple but, in opinion of the author, really
does not hit the nail on the head ... The presence of unpaired electrons and their spins,
and their pairing in molecules, undoubtedly act usually as convenient indicators of valence
and of the formation of valence bonds respectively, but even then in the author's opinion,
they conceal something which is more fundamental . . . what is fundamental is that we
have in the Is orbit in hydrogen an orbit whose character permits it to become consid-
erably more firmly bound when the hydrogen atom unites with another hydrogen atom."
166 KOSTAS GAVROGLU
Summarizing his overall approach, Mulliken placed himself in the

antipodes of Heitler and London
The concept of a discrete, whole number property of atoms called valence, is less fun-
damental from the point of view of quantum theory than a continuous conception of
chemical binding . .. The fact that valence electrons almost always occur in pairs in
saturated molecules appears to have after all no fundamental connection with the exis-
tence of chemical binding . It can be adequately explained on the basis of the fact that,
because of the Pauli principle and the properties of the electron sp in, each type of mol-
ecular orbit can be occupied by just two electrons."
Many a time such discussions have been portrayed as discussions

among the more sophisticated protagonists which did not really touch the
rank and file. I do not think so. The settling of these issues deeply affected
laboratory practices and research agendas and were never snubbed by the
chemists as a whole. The issues were suggestive of the various ways
available to map the undefined and undelineated middle ground that some
called physical chemistry and some molecular physics. In a manner anal-
ogous to the situation in spectroscopy, or even thermodynamics, the
question could be asked as to whose domain quantum chemistry was.
Was it an activity for physicists or chemists? Or was it a subject for
physical chemistry or atomic physics? How would the boundaries be
drawn, what would be the methodological priorities, the ontological com-
mitments and, above all, what would be the character and extent of the
practitioners' allegiances to physics and chemistry? Most importantly,
the question as to the character of theory in quantum chemistry domi-
nated the minds of many chemists. The problem of the chemical bond
contributed more than any other problem, so that quantum chemistry
could articulate its own autonomous language with respect to both physics
and chemistry, chart its own theoretical agenda and formulate its own
theoretical framework. The beginnings and the establishment of quantum
chemistry involved a series of issues which transcend the question of
the application of quantum mechanics to chemical problems. The out-
standing issue to be settled in the community turned out to be the
character of theory for chemistry and this is something which could
provide a useful insight to the problems raised in the discussions con-
cerning realism . I have tried to present a reading of the developments
in quantum chemistry so that at least some of the problems relevant to
the discussions of realism could be made a little clearer. And , perhaps,
working in the history of science, I have the luxury of posing the
problems by avoiding their philosophical discussion!
I wish to thank Bob Cohen, Theodor Benfey and Ana Simoes for
their valuable comments. Part of this work has been done during my
stay as the Edelstein International Fellow for 1992-1993 at the Beckman
Center for the History of Chemistry. I wish to thank Arnold Thackray for
his hospitality and for the many enlightening conversations.
Department of Physics ,
National Technical University,
Athens Greece.
NOTES
I Sidgwick (1927), iii.

2 Coulson (1970), p. 259 and p. 287.
3 Dirac (\ 929), p. 714.
4 Thomson (\ 924), p. 450.
5 Mulliken (1965), p. S7.
6 Archive for the Histor y of Quantum Physics (A.H.Q.P.) Interview with Walter
Heitler.
7 One of the main drawb acks of such an approach is the non-existence of the contribu-
tion of the ionic terms . Though they do not lead to any appreciable differences in the
case of the hydrogen molecule, their a priori absence in similar kinds of calculations
had been systematically examined by Van Vleck and Sherman (\935).
• A.H.Q .P. Interview of Walter Heitler.
9 Van Vleck (\970), p. 240.
10 Clark (\ 928), p. 362.
II Rodebush (1928), p. 513.
12 Fry (\928), pp. 558-559.
13 Pauling (\93Ic), p. 1367.
14 Wheland (\ 944), p. 31.
15 G. W. Wheland to L. Paul ing January 20, 1956.
16 Pauling to Wheland January 26, 1956. Wheland papers at the University of Chicago
Library .
J7 Pauling (\ 960), p. 95.
IX A.H.Q.P. Interview of Linus Pauling , p. 9.
19 Pauling (\939), p. 219.
20 Pauling (1955) , p. 92.
21 Kursanov et al. (\ 952); Tatevskii , Shakhparanov (1952); Moyer Hunsberger (\ 954).
22 Pauling (\ 960), 220.
23 Pauling, Private communication.
24 Mullikan (\ 989).
25 Mullikan (\ 935) , p. 376.
26 lbid., p. 375.
27 lb id., p. 378.
168 KOSTAS GAVROGLU
2X Mulliken (1931), p. 350 .

29 lbid., p. 353.
30 Ibid., p. 355.
31 lb id., pp. 359-360.
32 Ibid., pp. 386-387.
BIBLIOGRAPHY
Clark , G. L. (1928) . 'Introductory Remarks in the Symposium on Atomic Structure and

Valence ', Chemical Reviews 5, pp. 361-364.
Coulson, C. A. (1970). 'Recent Developments in Valence Theory - Symposium: Fifty
Years of Valence ', Pure and Applied Chemistry 24, pp. 257-287.
Dirac , P. A. M. (1929) . ' Quantum Mechanics of Many-electron Systems' , Proceedings
of the Royal Society A123, pp. 714-733.
Fry, H. S. (1928) . ' A Pragmat ic System of Notation for Electronic Valence Conceptions
in Chemical Formulas ' , Chemical Reviews 5, pp. 557-568.
Gavroglu , K. (1995). Fritz London (/900-1954) A Scientific Biograph y, Cambridge
University Press.
Gavroglu, K. and Simoes , A. (1994) . 'The Americans, the Germans and the Beginnings
of Quantum Chemistry', Historical Studies in the Physical Sciences 25, 47-110.
Heitler, W. (1955). ' The Theory of Chemical Bond ', Ark iv fur Fysik 10, pp. 145-
156.
Heitler, W. (1967). ' Quantum Chemistry: The Early Period' , Inter. Jour . of Quantum
Chemistry 1, pp. 13-36.
Hund, F. (1929). ' Chemical Binding ' , Transaction s of the Faraday Society 25 , pp.
645-647.
Kohler , R. E. (1971) . 'The Origin of the G. N. Lewis Theory of the Shared Pair Bond ',
Historical Studies of the Physical Sciences 3, pp. 343-376.
Kronig , R. (1935) . The Optical Basis of Valency, Cambridge University Press .
Kursanov, D. N., Gonikberg, M. G., Dubinin , B., Kabachnik, M. I., Kaveraneva, E. D.,
Prilezhaev a, E. N., Sokolov N. D. and Freidlina, R. Kh. (1952) . ' The Present State
of the Chemical Structural Theory ' . Translation by I. S. Bengelsdorf in Journal of
Chemical Education 78, pp. 2-13.
Lennard-Jones, J. E. (1929) . ' The Electronic Structure of Some Diatomic Molecules ',
Transactions of the Faraday Society 25, pp. 665-686.
Lewis , G. N. (1923). Valence and the Structure of Atoms and Molecules, The Chemical
Catalog Company.
McCrea , R. (1985) . ' How Quantum Physics Came to England' , New Scientist 17 October,
pp.58-60.
Moyer Hunsberger, I. (1954) . 'Theoretical Chemistry in Russia' , Journal of Chemical
Education 80, pp. 504-514.
Mulliken, R. S. (1928) . 'The Assignment of Quantum Numbers for Electrons in Mole-
cule s. Part 1', Physical Review 32, pp. 186-222.
Mulliken, R. S. (1928). 'The Assignment of Quantum Numbers for Electrons in Mole-
cules . II', Physical Review 32, pp. 761-772.
Mulliken, R. S. (1929). ' The Assignment of Quantum Numbers for Electron s in Mole-
cules. III', Physical Review 33, pp. 731-747.
Mulliken, R. S. (1931). 'Bonding Power of Electrons and Theory of Valence' , Chemical
Reviews 9, pp. 347-388.
Mulliken, R. S. (1932). 'The Interpretation of Band Spectra ' , Reviews of Modern Physics
4, pp. 1-86 .
Mulliken, R. S. (1932). 'E lectronic Structures of Polyatomic Molecules and Valence . II.
General Considerations', Physical Review 41. pp. 49-71.
Mulliken , R. S. (1933) . ' Electronic Structures of Polyatomic Molecules and Valence .
V' , Journal of Chemical Physics 1, pp. 492-503.
Mulliken , R. S. (1935) . ' Electronic Structures of Polyatomic Molecul es and Valence.
VI. On the Method of Molecular Orbitals', Journal of Chemical Physics 3. pp. 375-
378.
Mulliken , R. S. (1965) . 'Molecular Scientists and Molecular Science: Some Remini s-
cences' , Journal of Chemical Physics 43, S7.
Mulliken, R. S. (1967). ' Spectroscopy Molecular Orbitals and Chemical Bonding', Science
157,7 July, 17.
Mulliken, R. S. (1989). Life ofa Scientist, edited by B. J. Ransil, Springer-Verlag.
Pauling, L. (1928) . ' The Application of the Quantum Mechanics to the Structure of the
Hydrogen Molecule' , Chemical Reviews 5, pp. 173-213 .
Pauling, L. (1928). 'The Shared-electron Chemical Bond' , Proceedings of the National
Academy of Sciences 14, p. 359.
Pauling, L. (1931 a). 'The Application of the Theory of Homopolar Valency to Poly-
atomic Molecules ', B.A.A.S. meeting of 1931, pp. 249-254.
Pauling, L. (1931 b). ' The Nature of the Chemical Bond I' , Journal of the American
Chem ical Society 53 , p. 1367; ' . .. II ', Ibid., pp. 3225-3237; ' ... III ', Ibid . 54
(1932), pp. 988-1003.
Pauling, L. (l93Ic). 'The Nature of the Chemical Bond. Application Obtained from the
Quantum Mechanics and from a Theory of Paramagnetic Suceptibility to the Struc-
ture of Molecules ', Journal of the American Chemical Society 53, pp. 1367-1400.
Pauling, L. (1939). The Nature of the Chemical Bond, Cornell University Press.
Pauling, L. (1940). ' A Theory of the Structure and Process of Formation of Antibod ies',
Science 62, pp. 2643-2660.
Pauling, L. (1955). ' Modem Structural Chemistry' in Les Prix Nobel, Stockholm .
Pauling, L. (1960). The Nature of the Chemical Bond, Third Edition, Cornell University
Press.
Rodebush, W. H. (1928). 'The Electron Theory of Valence ' , Chemical Review s 5, pp.
509-531 .
Schweber, S. S. (1990) . ' The Young John Slater and the Development of Quantum
Chemistry', Historical Studies in the Physical and Biological Sciences 20, pp. 339-
406.
Sidgwick, N. V. (1927). The Electronic Theory of Valency, Clarendon Press.
Simoes, A. (1993). ' Converging Trajectories Diverging Traditions: The Chemical Bond
Valence Quantum Mechanics and Chemistry 1927-1937', Doctoral Thesis University
of Maryland.
Sopka , K. (1988). Quantum Physics in America , American Institute of Physics.
170 KOSTAS GAVROGLU
Tatevskii, V. M. and Shakhparanov, M. I. (1952) . ' About a Machistic Theory in Chem-

istry and its Propagandists ', Journal of Chemical Education 78, pp. 13-14.
Thomson, J. J. (1924) . 'Introduction to the Session on the Electronic Theory of Valency',
Transactions of Faraday Society 19, p. 450.
Van Vleck, J. (1928). ' The New Quantum Mechanics', Chemical Reviews 5, pp. 467-
507.
Van Vleck, J. (1970) . ' Spin the Great Indicator of Valence Behaviour', Pure and Applied
Chemistry 24, pp. 235-255.
Van Vleck, J. and Sherman A. (1935) . 'The Quantum Theory of Valence', Reviews of
Modern Physics 7, pp. 167-228.
Wheland, G. (1944) . The Theory ofResonance and its Applications to Organic Chemistry,
John Wiley.
MATTHIAS KAISER
EMPIRICAL VERSUS THEORETICAL PROGRESS

IN SCIENCE
I. INTRODUCTION
Progress is a value -term . It denotes some form of improvement between

at least two parameters. Thus it is also a two-place relation. The progress
of knowledge that we believe the sciences achieve over time involves
the relation between two intersubjectively shared belief-descriptions.
Philosophers traditionally assumed that the only beliefs worthy of being
compared in terms of a 'progress relation ' were those characteristic of
the presumably highest intellectual achievement: theories that explain
the underlying mechanisms of the world . Intellectual achievement and
progress were thus , albeit implicitly, den ied with regard to the empir-
ical basis of science.
I disagree with this assumption. The central thesi s of this paper is
the contention that scientific progress can only be adequately under-
stood if we distinguish systematically between theoretical progress
on the one hand and empirical progress on the other. It is main-
tained that empirical progress is no less of an intellectual achievement
than the progress of high level theories . Yet, its characteristics are very
different from what philosophers say about the progress of theories. It
is also different from what the author thinks should be said about the
latter.
The position that is sketched in the remainder of this paper has cer-
tainly a realist flavour to it. It is also a position that seems to fit in
well with the recent interest that many philosophers of science take in
the making of science. The laboratory, the experiment, the instrument,
the data in science are now focussed upon by several philosophers. A
fundamental view presupposed in this paper is the replacing of the
dichotomy between theory and data by a trichotomy between theory,
phenomena and data. This position is most clearly articulated in the
work of Jim Woodward and Jim Bogen (Bogen and Woodward, 1988;
Woodward, 1989; but also Kaiser, 1991a). On this view, a scientific
phenomenon is a conceptual abstraction, sprung out of the available data,
and far removed from sensible experiences. However, this tripartite
171
172 MATTHIAS KAISER
view seems also to go well with the views of Nancy Cartwright and
several other philosophers.
The evidence for the thesis to be defended here is basically a case
study from earth science. Parts of this study have been described else -
where, and shall not be repeated in detail. A brief sketch of the relevant
features of this case study shall however follow. The full characteris-
tics are developed in a book which appeared in German under the title
Aspekte des wissenschaftlichen Fortschritts (Kaiser, 1993b).
II. CONTINENTAL DRIFT AND PLATE-TECTONICS

AS A CASE -STUDY
Many philosophers agree in the assumption that the breakthrough of

plate-tectonics in earth science during the late 1960's constitutes an
example of progress in science. Interpretations of the history of this devel-
opment vary with the preferred view of scientific change. Where some
see a Kuhnian development (e .g., Hallam, 1973), and some a realist
unfolding of new facts (e .g., Ruse, 1981), others see a Laudan-type
rationalist story (e.g., Le Grand, 1988), or, yet others, a vari ant of
cognitivistically interpreted naturalism (e.g ., Giere, 1988). My reading
of the history of this case is realist since I hold that the major devel-
opments that account for this theory change are a number of new
phenomena that were newly discovered. However, I shall not defend
this view here in any detail. Rather I shall outline some of the devel-
opments that are important with regard to the topic of the paper.
The early history of continental drift theory (as started by Alfred
Wegener) was based on evidence from field work and interpretations
of these findings . It was beset with many errors and difficulties. As
Oreskes points out in an important paper (Oreskes, 1988) , the quality
of data pointing in the direction of drift changes dramatically from the
early 1950's on. There is the new evidence from paleomagnetism (which
I described in detail in Kaiser, 1991), and there is the new evidence
from the sea-floor. Important developments took place in both areas in
the 1950's and 1960's, developments that first re -opened the discus-
sion about drift after approximately 30 years of virtual silence, and that
later formed the basis for the acceptance of plate -tectonics. I shall now
summarize these developments with a view to the thesis of this paper.
Developments that contributed to the evolution of the theory of plate-
tectonics will thus only playa marginal role here .
EMPIRICAL VERSUS THEORETICAL PROGRESS 173
1. New Instruments and New Methods
In the newly emerging sub-field of earth science called paleomagnetism

a new instrument makes possible the measurement of a very weak mag-
netization of rocks: remanent magnetization. This magnetization is
typically acquired from the earth's magnetic field through cooling from
high temperatures (as, e.g., in lava flows) and kept constant in magnetic
declination and inclination through large changes of the surrounding
magnetic field, and even maintained in processes of sedimentation. The
significance of this lies in the fact that remanent magnetization can be
regarded as a truthful record of the earth's magnetic field at the forming
of strata or sediments. Thus, readings of remanent magnetization provided
the earth scientist with hitherto unparalleled data from geological history.
The importance of this new class of data was early recognized by a small
group of leading scientists, many of them from England. It was made
possible by the development of a highly sensitive instrument, the astatic
magnetometer (cf. figure 1), which the Nobel laureate P. M. S. Blackett
originally built for quite different purposes. The instrument promised
precise and quantitative access to many geological phenomena which
until then could only be inferred by very indirect qualitative data from
field observations, as, e.g. , fossil evidence in paleoclimatology. The
researchers could thus utilize the ever so slight magnetic traces of
minerals contained in rocks as direct evidence of past conditions in
earth history. It is part of the story that this instrument was later replaced
by others which were even more precise.
But remanent magnetization is not always easy to detect. Many rock
samples have acquired an additional magnetization from their sur-
roundings over long time spans, the so-called viscous magnetization. It
is less permanent than remanent magnetization , growing logarithmi-
cally with time , superimposed on the prior remanent magnetization, and
deflecting the measurements. The Icelandic researcher Ari Brynjolfsson
developed a method of magnetic cleaning by alternating magnetic fields
that managed to target the measurements to true remanent magnetization.
Other methods, instruments, and measurement techniques were devel-
oped during this time which in a similar manner gave scientists access
to totally new data -set s. This is particularly true for studies of the sea-
floor. Backed by large research grants ,' heavy machinery could now be
employed in order to chart the seafloor and conduct seismological, topo-
graphical, gravimetric, and magnetic measurements. A huge amount of
174 MATTHIAS KAISER
Fig. I. Astatic magnetometer, showing Helmholtz coils, for removing the hor izontal
and vertical components of the geomagnetic field, the magnetometer, and specimen holder
(reproduced from : S. K. Runcorn , 'The permanent magnetization of rocks' Endeavour
J4 (1955), 152-159).
new data was produced that to start with had no obvious link to theo-
retical disputes in earth-science.
One last development deserves mention in this context. Data from
the earth's past would not be very helpful if we could not associate the
appropriate time period with them . This has always been a problem in
earth science . For a long time one was aware of several methods to deter-
mine the relative age of strata, or other qualitative determinations. But
in the late 1950's a new method emerged that allowed relatively precise
quantitative measurement of geological age: the so-called 'radioactive
clocks' . The key principle of this method is the longevity of some radioac-
tive substances. Slight traces of such substances are contained in all rocks.
Given that we know the rate of decay of these substances, we can deter-
mine the ratio between parent element and daughter elements, and thus
calculate backwards to the beginning of the decay which coincides with
the time of the forming of the rock. The potassium 40 - argon 40 method
(as well as several similar methods) proved to be highly important for
the understanding of the geological history of our planet.
2. New Phenomena Emerged
A small group of scientists, among them Blackett and Runcorn, started

in the early 1950 's to collect evidence of remanent magnetization in
different parts of the world . At this time the theory of continental drift
was considered but a mere aberration of the past, certainly not a current
alternative for theory choice, studies of remanent magnetization in rocks
(and sedimentations) set out unrelated to the hypothesis for which it
was to become deci sive later. However, the data revealed evidence of
a clearly anomalous kind in the light of accepted theory at the time.
First of all, the data indicated magnetic reversals, meaning that the
polarity of the earth's magnetic field had changed several times from
North to South and back. Secondly, if the readings were trustworthy,
then the data indicated records of dramatic changes either in the
earth 's magnetic field (e.g., pole wandering), or in the relative location
of continents and islands with reg ard to the magnetic north pole , or
even in the nature of magnetism itself (bipolarity) (cf. figure 2). Due
to the complexities of measurement involved and the anom alie s with
regard to textbook knowledge , many of the earth scientists who were
aware of this new research (and those were but a few), typically without
knowing much about its theoretical underpinnings, felt legitimized in
doubting the trustworthiness of the data. Yet, the few who were actively
engaged in this research soon found the means to support new phenomena
that were at variance with established dogma in earth science. It took
some time before these findings were incorporated into the discipline.
Historians of science seem to agree on the two main reasons for this.
First, paleomagnetism was still an esoteric field of research in earth
science, unknown to most practitioners and ignored by the leaders of
the profession. Second, the absence of a good theory that could provide
a coherent account of the causal mechanisms explaining these phenomena
made them difficult to swallow for other fields in earth science, as ,
e.g ., geophysics.
176 MATTHIAS KAIS ER
0
0 0
0
0
0
0
0
0 0
+..e.,..:'i....
0 0
0 0
0
••
CX) •••
...( ...
..
...
• • •
•
Fig. 2. Directions of magnetization of Torridoni an sandstones. Open circles - upward

directions . Closed circles - downward directions (reproduced from: S. K. Runcorn, ' The
permanent magnetization of rocks ' Endeavour 14 (1955), 152-159).
But the discovery of new phenomena was not restricted to paleo-

magnetism. Basically it was the result of the mutual support that different
fields of study lent to each other.
What seemed a mere possibility in the light of one line of study
became a near certainty in the light of the combination of several
very different studies. And phenomena that, when seen in isolation,
seemed incoherent fell into a coherent pattern when combined with
other new phenomena. It is absolutely essential to my (realist) under-
standing of the theory change that occurred in earth science during the
late 1960's that the changes in the empirical basis of earth science are
duly recognized.
The most important new phenomena were the following:
(i) drift: the relative position of large land masses on the earth's crust
has changed during geological history ;
(ii) magnetic reversals: the magnetic field of the earth has changed
its polarity from north to south and back several times over the
geological past;
(iii) pole-wandering: the location of the magnetic (north) pole had
moved along a certain path during geological history;
(iv) transform-faults: in addition to the known 'ridges'and 'trenches'
(subduction zones) of the sea floor there exists a third class of
faults in which the displacement suddenly stops or changes form and
direction, Several types of transform-faults could be shown to exist
(cf. figure 3);
(v) succession and duration of magnetic reversals: one could estab-
lish a timetable where 'events'of magnetic reversals were placed
in larger periods (epochs) of geological history.
This is the list of empirical phenomena that together provided the basis
for the acceptance of continental drift theory. It should , however, be men-
tioned that the theory which was supported by these phenomena also
changed considerably in relation to the very first versions of it. The
new theory of plate-tectonics had little in common with Alfred Wegener 's
.-- --+
-
OIl
-
.. • -
•
.-- ---.
A B
Fig. 3. Diagrammatic illustration of: A: The directions of movement along a trans-

current fault respon sible for the off-setting of the black band ; B: The direct ions of
movement along a transform fault connecting off-set portions of a mid-oceanic ridge
(black) (reproduced from: A. Holmes , Principles of Physical Geology, 3rd edition, Van
Nostrand Reinhold (UK) Co. Ltd., Workingham, 1978).
178 MATTHIAS KAISER
theory of continental drift other than postulating some kind of drift. It

was also different from Arthur Holmes 's early and merely hypothetical
speculations about drift. Basically the theory was shaped as a merger
between Harry Hess 's ' geo-poetry ' of 1960 (in which a causal process
of displacement was described), F. Vine and D. Matthew 's theory of
sea floor magnetization, and Tuzo Wilson's integration of transform-faults
into the class of contact-zones between tectonic plates.
3. How Phenomena were Established

The claim that most if not all of these new phenomena were established
in relative independence from explanatory causal theory, like plate-
tectonics, is essential to my thesis . In addition, none of the phenomena
mentioned were in any sense the result of some direct observation. They
were constructed out of existing data by utilizing a highly developed
theory, though not the very theory for which they constitute the empir-
ical bas is. The phenomena are abstractions in the sense that they
constitute the facts of the world to which the data point. Where data
are incomplete, local, and beset with complexities of their own (like, e.g.,
the super-imposition of several types of magnetizations in a given rock
sample), phenomena arc relatively stable, simple, and universal. Though
why-questions may arise at different points in research, it is in relation
to phenomena such as those mentioned that why-questions are posed
that call for causal explanation by means of high-level theory. It is thus
important that phenomena are established by means of reliable data,
preferably by means of very heterogeneous data-sets, and utilizing
reliable theory.
In this paper I have no intention of showing how the phenomena
mentioned above actually were derived from different sets of data. This
would by far exceed the given space limitations. However, I shall try
to give a general outline of some typical developments which may shed
some light on the central thesis defended here.
The phenomenon of drift was in some sense first indicated by pale-
omagnetism. Rock samples from different continents showed remanent
magnetizations that were clearly at variance with the present magnetic
conditions on earth. They also showed some relative uniformity when
restricted to samples of the same age and same continent. There also
emerged a dynamic uniformity when samples of different ages from
the same continent were compared. This suggested the hypothesis that
the continents may have undergone a change of relative location.

Yet, other explanations were still reasonable, for instance the assump-
tion of pole-wandering. Blakett and Runcorn - the latter was at that
time still opposed to drift - realized that if data from Europe were
compared to data from America one would be able to settle this question.
When this finally was done, the data revealed a surprise to everyone:
the data -set yielded a sensible uniformity only if one combined drift
with pole wandering (cf. figure 4). In some sense, this was a first test
of a prediction made on the basis of the drift hypothesis. Note that drift
W 1I(f t
,~ ! S- 1.0'
W a"
Fig. 4. Stereographic projection of the northern hemisphere from the N Pole to Latitude
16°, showing palaeo-magnetic polar-wandering curves for Europe (solid black circles) with
an offshoot for Siberia (open circles with a central dot); North America (open squares);
Africa (crosses); India (open triangles); and Australia (open circles). Geological ages of
samples from which the data were obtained are indicated by letters: E, Eocene K,
Cretaceous; J, Jurassic; Tr, Triassic; P, Permian; and C, Carboniferous . It should be noticed
that to avoid congestion these curves are drawn through the mean positions of clusters
covering considerable areas . To represent the data more accurately the curves should be
broad bands instead of lines (reproduced from: A. Holmes, Principles ofPhysical Geology,
3rd edition, Van Nostrand Reinhold (UK) Co. Ltd., Workingham, 1978).
180 MATTHIAS KAISER
in this context does not denote some (causal) theory, like Wegener's
or Holmes's theoretical outline of continental drift. It only refers to
a supposed fact of nature, the relative displacement of large land
masses .
If such a phenomenon is to be credible, it is reasonable to expect
that data from other sub-disciplines from earth science will point in the
same direction. E. Irving was the first to suggest that the paleomag-
netic data should be linked to paleoclimatic data. If the continents
occupied widely different positions on the earth 's surface, then the dis-
tribution of climatic zone s should also be expected to have changed
considerably. This suggestion implied an important methodological
point. Whereas on the one hand the data from paleomagnetism were
comparatively uncertain due to the complexities of measurement and
interpretation, data from paleoclimatology were relatively straight-
forward. And whereas , on the other hand, the data from paleoclimatology
were relatively imprecise and rough, the data from paleomagnetism
were relatively precise and quantitatively accessible. Field data from
paleoclimatology did indeed indicate that drift could be a sensible hypoth-
esis (e.g., Permo-Carboniferous glaciations, the distribution of salt
deposits, etc.). This was in part true already at the times of Wegener,
but in the 1950 's it was even more pressing, particularly for young
researchers who tried to be open-minded about the interpretations of their
findings .
But not only paleoclimatology contributed to the phenomenon of
drift. Studies of the sea floor had an even more important impact. The
magnetic charts of the sea floor showed an impressive pattern of varying
magnetization (cf. figure 5). It showed stripes of magnetization that
seemed to follow along certain regular lines, until they suddenly were
disrupted by fracture lines. These fractures could displace the pattern
by regular several hundred miles. In 1963, in relation to the Carlsberg
Ridge, F. J. Vine and D. H. Matthews suggested interpreting these
stripes as result of magnetic reversals, in combination with sea floor
spreading.
The phenomenon of magnetic reversals was also first shown in pale-
omagnetism. Vine and Matthews's paper permitted the use of the sea floor
as a relatively continuous magnetic tape on which these reversals were
recorded. The basic feature of it is that stripes of identical magnetiza-
tion could be found along both sides of oceanic ridges in a strictly
symmetrical manner. One could study the (magnetic) history of the earth,
".
...
' 0' >0"
'"
x~ , .
'"
-:......"."...:..- ,
" "
...
1000. C".-_
r-"
''' --~ ('"~
, '
< ;~
.
"",."
( '"
.... - ', '. oz·
'.
... :,' . ..' . \~ .. ..,

,
,
1... ,
... _... ---.. .......'.

: '
,
...
...
oz·
,.. lO" ...

Fig. 5. Magnetic anomalies over the Reykjanes Ridge, south-west of Iceland, showing
' magnetic stripe s ' (positive anomalies black) and their bilateral symmetry (After J . R.
Heirtzler et a\., 1965 . Reproduced from Heirtzler, 969 , with perm ission of John Wiley
and Sons) (Reproduced from : A. Holmes, Principles of Physical Geolo gy, 3rd edition,
Van Nostrand Reinhold (UK) Co. Ltd., Workingham, 1978).
given that one had a good classification of fractures of the seafloor.

Tuzo Wilson's introduction of transform faults provided just that. At
this stage the sea floor studies mentioned were doubly relevant for the
phenomenon of drift. First, they could confirm paleomagnetists' assump-
tion of magnetic rever sals, and, second, if the continents were carried
182 MATTHIAS KAISER
on top of a spreading sea floor, as Hess's ' geo-poetry' suggested, then

the postulated convection currents in the mantle could provide drift
with a mechanism of movement.
The really compelling breakthrough in this development happened then
in connection with the establishment of a time-scale for magnetic rever -
sals . Cox, Doell, and Dalrymple published a first rough timetable of
magnetic reversals in 1963. They distinguished between major magnetic
'epochs' (Bruhnes, Matuyama, Gauss, Gilbert), and minor magnetic
'events ' (Olduvai, Mammoth) (cf. figure 6). This was made possible
by utilizing the radioactive clocks mentioned earlier. The time-table
was soon perfected by identifying other new magnetic events. One such
event proved particularly decisive , the Jaramillo event. Given this new
event, existing data from the research vessel Eltanin, specifically the
Eltanin 19 profile, could now give rise to an interpretation of a uniform
speed of sea floor spreading.
At the same time other, and quite different, studies fell into the same
pattern based upon the phenomena mentioned above. Opdyke had com-
pleted a study of sea floor sediment magnetization, based not on the
horizontal magnetization stripes, but on vertical probes . The same mag-
Brunhes Gauss GIlbert
:: :: :
II I.. , II' II I
: ~: :: : ;
~> ::»' » ..... JiiiiiI
Fig . 6. Block diagram illustrating how ' magnetic stripes ' on the ocean floors can be
explained by spreading of the oceanic lithosphere away from oceanic ridges on either
side, according to the Vine -Matthews hypothesis (Adapted from A. Cox et a\., 1967)
(reproduced from: A. Holmes, Principles of Physical Geology, 3rd edition, Van Nostrand
Reinhold (UK) Co. Ltd ., Workingham, 1978).
netization pattern could be employed here (cf. figure 7). Later Sykes pub-
lished her seismological studies which also supported the uniform pattern
of dyn amic change of the earth ' s crust that emerged from these other
lines of research.
The phenomena mentioned were thus firmly established in earth
science by the end of the 1960's. Together they formed the empirical
basis for plate-tectonics. Many of them could be derived not only from
very different sets of data, but also from very different lines of research .
All of these phenomena constitute a complex intellectual achievement,
and they are very much the result of advanced instruments and complex
theory . Yet each of these phenomena enjoyed a relative independence
with regard to the theory of plate-tecton ics itself. Even if we were to
revise our conception of how the continents are moved , their relative dis-
EPOCHS EVENTS
III
!:! '"
em
Z o
:l 0 100
'"
..
\[
<Xl
Jaram,Ho
::;
. ~~...
-c
;2 .::;
~~
~ 10. -1300
C
~
O I(l U \lal ~~
10. oj, 16. -1:00
]
J
16. ; 500
' ~, X
.
w III X -,
<.:l
j l
III
.. 3 :l
<.:l
Mam mo~h .
\ 100)
!I 11
~700
'-........ <,
: -{ooo
X
Clay 101 1 '~ 900
:.
~:
S lltv clay
~ ~ I -{HlOO
Silt
Dldlomaceous clay 1:1 I c:. 1:::1 1 -11100
. ... 1200
CdlCareOu$ ooze
Fig. 7. ' Magnetic stratigraphy ' es tablished within deep- sea cores from the Antarctic ,
and correlated with radiolarian faunal zones (greek letters) (After N. D. Opdyke, B. Class,
J. D. Hayes and J. Foster, 1966. Copyright the Americal Association f or the Advancement
of Science) (reprodu ced from: A. Holmes, Principles of Physical Geology, 3rd edition,
Van Nostrand Reinhold (UK) Co. Ltd., Workingham, 1978).
184 MATTHIAS KAI SER
location would still be an empirical phenomenon to take into account,

as well as the wandering of the poles or the magnetic reversals. The
discovery of each of these phenomena constitutes an instance of what
I call 'empirical progress' in science.
III. FROM DATA TO PHENOMENA
Elsewhere (Kaiser, 1991a, 1993b) I have described in some detail how

I conceive of the path from data to phenomena. I shall therefore restrict
myself here to the provision of a rough outline of the basic view, and
refer to my other works for further elaboration. I would also like to
mention that I conceive of Woodward and Bogen 's work as contribu-
tions to this topic along similar lines. Indeed , lowe my terminology to
them.
The fundamental claim is that phenomena are highly theoretical con-
structs based upon data. In my view data must always be intersubjectively
accessible. This means that in principle the raw data of a study could
be handed over to another researcher for inspection. In their most rudi-
mentary form, data are objects of our experience, they are things or
recorded singular events. They may be the rocks of the paleomagnetist,
the cell-cultures of the micro-biologist, the filled-in questionnaires of the
social scientist, or the satellite pictures of the meteorologist. Data, I
say, are experientally anchored. It is important to differentiate between
the experiental anchor of a data set and the descriptions of it. Descriptions
are interpretations in the light of theory and allow for error. Exper iental
anchoring allows at most for ostention, like 'this rock here'. Philoso-
phers have rightly pointed out that we may be shown to be wrong in
our descriptions ('this is not really a rock') , given that we got the ref-
erence right.
However, a scientist seldom publishes her raw data. Typically we look
for descriptions that are both non-trivial and provide a maximum of infor-
mation. We redescribe the raw material we deal with in research. I suggest
that we conceive of this process as the provision of a stepwise struc-
turing. A structure consists basically of a domain and a number of
operations performed on the objects of the domain . What we do when
dealing with data is both redescribe the kind of objects we deal with,
and perform new operations (measurement, etc.) on them. Correct class-
ification is a prerequisite of good science. It is also in many areas, such
as in geology, a task that involves test-procedures or measurements of

a complex nature. In some areas the intended classifications presup-
pose some successful prior classification that subsequently becomes
irrelevant and is therefore thrown out again . For instance, after having
classified a rock sample as ' red-sandstone sedimentation', the paleo-
magnetist may then go on to classify the data as a sample of 'remanent
magnetization '. Later a specification of time may be added, etc . In
this stepwise process, both the description of the domain and the
operations performed on the sample change . They reveal higher levels
of abstraction.
It is an important claim of my earlier paper on this topic that the
process of stepwise abstraction is conducted under the constraint of
theoretical inference tickets. These inference tickets come in two kinds:
inference tickets for redescription and operational inference tickets. The
term denotes the scientific rationale for describing the data set in a new
manner, or for applying new kinds of operations, or measurements, to
it respectively. Typically, some theory is referred to in these inference
tickets, e.g. , statistical theory or the physical theory of magnetism. At the
most basic levels, the inference tickets employed in the dealings with
particular data often need not be mentioned in the final publication. They
are simply regarded as part of the standard procedure, and thus part of
the training in the discipline. The higher in the level of abstraction, the
more explicit the use of the inference tickets becomes . In the 'method
part' of a scientific paper, most such inference tickets only receive brief
mention , but some may be described in greater detail. Philosophically
speaking, inference tickets provide the grounds for the contribution made
to the phenomenal level. Thus, criticism in science is often levelled at
the legitimacy of the inference tickets that are used.
One ends up with structures that acquire the status of empirical
phenomena. These phenomena differ greatly in description and gener-
ality from the raw material from which they are inferred. Epistemo-
logically one may think of this inferential process as similar in kind to
the Aristotelian notion of induction. If one prefers, the term idealiza-
tion seems equally appropriate here. Phenomena refer to those facts of
nature which form the basis of higher level theories. The task of the
theories is then to provide good causal explanations that link different
phenomena together.
186 MATTHIAS KAISER
IV . EMPIRICAL PROGRESS
The basic progress relation can be conceived of as being of the following

kind:
(i) progress =: {(T', T"), ••.}
Depicting progress in this manner amounts to saying that progress is a
two-place relation linking two (scientific) theories. It should be an asym-
metric and transitive relation , if one sticks to some very basic intuitions
about the term. Since progress is a value term , the yardstick for com-
paring theories with regard to improvement is supposed to be constant,
and is thus suppressed in the formulation of the relat ion. For a realist this
yardstick is truth or truthlikeness. Ilkka Niiniluoto 's work (Niiniluoto,
1984, 1991) provides a very good analysis of progress in terms of
truthlikeness. If one holds other views about science, other aims, like
problem-solving or empirical adequacy, may serve equally well under
a purely formal point of view. The problem becomes more tricky if one
holds that the aim is itself a matter of historical development in science
(but cf. Kaiser, 1991b). I shall, however, not go into this matter here.
Progress conceived in this manner gives us a good idea of what it is
we mean by the term 'progress'. It states what genu ine progre ss amounts
to, what we actually mean by the term. In this sense, it is a semantical
explication of an idea, very much in line with the semantic conception
of truth. We disreg ard what indications we may have for ascertaining
whether or not such progress actually obtains between two theories.
Any such estimation will be subject to human fallibility. We may believe
that a certain theory change was progressive, but it may tum out that
it was not. The ideal-semantic version of progress on the other hand
elucidates what true progress is.
If, however, we try to capture those ascriptions of progress that deal
with our estimations of progress, the progress relation changes from a
two-place to a three-place relation . Basically it now looks like this:
(ii) progress =: {(T', T", E), •..}
Here 'E' denotes the evidence that is cited as a ground for believing
that T' is better than T" . If the available evidence changes, then our
estimations of progressiveness may change with it. Whereas (i) repre -
sents an ideal-semantical notion of progress , or a metaphysical notion
if you like, (ii) is an epistemological version of progress.' It is typical
for the realist attitude to distinguish these two versions. This idea we find
already employed in Popper's work. Truth is one thing, indications of
truth are quite another. Similarly, true progress is one thing, our esti-
mations of progress another.
For the historian of science, and indeed for scientists, the epistemo-
logical version of progress is of utmost importance. In it we try to capture
those arguments that may be regarded sufficient for matters of dispute
over theory preference. In philosophy of science I regard it as still an
unsolved problem what kind of evidence one should expect for an
adequate notion of epistemological progress in science. I shall not enter
this discussion here, since it is not the main objective of this paper.
The lesson to be learned from the case study mentioned earlier is, I
claim, that we should widen the scope of the above explications of
progress. I think there is sufficient historical evidence from science to
claim that we should integrate the following notion of progress into
our picture:
(iii) empirical progress =: {(P', P", E), ...}
Here I replaced the basic entities that are to be compared in term s of
improvement. Instead of dealing with theories, we now deal with two
(sets of) phenomena. The rationale for this move lies in the fact that
the conditions under which we would like to replace one set of phe-
nomena with another are quite different from those that apply in the
case of high-level theories. Basically, phenomena are judged on the basis
of data and the inference tickets used in their reformulation. Theories,
however, are judged on the basis of phenomena. Fundamental to this way
of viewing the matter is the conviction that there is no deductive rela-
tionship between high-level theories and low -level data. Our general
theories simply never reach the level of things in our world or their imme-
diate description. Since I have argued that phenomena constitute an
impressive intellectual achievement of science, it is natural to conclude
that progress in phenomenal change should complete our conception of
scientific progress.
The task then remains to explicate under what conditions such empir-
ical progress obtains. I have done this in some detail in the work
mentioned, and I shall now present the basic tenets of my view.
It seems to me that empirical progress occurs in at least two, possibly
in three ways. First there are those cases where we somehow manage
to establish new phenomena, i.e., phenomena which did not earlier figure
188 MATTHIAS KAI SER
within the empirical basis of our theories. The discovery of new types
of galaxies or stars may serve as an intuitive example from astronomy.
Secondly, there are those cases where we manage to ' fine-tune ' some
known phenomena, i.e., specify them in new and better ways. This
includes cases of merging or dividing known phenomena . Thirdly, there
are cases where new methods or instruments provide us with access to
hitherto inaccessible areas of reality. The introduction of radio telescopes
or electron-microscopes may serve as prominent examples here. I am a
little uncertain about this third area of empirical progress, since it seems
hard to specify in more precise terms. From an intuitive point of view,
however, several aspects indicate that it is reasonable to include this as
empirical progress , if for no other reason than that many scientists them-
selves commonly regard such developments as progressive breakthroughs
in their discipline .
In sum, I want to suggest the following general definition of empir -
ical progress:
In a given discipline a set of phenomena P' is to be judged
as empirical progress of the general form {(PI, P", E),
...} in light of the available evidential situation E if:
(i) P' contains the discovery of new phenomena in
relation to P", or
(ii) P' corrects known phenomena in relation to P", or
(iii) P' represents new methods or instruments in relation
to P " by means of which new types of phenomena
become empirically detectable which hitherto were
beyond the reach of scientific investigation.
It will perhaps be noted that I relativize empirical progress to disciplines.
This has to do with my general view of epistemological progress in
science. I have argued elsewhere (cf. Kaiser, 1991 b) that at least some
methodological standards of theory acceptance that we find in scien -
tific practice are designed for the particular purposes of a given discipline,
and may be at variance with standards from other disciplines . The degree
to which they are in conformity with those very general standards that
the methodologist is normally interested in, is a matter of empirical inves-
tigation . In methodology I consider these disciplinary methodological
standards as epistemologically prior. In any case, taking a close look at
the accepted standards of disciplines allows us to take seriously the
justifications that the scientists themselves inside a discipline provide for
a given scientific change. We need not artificially look for reasons that
may not be there, or may not have played a role at all. Thus, judge-
ments of progress do not result in a proof or disproof of scientific
rationality. Instead we start with an initial assumption of scientific ratio-
nality . I see scientific rationality and scientific progress as intertwined,
as two sides of the same coin, if you wish. One should note, however,
that this view gives one room to deviate from an insider judgement.
We may still disagree with the scientist with regard to a particular
scientific change. For instance, we may find that the methodological
standards accepted in the discipline do not validate the change. Or we
may find that the standards do indeed validate this particular change,
but the scientists were unaware of it, and accepted the change for
irrelevant reasons. Of course, one may also have reason to criticize the
accepted standards of a discipline, but this is a different matter that I
shall leave aside here (but again cf. Kaiser, 1991b). It is my claim then
that in judging cases of alleged empirical progress in science (as well
as indeed cases of theoretical progress), we have to utilize those method-
ological standards that characterize a discipline at a given time.
All three cases mentioned in the above definition should be clarified
further. To some degree this can be done by utilizing the model of the
relationship between data and phenomena that I briefly described above.
I shall try to elucidate some of the main characteristics of it.
What precisely could it mean to say that we have discovered a new
phenomenon? In terms of my model , I think it can mean one of three
things. First, the new phenomenon P may emerge as the abstraction
of a new set of data , utilizing legitimate inference tickets. This is, for
instance, the case with the magnetic reversals of the earth's magnetic
field . Reversals were shown to have occurred by having available data
(and methods) of a totally new kind, i.e., data of remanent magnetiza-
tions. Secondly, a new phenomenon may be established by higher
abstraction from given phenomena when new inference tickets back up
such a move. In our example, the emergence of a precise timetable of
reversals was a development that could be utilized as inference ticket
to infer a uniform spreading of the sea floor. Thirdly, a new phenomenon,
P, may be the result of merging several lines of abstractions from data
to phenomenon , where each individual line employs some inference
tickets that yield uncertainty with regard to P. Yet when 'merged' , these
different lines of inference may compensate for these uncertainties. To
some degree this was already the case by relating the paleomagnetist's
190 MATTHIAS KAISER
with the paleoclimatologist's findings . Later, when the findings from

the sea floor were also included, the phenomenon of drift became a virtual
certainty.
Now, what could it mean to say that P' corrects some phenomena in
relation to P"? Again I think three cases can be distinguished. First, there
are cases where a discipline realizes that certain inference tickets are
much more problematic than assumed earlier. The evidential situation
E may change in such a way that it can be shown that certain methods
just cannot be employed as intended. For instance, Wegener believed that
certain measurements, made by himself and others, showed a westward
drift of Greenland. Not only could one soon show defects in the carrying
out of those particular measurements, but one could also show that the
theoretical limit of such measurements was such that one either had to
assume a tremendous rate of movement, or one had to compare them with
nonexisting measurements from some time before human history. Thus,
the data could not yield the phenomenon in question.' Secondly, there
are cases where known phenomena have been described differently by
utilizing new inference tickets or new sets of data or both. Most typically
this happens in those cases where phenomena which originally were
described in qualitative terms are subsequently described in quantita-
tive terms. The time-scale for magnetic reversals was such an example
representing an improvement of the general phenomena of reversals as
such . A previously known phenomenon was not replaced, but represented
in much more precise terms. A third case of correction may occur through
the combination of these processes. One may replace a known phe-
nomenon by utilizing a new classification (structure) which is better
supported by the known data and inference tickets. The introduction
of transform faults is, I suggest, an example. The known fracture zones
of the sea floor were originally described as either ridges or trenches.
Wilson showed that transform faults are not really a class of the known
transcurrent faults, and thus improved the set of phenonomena that
together described sea floor fracture zones .
It is much more difficult to specify the third alternative of empirical
progress, the introduction of new methods or instruments that reveal new
'layers' of reality. I shall therefore abstain from attempting this here,
and rather point out that leaving out this alternative would make the
definition less in accordance with the intuitive judgements of scien-
tists." In the case study, it seems that the astatic magnetometer, the
radioactive clocks, etc., were recognized as important breakthroughs
inside the speciality even before they actually produced data that revo -
lutionized the discipline.
V. A CONCLUSION
I believe that there are many good reasons to differentiate between empir-
ical and theoretical progress in science. The reasons are both of an
empirical and a philosophical kind. As to the first, it seems to me that
the case of the recent revolution in earth science, indicates that such a
differentiation is in line with many important intuitions of scientists. It
also structures the whole history of this revolution in a manner which
makes it more easily amenable to philosophical descriptions. The theo -
retical change that occured after 1967 is incomprehensible if we ask what
particular set of data brought about this sudden conversion. There was,
for example, nothing in the Eltanin 19 profile that would warrant such
a dramatic conversion, if the profile is not seen in relation to the number
of new phenomena that already were established or largely supported (cf.
also Kaiser, 1993a). It would be wrongheaded to assume that, e.g ., the
phenomena of magnetic reversals, transform faults, and drift were
accepted then because of the acceptance of the theory of plate-tectonics.
The theory was accepted because of these phenomena. Phenomena enjoy
a considerable independence with regard to the theory they support. Still,
the theory itself was more than just the sum of these phenomena. The
theory consists fundamentally in the provision of a causal mechanism,
the sea floor as conveyor belt of the moving continents, and physical
models of how plates collide. If one agrees with my idea that theoret-
ical progress in science (in its epistemological version) consists basically
in the provi sion of causal mechanisms (processes) with explanatory
relevance, then this change of theory constitutes a good example of
theoretical progress. It was made possible by the previous empirical
progress that consisted in the discovery of important new phenomena.
As to the philosophical reasons for differentiating between empir-
ical and theoretical progress in science, it seems to me that the most
important reason has to do with the recognition of the essential distinc-
tion between data and theory. We have to appreciate how phenomena
are made out of data, and how theories connect not to data but to phe-
nomena. In terms of epistemology, the ' upward ' movement from data
to phenomena is met with the ' downward' movement from theory to phe-
nomena. Furthermore, realizing the important intellectual achievement
192 MATTHIAS KAISER
that goes into the construction of phenomena, the theoretical inference

tickets for abstractions, we realize that if we have reasons to assign the
value-term 'progress' to theory change, then we also have reasons to
assign this very same term, though with different specifications, to phe-
nomenal change.
With regard to the realist position it seems worthwhile to point out
that the view sketched in this paper fits in well in several respects.
Truth still plays the dominant role . Our division between true progress
and estimated progress is in line with a realist attitude. The realist rejec-
tion of the traditional division between observational terms and theoretical
terms in scientific theories is backed up by an explanation of why this
distinction is inadequate. Furthermore, the so-called 'entity realism'
may be considered a sub-case of some phenomenal realism. Most entities,
like electrons or quasars, are phenomena of nature that are inferred along
similar lines as described here. However, some phenomena may not have
entity character. Magnetic reversals are not entities or things, but rela-
tively stable, universal (in time and space) re-occurring events. Like
electrons that one can spray and then detect their effect, the effects of
the reversals can be detected. I thus conclude that a realist has many good
reasons to adopt the differentiation between empirical and theoretical
progress in science which has been presented in this paper.'
The National Committee for Research in Ethics in Science

and Technology (NENT),
Oslo, Norway.
NOTES
1 The money came from the US Office of Naval Research (ONR) and the National
Science Foundation (NSF) , and must be seen in connection with the military 's recogni-
tion of the importance of submarine activity for future warfare between the super-powers.
2 The same distinction between true and estimated progress is utilized in the work of
Ilkka Niiniluoto. lowe the terminology to him. Cf. Niiniluoto, 1984, 1991.
3 This is not really an example of empirical progress since few if any practioners of
the field lend any credibility to the phenomenon in question to begin with.
4 Among other things this seems to deviate from the intuitions that guide the Nobel prize
committee.
5 I gratefully acknowledge helpful criticisms and comments by Craig Dilworth and
David Olridge.
BIBLIOGRAPHY
Bogen, J. and Woodward, J. (1988). 'Saving the Phenomena', The Philosophical Review
97, pp. 303-352.
Giere.R. (1988). Explaining Science - A Cognitive Approach, Chicago/London .
Hallam, A. (1973). A Revolution in the Earth Sciences, Oxford .
Kaiser, M. (\ 991a). 'From Rocks to Graph s - The Shaping of Phenomena ' , Synthese
89, pp. 111-133.
Kaiser , M. (1991 b). 'Progress and Rationality: Laudan 's Attempt to Divorce a Happy
Couple', Inquiry 34, pp. 433-455.
Kaiser, M. (1993a). 'Philosophers Adrift ? - Comments on the Alleged Disunity of
Method', Philosophy of Science 60, pp. 500--512.
Kaiser , M. (1993b) . Aspekte des wissenschaftlichen Fortschritts, European University
Studies XX/398, Frankfurt a. M.lBeriin/Bem/New York/Paris/Wien .
LeGrand , H. E. (1988). Drifting Continents and Shifting Theories , Cambridge .
Niiniluoto, I. (1984). Is Science Progressive ", Dordrecht.
Niiniluoto, I. (\ 987) . ' How to Define Verisimilitude', in T. A. F. Kuipers (ed.), What is
Closer-to-the- Truth'l, Amsterdam .
Niiniluoto, I. (1991). ' Realism, Relativism, and Constructivism', Synthese 89, pp. 135-162.
Oreskes, N. (\ 988). 'The Rejection of Continental Drift' , Historical Studies in the Physical
and Biological Sciences 18, pp. 311-348.
Ruse, M. (\981). ' What Kind of Revolution Occurred in Geology ?', in P. D. Asquith
and I. Hacking (eds.), PSA 1978. Proceedings of the 1978 Biennial Meeting of the
Philospohy of Science Association, Volume Two , East Lansing, Michigan.
Woodward, J. (1989). 'Dat a and Phenomena' , Synthese 79, pp. 393-472.
CHIN-TAl KIM
SCIENCE AS IDEOLOGY
The idea of science as theoria has since Aristotle dominated the philo-
sophical consciousness of the West. Theoretical science, according to
Aristotle, is the pursuit of knowledge of universal and necessary truths
about the world for its own sake; if there is any interest that should
motivate theoretical inquiry, it is the interest in knowing, nothing else.
Aristotelian defenders of science will therefore eschew and debunk as
ideology I any cognitive enterprise that either has an explicit non-theo-
retical end or evinces an implicit non-theoretical interest. We should
ask the following question to assess the idea of science as theoria: Can
theoretical interest be as sharply distinguished from the non-theoretical
as Aristotle thought? And, given a distinction, can such interest alone
adequately justify theoretical inquiry? The answer to each question should
be negative.
The Aristotelian thesis that knowledge of universal and necessary
truths about the would should be sought for its own sake implies that
such knowledge is intrinsically valuable or worth seeking for its own
sake .' The thesis cannot mean that theoretical knowledge is inapplic-
able, hence cannot be sought for anything else. Such knowledge, by
Aristotle 's own admission, has wide relevance and applicability. If
the thesis, on the other hand, implies that such knowledge is actually
sought by all humans for its own sake and for the sake of nothing else ,
it is likely to be false. We shall not here address the question as to
how being worth seeking is to be differentiated from being actually
sought by all humans. We only note that, whatever the proper analysis
of being worth seeking for own sake may be, such knowledge must
be thought worth seeking for its own sake by defenders of science as
theoria.
Is knowledge of universal truths about the world the only thing that
is intrinsically valuable? The affirmative answer to this question is dif-
ficult to defend. Are life, beauty and happiness, for instance, not also
things that are intrinsically valuable? In point of fact Aristotle himself
195
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.), Realism and Anti -Realism in the
196 CHIN -TAl KIM
subscribes to a pluralism of intrinsic values. If a listing of plausible

candidates for being intrinsically valuable other than such knowledge
is not an effective argument against the claim that it is the only thing
having intrinsic value, then by the same token the claim to that effect
cannot be grounded on intuition either as in Aristotle.
Sensible defenders of the idea of science as theoria are not likely to
say that their commitment to this idea commits them to the view that
theoretical knowledge is the only thing that is intrinsically valuable.
Let us assume that it is not. Suppose that something which is not theo-
retical knowledge is intrinsically valuable also. Its pursuit will have as
much justification as the pursuit of theoretical knowledge unless some
of its properties other than its intrinsic value are relevant to the justifi-
cation of its preferential pursuit.
Defenders of the idea of science as theoria must claim that intrinsic
values come ranked, that the value of theoretical knowledge is to be
ranked as the highest or as one of the highest at least. There indeed is
no inconsistency in saying that two things are both intrinsically valuable
but one of them is contextually preferable to the other. A promising
argument for the preference of theoretical knowledge will have to rely
on the consideration that more than anything else it contributes to the
realization of some greater good such as personal happiness or the welfare
of society or humankind. It can be said that the judgment as to what
proportions of various ends should go into the constitution of the greater
good must be an application of relevant knowledge; that the realization
of those proportions of those ends requires technology; that the most
effective technology is again an application of knowledge. Making these
points however, is trying to justify the preferential pursuit of theoret-
ical knowledge by appealing to its utility , abandoning the Aristotelian
dogma that such knowledge has intrinsic value only, but no extrinsic
value.
The situation that unfolds itself is as follows: Theoretical knowledge
is likely to be only one of many things that have intrinsic value if there
are such values, though it may be an indispensable element of and a
necessary instrument for realizing some greater good that transcends
whatever intrinsic value it may have. So even if the initial pursuit of
knowledge can be explained by its having intrinsic value, a full justifi-
cation of its pursuit should be ultimately grounded in its instrumental
relationship to some greater value. If ideology is to be thought of as
pursuit of knowledge for use in realizing a non-theoretical end, the above
SCIENCE AS IDEOLOGY 197
consideration shows that theoretical science cannot be clearly differen-

tiated from ideology.
II
The conclusion of the foregoing discussion still leaves intact the idea
of theoria as a privileged sort of human endeavor and achievement -
search for and discovery of truths. Apart from the question of whether
theoria is free of values other than its intrinsic value, there is the impor-
tant question as to what it is and how it is possible.
There is an interpretation of theoria according to which it is a mental
representation of things independent of human consciousness in existence
and characteristics, and the cognitive apparatus of humans has a natural
affinity with potential objects of knowledge so that the apparatus can
be made to serve the end of representing things faithfully if only due
methodological care is taken.
Much of the history of Western epistemology has been a battle against
these assumptions. The critique of such realism typically took the form
of denying its metaphysical thesis by denying its epistemological thesis:
no knowledge of objects is possible if they are metaphysically external;
but there is possible knowledge of physical objects; therefore, physical
objects are not metaphysically external. The core of the reasons
supporting the first premise is that there is no cognitive access to
metaphysically external reality except through subjective vehicles of
representation, ideas or concepts, so that there is no possible knowl-
edge of the agreement of subjective representations with external reality.
The denial of the metaphysical externality of the physical opened up a
range of possible views, e.g ., forms of idealism and phenomenalism.
What is common to the older forms of anti-realism is the thesis that
the physical world at least in part is a human construction though they
differ on the precise nature of the elements of the construction and the
principles by which they are organized.
Kant's transcendental idealism deserves special consideration because
it was the most thorough response to the problem of representation raised
by metaphysical realism, of which Cartesian dualism gave a clear
symptom, and because a perception of its difficulties gave rise to trends
of thinking some of which are alive today. While denying the human
possibility of representing a world of things -in-themselves, thus siding
with the subjective idealists and the phenomenalists, Kant held that
198 CHIN-TAl KIM
there are a priori forms of sense intuition and a system of a priori

concepts and principles that must pre-structure, 'make possible,' all
human representations. 'Objectivity ' can have one of a least two dif-
ferent senses: structural correspondence between things-in-themselves
and their representations, or the universal and necessary validity of
subjective forms , concepts and rules of representation that make the
representation possible. Kant replaced the first sense with the second , his
only concess ion to metaphysical realism, " transcendent realism," being
that there must be some unknowable things-in-themselves that somehow
are responsible for sensations, the matter of representations.
But the notion of the synthetic a priori, which can be interpreted as
the invariable species-universal presuppositions of knowledge - has grad-
ually eroded. Kant 's idea of a transcendental subject constructing nature
without being in it leaves the subject outside the domain of possible
scientific knowledge. And his transcendental theory of experience was
not preceded by a reflective investigation of its own possibility. The
idea of science as a comprehensive inqu iry about the world demands
that the human cognitive apparatus and its use should themselves be
scientifically accessible. The human subject of knowledge needs to be
viewed as a being in nature to make scientific knowledge of it possible.
Since Kant's time there has been a steady movement toward de-
transcendentalization and naturalization of the human subject and its
cognitive projects and acts . With the naturalization of the subject, its
cognitive competence comes to be thought of as a property whose
realization throughout the species and persistence through time is to be
taken as matter of contingency to be discovered, not to be postulated a
priori.
The placing of the human subject in nature provides a basis for placing
it in history as well. Being in nature makes history possible. Humans
create history and culture while existing in nature. Once the human
subject is placed in nature and history, however, there arises a dialectic
of the two. Humans are situated in nature yet nature in turn becomes
an object of their interpretation and construction. That with which humans
interpret nature, theory, is a product of nature in a broad sense but it,
more proximately, is a product of culture: its genesis, its acceptance,
and its use cannot be fully explained apart from the history of human
inquiry into nature. The history of natural science, and indeed that of
all cognitive pursuit, becomes a proper reference in its interpretation,
critique and justification.
A revised Kantian view that the knowable nature of the world must
depend on a variable theory constructed by humans existing in nature
and history underlies modern -day anti-realism. The current realism/anti-
realism controversy differs from the older, metaphysical controversy.
Today's realists and anti-realists accept a broadly naturalistic metaphysics
that places humans within physical nature, so that the old question as
to whether the physical is metaphysically external and, if so, how human
consciousness can ever have cognitive access to it is no longer their issue.
Also the realists and anti-realists alike consider the history of scientific
inquiry relevant to an epistemological critique of science. Scientific
theories are recognized by both part ies, willingly or reluctantly, to be
cultural artifacts of historical beings - beings subject to limitations of the
human mind and the senses and influences of language, tradition, inter-
ests, values , prejudices and societal forces.
Despite such agreements, the realists and the anti-realists still differ
in the theory of science. Today's realism, although it has shorn the
physical of its metaphysical chorismos from human consciousness,
adheres to the idea of truth as correspondence and knowledge as repre-
sentation: a physical theory is true if and only it represents the physical
world. If we consider the fact that the metaphysical externality of the
phy sical world was rejected primarily to ensure the possibility of its
representational knowledge, we can understand the continuing attach-
ment to the idea of knowledge as representation that seemed to have been
legitimated.
The realists, however, recognize that the representational adequacy
of theory can only be conceived as an ideal and that still developing
theories can only be conceived as approximations to true representa-
tion. In order to prevent this ideal from becoming empty, realists have
to devise an epistemology and a methodology of science that promise
to ensure a continuous progress of science to ever closer approxima-
tions to adequacy. But a gap will always remain between a semantic ideal
of true representation on the one hand and an epistemological and
methodological account of its possible achievement on the other. And the
question arises as to how the epistemology and methodology them-
selves are to be grounded. It will be a reversion to transcendentalism
to try to construct them a priori from an a-historical idea of scientific
knowledge. It will be only half-heatedly historical to try to construct them
on the basis of an analytical and critical reflection upon the current
state of science with an uncritical assumption that current science is a
200 CHIN-TAl KIM
culmination of a continuous process of evolution. A fully self-conscious

epistemology and methodology of science should be grounded in a
critique of the history of scientific practice with different epistemolo-
gies and methodologies that have legitimated and regulated its varieties
- a critique that shows that the past norms of scientific practice have
indeed been superseded. Part of the burden of such a critique will be
to show that any significant change in the history of scientific practice
is a case of genuine progress that preserves the continuity of science.
Confronted with the historical fact of changes in science, realists
may say that there is a constant model of science with a body of meth-
odological canons and try to explain theory changes, rejections and
replacements as being due to new observations or corrected observational
errors, that it is the content, not the structure, of theory that changes.
The realists can hope to place on this basis their claim that the history
of scientific inquiry is a continuous progress toward a true theory. Today 's
realism thus centers around the key theses of a possible correct repre-
sentation of the real and the adequacy and diachronic constancy of the
criterion of adequacy of theory.
Current anti-realism expresses a skepticism about each of the realist
theses on various grounds. Going beyond rejecting the old metaphys-
ical view of the physical world as a reality independent of human
consciousness yet capable of being accurately represented, they impugn
the concept of representation itself. One of the grounds of such impugning
is the relativity of the criterion of correspondence to a conceptual scheme,
which will imply the untenability of the view that the world has one
intrinsic structure to be uniquely represented. Each conceptual scheme
may still avail itself of the idea of truth as correspondence but that will
not mean that realism with its original intent can be preserved, for a
realism that deserves the label must claim that there is only one true
theory. The realist idea of truth as correspondence should be either
relativized to theories or replaced with an epistemic concept that can
countenance a plurality of incompatible theories, such as coherence of
beliefs .
Anti-realism also draws support from reflections on the history of
scientific practice. There is no norm of science constant through the
history of science, nor can one be instituted a priori. A rigid distinc-
tion between the structure or form and method of science on the one hand
and its content, or between theory-free observation and theory on the
other, is untenable. The Aristotelian theory of science and Aristotelian
methodology consisting of syllogistic and intuitive induction, for instance,

cannot be separated from his essentialist realist metaphysics. And
although an ultimate convergence of theories can be an object of hope,
even faith, the history of scientific inquiry should not be assumed to
be a continuous progress such that a later phase of science supersedes
all of its earlier phases . Any discontinuities or revolutionary changes,
'paradigm shifts,' in the history of science should be acknowledged to
be such after the exhaustion of an honest effort to find hidden continu-
ities. The future states of science exemplifying new norms cannot be
predicted.
A key issue between current realism and anti-realism thus turns out
to be how the historical career of science should be brought to bear
upon the explanation and justification of scientific practice. The verdict
on them is still pending. But it would not be unfair to say that the realist
burden of defense against the anti-realist attacks is heavy.
III
Current realism and anti-realism have different implications on the issue

of the value-neutrality or otherwise of science . Even the realists cannot
honestly deny that non-theoretical interests have partly motivated the
pursuit of knowledge and the conduct of inquiry, or that science has
served non-theoretical interests, but they would say that an objective
critical reflection is always possible that can identify intrusive non-
theoretical interests and correct their negative influences . Reason is prone
to influence by interest but is ever able to reclaim and exercise its
autonomy with the strength of its self-interest.
Norms of science appear to differ from ethical norms . The existence
of a plurality of conflicting ethical norms and people 's choice of different
ones tends to be taken for granted as a human condition, while a
plurality of conflicting norms for doing science discomfit scientific
sensitivity. The explanation is that the goal of science is thought to be
a true representation of the world, and that a plurality of conflicting truths
of the same world is thought impossible, while it is thought that even
with a universal acceptance of the same factual account of the world
people can differ on the norm for action. That is, science is still thought
'objective' while ethics is thought to have an uneliminable 'subjective'
dimension. This differential intuition about science and morals may
appear to support realism. However, the anti-realists can counter: There
202 CHIN-TAl KIM
is not one world with many representations or 'versions' but there are
many world-representations or word-versions; the world can be no more
than a world-representation that will ideally be accepted by a global
scientific community at the limit of possible inquiry. But that world is
something that cannot now be predicted. It is arguable that a convergence
of human morals also is a desideratum and that such a convergence
will be a function of a convergence of the human sciences. The diver-
gence of human morals in significant part is due to the conflicts among
the theories of human nature and human society. The point is that science
and morals do not radically differ.
It was argued, in Section I, that even science conceived as theoria
is value-oriented. Two points were made: first, science as theoria is
conceived by its advocates to have intrinsic value, so that its pursuit
expresses a commitment to a value - the epistemic value if you will ;
second , the pursuit of science must be ultimately justified by the instru -
mental connection that the product of theoretical activity has some good
that transcends its epistemic value . If value-neutrality is to be understood
as the total absence of commitment to a value, even science as theoria
is not value-neutral. And if value-neutrality is to be understood as the
justifiability of the given practice apart form a consideration of its instru-
mental connection to a value that transcends whatever intrinsic value
the practice may have, even science as theoria cannot be considered
value-neutral.
With the discarding of the idea of science as representation of a
metaphysically independent world by a subject wearing a protective tran-
scendental shield , the issue of the value-neutrality or otherwise of science
becomes more complex. The human subject can no longer be viewed
as a pure knower, with no interest but the theoretical; it must be viewed
as a being subject to forces and influences of nature and history that it
tries to understand and cope with. The question no longer is "Is science
value-oriented or value-neutral?" but "What value commitments explain
or justify scientific practice ?" The avowal, even an honest one, by the
practioners of science, that they have nothing but a general commit-
ment to the epistemic value and a specific commitment to the regulative
norms of their practice , should not be taken as sufficient evidence that
their practice is neutral as to all other values. There are reasons for
caution other than a psychological one.
A practice is not a totally isolable segment of a web of practices.
The total web of practices within a culture is permeated with value-
commitments. The influence of one practice upon another, that of religion

upon science in the Middle Ages, for a notable example, attests to an
osmosis of values and beliefs from one practice to another. The fact
that natural science has steadily gained autonomy from religion need
not be denied, but this fact implies that the epistemic values thought
internal to it have gained precedence over values of other sorts, and/or
that the interests of the political and social forces opposing religious
authority and institution have coincided with the interests of science,
not that science has become value-neutral. Even if for argument's sake
science is supposed to have nothing but an internal epistemic value,
the practice of science must still be justified in reference to a system
of values that transcend it.
Science must be ideological at least in a broad sense. Indeed all human
practices are . The idea of a value-neutral practice expresses an igno-
rance of the organic interconnections among the domains of the cultural
and historical existence of humans. It is one thing to be immersed in a
practice; it is another to have a reflective understanding of it in its exis-
tential and historical context. Fully self-conscious science is one that
acknowledges and takes seriously its ideological nature . Science distin-
guishes itself from other human practices by its commitment to the
achievement of rational consensus on its goal and method. Anti-realism
gives support to a conception of science as a practice having value
commitments and subject to values by noting that it is a practice in a
system of practices that values permeate. Just as anti-realism must impugn
the distinctions between the given world and its variable representa-
tion, between the constant form of science and its variable content,
between a diachronically valid norm of science and its errant application,
it must also impugn the distinction between the value-free knowledge
of the world and value-oriented distortions of its true picture. The issue
between realism and anti-realism here is not so much whether non-
epistemic values have actually influenced scientic practice as whether
science should be so conceived as to require its freedom from such values.
The issue is a humanization of science.
Current anti-realism helps uncover the ethnicity or, better, cultural
specificity, of science and raises an issue about its universality. Even
the realists would find little difficulty in accepting the fact that customs,
morals, laws, institutions, religions, and philosophies (taken broadly as
world views) are relative to cultures. But they retain their sanguinity
about the universality of science. But the science whose universality they
204 CHIN-TAl KIM
claim is their science, a practice that has been shaped in Western culture
and civilization. All of the problems and disputes about Western science
are fully significant only in the conceptual space of that culture and
that civilization. Anti-realism seems to offer an antidote to cultural
chauvinism and ethnocentricism. If even within the same Western civi-
lization the idea of science is, and has been, an issue, then the alleged
universal validity of that idea across civilizations should all the more
be an issue .
Upon admission of the historical and cultural identity and bound-
aries of Western science, two extreme attitudes should be avoided. One
is that of assuming the total impermeability of non-Western civiliza-
tions to Western science and resolving to do Western science in an
isolationist way; the other is assuming the superiority of Western science,
indeed of Western civilization, and trying to convert the non-West to
Western ways of thinking with missionary zeal. Neither attitude seems
justifiable. A third option, though apparently constructive, is equally
untenable, namely assuming the rationality of all civilizations and the
possibility of a universal meta-di scour se in which Western science and
non-Western analogues can be comparatively interpreted and assessed.
There is no such meta-discourse even within a civilization. What is taken
to be such a discourse may only be a rhetorical modification of a favored
object discourse devised for its justification.
The correct attitude, I propose, is that of promoting studies of civi-
lizations by one another and assimilation of the foreign by each, and
hoping for expansions of inter-civil izational spaces for conversation. Such
promotion requires as much mutual openness as self-reflection, as much
pride in one's own tradition as respect for another, as much valuation
of unity as recognition of the historical facticity of diversity, and as much
appreciation of agreements after conversation as tolerance of differ-
ences that persist.
Case Western Reserve University ,

U.S.A.
NOTES
I The label 'pseudo-science' tends to be applied to cognitive practices that in the judgment
of the critics who use the label employ a wrong method or have false substantive assump-
tions or presuppositions. The pejorative label, however, has no specific implication that
what is so labelled is motivated by a non-theoretical interest, though it may well be.

The label ' ideology' on the other hand explicitly impugns the motive of a cognitive practice
as impure .
2 See Aristotle, Met., Bk. I, 982'15. The word 'theoretical' in Aristotle primarily concerns
the motive of inquiry. But a clas sification of knowledge by motive for its pursuit fails
to inform us of the objects for different classes of knowledge. First Philosophy, mathe-
matics, and philosophy of nature ('physics') are the major branch es of theoretical science
as Aristotle conceives it. Aristotle fails to make clear whether these sciences are theo-
retical because of the nature of their subjective matters or the concept of theoretical science
implies that it should be about those objects that sciences concern.
MARX W. WARTOFSKY
THREE STAGES OF CONSTITUTION : HISTORICAL

CHANGES IN THE ONTOLOGICAL STATUS
OF THE SCIENTIFIC OBJECT
In this paper, I muse very briefly on realism and constructivism not

simply as conceptual or philosophical alternatives in the interpretation of
the relation between theoretical knowledge and its object, but rather as
philosophical articulations of historically different modes of scientific
practice , in which the relations between knower and known are in fact
different.
What I would like to pose for discussion is a thesis on the historical
changes in the relation between the subject and the object in science. That
is to say, that there are in fact situations in scientific inquiry which
are, so to speak, "realist", in which the object of knowledge is entirely
independent of the subject, as were the stars and celestial phenomena
in classical astronomy. There are on the other hand objects of science
which exist only in being constituted in the course of experimenta-
tion itself - i.e., either as artifacts of research activ ity, or as transfor-
mations of such objects as these are changed by scientific-technical
practice.
It seems to me that there is a certain dialectic that develops in this
regard ; that, in the natural sciences, where we began by eliminating all
supernatural and mythic anthropomorphisms, we now return more and
more to a humanized naturalism. That is to say, classic realism, which
proposed an absolute breach between the subject and the object [in
asserting the ontological independence of the object], has to be modified ,
but not because of any philosophical decision . It is more a matter of a
change in the mode of scientific production, i.e., in the relations between
scientific theory, on the one hand, and experimental practice and obser-
vation, on the other. We are now in need of a theory of scientific
production as theoretical production and that means not only the pro-
duction of ideas, as some hypostatized, objectified "third world" entities
of the sort Popper proposes, in Objective Knowledge, but rather the repro-
duction of the world in theory, as an actual transformation of matter
into object, of the an sich into the fur uns.
This production is a process of scientific construction that proceeds
207
R. S. Cohen . R. Hilpinen and Qiu Ren zong (eds.) , Realism and Anti-Realism in the
Philosophy of Science, 207-21 8.
208 MARX W. WARTOFSKY
historically from the level of pure speculation, to the level of passive

reflection, to the level of active transformation of the object of knowl-
edge by intervention. Nowadays, one does science not merely by inter-
preting the world, but by changing it.
I. THE PROBLEM AND ITS GENESIS :

THE BOHR-EINSTEIN DEBATE
One of the crucial debates in contemporary physical theory that has

been as explicitly 'philosophical' as it has been 'scientific' - I use
scare-quotes here because, as you may know, I don't believe there is
any sharp demarcation between the two - is the Bohr-Einstein debate
over the completeness or incompleteness of Quantum-mechanics. The
disagreement centered, originally, over the physical interpretations of the
wave-function, 'If, specifically, as to whether its statistical form was to
be taken as a description of the average behavior of a sub-atomic system,
and therefore as an incomplete account of the individual phenomena or
states making up the ensemble; or whether it was instead to be taken
as a complete description of an objectively probabilistic phenomenon,
in which an exhaustive measurement of the simultaneous values of the
classical parameters - the canonical conjugates - of individual systems
or of particles was in principle excluded, by the Heisenberg uncertainty
relation, and by the formalism of Quantum-mechanics, e.g., the non-com-
muting operators. In terms of physical interpretation, the question was
one concerning the locality or non-locality of the individual subsys-
tems or particles constituting the ensemble which the wave-function
described, and hence, also the time-evolution, in terms of continuous
functions, of such systems. Locality and causal ity (defined for a certain
ontology of the physical world) were apparently put in question by the
Bohr interpretation. In their place, what was proposed was the local
non-separability of quantum-systems. In the classical case, if two parts
of a system - two particles, e.g., electrons or photons which have
interacted - separate and go in different directions, it is possible to
correlate the states of the two particles at a distance from each other,
given the initial energy or momentum of the system, in the sense that
a measurement on one of the particles will not have any effect upon a
measurement on the other. Given the conservation law. the energy or
momentum of the unmeasured particle can be inferred from measurement
on the other. In effect, each is locally separable from the other. If, on
THREE STAGES OF CONSTITUTION 209
the other hand, the two systems are locally non -separable, any mea-
surement on one will affect a measurement on the other, so that the
two subsystems are then not independent of one another. Metaphorically,
we may say that one subsystem 'knows' instantaneously what is
happening to the other (what Einstein dismissed as " telepathy"); or
alternatively, what would require action at a distance, in contravention
of the constraints of special relativity, where signal velocity cannot exceed
the velocity of light. This quantum-correlation of the measurements taken
upon two such subsystems at a relatively large distance from one another
violates the notion of locality, which Einstein held to be crucial to a
complete physical description.
Bohr 's response to the charge of incompleteness was to take
Einstein to task for what he called "the essential ambiguity involved
in reference to physical attributes of objects when dealing with phe -
nomena where no sharp distinction can be made between the behavior
of the objects themselves and their interaction with the measuring
instrument".
2 . STRIFE ABOUT REALISM
From this and from the way in which the discussion of the so-called
'Copenhagen Interpretation' of Bohr and Rosenfeld progressed, in later
years, it would seem that there are ambiguities about 'physical reality'
on both sides , however. Bohr, and especially Rosenfeld, eschewed the
charge of ' positivism' and of Berkeleyan phenomenalism that Einstein
leveled at what he called the 'orthodox' interpretation of Quantum-
mechanics - namely, that it reduced reference to nothing but the
phenomena of measurement, ' observables' , and thus denigrated any talk
of ' physical reality ' beyond such observables, as empty or unwarranted
metaphysics. But, Bohr objected, what we cast in doubt is not our
access to 'physical reality' but rather a certain characterization of that
reality, or of our relation to it in a given mode of scientific practice. In
effect, Bohr rejected the accusation that he was not a realist in physical
theory.
In short, we may speak here of two alternative proposals for consti-
tuting this 'physical reality' as an object for science; and we will also
have to speak of an ambiguity, in the distinction between the reference
of the term 'scientific object' and the term ' physical reality' or 'physical
object' as it is used by Einstein and by Bohr. It is true that Bohr some-
times sounds like a reductive or eliminative positivist, asserting that

reference ends at the measurement itself, that what is, is the value of a
variable, determined by measurement, so that the entity, in physics -
or the 'physical' object, so-called - is no more than what the phenom-
enon of measurement yields. It is this point of view that Einstein reacts
against in his argument concerning the time of radioactive decay, which
he offers as a crucial argument for the continuing incompleteness of
Quantum-mechanics. The issue, in that case, concerns whether it makes
sense to talk about the time-instant of the disintegration of a radioac-
tive atom, or whether, on the contrary, it makes any sense to deny that
there is such a time instant at which disintegration taken place. Einstein
argues that the quantum theorist is forced to take a theoretically implau-
sible position, in this case - another argument of the original EPR sort
- and to deny that the notion of time-instant of radioactive disintegra -
tion has any physical significance, or any physical correlate in reality.
Einstein writes, "If .. . the individual atom has a definite disintegra-
tion time, then as regards the individual atom, its description by means
of the '\If-function must be interpreted as an incomplete description"
(Schilpp, p. 668). In the mouth of a hypothetical quantum-theoretical
critic, Einstein puts the following positivist language: "This considera-
tion stands or falls with the assertion that there actually is such a thing
as a definite time of disintegration of the individual atom (an instance
of time existing independently of any observation). But this assertion,
from [the Quantum-Mechanical] point of view, is nor merely arbitrary
but actually meaningless. The assertion of the existence of a definite time-
instant for the disintegration makes sense only if I can in principle
determine this time-instant empirically. ... The entire alleged diffi-
culty proceeds from the fact that one postulates something not observed
as real' (Ibid., pp. 668-669).
Einstein replies : (to his own version of the 'positivist'): "What I dis-
like in this kind of argumentation is the basic positivist attitude, which
from my point of view is untenable, and which seems to me to come
to the same thing as Berkeley 's principle, esse est percipi" (Ibid.,
p. 669).
But is Bohr 'positivistic' in this sense? And, on the other hand, is
Einstein 'realist' in the sort of metaphysical realist way his statements
sometimes suggest? In the Bohr quotation earlier, he distinguishes
between "the objects themselves" and "the measuring instruments",
where, presumably, reference is distinct, between one and the other, in
THRE E STAGES OF CONSTITUTION 211
straightforwardly classical terms . And yet, of course, Bohr's reference

to "the objects themselves", is incomplete, since the "object" can be
described only within the limits of quantum measurement. But the mea-
surement apparatus is also a physical object, which in the activity of
making measurements interacts with the first object in some way so
that the two together form a system. Yet, we do not 'observe' this
system by means of still another measurement apparatus, which measures
it, but rather interpret the measurements of the apparatus in terms
of the framework of the (necessary) interaction between object and
apparatus.
Thus, the ' physical object' or ' physical reality' as the object of
measurement must 'exist' in some sense, to be measured, but cannot
be known apart from the conditions of measurement and the theory of
that measurement, which interprets it physically. But surely, in talking
about 'physical reality ', Bohr is not simply indulging in pre-quantum-
theoretical and hence anomalous usage, resorting, so to speak, to
folk-language. Or is he? Bohr, of course, insisted that the facts of mea-
surement - the marks on strips of tape, the clicks of counters, the tracks
or patterns on photographic plates, etc. - had to be dealt with in terms
of classical physics. If, on the other hand , Quantum-mechanics is a fun -
damental theory and complete, then the classical talk is relativized to
these quantum-foundations, and the dynamic variables of the classical
macroworld are simply the 'phenomena' yielded (as we approach the
classical limit), by the essentially statistical or objectively probabilistic
'reality ' at the level of the quantum microworld.
Thus, the very notion that the quantum-description is incomplete
depends on smuggling classical criteria of completeness into a world
- or into an ontological level of the phys ical world - where they are
anomalous. Quantum phenomena, in this sense , are completely described
the way they really are in themselves, for their being-in-themselves is
the being-for-one-another of their constituent subsystems, that is, in
the interaction. There's nothing positivistic about that. Rather, it is just
ontologically queer, and it may seem queer because we've only just
opened up, in the last sixty or seventy years, the real domain of the
quantum, which would have been there all along, had we gotten to the
territory earlier. Ontologically speaking, we may say - and of course it
had been said - that we're dealing with a world fundamentally charac-
terizable, physically, as a set of (graded) potentialities - the probabilities
assigned to the eigenstates by the wave-function - waiting to be actu-
212 MARX W . WARTOFSKY
alized by interactions. Thus, Bohr may claim to be a kind of realist, given

the latitude of classically queer reformulation of the ontological status
of the scientific object.
The confrontation of the two views thus centers on a sharp differ-
ence concerning not only the characterization of the nature of the object
of inquiry - in terms of criteria of locality and separability - but on alter-
native construals of the ontological status of the scientific object, i.e.,
whether it exists objectively, as independent of the interaction of mea-
surement, or whether it is constituted in this interaction itself. The first
position is realist, with whatever qualifications or ambiguities there
may be in Einstein's formulation of it. The second is con structivist, in
the sense that the object is taken to be constituted as an artifact of the
practice of inquiry itself. Now it may be said that artifacts are no less
'real' than non-artifacts, and that 'nature' or 'physical reality' contains
both . But the same thing cannot be both . The only escape between the
horns, here, is to distinguish between the object as construed in a physical
theory, as a ' theoretical' artifact, and its reference to or correspondence
with some physical object, established by some criterion of ' fit' between
predicted values and measured values, which justifies us in taking the
theoretical object as a true description of the ' real' object; or, more holis-
tically, which justifies us in accepting the theory which yields such
successful predictions as a true theory, or as a putatively true theory.
But it is precisely this claim which is thrown in question in the context
of quantum measurement.
Let me turn then to the way in which this problem, posed in the
Einstein-Bohr debate, bears on my thesis about historical changes in
the constitution of the scientific object.
3. PRELIMINARY STATEMENT OF THE THESIS

AND CLARIFICA nON OF TERMS
The thesis I propose and examine in this paper is that the ontological
status of the scientific object changes historically, in the sense that it is
constituted differently by changing modes of scientific praxis; and further,
that we may mark, in a general way, three distinctive stages of scien-
tific praxis, the first of which I will characterize as an anthropic or
projective mode, the second, as a deanthropic or objective mode, and
the third as an interactive or constructive mode, I argue that in each of
these stages, there are both practical and theoretical changes in the
relation between scientific inquiry and its object.
Let me offer a rough sketch of the three stages, and a quick charac-
terization of the modes of scientific praxis that are associated with them .
The first stage, I called anthropic or projective. I take this to be the
'classical' stage of Greek science, in which the earlier modes of the
mythopoetic understanding and explanation of nature are replaced by
explanation in terms of ostensibly 'natural' categories, i.e., in terms of
physical elements, and depersonified natural processes. This mode of
scientific praxis I will characterize as a speculative construction, in which
modes of technological and social action are depersonified, negated,
abstracted, mediated - aufgehoben is the word - and reconstrued as
accounts of physical process. In this context, I will distinguish the anthro-
pomorphic from the anthropic, and suggest that this anthropic mode is
also mediated by an objectivist mode of explanation, which already
prepares the ground for the second stage. The second stage is the stage
of the classical realism of the scientific revolution - the objective mode
in which the physical world , as the object of knowledge, is construed
as an independent reality which can be known by passive contempla-
tion or observation, and in which the knower stands to the known as a
spectator. There is no interaction with the object, except as it stands in
causal relation to the knower, in perception, and as it has a rational or
mathematical structure which is correlative with human reason. Here, the
development of experimental science, as it involves the manipulation
or transformation of nature in the practice of empirical tests, creates a
mode of interaction, which poses difficulties for such a contemplative
realism, and points to the third stage. The third stage, which I have
characterized as interactive or constructive introduces the notion of the
scientific object as an artifact, produced by or transformed by the mode
of inquiry and hence, not only theory-dependent, but praxis-dependent,
in the preparation of the contexts of measurement. The conceptual
alternatives of realism and constructivism are seen here as results of
different modes of scientific practice, in the second and third stages,
respectively.
Pre-scientific mythopoetic understanding of nature is distinctively
anthropomorphic, in the explicit sense that natural forces, events,
processes are taken to be the activities of persons, or person-like entities
- gods, or spirits, or even, more abstractly, principles; and the actions
of these personified forces are understood in terms of the motives,

purposes, or effective responses of such beings . Anthropomorphic expla-
nations, in general, project human modes of production and social
interaction upon nature, reading nature in the image of human activity,
and empathetically exercising 'control' over the physical world, or
bending to its constraints and necessities in terms of models of human
making and doing.
The first stage of scientific practice is, in effect, a negation of
this anthropomorphism, but one which retains essential feature s of it
in a transformed way. I am thinking here of Greek science, from
the pre-Socratics through the classical modes of Parmenideanism,
Democriteanism, Platonism , Aristotelianism. The break with pre-scien-
tific modes includes the rejection of the personified nature-gods in
favor of more abstract principles: the natural 'qualities' or fundamental
elements, in motion or change : the wet, the hot, the dry, the cold, in com-
bination , or in just proportions; matter in motion , in its uncuttable form
as atoms; potencies realizing themselves ; natural kinds doing what comes
naturally, and behind all, form , structure , mathematical essences bodied
forth in the show of themselves that the senses perceive , but which reason
sees through, to the objective, hidden reality behind the appearances.
The Greek mode is complex , as we know. Along with the retention
of anthropisms, in their abstract, mediated form, there is also the dean-
thropic , objectivizing side of Greek thought.
This alternative, deanthropic aspect in Greek science, lies in the
notions of blind necessity, implacable and immune to human entreaty
or action, to be recognized as an otherness which stands beyond the power
of human agency. It also lies in the notions of truth and being as con-
stituting a reality which reason can attain to, and may contemplate, but
to which it is related in only a spectative, passive way, though by virtue
of a correspondence between the structures of reason and those of
being.
Thus, in Greek science and philosophy, there is a tension between
anthropic and deanthropic modes, between projective, sympathetic
models of a physical nature in the image of human action, and objec -
tive models of a transcendent reality which can only be contemplated,
but not acted on, or transformed.
The mode of speculation which reflects on human practices, and
retains their normativeness in its construction of the scientific object, thus
combines with a mode of 'pure' speculation, divorced from all activity
or intervention in nature, active instead as 'pure ' thought, theoria , con-

templating the universal forms or principles, the "archei" of things; or
observing nature, for the sake of classifying the many in some unified,
hierarchical order which will reveal the essential rational structures of
being.
The second stage constitutes the scientific object as an ontologically
independent, self-subsisting physical world, apprehended by the senses,
but known by thought. It is the deanthropic moment of the first stage
in its full realization and dominance. In it, the last obvious element of
anthropism is eliminated - i.e ., the notion of teleology - when
Aristotelian " final causes", and any notions of providential causality
are assigned either to supernatural miraculous intervention, or to the
asylum of ignorance. Determinism is, typically, taken to be an objec-
tive feature of the relations among things and things remain self-identical
through transformation. Space-time is the universal matr ix of the physical
world: Space is homogeneous and isotropic, each part like each other,
and the whole the sum of its parts, in a uniformly metrical extension.
Time flows equably. Mea surement is therefore dependable, and locally
indifferent: Objectivity is not the view from nowhere. It is the view
from anywhere and everywhere; The object of science is, classically,
impassive. It bears contemplation with no ill effect. Observation is
transparent: nothing muddies the view or distorts it, which cannot be
removed by correction or cancellation. Everything is clear and distinct;
or can be made so. The order and connection of things is the order and
connection of thoughts. The truths of nature are simply the explication
of these correspondences of thought to thing. We are in good old meta-
physical realism country, and the primary properties - number, shape,
size, position and time, tell all and hide nothing.
Now this is clearly an idyllic, untroubled version of Cartesian or
Galilean science, a movie version of the novel. Yet, it reveals some crucial
features . The scientific object is a space-time object and as such fully
accessible to measurement. Measurement, moreover, exhausts the account
of all of the properties of this scientific object, so it is wholly knowable,
in principle, though not wholly known. The theory describing it in these
terms is essentially complete, therefore. The formalism - ultimately,
differential equations - is congruent with the continuous magnitudes of
space and time , and leaves no pockets of discontinuity in the manifold,
due to gaps in the language, and what it can name.
Neither observation nor the activity of measurement disturbs the
216 MARX W . WARTOFSKY
scientific object in any way : it is, ontologically, independent of the

knower. Therefore, realism, in its classical form, is an accurate account
of this relation.
But isn't this just a realist version of physics? How then could realism
fail to be accurate , since its appropriateness is circular. Not quite. Take
the paradigmatic object of science of this scientific realism: it is, in the
first place, the heavens. Astronomy has, as its object, that with which we,
in practice, cannot interact. All we can do is look at the heavens from
a distance, contemplate them . They are the ideal object for a realist
epistemology, since measurement cannot disturb the stars , though dis-
turbances of measurement can take place, and need to be corrected.
This seems to be common-sensically incontrovertible. (At least in the
17th century, and before Goodman.)
In short, in the hopeful picture of stage two, characterized as dean -
thropic and objective, a robust realism would seem to be the epistemology
of choice. The paradigm case of astronomical observation and mea -
surement clearly leaves the object of scientific knowledge undisturbed.
This ideal case , is, on the one hand, the contemplative model of Greek
science, mathematized and experimentalized, and brought up to date.
In that sense , 17th century realism recapitulates the intellectualist model
of Greek thought, in which detached spectation without intervention is
the ideal of a gentlemanly science. Astronomy, mechanics, mathematics
are a clean game: "No shadows. Everything on the surface, everything
clear" as Neurath liked to say.
Steam, coal and urea are, by contrast, dirty games , as are fermenta-
tion, fertilization of the soil, syphilis, and surgery without anesthesia.
Thermodynamics, industrial chemistry, pharmacology are messy sciences,
in which the scientific object is not so clearly detached from the con-
ditions of its observation, as are the heavens. At the very least, conditions
of experimentation are such that the experimental situation is not there
for the contemplation; it has to be prepared. The experimenter rearranges
nature, tortures it, subjects it to transformations, in which its patterns,
its truths, its secrets are revealed. Yet, here too, it seems we may ideally
translate the messier phenomena into our clockwork heavenly model. But
there are shadows, no larger that a man's hand. And thus we turn (or
turn back) to stage three.
The issue between Einstein and Bohr, in their confrontation over the
completeness or incompleteness of quantum-mechanics, is the way in
which the difference between stage two and stage three develops, in actual
fact. Einstein is no simple realist, though he preserves the ideological

core of stage two thinking. Bohr holds out against the adequacy of stage
two , and for stage three as the fundamental model of the ontological
status of the object of physical science. But Bohr, in principle, accepts
Einstein 's classical realist requirements as holding for the macroworld.
For that level of physical reality, our classical picture is not only
adequate, but essential. As Arthur Fine points out, Bohr is the conser-
vative and Einstein the radical, in terms of willingness to change or
reformulate the classical physical world picture to be adequate to our
experience. But for Einstein, there is a deep intuition that the changes,
if any, should retain locality and separability, hence also causal deter-
minism, and that there should be continuity in these ontological criteria
across relativistic, classical and quantum mechanical frameworks. Bohr
is ready to opt for discontinuity, and that's what complementarity is all
about.
Now, all of this focuses on questions of measurement, the conditions
of measurement, and the ineliminability of the effects of measurement
upon the object of knowledge. The scientific object, therefore, is essen-
tially an interactive one , or one in which what is taken to be is projected
by the conditions of measurement. ' Taken to be' here is not to be read
as simply an act of mental representation, or as the way we happen to
understand something, in our reflection upon it. Rather, ' taken to be'
connotes an activity of construction, in which "taking the object to be
x" is an actual transformation of something, or a projection of some-
thing - a projective tran sformation then - into x. A measurement, at
the quantum level, requires a preparation: not simply the existence or
presence of a measurement apparatus, e.g. , a diaphragm, an emulsion,
a screen, a cloud chamber, a geiger counter which registers the quantum-
events, but also the generation of quantum-events under controlled
conditions, or under conditions of selection and screening, on the basis
of physical theory, e.g., the use of polarizers, of optical switches, of
slits; of linear accelerators, under the tutelage and direction of partic-
ular experimental hypotheses: Nature theorized, so to speak, but not
inside the head; rather in the objectified, externalized, constructed world
of experiment: the arena in which the scientific object comes into being,
emerging on the stage of scientific observation, to tryout the role written
for it in the scripts of the theorists, the mathematicians, the imaginers
of 'thought experiments' .
In this sense of construction, the scientific object is not a 'theoret-
ical construct ' as some conceptual entity, 'existing', so to speak , as the

'meaning' of the theory, or as some sort of dispositional mental entity,
e.g., what would be the intentional object of thought thinking the theory.
Instead, the scientific object is an actual experimental transformation
of what, for want of more neutral terms of reference, we may call the
natural world. In this sense, we may speak of changing the world by 'the-
orizing' it.
Department of Philosoph y,
Baruch College & the Graduate Center,
The City University of New York,
New York, U.S.A.
JOHN WATKINS
SCIENTIFIC REALISM VERSUS

COMMON-SENSE REALISM?
Adopting or rejecting scientific realism may make a big difference to

one's world-view, and so may adopting or rejecting common-sense
realism; it therefore becomes an important question whether these two
kinds of realism are mutually exclusive, as has often been supposed,
so that one cannot adopt them both.
What do they respectively assert? Scientific realism is primarily con-
cerned with the invisible infrastructure of things. Let a generally accepted
scientific theory say that the sun is largely composed of hydrogen atoms,
and exerts a gravitational pull on the earth and the other planets that keeps
them in orbit round it. Scientific realism says that if that theory is true ,
then the sun is largely composed of such atoms and does exert a grav-
itational pull. Now common-sense realism, as I understand it, does not
deny that the sun has an internal structure, though it leaves it to science
to discover what this is, rather as it leaves it to science to estimate the
sun's size and distance from the earth. But common-sense realism does
insist that the macroscopic sun which we perceive exists independently
of our perceptions of it: all perceivers of it might perish without it per-
ishing. It is a general presumption of common-sense realism that where
there are intersubjectively shared perceptual experiences of a macro-
object, these experiences constitute perceptions of something "out there",
something which exists independently of these perceptions of it. In the
case of rainbows and mirages it has been obliged to withdraw that pre-
sumption; but it stands by it with respect to mountains, buildings,
elephants, oranges, and beetles.
That common-sense realism should be repudiated in favour of sci-
entific realism, is asserted by the doctrine of the Omnicompetence of
Physics. This doctrine, which has sometimes sailed under the banner
of 'Unified Science', I claims that all meaningful questions about matters
of fact are in principle answerable by a completed physics; if there are
questions which appear to be unanswerable by physics this is only
because physics is as yet far from being completed. It is a doctrine which
draws a large cheque on the future. When philosophers support their
metaphysical predilections with forecasts about future scientific devel-
219
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.) , Realism and Anti -Realism in the
220 JOHN WATKINS
opments, they engage in a sort of circularity, since their forecasts are

obviously inspired by their metaphysical predilections. But this appeal
to future developments renders the present doctrine irrefutable; its
defenders can always say of a question apparently unanswerable by
physics that it is so only because physics is not yet complete. Nevertheless
I will now mention some meaningful, factual questions, which it seems
very doubtful that physics could ever answer.
Question I: What divides the past from the future? The obvious answer
is the present, or the Now as Einstein called it. Is this something that
physics will ever comprehend? In his ' Intellectual Autobiography' Camap
reported some conversations with Einstein, in one of which
Einstein said that the problem of the Now worried him seriously . He explained that the
experience of the Now means something special for man, something essentially different
from the past and the future, but that this important difference does not and cannot occur
within physics. That this experience cannot be grasped by science seemed to him a
matter of painful but inevitable resignation. .. . [T[here is someth ing essential about
the Now which is just outside the realm of science' (1963, p. 38).
Carnap disagreed ; although he had not pressed the point against Einstein
at the time, he now commented: "Since science in principle can say all
that can be said, there is no unanswerable question left." But that is
merely to summon up the Omnicompetence doctrine to sweep away
this counter-example to it. I think that Einstein was right. For each of
us, at any moment in our waking lives, nothing could be more real than
the now, or specious present. But how could science ever capture this
elusive and shifting datum? Science can locate events, to employ
McTaggart 's distinction, only on the B-series, with its earlier-later
relation . If it ties to to a momentary present, to is then fixed, and a moment
later has ceased to be now.
Question 2: How did consciousness arise? On this issue I like the
following exchange between two distinguished evolutionists. Sewall
Wright: "Emergence of mind from no mind at all is sheer magic" (1964,
p. 278); Dobzhansky: "If this is 'sheer magic', it is a kind of magic
the world is full of" (1967, p. 31). I think that both men were right, if
by 'magic ' is meant 'unexplainable in principle by physics '; there is
something magical about the trick so effortlessly performed each time
a newborn infant grows into a normal child with a mind of its own;
we certainly seem to have emergence of mind from no mind here . The
conviction that that is impossible has driven some tough-minded thinkers
to very strange conclusions. Ernst Haeckel, the nineteenth century evo-
SCIENTIFIC REALISM VERSU S COMMON-SENSE REALISM? 221
lutionist who did for Darwin in Germany very much what T. H. Huxley
did for him in England, had a mind that was usually as tough as old boots ;
but it turned pretty soft in a chapter in his 1899 on the 'embryology of
the soul'; he there announced that when a sperm and an ovum come
together at conception, they each bring with them their own "cell-soul";
these two mini-souls then coalesce into one. I regard this as doubly
absurd, first because gametes surely have no mentality, and second
because I hold it to be a fundamental feature of the mental that separate
centres of mentality cannot coalesce. But Haeckel had not yet gone far
enough: these gametes would have needed to derive their soulfulness
from an earlier source if soulfulness was not to have emerged ex nihilo
with them . So Haeckel added that unicellular protozoa, the earliest and
most rudimentary form of life, have a simple cell-soul (p. 53). But these
protozoa would in their turn have needed to have acquired their soul-
fulness from a still earlier source if soulfulness was not to have emerged
ex nihilo with them. Once again Haeckel obliged: 'even the atom is
not without a rudimentary form of sensation and will' (p. 80) . William
James had gone straight to the conclusion to which one is driven if
one admits consciousness and denies emergence: "If evolution is to
work smoothly, con sciou sness in some shape must have been present
at the very origin of things" (1890, p. 152, italicized in the original).
I am glad to say that James repudiated this panpsychist conclusion
(p. 164).
Question 3: How did humour arise? So far as I know, only human
beings (and by no means all of them) possess humour. Some animals
possess rough equivalents of those of our smiles and laughter that go with
high spirits and playfulness (see Darwin, 1872, pp. 196f); but it seems
that they all lack even a rudimentary or incipient form of anything like
our chuckling over something funny . Will physics ever explain why there
are jokes in our lives but not in theirs? It occurred to me that when
Darwin was writing The Descent ofMan he would have been glad to find
evidence of humour in other animals, since it was a main thesis of that
work that 'the various emotions and faculties, such as love, memory,
attention, curiosity, imitation, reason, & c., of which man boasts, may
be found in an incipient, or even sometimes in a well-developed con-
dition, in the lower animals' (p. 193), and he would surely have liked
to add humour to the list. The second (1874) edition has a very thorough
index, running to 77 pages, and I have combed it carefully, There are
no entries for ' jokes' , 'laughter' , or 'wit'; the only entry I could find
222 JOHN WATKINS
that seemed relevant is the following: 'Humour, sense of, in dogs, 743' .
I turned eagerly to page 743, but Darwin is there dealing with the bril-
liant plumage of certain male birds, and there is nothing about either dogs
or humour. So I turned to the long entry under 'Dogs' . Among its many
sub-entries are ones on dog's reasoning faculties, moral qualities, and
possession of conscience, but none on their sense of humour. I began
to wonder whether he had inserted that page 743 entry as a joke.
Question 4: How did physics arise? I will here take it for granted
that its growth has involved intellectual creativity and was not just an
inductive accumulation; so we could frame question 4 more generally:
will science ever explain how some particular product of human cre-
ativity, whether in science or art, came to exist? Kant had a view about
creativity which seems right and which implies that science cannot
explain such comings-to-be. Let x be the occurrence of some unprece-
dented event. Thus x could be Michelangelo painting the creation of
Adam; or it could be the first explosion of an atom bomb. A scientific
explanation of x, if there is one, consists of laws of nature in conjunc-
tion with a set (perhaps a very large set) of initial conditions. If true, such
an explanation shows both how x was in fact produced and, at least in
principle, how it could have been produced at an earlier (or later) time;
for it implies that x occurs whenever such a set of initial conditions is
assembled. In short, a scientific explanation of x would provide a recipe
for the production of x.2 It might have been impractic able to assemble
such a set of initial conditions at any earlier time; but a recipe for a
cake doesn't stop being a recipe just because some of the ingredients it
calls for are presently unobtainable.
So a scientific explanation of how Michelangelo's picture of the
creation of Adam came into existence would be a recipe (perhaps a
very cumbersome one), not for duplicating the picture subsequently which
would be philosophically uninteresting, but for its first creation. Now
Kant insisted that there can be no such recipe for a product of genius;
that is something 'for which no definite rule can be given' (1790, §46).
I think he was right.
We cannot refute the Omnicompetence of Physics doctrine, but we can
and should put a large question-mark over it. But that still leaves open
the question of the reconcilability of science and common sense.
Let us hark back for a moment to the seventeenth century, when a
serious split developed in the realist camp. It seemed to many leading
thinkers on both sides of the divide that there is an irreconcilable conflict
SCIENTIFIC REALISM VERSUS COMMON-SENSE REALISM? 223
between science and common sense: adopting the new science meant
repudiating one's common-sense convictions even where these are not
under specific challenge from science (in the way in which, for instance,
the geocentric hypothesis was). Nearly all those in the vanguard of the
scientific revolution, such as Galileo and Boyle, together with their philo-
sophical allies, such as Descartes and Locke, believed that the new
science had overthrown the common-sense world-view: must not the
world be essentially different from what it appears to be if it consists,
in reality, of hard, massy, odourless, and colourless corpuscles, obeying
only the laws of mechanics? Its apparent colours, sounds and smells must
be subjective qualities projected by us onto external objects. These men,
Berkeley declared, "lead us to think all the visible beauty of the creation
a false imaginary glare" (1713, p. 211). He sought to turn the tables by
arguing that it is not the world but science that is not what it appears
to be; it appears to give us information about realities behind the phe-
nomena; but since there is nothing behind the phenomena for it to get
hold of, the most science can do is to give us rules for correlating and
predicting phenomena.'
I say that what generated this conflict was a wrongful assumption
shared by both sides, namely that what is real is all at one level; at the
bottommost level, according to the scientific realists, and at the surface,
according to Berkeley (for whom the only other realities were souls
and God, which are not science's business). Remove this assumption,
allow that reality is multi-leveled, and the conflict dissolves: quarks,
atoms and oranges may coexist.
The thesis that only entities at the bottommost level are real had a
certain plausibility so long as classical atomism survived intact; it
declared any complex thing to be, at bottom, just a swarm of particles.
But classical atomism has of course been swept aside by modern physics;
as Patrick Suppes put it, "it is not swarms of particles that things are
made of, but particles that are made of swarms" (1984, p. 122). Popper
had previously made a similar point with clouds and clocks (1972, chap.
6): in the heyday of classical mechanics, with its atomism and deter-
minism, seemingly "cloudy" things could be assumed to be made up
of bits that behave like clockwork; but now, in the aftermath of quantum
theory, it seems to be the other way round, with seemingly clocklike
things really being statistical aggregates of "cloudy" bits. Suppes con-
tinued: 'we cannot have a reduction of subject matter to the ultimate
physical entities because we do not know what those entities are' (p. 123).
224 JOHN WATKINS
It is not just that we do not at the present time know what they are;
the assumption that science will ever reach a bottommost level, or even
that there is such a level for it to fail to reach, is in doubt. As John A.
Wheeler put it: "One therefore suspects that it is wrong to think that as
one penetrates deeper and deeper into the structure of physics he will
find it terminating at some nth level" (1977, pp. 4-5).
So a scientific realist who restricts the title 'real' to the ultimate
components of matter would have to admit that nothing so far postu-
lated by science is real. If we shy away from that, and allow that atoms
are real despite having graduated to a non-ultimate level, then we must
surely allow that molecules are real; and so we can go on working up
towards the macro-level. And why should we not eventually allow that
diamonds are real? They are, after all, a lot more stable than, say, radon
atoms. I say that just as common-sense realism is under no internal
compulsion to deny that things have an invisible infrastructure, so sci-
entific realism is under no internal compulsion to deny the reality of
macro-objects.
It is a mistake to suppose that a scient ific explanation of properties
of things at one level , in terms of the properties and relations of things
at a deeper level, explains away the former properties, or reduces them
to the deeper ones . Bohr's model of the hydrogen atom, with its electron
circling a positively charged nucleus, explained something not explained
by Rutherford's model , namely the stability of the atom. There was not
the slightest tendency for its stability to be thereby explained away
(though a deep explanation may have revisionary implications for its
explanandum; thus Bohr's model did suggest that an atom is not, after
all, absolutely stable , but might conceivably be split). The account that
nuclear physics gives about what happens at the micro-level when an
"atom bomb" explodes has no tendency at all to deprive the macro-
event of its reality. Likewise, a social science explanation of a social
phenomenon, such as inflation, as the unintended outcome of the deci-
sions and activities of individual people has no tendency at all to deprive
the social phenomenon of its reality. Reality is multi-leveled.
A plausible but invalid argument from scientific realism to the
unreality of macro-objects was put forward by the late Grover Maxwell,
in his 1968: - You are looking at an orange, say; then please make an
inventory, in the light of the best available scientific knowledge, of the
kinds of entity involved in the causal processes that issue in your orange-
like perceptions. Such an inventory will include all sorts of things like
SCIENTIFIC REALISM VERSUS COMMON-SENSE REALISM ? 225
photons, neurons, atoms, and so on, but no orange : 'the stimulation of

the retina is accomplished solely by photons emitted from the atoms com-
prising the surface of the material object in question. The material object
just is this collection of submicroscopic particles and the relations that
subsist among them' (p. 151); in seeing an "orange". then, you are seeing
something that isn't really there, an illusion generated by the interplay
of submicroscopic particles which you don't see and are really there .
This argument fails to take account of the fact that animals ' brains
decode the input from their sense-organs. My television set has often pre-
sented me with images of Mrs Thatcher. An inventory of the impulses
reaching the TV aerial, and of the processes within the set, will include
photons and electrons, but no Mrs Thatcher, Does that mean that nothing
real corresponded to these Thatcherite images? No; a television set
decodes signals encoded at the transmitting end, and these images cor-
responds pretty well (apart of course from their two-dimensionality) with
certain realities at the other end. Now consider a dog sniffing a piece
of meat. The meat is giving off signals; not deliberately encoded signals,
to be sure, but signals that the dog 's brain is able, as a result of natural
selection, to decode. The dog turns away , because the meat smells bad.
And this corresponds pretty well with certain realities at the other end;
the meat is bad .
Even so , it will be said, I must admit that colours are subjective
qualities with which we daub physical objects. In my youth I would have
agreed with this, but now I am not so sure . Suppose that you are told
that there is something, call it x, which usually appears to be homoge-
neous; however, when x is subjected to a certain treatment, call it PI'
it is decomposed into components a, b, c. . .. ; and when a, b, c. . . .
are subjected to a certain treatment, call it Pz, x is reconstituted, this whole
process occurring independently of perceivers. Would you not conclude
that a, b, c, . . . are not mind-dependent but objective? But this account
fits an experiment of Newton's where x is sunlight, PI and P z are suitably
positioned prisms, and a, b, c, . . . are the colours of the spectrum.
The ontology of common-sense is revised and enriched, but not
abolished, by science.
Centre for the Philosophy of the Natural and Social Sciences,

The London School of Economics ,
London, U.K.
226 JOHN WATKINS
NOTES
I Thus Otto Neurath included physicalism and unified science in the so-called 'scien-
tific conception of the world ' (1973, p. 417).
2 See Briskman 1980.
3 See Popper, 1963, chaps . 3 and 6.
BIBLIOGRAPHY
Berkeley, George (1713). 'Three Dialogues Between Hylas and Philonous', in A. A.

Luce and T. E. Jessop (eds.), The Works of George Berkeley, 9 vols, 1948-57, Nelson,
London.
Briskman, Larry (1980). 'Creative Product and Creative Process in Science and Art ',
Inquiry 23, pp. 83-106.
Carnap, Rudolf (1963). 'Intellectual Autobiography', in Schilpp (ed .), The Philosophy
of Rudolf Carnap (The Library of Living Philosophers), Open Court, La Salle, pp.
3-84.
Darwin, Charles (1871). The Descent of Man and Selection to Sex, Murray, London,
first ed. 1871, second ed. 1874.
Darwin, Charles (1872) . The Expression of the Emotions in Man and Animals, 1872;
reprinted University Press, Chicago, 1965.
Dobzhansky, Theodosius (1967) . The Biology of Ultimate Concern, New American
Library, New York.
Haeckel, Ernst (1899) . The Riddle of the Universe, trans. by Joseph McCabe, Watts,
London, 1904.
Kant, Immanuel (1790) . Critique of Judgem ent , first ed .; trans. by J . C. Meredith,
Clarendon Press, Oxford, 1928.
Max well, Grover (1968). ' Scientific Methodology and the Causal Theory of Perception ',
in Lakatos, Imre and Musgrave, Alan (eds .), Problems in the Philosophy of Science,
North-Holland, Amsterdam, pp. 148-160.
Neurath, Otto (1973). Otto Neurath: Empiricism and Sociology, Robert S. Cohen and Marie
Neurath (eds .), Vienna Circle Collection, vol. I, Reidel, Dordrecht.
Popper, Karl R. (1963). Conjectures and Refutations, Routledge & Kegan Paul, London.
Popper, Karl R. (1972). Objective Knowledge: An Evolutionary Approach, Clarendon
Press, Oxford.
Suppes, Patrick (1984). Probabilistic Metaphysics, Blackwell, Oxford.
Wheeler, John Archibald (1977). 'Genesis and Observership', in Butts, Robert E. and
Hintikka, Jaakko (eds .), Foundational Problems in the Special Sciences, Reidel,
Dordrecht, pp. 3-33.
Wright, Sewall (1964). 'Biology and the Philosophy of Science ', The Monist 48, pp.
265-290.
JOHN WRIGHT
METAPHYSICAL REALISM AND THE

EXPLANATION OF THE SUCCESS OF SCIENCE
One argument for Scientific Realism is from the success of science.

But it is not clear whether this argument supports Metaphysical Realism
with respect to the entities of science, or merely some weaker form
of Realism such as Hilary Putnam's 'Internal Realism'.' In this paper
it will be argued that if the argument supports any form of Realism, it
supports Metaphysical Realism . More specifically, it will be argued that
if some notion of truth is needed to explain the success of science, then
that notion of truth must be a non-epistemic, Realist notion .
We can , on the face of it, draw a distinction between the strong
predictive success of a theory and weak predictive success. A theory
is strongly predictively successful if it successfully predicts novel
regularities - that is, if it successfully predicts regularities that were
not used in actually formulating the theory. If a theory only succeeds
in predicting more instances of regularities which were used in its
formulation , then it is only weakly predictively successful.' The main
claim argued for in this paper is that strong predictive success cannot
be explained by an epistemic or verificationist conception of truth.
This form of success can only, I think, be explained by a Realist
conception.
1. SOME QUALIFICATIONS TO THE ARGUMENT
In this paper it will be assumed that in order to explain the strong pre-
dictive success of a theory, we need to say that at least a part of a
theory is true, or close to the truth in some sense of truth.' However, three
qualifications to this claim need to be strongly emphasised.
(a) It is not claimed, and neither is it for the argument presented here
necessary to assume, that if a theory is strongly predictively suc-
cessful then all parts of it are true . It only needs to be assumed
that certain parts, very directly involved in the derivation of the
prediction, are true. 4
(b) Strictly speaking, it is not even necessary to assume that the parts
of a theory directly involved in the derivation of the prediction are
227
R. S. Cohen . R. Hilpinen and Qiu Renzong (eds .), Realism and Anti-Realism in the
Philosophy of Science. 227-243.
228 JOHN WRIGHT
true tout court. It can be sufficient to say that they are, in some sense,
'close to the truth'."
(c) Even the claim that some part of a strongly successful theory is close
to the truth is not claimed to be certain. It is merely claimed it can
be something that is reasonable to believe, but which may be refuted
by future experimental results.
Although I will not argue for the claim that some notion of truth is
needed to explain the success of science, one possible objection to this
claim will be considered. This is what may be termed ' the redundancy
objection' . It is that in order to explain the success of a theory it is not
necessary to state that the theory is true, it is enough merely to state
the theory. For example, in order to explain the success of the theory
that electrons have negative charge it is not necessary to state that the
theory is true; it is sufficient to simply say: ' Electrons have negative
charge'. Call this the redundancy explanation of the theory's success.
Even if it is granted that some examples of success can be given a
redundancy explanation it does not follow that this type of explanation
is adequate for all phenomena connected with the success of science.
Some phenomena cannot , it will be argued, be given a redundancy expla-
nation. Suppose, for example, that scientists are observed to manipulate
some ink marks on paper in a particular way, and at the end of this
process they produce the observational prediction, 'There will be a yellow
spark in region R'. Let us assume that this is a 'novel' observational
prediction. Subsequently, the scientists do indeed observe a yellow spark
in region R. We will call this phenomenon - that is, the phenomenon
of the manipulation of ink-marks yielding a novel prediction that is
subsequently verified - phenomenon S. If the novel observational pre-
diction is a surprising one , then phenomenon S will surely require
explanation. But it needs to be emphasised that to say that phenom-
enon S requires explanation is not to say that it is tokens of sentences
that are the primary bearers of truth. To admit that a link between the
manipulation of tokens and success needs explanation is compatible
with saying that the bearers of truth are, for example , propositions. One
possible explanation of phenomenon S is that the ink symbols manipu-
lated by the scientists expressed true propositions. (Perhaps some of them
expressed the true proposition that electrons have negative charge .) This
enabled the scientists to derive the true observational prediction that there
will be a yellow spark in region R. Here I will not be concerned to
argue either that this is the only, or even the best, explanation of the
MET APHYSICAL REALISM 229
scientists' success . But let us ask, 'Can it be given a redundancy expla-

nation?' Clearly, it is no explanation of phenomenon S to merely say
'Electrons have negative charge ' . That does not explain why the scien-
tists who manipulated those ink marks made a successful prediction.
It might be suggested that a redundancy explanation would be adequate
if that which required explanation were re-expressed: 'Why were the
scientists who manipulated symbols that expressed the proposition
'Electrons have negative charge' successful? It may appear that a suf-
ficient explanation of this re-expression of the explanandum is that
electrons have negative charge . However, this line of thought does not
show 'true' to be redundant in the explanation. Suppose that drugs A and
B are both necessary, and jointly sufficient to cure a particular disease .
One explanation of why patient P was cured of the disease is 'Because
he was given drugs A and B' . Now, suppose we change the explanandum
to ' Why was P, who was given drug A, cured?' It might be thought it
would be sufficient to say ' Because he was given drug B' . But this is
not sufficient: we still need to add, as a part of the explanation, that P
was given drug A. After all , the fact that P had been given drug A
might have been included in the explanandum as an irrelevant bit of extra
information. So it is not otiose to include it in the explanans, since
doing so tells that it is an explanatorily relevant factor. Its mere inclu -
sion in the explanandum does not tell us that.
Now, suppose the explanandum is: 'The use by scientists of tokens
that expressed the proposition that electrons have negative charge led
to success' . Is a sufficient explanation of this: 'Because electrons have
negative charge ' ? The considerations of the above paragraph indicate that
it would not be sufficient if the fact that the tokens expressed the propo-
sition that electrons have negative charge is relevant in leading to the
success. And it surely is relevant. If the tokens used by the scientists
had expressed something different - if, for example, they had referred
to protons instead of electrons or they had made claims about electrons
that were false - then they would have yielded observational predic-
tions which we have no reason to suppose would have been confirmed
by tests. The semantic properties of the tokens are relevant in producing
the empirical success that follows the manipulation of these tokens.
Consequently, the redundancy objection - that the semantic properties
of a theory need not be mentioned to explain the success of the theory
- fails if what is to be explained is the success that follows from the
manipulation of tokens.
230 JOHN WRIGHT
Of course, to say that the redundancy objection fails is not to say

anything about which semantic notion is needed to explain the success
of science. In particular, it is not to say whether it is a Realist, or a
verificationist conception of truth (or truthlikeness) that is needed.
There is another objection to the general enterprise of arguing for a
realist conception of truth on the grounds that only it can explain the
success of science. This is the objection that the strong predictive success
of science does not require explanation. The claim is that since so many
scientific theories are advanced, it would rather be surprising if some
of them did not, simply through lucky chance, turn out to be strongly
predictively successful. But there are, I think, two reasons why this objec-
tion is unconvincing.
(a) Some types of strong predictive success simply have too many
positive instances to be merely due to lucky chance. For example,
scientists are able to correctly predict the ways in which artificial,
trans-uranium elements will chemically combine with thousands of other
substances. It is beyond the bounds of plausibility that such a large
number of successes could simply be a fluke .
(b) If on a particular occasion, a novel prediction made by a theory
is confirmed, scientists will re-test the theory to ensure that the apparent
success is not an anomalous result but is a repeatable effect. So what
requires explanation is that the ability of a theory to correct predict a
novel observation is a repeatable effect. Establishing that a result is a
repeatable effect makes it less likely that it is merely due to chance.
B. van Fraassen has offered an evolutionary explanation of the success
of scientific theories." Van Fraassen says that a theory will be retained
in the body of accepted science only if it successfully predicts obser-
vat ions . If it is not succes sful it is removed. As a result of this process
science has come to contain only successful theories. However, it is clear
that van Fraassen's suggestion at most only provide s an explanation of
weak predictive success. It does not explain how a theory correctly pre -
dicted observations that were only made after the theory was formulated,
that is, it provides no explanation of strong predictive success.
2. IS AN EPISTEMIC NOTION OF TRUTH ABLE

TO EXPLAIN THE SUCCESS OF SCIENCE ?
Broadly speaking, there are, I think , three epistemic notions of truth . I

will call these Strict Verificationism, Liberal Verificationism and the
Epistemic Ideality Conception of Truth. Roughly, Strict Verificationism
METAPHYSICAL REALISM 231
asserts that a theory is true iff it can be verified by the techniques we

currently possess. Liberal Verificationism says that a theory is true iff
we have , or could acquire, the ability to verify it. The Epistemic Ideality
Conception of Truth says that a theory is true iff it is a part of an epis-
temically ideal theory . I do not claim that these three notions exhaust
all the conceptions of epistemic truth that have ever been advanced by
anyone . But I do think they represent currently popular conceptions.
2(a). Strict Verificationism and the Explanation

of Strong Predict ive Success
A more careful statement of Strict Verificationism is:
A sentence S, as uttered by a speaker U at time t, expresses

a true proposition at time t iff it is possible for U to estab-
lish its truth using only techniques available at t. 7
Note that a Strict Verificationist need not say that if we acquire new
techniques of verification then some propositions may change their truth-
value . He/she can say that, with the acquis ition of such new techniques,
a previously unverifiable sentence may come to express a new propo-
sition . So, although the sentence 'Diamonds and coal are made of the
same substance ' was, perhaps, once unverifiable, the Strict Verificationist
can say that with new techniques of verification that sentence came to
express a different proposition that was both verifiable and true . No
propos ition changed its truth-value .
There are at least some instances of the strong predictive success of
science that cannot be explained by Strict Verificationism. Let us say that
M is a theoretical property. At some time , say 1950, M was only
detectable by a particular range of tests t l . . . ~. Now, suppose T is a
theory of the nature of M that makes the novel prediction that M is
reliably indicated by phenomenon en' Subsequently, in 1951 (say), this
novel prediction is verified.
In order for the argument against Strict Verificationism to go through
we only need to say that the part of T which asserted
(1) M is reliably indicated by en
expressed a true proposition prior to its verification. What will be argued

is that the sentence (I) needed to be true in a sense not available to the
Strict Verificationist.
232 JOHN WRIGHT
Note that (1) does not refer to unobservables, or to anything else not
very directly involved in producing T's success .
Now, let us consider what an explanation of T's strong predictive
success would look like. We are assuming that strong predictive success
is to be explained by saying that some part of the successful theory T
is true. Let us refer to this part of T as T D. Then the explanation of T 's
success will consist in the derivation of (1) from T D.
Now, if we are to explain the strong predictive success of T by saying
that the special part T D is true, then it must be asserted that T D was already
expressing a true proposition prior to the verification of (1) in 1951. If
T D only came to express a true proposition as a result of the observa-
tions confirming (1), then its truth could not be used to explain those
observations. But if T D was expressing a true proposition prior to 1951,
then so must (1) have been expressing a true proposition, since it is
entailed by T D. As can be seen, these considerations do not require us
to make any specific assumptions about what part of T is true; they do
not require us to say anything about T D. The only specific sentence that
needs to be assumed to be true is (1) .
Now, suppose that an object X was found in 1950 which was known
to cause en, but which could not be verified by any means available in
1950 to have M. From 'X has en' and the claim that (1) expressed a
true proposition in 1950 it logically follows that
(2) X has M
expressed a true proposition in 1950. But (2) was not verifiable at that
time. So, the Strict Verificationist cannot say it expressed a true propo-
sition at that time.
The two claims required to derive the claim that (2) expressed a true
proposition in 1950 are that (1) expressed a true proposition in 1950
and that 'X has en' expressed a true proposition in 1950. Since the
advocate of Strict Verificationism is unable to say (2) expressed a true
proposition, he must reject one of the two statements that together imply
it. But, we are assuming, 'X has en' was known by observation to be true.
Therefore, what the advocate of Strict Verificationism must deny is that
in 1950 (1) expressed a true proposition. But this claim must be asserted
if the strong predictive success of T is to be explained. So, there are at
least some cases of strong predictive success that cannot he explained
by the Strict Verificationist.
Since the above argument against Strict Verificationism is rather
intricate, it is worth giving a briefer, more intuitive account of it. If the

success of a theory T in correctly predicting some novel phenomena is
to be explained, it needs to be asserted that at least some part of Twas
true prior to the observation of the novel phenomena. In the above
example, it was argued that the sentence 'Property M is reliably indicated
by en' is such a sentence. But if this was true prior to its verification,
it is easy to imagine situations in which there is some object X that
has M (since it has en) but cannot be verified to have it by available
techniques. Since the Strict Verificationist cannot allow this , he/she
cannot say that 'Property M is reliably indicated by en' was true. But
since the truth of this sentence is required to explain the strong success
of T, in this case the advocate of Strict Verificationism is rendered unable
to explain T's success.
2(b) . Liberal Verificationism and the Explanation

of Strong Predictive Success
The argument of the previous section attacks only the strict form of
verificationism which asserts that a statement made at a particular
time is true iff its truth can be established by the techniques of verifi-
cation available at that time . But, it may be objected, an anti-Realist need
not accept this. Although the techniques for verifying 'X has M' did
not, by hypothesis, exist in 1950, it would be easy enough for a scien-
tist working at that time to acquire the ability to verify it. All he has
to do is to confirm inductively that whenever an object has been estab-
lished by other methods to have M, it has the capacity to produce en'
Once the scientist has inductively confirmed this , he will be justified
in making the inference from 'X causes en' to the conclusion 'X has
M'. And perhaps the anti-Realist will say that in order for a sentence
to be true at time t, it is sufficient that a scientist at that time be able
to acquire the ability to verify it. We will call such a doctrine 'Liberal
Verificationism' .
The difficulty with Strict Verificationism was that it implied that
certain sentences, the truth of which was required to explain the strong
predictive success of science, were not true. But Liberal Verificationism
is confronted with a different difficulty. The Liberal Verificationist is able
to say that the required sentences are true, but (it will be argued) this
notion of truth does not give us an explanation of strong predictive
success. Roughly, the difficulty is that the Liberal Verificationist's account
234 JOHN WRIGHT
of what it is for a strongly predictively successful theory to be true simply

gives back to us the very thing that it is meant to explain.
A theory is verifiable, on the Liberal Verificationist's account, in virtue
of the fact that a user of the theory has, or could acquire, techniques
that would enable him to verify it. Let T be a theory that is not verifi -
able using techniques currently available but which is verifiable in the
Liberal Verificationist's sense. More specifically, assume that we could
acquire some experimental technique E, which would yield observa-
tions 0, and the obtaining of these results would verify T. Now, since
E is a technique we do not currently possess, E -7 0 is a novel obser-
vational regularity. And if E -7 0 verifies T, then it is a novel regularity
which is predicted (perhaps in conjunction with other theories) by T.
Since T is (we are assuming) verifiable in the Liberal Verificationist's
sense the novel observational regularity E -7 0 would in fact be obtained .
And I am assuming that the best explanation of this is that T - or at
least the part To of T directly responsible for the derivation of 'E -7
0' - is true. But now, on the Liberal Verificationist's account of truth,
T is true in virtue of the fact that it is possible for us to acquire the
techniques that would verify it. In particular, T is true in virtue of the
fact that we could acquire technique E, and that if we applied E, we
would obtain O. That is, the Liberal Verificationist's account of that in
virtue of which T is true mentions as one of its conjuncts the very thing
that is to be explained by saying that T, or some part of it, is true. But
if the truth of T, or a part of it, is to explain certain results, then it
cannot be the case that it is true in virtue of the fact that those results
would be obtained. This is merely an instance of the general principle
that if some state of affairs S holds in virtue of condition C, then the con-
dition C cannot be explained by state of affairs S. This can be illustrated
by a few simple examples . There is a disease of horses called Glanders.
A horse with Glanders exhibits certain symptoms such as swelling, the
production of mucus, etc. But we cannot say that the horse has Glanders
in virtue of exhibiting these symptoms, since there are a number of
diseases which could be associated with those symptoms. Rather, the
horse has Glanders in virtue of being infected with the Actinobacillus
Pfeifferella mallei. But now, suppose we want an explanation of why
the horse is infected with Pfeifferella mallei. It is clearly no explana-
tion to say 'because the horse has Glanders '. Rather, what is wanted is
an account of how the horse came to be infected. This account may be
because the horse drank infected water, or because the horse came into
contact with another infected horse . Or consider another example. We

can say that a uten sil is a spatula rather than a knife in virtue of having
a broad flat blade with neither edge sharper than the other. But it clearly
does not explain why neither edge is sharper than the other to say it is
a spatula.
There is a simple conceptual reason why the condition C in virtue
of which some state of affairs S holds cannot be explained by S. To
cite some condition C in virtue of which S holds is to say what it is
for S to hold . We can perhaps say that C constitutes S. But to explain
C is to mention some other event or state of affairs to which C is related.
Usually, this relation is a causal relation: If C* explains C, then C* is
- at least very often - a cause of C. But in general the explanation of
why C holds must refer to some other state of affairs or event that features
in an account of how C came to be. If we are simply given the state of
affairs that C constitutes, then C remains unexplained. The difficulty with
Liberal Verificationism is that it says our theories are true in virtue of
the fact that they would be confirmed. But then the Liberal Verificationist
cannot explain this confirmation by saying our theories are true.
In the previous section it was argued that Strict Verificationism cannot
explain strong predictive success because it is incompatible with the truth
of certain sentences required for that explanation. But we can now
observe that there is another difficulty that Strict Verificationism shares
with Liberal Verificationism. If Strict Verificationism says that a theory
is true in virtue of being verifiable by currently existing techniques,
then it cannot explain the empirical success of the theory by saying it
is true.
It may be objected that the argument given above against Liberal
Verificationism relies on the assumption that the strong predictive success
of a theory is to be explained by saying that the theory is true, whereas
all that may be needed is the assumption that the theory is in some
sense close to the truth. However, it is hard to see how this could aid
an advocate of an epistemic conception of truth. If truth is defined in
epistemic terms, then so must the notion of closeness-to-the-truth." Hence,
a verificationist would explicate the claim that a theory is close to the
truth in terms of our ability to (perhaps partially) verify the theory . But
then the claim that the theory is close to the truth would fail to explain
its subsequent confirmation for exactly the same reason as would the
claim that the theory is true (in the verificationist sense of ' true ' ).
236 JOHN WRIGHT
2(c). Truth as the "Epistemically Ideal" and the Explanation

of Strong Predicative Success
In a number of places, Hilary Putnam has argued that if a theory is

epistemically ideal, it could not fail to be true. According to Putnam, a
theory is epistemically ideal if it is 'complete, consistent predict[s]
correctly all observation sentences (as far as we can tell), meet[s]
whatever "operational constraints" there are, [and is] "beautiful",
"simple", "plausible", etc.' ? In this section it will be argued that we do
not explain the strong predictive success of a theory by saying that it
is epistemically ideal.
To explain why a theory T is strongly predictively successful is to
explain why it was able to predict successfully some regularity N,
instances of which were not used in its formulation. On the face of it,
this is not explained by saying that T is epistemically ideal. The only part
of what it is for a theory to be epistemically ideal that might have a
role in explaining this is that an epistemically ideal theory correctly
predicts all observation sentences. So the proposed 'explanation ' would
amount to: 'T successfully predicted the novel regularity N because T
correctly predicts all observation sentences'. Intuitively, that does not
seem to be an explanation of T 's predictive success: It seems merely
to give back to us that which requires explanation.
Against this, it may be objected that the proposed explanation is
legitimate since it can be represented in deductive-nomological form.
(i): For any theory, t, if t is true, then a novel prediction made

by t will be successful.
(ii): T is a true theory that makes novel prediction N.
So: T's prediction of N will be successful.
If truth is interpreted as epistemic ideality, then the generalisation

(i) will certainly be true. And (i) and (ii) together deductively entail
that which is to be explained. So the proposed explanation has the logical
form of a deductive-nomological explanation. But if it is actually to be
such, the generalisation (i) must be a law. In this section it will be
argued that if truth is given a Realist interpretation, then (i) will be lawlike
and explanatory, but if truth is interpreted as epistemic ideality, it does
not explain the strong predictive success of science .
One test of the lawlikeness of a statement is that S is lawlike if we
can infer from the fact that S holds true in actual situations that it would
continue to hold true in (at least some) non-actual situations. If this
test is applied to (i), and truth is interpreted Realistically, then (i) is
lawlike and intuitively seems to be explanatory. To say that a theory is
true in the Realist sense is at least to say that it is true independently
of our forms of perception or of our means of testing theories. A theory
that is true in the Realist sense would remain true even if we had dif-
ferent sensory organs, or developed ways of testing theories very different
from those we currently possess or will develop in the future. So, such
a theory would still be empirically adequate even if we developed sensory
organs, or modes of testing, different from those we have or will have.
That is, if a theory is true in the realist sense, not only will it actually
be empirically adequate, but we may assert that it would still be empir-
ically adequate in possible counterfactual situations, such as those in
which we developed modes of testing very different from those we
actually will develop. Consequently, we may assert that if a theory is true
in the Realist sense it has a lawlike tendency to be empirically adequate,
and hence that if ' true' is interpreted realistically, (i) will be lawlike.
Therefore saying that a theory is true in the Realist sense provides us
with an explanation of its empirical success. On this view, it is the
realist idea of truth being independent of our modes of perception or
testing that confers upon the Realist notion of truth its explanatory power.
To say that the truth of a theory is independent of all types of test
not only provides an explanation of individual events of a theory passing
tests, it also provides an explanation of why a theory regularly passes
types of tests, since if a theory is true in the Realist sense it would
continue to be empirically adequate even if we subjected it to types of
test very different from those we will ever develop.
Let us now consider the question of whether the notion of epistemic
ideality provides us with an explanation of strong predictive success. First
we should note that, quite plausibly, some counterfactual statements are
supported by the claim that T is epistemically ideal. Putnam says that
a theory is ' epistemically ideal' only if it 'correctly predicts all obser-
vation sentences (as far as we can tell)' . Now, the observations predicted
by a theory will depend, at least in part, on what means of testing the
theory there are . The development of new apparatus and experimental
techniques can lead to new means of testing a theory and to new obser-
vational predictions made by a theory. Similarly, if we had sensory organs
different from those we actually possess, then what would count as an
238 JOHN WRIGHT
'observation-sentence' would, quite plausibly, be different from what it

actually is, and so the observational predictions made by a theory would
be different from what they actually are. Let us initially assume that a
theory ' correctly predicts all observation sentences' if it successfully
passes all the tests to which we actually have, or actually will subject
it. Later we will examine the consequences of rejecting this assump-
tion. To say that a theory passes all the tests to which we actually have
or will subject it certainly gives us good reason to suppose that if we
had performed on the theory tests of the same type as those, but at
locations of space and time different from the ones we actually used,
the theory would still have passed those tests. So there are, very plau-
sibly, some counterfactual statements supported by the claim that a theory
is epistemically ideal. 10 But it will be argued that this does not enable
the notion of epistemic ideality to explain strong predictive success.
At this point it is worth considering again exactly what it is that
needs to be explained . A theory is accepted as being strongly predictively
successful, not merely if it correctly predicts one instance of some novel
regularity N, but if scientists have become satisfied that the production
of instances of N is a repeatable effect. So, what needs to be explained
is not just the fact that the theory has passed one, or even a number of
instances of test N, but that the theory has a lawlike tendency to pass
tests of this type. That is, what needs to be explained is not just a
number of individual events, but a regularity. We refer to this regu-
larity when we say that a theory's ability to pass a type of test is a
repeatable effect.
If T is epistemically ideal then we have good reason to suppose that
if it had been subjected to instances of test N at any location of space
and time, it would have passed those tests. This supports the idea that
T has a lawlike tendency to pass instances of N, and this in turn makes
it possible to explain instances of T passing tests of type N by saying
it is epistemically ideal. But this is not sufficient to explain everything
that needs to be explained about the strong predictive success of T. On
the contrary, it is T's lawlike tendency to pass tests of type N that needs
to be explained. The lawlike component implicit in the claim that a theory
is epistemically ideal does not explain the strong predictive success of
the theory; it is rather an expression of what it is that needs to be
explained.
We can now see that our original intuitive assessment of the epis-
temic ideality of truth is correct. Like Liberal Verificationism, it does not
explain strong predictive success, but merely re-states the very thing
that requires explanation.
To say that a theory is true in the realist sense, that is, true indepen-
dently of all perception or means of verification, supports the claim
that it would have a lawlike tendency to pass whatever tests to which
it was subjected. In particular, it supports the claim that it would have
a lawlike tendency to pass whatever new type of test to which it was sub-
jected. So, saying that a theory is true in the realist sense does provide
an explanation of strong predictive success.
Earlier on we tentatively defined an epistemically ideal theory as
one that passes all tests to which it is actually subjected. We also noted
that if a theory passed those tests then it would be reasonable to assert
that it would also pass tests of the same type performed at different
locations of space and time . But it may be objected that this concep-
tion of an epistemically ideal theory is unnecessarily restrictive. In the
remainder of this section it will be considered whether a less restric-
tive definition of 'episternically ideal' might enable us to explain strong
predictive success.
Might an epistemically ideal theory be defined as one that would
pass all possible types of test? There are, I think, at least three things
wrong with this suggestion.
(1) The notion of all possible types of test is extremely unclear. In the
previous section we noted that the observational predictions made by
a theory depend at least upon experimental apparatus and the sensory
organs of users of the theory . But what would count as a possible
type of experimental apparatus, or a possible type of sensory organ?
It is doubtful that these ideas are clear enough to aid our under-
standing of the concept of truth.
(2) On this suggestion a theory would be epistemically ideal iff it would
pass all possible tests . But there almost certainly are tests which,
although possible, could never be performed by us. So defining epis-
temic ideality in this way has the effect of making truth inaccessible
to us. And one of the motivations for adopting an epistemic con-
ception of truth is that it ensures that truth is accessible to us. So
the proposed definition of epistemic ideality deprives the epistemic
conception of truth of one of the features which has been thought
to make it attractive.
(3) It is not clear that, as defined, the epistemic definition of truth would
be any different from the Realist conception. It was noted in (2)
240 JOHN WRIGHT
above that if truth was defined in this way it may transcend verifi-
cation by us." On this suggestion, a theory is true iff it would pass
all possible tests . But might not tests capable of identifying the
(Realistically interpreted) truth-value of a theory be possible? If such
tests are possible this version of the epistemic conception of truth
is no different from the realist conception.
Other ways of defining the notion of 'episternically ideal' might be
suggested, For example, it might be suggested that a theory is epistem-
ically ideal if it would pass all physically possible tests, or all practically
possible tests. But such definitions have difficulties that are familiar. The
notion of a ' physically possible' test, and especially the notion of a prac-
tically possible ' tests, are not clear. The notion of a practically possible
test is defined with reference, not only to the laws of physics, but also
with reference to facts about human physiology and the availability of
resources. But it is not very plausible that our concept of truth should
be defined with reference to facts about the availability of resources or
the laws of physics.
In any case, there is a more fundamental difficulty with the proposal
to define a theory as epistemically ideal iff it would pass all possible
types of test - whether those tests are construed as logically, physically
or practically possible . Perhap s saying that a theory is true in one of these
senses would explain the strong predictive success of the theory, but a
new problem would arise : why would the theory pass all possible types
of test? That this is something that would very much require explana-
tion can be brought out by considering a simple example. Suppose T
predicts that the value of a particular physical constant k is 1.06758.
Suppose it is also claimed that T is epistemically ideal in the sense of
passing all possible types of test. This means that all possible types of
test would give the value of k as 1.06758. This remarkable agreement
across types of test would surely require an explanation. As far as I
am aware, no advocate of an epistemic conception of truth has even
attempted to provide an explanation of why a theory should pass all types
of test.
So, the advocate of an epistemic conception of truth would only
acquire the ability to explain the strong predictive success of science
by embracing a claim which is at least as much in need of explanation
as the phenomena it explains. The Realist , of course , has a simple and
natural explanation of the agreement between all possible tests : k has
its value independently of human perception or techniques of verifica-
tion and this is why all the techniques of verification are led to assign
the same value. Of course, one very important problem remains for the
Realist: given the underdetermination of theory by actually obtained data ,
how have we managed to find theories that are true in the Realist sense?
However, that is an epistemological problem ." It is very different from
the ontological or semantic problem : How is the claim that our theories
are true to be interpreted if it is to enable us to explain their strong
predictive success?
CONCLUSIONS
In this paper it has been argued that neither Strict Verificationism nor
Liberal Verificationism nor the 'Episternic Ideality ' conception of truth
can explain the strong predictive success of science. It has also been
argued that a Realist conception of truth, according to which truth is inde-
pendent of our perception, modes of testing or techniques of verification,
is able to explain strong succe ss. Whether this will be regarded as a good
argument for a Realist conception of truth may depend on the attitude
adopted to abduction as a form of inference. It is beyond the scope of
this paper to assess the strength of abductive arguments. However it is
possible to construe the argument for Realism from the success of science,
not as abductive, but as an instance of modus tollens in which the con-
sequent is probabilistically falsified . It should also be noted that a recent
study has defended the legitimacy of inference to the best explanation. 13
The University of Newcastle,
Australia.
NOTES
1 Putnam ' s notion of Metaphysical Realism is developed in much of his writing since
1976. See espe ciall y his Reason . Truth and History and The Many Faces of Realism.
2 I will not attempt to clarify the notion of strong predictive success, or of a theory making
novel predictions, in this paper . But a good discussion of the notion , and a defence of
its viability, can be found in J. Worrall 'Scientific Discovery and Theory-Confirmation '
in J. C. Pitt (ed.) Change and Progress in Modern Science , especially pp. 301-331. One
author who has emphasised the importance of novel predictions in lending credibility to
a theory is Alan Musgrave in his ' The Ultimate Argument for Scientific Realism ' in
Nola (ed.) Relativism and Realism in Science, esp. pp. 229-252.
242 JOHN WRIGHT
J This claim is argued for in my Science and the Theory of Rationality (Aldershot,
Avebury, 1991), especially pp. 10-25.
4 A way of singling out the parts of a theory directly involved in the derivation of a
prediction is given in my Science and the Theory of Rationality (loc. cit.).
5 I recognise, of course, that the notion of 'closeness to the truth' is an unclear one .
But (a) it is not necessary, for the argument presented here, to say that theories are close
to the truth . Rather, it is claimed that even if it is said that theories are in some sense
merely ' close to the truth ', still some form of the argument presented here would remain
sound . (b) The notion of closeness to the truth is one used by many authors including
K. Popper, W. H. Newton-Smith and R. Boyd, and may be necessary to give an adequate
account of scientific progress . (c) Recent work on the notion of closeness to the truth gives
cause for optimism regarding our ability to understand this notion (See G. Oddie, Likeness
to Truth (Dordrecht : Reidel, 1986) and I1kka Niiniluoto, Truthlikeness (Dordrecht: Reidel ,
1987). See also Theo A. F. Kuipers (ed.) What is Closer-to-the-Truth ? (Amsterdam :
Rodopi, 1987) in which a number of approaches to the concept of truthlikeness are
discussed .)
6 See B. S. van Fraassen The Scientific Image especially pp. 39-40.
7 There is reason to suppose that, at least at one stage, Dummett saw anti-Realism as
involving Strict Verificationism. For example, in his 'The Social Character of Meaning '
he says that the adoption of a technical means of identifying gold would change the
sense of 'gold' and it is clear from the context that this might also involve changing its
reference. But this means that some sentences of the form 'This is gold ' would change
their sense, and maybe even their truth value , with the acquisition of a new technique.
See Dumrnett 's Truth and Other Enigmas, p. 429 . Also, Dumrnett's suggestion that
according to anti-Realism , a statement has a truth-value only if we possess an effective
method for deciding it, suggests Strict Verificationism . However, Dummett also some-
times says that we can display knowledge of what it is for a sentence to be true if we
can 'recognise it wherever it obtains, or [get] ourselves in a position to do so' (my
italics). (See, for example , Truth and Other Enigmas, p. 225.) The italicised phrase suggests
Dummett would allow that we could know what it is for a statement to be true if we
could acquire a technique for verifying it. This would make Dummett what 1 have called
a 'Liberal Verificationist' (see below) . Dummett's frequent use of the word 'capacity ',
in his claim that, according to anti-Realism, a statement is true or false if we have a
capacity to determine whether it is true or false, can perhaps be interpreted as supporting
either Strict or Liberal Verificationism.
8 This is obviously so if closeness to the truth is identified with Popper's verisimili-
tude. Popper's definition of verisimilitude explicitly uses the concept of truth . (See
Popper's Conjectures and Refutations, p. 392) According to the 'similarity' approach to
truthlikeness advocated by Niiniluoto, a claim S is 'truthlike' iff the state of affairs allowed
by S is similar to that allowed by a true sentence. Similarly, Oddie says that a claim is
like the truth if the state of affairs it allows is close to that allowed by a true sentence.
So, if true were to be defined in epistemic terms , so would closeness to the truth.
9 See Putnam's Meaning and the Moral Sciences, p. 125.
10 lowe this observation to Professor Gregory Currie. But we should also note that while
the claim that epistemically ideal theories are projectible in this way seems very plau-
sible for theories of physics, it is not so plausible for theories of biology . For example,
that all organisms of a particular type that we have so far observed have immunity to
bacterium Y doe s not lend great support to ' Organisms of that type in some unob served
location also have immunity to bacterium Y' .
11 In at least some of his writ ings, Dummett has seen the possibility of verificat ion-
transcendent truth as one of the hallmarks of Realism. See, for example , his paper 'Realism'
in Truth and Other Enigmas. Crispin Wright , in his discussion of Dumrnett 's work, has
even seen the possibility of verification-tran scendent truth, rather than bivalence, as con-
stituti ng the core of Dummetr's conception of Realism. (See Cri sp in Wright' s paper
' Dummett and Revisionism ' in Barry Ta ylor (ed.) Michael Dummett: Contributions to
Philosophy, pp. 1-31, especially p. 4.)
12 I have attempted to give one possible answer to this problem in Science and the Theory
of Rationality .
IJ For a defence of abductive inference, see P. Lipton Inference to the Best Explanation.
For the possibility of construing the argument for Realism from the strong predictiv e
succe ss of science as an instance of modu s tollens, see my Science and the Theory of
Rationality, pp. 14-15 .
BIBLIOGRAPHY
Dummett, M. A. E. (1978) . Truth and Other Enigmas, Duckworth, London .

Dummett, M. A. E. (1991). The Logical Basis of Metaphysics, Duckworth, London .
Fraas sen van B. (1980) . The Scientific Image, Clarendon Press, Oxford .
Kuipers, T. A. F. (ed .) (1987) . What is Closer-to-the-Truth?, Rodopi , Amsterdam.
Lipton, P. (1991) . Inference to the Best Explanation, Routledge, London.
Niiniluoto , I. (1986) . Truthlikeness, Reidel, Dordrecht.
Nola, R. (ed.) (1988) . Relativism and Realism in Science, Kluwer Academ ic Publi shers,
Dordrecht.
Odd ie, G. (1980). Likeness to Truth, Reidel, Dordrecht .
Pitt, J. C. (ed.) (1985). Change and Progress in Modern Science, Reidel, Dordrecht.
Putnam, H. (1978). Meaning and the Moral Sciences, Routledge and Kegan Paul , London .
Putnam, H. (198 1). Reason . Truth and History, Camb ridge University Press .
Putnam, H. (1987) . The Many Faces of Realism: The Paul Carus Lectures, Open Court,
La Salle , Illinois.
Tayl or, B. (19 87) . Michael Dumm ett: Contr ibutions to Philosophy, Mart inus Nijho ff
Publi shers, Dordrecht.
Wright J. (1991) . Science and the Theory of Rationality, Avebury, Aldershot .
WU QIONGBING
IS NATURAL SCIENCE FREE FROM MORALITY?
I. INTRODUCTION
It has been frequently asserted that science is distinguished by its objec-

tivity. Science, real science, deals only with what is, and has no concern
with what ought to be. Science is antiseptically devoid of any involve -
ment with human values. Science, in its way of looking at the matter,
is so purely objectively and narrowly factual in its concerns that it can,
and indeed should, be wholly insensitive to the emotional, artistic, and
ethical values of human life. The activities of science are morally and
socially value-free and so are the laws and facts of science. As regard
the subject of scientific activities, the scientist goes about his work in
a rigidly impersonal and unfeeling way, unmoved by any emotion other
than the love of knowledge and the delights of discovering the secrets
of nature . A scientist makes no value judgement in his whole scientific
research, he is someone painstakingly obtaining objective data, testing
every side of a question and disregarding personal interests. Although
the applications of science may be good or evil, these have nothing to
do with science itself. The two-edged sword of science is fashioned for
whomsoever will pick it up and wield it, the scientist carries no special
responsibility for those applications, save as a normal citizen.
This image of science has gained for itself a distinctive label as the
thesis of the value neutrality of science. It has been widely accepted
not only by the scientifically uninformed and scientific outsiders, but also
by many authorities within the scientific community itself.
The view of the value neutrality of science also has been subjected
to quite a lot of attacks. For instance, some people point out that in the
whole course of the intellectual history of human beings, from Aristotle
and his predecessors to Descartes, Newton, Kant, and Einstein, science
has been a part of the cultural tradition in its larger sense . Throughout
the whole course of the development of our civilization, science has
always merited the historic epithet of "natural philosophy". No matter
how much our way of describing the facts may change, there is little
doubt that this basic circumstance of the formative role of science in
245
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.) , Realism and Anti -Realism in the
246 WU QIONGBING
molding all of our areas of thought will remain invariant. Natural science
and human or social science, as parts of human culture, are unavoid-
ably inseparable and interdependent, they permeate each other. Some
people propose that scientific activity is a kind of human activities, and
since human activities are purposeful and goal-oriented, scientific activity
inevitably becomes morally and ethically coloured. No matter how the
purpose is understood, whether by appealing to the logic of scientific
development, or to the needs of society, or to the desires of the employer,
the act of cognition is immersed in a value -rich rather than value-neutral
atmosphere. Some people sustain the great role of values within science
by indicating the intellectual values, such as simplicity, regularity, uni-
formity, comprehensiveness, systematicity, etc. , in scientific knowledge,
or by examining the ethical values, such as independence, originality,
dissent, tolerance, in scientific activities. Some analysts, on the other
hand, try to bridge between "is" sentences and "ought" sentences by
language analyses. In our country, many authors criticize the view of
value-neutrality of science on the basis of Marxist philosophy, by ana-
lyzing the course of the historical development of science, by showing
the interaction between science and other dimensions of society, that
is, the great role science plays in society and the effects of society
imposed on science, etc .
There is something reasonable in all these ideas. However, it is regret-
table to see that the view of value-neutrality of science still remains in
fashion despite so many diverse criticisms. Why?
II. THE KEY TO THE PROBLEM
In my opinion , this is because most of the criticisms fail to reach the

heart of the view. There are various types of values, such as cognitive,
aesthetic, theological, political, ethical, economic, etc. In this issue, values
mainly refer to social-ethical values concerning rightness and wrong-
ness of conduct. I thinkthat whether science is value free or not primarily
relies on the way we understand the concept of science, on how we
look at science. So in criticizing the view of the value-neutrality of
science, we should first find out the way it understands science. In
order to get the nub of the issue, I had better trace back to the origin
of this view.
The issue of value neutrality comes into existence with the advent
IS NATURAL SCIENCE FREE FROM MORALITY ? 247
of modern philosophy and science, especially the scientific revolution

of the seventeenth century. It is the product of the so-called mechan-
ical picture of the world (especially classical Newtonian science), and
the scientific revolution, most especially the epistemological and method-
ological revolution in science and philosophy inspired by its main
architect, Rene Descartes (1596-1650).
The mechanical world-picture regards nature as a vast machine
governed by quantitative laws and relationships (nature is written in
the language of mathematics); the objective features of the world turn
out to be those features - matter, motion, and physical magnitude - which
constitute the nut s and bolts of the machine, together with the laws
governing it. Only such features of experience are truly objective. Thus
a rational methodology for inquiring about the machine's working -
i.e., for acquiring knowledge - must take into account only those features
which can be quantified, i.e., written in nature's language. The very
essence of the world is given by the objective properties just mentioned,
together with the mechanical laws which govern them (These essential
features of the world are dubbed ' primary qualities ' by Galileo and
Locke). So, in order to acquire reliable (i.e., 'objective ') knowledge of
the world by the use of reliable ('rational') methods of inquiry, we must
make a sharp distinction between what is objective and what is subjec-
tive. Science reflecting the objective world certainly belongs to the
category of objectivity, value is subjective, and thus does not belong in
an objective account of the world. Science and morality cannot conflict
(they are 'complementary ') since they have nothing to do with each other:
they govern different spheres of experience (those related to the differ-
ences between 'man as object' and as actor) . Science, which is filled with
certainty and characterised by quantitative qualities, is, and must be, value
neutral. Value judgements are merely subjective and irrational acts of the
will, and scientists should, and indeed must , take no account of anything
'subjective' in their way of arriving at objective truth. Scientists have
declared the neutrality of the knowledge of science since the 17th century,
and extolled this kind of neutrality as a credit relating to the objectivity
of the knowledge of science and to the powerful authority of science
itself.
The way we look at science, in my opinion, is the key problem in
this issue . Thus before I express my ideas , I would like to examine the
concept of science first.
248 WU QIONGBING
III. OUR CONCEPT OF SCIENCE
What is science? The English word 'science' is derived from the Latin
world 'scientia', which simply means knowledge or intellect. The German
word ' Wissenschaf t' means systematic, organized knowledge. Thus,
science is generally considered to be theoretical knowledge concerning
the natural world both in living and non-living forms. The course of
the historical development of science and the new characteristics
emerging in modern science have proved that this kind of understanding
of science is far from enough to embrace the contents of science.
According to the Modern Science Dictionary (Franklin Publishing
Company, Palisade , New Jersey 1975), science is "A systematic arrange-
ment of scientific data, secured by controlled observation, experience,
or experiment, including the method of obtaining and testing these data".
So science is not only a kind of orderly knowledge but also a kind of
cognitive activities to achieve this product of cognition. In fact, modern
science is notably different from that before the early period of the 20th
century. Science has been speedily regularized, industrialized and per-
formed on a large scale . Now our cognition in science is still changing,
for science has played such a great role in society that many disciplines
of metasciences, whose research objects are science itself, emerge: history
of science, philosophy of science, sociology of science, psychology of
science , etc. These disciplines give science different definitions respec-
tively in terms of their own interests in science - This is determined
by the multisidedness of science itself - such as 'an instrument to resolve
problems' which focuses on its instrumental property, or ' organized
knowledge' in its archival dimension, or 'a series of procedures for
obtaining objective knowledge concerning the natural world', in its
methodological character, or 'discoveries made by those who own
special research abilities ', in its professionalism, etc. These definitions
only describe some particular characteristics of science from different
angles.
In respect to the general definition of science, this is just as J. D.
Bernal said, that such important changes in science have taken place in
the whole of human history that it is very difficult to give it a suitable
definition. The nature of science cannot be fixed by definition forever
through our historical research into science. Indeed, while we cannot give
science an absolute and perfect definition for science itself is always
changing , but this does not mean that we cannot give it a relative one.
Because our understanding of science in any time must have some marks
of the developmental characteristics of science in that era, then our
science concept should be based not only on the historic analyses of
the development of science but also, and it is more important, on the
features of our modern science. Unprecedented features have taken
place in the development of modern science: 'small science ' turns into
' big science ', science becomes a profession, scientists are engaged in
organized scientific activities and receive remuneration, etc . Bernal
pointed out in his work Science in History (C. A. Watts, London 1954,
pp.5-6):
Scienc e may be taken, (I, I) as an institution; (I , 2) as a method; (I, 3) as a cumulative
tradition of knowledge ; (I, 4) as a major factor in the maintenance and development of
production; and (I, 5) as one of the most powerful influence s moulding beliefs and atti-
tudes to the universe and man.
These descriptions synthesize primary respects of science which seem

to be different. Here, by making reference to Bernal's descriptions, I
try to give science a descriptive definition : Science is: (1) a growing body
of knowledge - not knowledge in a sense implying certainty but rather
knowledge fallible in principle, and knowledge consists of scientific state-
ments, explanations and theories ; (2) above all a special social activity,
i.e., research, governed by specific rules or norms of intellectual conduct;
(3) a social institution: it is interplay of tradition and rational criticism
contained in the network of institutions such as universities, scientific
societies, and journals, which constitute the scientific community. The
existence of such a community is a precondition for the possibility of
that special activity which we call ' research'.
From the previous section, we can see that the value -neutral view
derives from the so-called mechanical picture of the world, based on
the classical Newtonian science of the 17th century . It rests on the so-
called ' conception of small science' , which refers to traditional science
aimed at the accumulation of knowledge and characteristic of individual
free research. The argument that science is ethically neutral originated
precisely from this understanding of science. The neutral view essentially
identifies science with classical natural science which inclined to mech-
anism and tried to explain all phenomena and objects in terms of
Newtonian physics. As a result, the guidelines of scientific cognition
at a certain stage in the development of science (although even at that
time their authority was by no means absolute) are misrepresented as
250 WU QIONGBING
specific features of scientific cognition in general, and a conclusion is

reached that science inevitably brings a disinterested narrow view of
reality and kills creative imagination: the scientist merely describes objec-
tively what happens in his research work.
This comprehension of science is too narrow to be suitable for our
present concept of science. It exaggerates some properties of the devel -
opment of science in the Newtonian period and identifies them as the
fixed nature of science, paying no attention to the development of modern
science. In fact, today's science becomes more and more dependent on
society and the public; science has not been merely the individual pref-
erence of the scientist, or a pure search for truth just for the sake of
science itself. To start, scientific research must have some value or util-
itarian demands, and not merely aim at the accumulation of knowledge.
And the solitary scientist laboring in isolation in his study or labora-
tory has given way to the institutionalized laboratory. The scientist
nowadays usually functions not as a detached individual unit , but as
part of a group , as a 'member of the team' . Science 's research objects
are not confined to the physical world, for it is increasingly turning
man into an object of research (it is in this sphere that the 'headaches'
of today 's science are concentrated which may not be understood without
socio-ethical analysis). Now the intension and extension of the concept
of science have expanded quite a lot , the image of science which the
neutral view stems from is far from enough to embrace the present
contents of science, for it is just the product of a certain historic period
and has become out-dated.
We should not always keep our eyes on the past and pay no atten-
tion to the present. Here I would like to emphasize our concept of science ,
that is, a growing body of knowledge, a special social activity, a social
institution, and propose the very problem which will be discussed in detail
in the following sections: Is natural science free from morality?
IV. TWO DIMENSIONS OF THE PROBLEM
The relationship of science to value has two dimensions: the issues of

the 'value of science' and of 'value in science'. The former takes place
in the course of interaction of science and other dimensions of society,
concerning the relations between the uses (and abuses) of science and
our moral beliefs and principle. The latter happens within science itself,
concerning whether the scientist makes value judgements in his scien-
IS N ATURAL SCIENCE FREE FROM MORALITY ? 251
tific research and whether there exist value judgements in the knowl-
edge of science.
It is well known that science as a whole is closely related to human
morality, in its interaction with society. The starkest forms appear in
the moral dilemmas in nuclear physics, in issues of science and the global
problems of human genetics, in freedom of research and the scientist's
socio-ethical responsibility, etc . These problems are still under debate.
Some authors criticize the neutral view by indicating these widely known
facts, but I think that this kind of criticism does little to help under-
mine the neutral view since the neutral view also acknowledges that
science may interact with society : the uses of science may be good or
evil , science may be consistent with or conflict with our moral beliefs
and principles. The question is: does it insist that scientists carry no
special responsibility for the application of science ?
That scientists have to be responsible for their research findings or
not , in my opinion, relies to a large degree on how to interpret the
ethical values. If we understand values in terms of a subjective inter-
pretation (from which the neutral view derives), that values originate from
the interests and desire s of the subject (human beings) , then the scien-
tists have nothing to do with the uses or abuses of their research findings.
If, in terms of an objective interpretation that values are the properties
of the object, then 'counterculture' seems to gain its rational base, scien-
tists should be responsible for the opening of 'Pandora's box ', Einstein
should be responsible for the nuclear disasters in Hiroshima and Nagasaki
in August 1945. Here I prefer the Marxist definition of value: value is
a specific form or manifestation of the relation between subject and object
in which the properties of the object are appraised in terms of their ability
to satisfy the needs of the subject. The interaction of object and subject
is the decisive element in the determination of value, and values are
likewise indices of subject-object interactive states. Then the scientists
have to undertake, together with society, the socio-ethical responsibility
for the uses of science; here the social responsibility of scientists is in
itself a condition which, though indispensable, is unable to exclude the
possibility of the misuses of scientific findings. The problem can only
be resolved by the entire course of progressive social development.
All of the various moral problems of what is done with scientific
discoveries are not problems that arise within science, and are not ethical
choices confronting the scientist himself. I will not be concerned with
them.
252 WU QIONGBING
In this article I will concentrate on the issue of 'value in science',

namely, the ethical problems in science. According to our definitions
of science, we may discuss these ethical problems respectively in three
dimensions. But I think, the ethical problems in science as a social
institution and as a special activity are bound up with society, for the
members of scientific institutions and the subjects of scientific activity
are scientists themselves; thus these ethical problems are just what they
have to face to in their scientific research. The existence of such a
scientific institution is a precondition for the possibility of that scien-
tific activity - just like a 'form of life' - in our era. It is difficult (and
even impossible) for me to distinguish these ethical problems that arise
in scientific institutions from those that arise in scientific activity.
Therefore, here I discuss only the ethical dimensions in science qua social
institution and special social activity as one part, and that in science
qua systems of knowledge as another part.
V. ETHICAL DIMENSIONS IN SCIENCE QUA

AN ACTIVITY AND AN INSTITUTION
Science as a special social activity and a social institution is a human

enterprise, it is carried out in society, and is performed for and by
human beings, and scientific research therefore inevitably exhibits some
ethical complexion. Scientists - like other members of distinct profes-
sional groups, such as lawyers, doctors, soldiers - do have specific
behaviour patterns which constitute the 'rules' and 'norms' of the
scientific community. A professional scientist must be familiar with these
rules and must be ready to abide by them in practice. The fact is that
these rules are not all formerly codified and precisely formulated. R.
K. Merton first put forward a set of norms in 1942 which may be
expressed as follows :
Communalism: Science is public knowledge, freely available to
all. That is to say, the results of research do not belong to individual
scientists, but to the world at large. Scientific discoveries should be
communicated immediately to the community by publication in the
open literature, which anyone may draw upon for their own further
use.
Universalism: There are no privileged sources of scientific knowledge.
Truth should be judged in terms of intellectual criteria, criteria that are
considered valid in the particular branch of science, and not in terms
IS NATURAL SCI ENCE FREE FROM MORALITY ? 253
of the attributes - nationality, race, religion, class , age - of the person

who produces it.
Disinterestedness: Science is done for its own sake, scientists should
undertake their research, and present their results with no other motive
than the advancement of knowledge. They should have no personal stake
in the acceptance or rejection of any particular scientific idea.
Originality: Science is the discovery of the unknown. That is to say,
scientific research results should always be novel. An investigation that
adds nothing new to what is already well known and understood makes
no contribution to science.
Scepticism: Scientists take nothing on trust. That is to say, scientific
knowledge should be tested both empirically and logically, and should
not be accepted on the basis of authority.
The Mertonian norms are much debated and are often ' honoured as
much in the breach as in the observance' . They describe a set of ideal
patterns of behaviour. What the neutral view takes interest in is the
norm of disinterestedness, it claims that scientists have by nature, or
acquire by nurture, a 'scientific attitude ' which makes them peculiarly
honest, objective, independent of mind, sceptical, rational, etc . Just as
laws exist because of the existence of behaviours which the laws explic-
itly forbid, there must be some behaviours violating the ideal scheme
which the neutral view asserts.
In this section I want to question the 'objective attitude' of scien-
tist s which the neutral view describes, that is, that scientists' activities
and efforts are directed towards the extension of scientific knowledge ,
and that regardless of their personal interest, scientists make no value
judgement in their research work. On the contrary, scientists have to
be confronted with many ethical problems and have to answer them
(i.e., make value judgements) in their whole research, just from the initial
inception of the research to the ultimate reporting of its completed
findings. A scientist is a flesh and blood man instead of a detached and
antiseptic one . By examining the ethical problems emerging in the whole
cognitive course of science, and through some case analyses, I want to
prove that scientific activities in scientific institutions cannot be free from
morality ; the scientist really has to make value judgements in his research.
Ethical problems, that is, issues regarding the rightness and wrongness
of conduct - arise out of people 's dealing with each other, and pertain
necessarily to the duties, rights, and obligations that exist in every kind
of interpersonal relationship. The limits of this paper prevent me from
254 WU QIONGBING
discussing in detail all those ethical problems at every junction of sci-

entific research, such as the motives of engaging in scientific work, the
choice of research goals, the staffing of research activities, the selec-
tion of research methods, the specification of a standard of proof, the
evaluation of research findings, the dissemination of scientific infor-
mation, the allocation of credit for research accomplishments, etc. Here
I primarily discuss some ethical problems in the scientists ' motives,
the evaluation of research findings, the allocation of credit for research
accomplishments, which are most acute in science, and just mention
roughly those of other junctions.
(1) Ethical Problems Regarding the Motives of Scientists
Ethical problems regarding the motives of scientists, in my opinion,

are the most critical of all, most of the ethical problems arising in
scientific activities are directly or indirectly associated with them. Does
the scientist go about his work in a rigidly impersonal and unfeeling way,
unmoved by any emotion other than the love of knowledge and the
delights of discovering the secrets of nature? Perhaps he would if he were
isolated from the external world. In his speech in the celebration meeting
of the sixtieth birthday of Max Planck, Einstein proposed that there are
mainly four kinds of motives of inquiry (Max Planck : Where Is Science
Going ?, with a preface by Albert Einstein, London, George Allen &
Unwin 1933, p. 18):
Many kinds of men devote themselves to Science, and not all for the sake of Science
herself. There are some who come into her temple because it offers them the opportu-
nity to display their particular talents. To this class of men science is a kind of sport in
the practice of which they exult, just as an athlete exults in the exercise of his muscular
prowess. There is another class of men who come into the temple to make an offering
of their brain pulp in the hope of securing a profitable return. These men are scientists
only by the chance of some circumstance which offered itself when making a choice of
career. If the attending circumstance had been different they might have become politi-
cians or captains of business. . . .
What has led them to devote their lives to the pursuit of science? That question is
difficult to answer and could never be answered in a simple categorical way. Personally
I am inclined to agree with Schopenhauer in thinking that one of the strongest motives
that lead people to give their lives to art and science is the urge to flee from everyday
life, with its drab and deadly dullness, and thus to unshackle the chains of one 's own
transient desires, which supplant one another in an interminable succession so long as
the mind is fixed on the horizon of daily environment.
But to this negative motive a positive one must be added . Human nature always has
tried to form for itself a simple and synoptic image of the surrounding world . In doing
this it tries to construct a picture which will give some sort of tangible expression to
what the human mind sees in nature . That is what the poet doe s, and the painter, and
the speculative philosopher and the natural philosopher, each in his own way. ...
In Einstein's view, there are many kinds of motives for scientists to

devote themselves to science . Some are for the purpose of gaining self-
satisfaction from their activities , some are for obtaining profitable returns,
some are for fleeing from the dullness of everyday life, some are for con-
structing a simple and synoptic image of the natural world, etc., which
indicate the complication of scientists' motives. In ancient times, science
was an individual preference and an activity of a cultured, moneyed
and leisure class ; the English term 'scientist' did not emerge in Britain
until the mid-nineteenth century; the phrase actually used was "a culti-
vator of science". At that time , the motive of the scientists was more
to enjoy themselves in the research than any others. Just as the neutral
view declares, such scientists go about their work for the love of knowl -
edge and the delights of discovering the secrets of nature. However,
that is far from the only motive of scientists. In modem society , things
are more complicated. Scientists are engaged in their work to enhance
their social position, or for money, or for patriotism, or for individual
interest, or for all of these, etc., which constitute the complex of scien-
tists' motives . For many centuries, one of the strongest motives driving
scientists to take up scientific research is, in my opinion, to get the
acknowledgement of other people (maybe expressed in other words, such
as glory, credit, priority, etc.), which is almost always neglected. For
example , the dispute in the history of science between Newton and
Leibniz and their followers regarding priority in the invention of the
calculus is such a case. Modern scientific research and scientific insti-
tutions are sponsored by governments or by industrial or financi al
magnates, and there are fierce competitions within scientific institution,
which impose quite a lot of pressures on scientists. They have to do
their best to acquire the research results in order to gain acknowledge-
ment inside and outside science . For that kind of recognition will make
them more easy to realize other ideals (i.e. to satisfy other motives).
The complication of scientists' motives leads to many complicated
ethical problems. For scientists are flesh and blood men, they have all
the general properties of a human being, and it is impossible for all
scientists to realize their purposes in a 'normal' manner. In America,
the integrity of the scientific research process is seriously challenged
256 WU QIONGBING
by the revelations of a series of misconducts of scientists. Misconduct

includes such deviant behaviour as fabrication, falsification, or plagia-
rism, in proposing, performing, or reporting research. Different types
of misconduct can be illustrated roughly as the following (cf. False
Prophets by Alexander Kohn, 1986):
Forgery: reports experiments that have never been carried out , but
which the researcher may feel he needs for the support of his or her
hypothesis. In some cases forgers have not only invented experimental
data, but have related these non-existent data to non-existent manuscripts
allegedly published or accepted for publication.
Plagiarism: consists of using the data or ideas of other investigators
without reference to the source, or even verbatim copying of a text written
by someone else . A number of plagiarizers obtained the necess ary data
either from a grant proposal or from a manu script submitted for
publication.
Trimming: which is based on amplification of an experiment: the
investigator describes accurately the nature of experiment and its controls ,
but reports a greater number of trials than have actually been performed,
' adds' or ' removes' animals to or from experimental or control groups ,
or misrepresents the variance, although using genuine numbers and
means.
'Cooking ': the researcher omits aberrant values, misreports actual
conditions of the experiment or alters ancillary data . A kindred mis-
demeanour is omission of whole experiments which yielded negative
or contrary results to the hypothesis under test.
There are many examples of such 'deviant behavior ' in scientific
research in which 'objectivity' was lost. Misconduct, regardless of
whether conscious or unconscious, was recognized centuries ago , is
evident today, and will no doubt continue into the future . I think that
such kinds of conduct directly or indirectly result from, or are related
to, the complicated motives of scientists, especially their purpose to
gain recognition.
A torrid controversy in America, begun in 1989, about the discovery
of cold nuclear fusion may be a good case in point (cf. American Scientist,
80). The principle of fusion is simple. It is the joining or fusing of
two light nuclei (usually with a mass number below eight) to make one
larger nucleus, and energy is released in the process. But fusion needs
a temperature of about 100 million degrees Celsius to overcome the elec-
trostatic repulsion. Cold fusion, however, requires only room tempera-

ture and no other extraordinary condition. It offers the potential for an
inexpensive, inexhaustible and clean source of energy and thus received
a great deal of attention. The controversy took place between two research
groups independently working toward cold nuclear fusion - Elec-
trochemists B. Stanley Pons of the University of Utah and Martin
Fleischmann of the University of Southampton in England directed one
group (PF group), and Physicist Steven Jones of Brigham Young
University directed the other group (J group). Although the basic exper-
iments and the theory they were based on were similar, the two groups
obtained radically different results. The PF group detected more heat and
no neutron emissions, the J group detected no heat but neutron emissions.
In order to win the prestige of being the fir st to publish results on cold
fusion, they all announced their findings hastily to the press before getting
strong support or evidence. Therefore, cold fusion was doomed from
the start when a race to be first took precedence over the desire to be
right.
The analysis of what followed will demonstrate that the scientist is
a real person, he has to confront, whether willingly or not, many ethical
problems in his whole research process. Even scientists may lose 'objec-
tivity' in the pursuit of truth .
(2) Ethical Problem s Regarding Research Goals

The problems arise in such a way as to how to allocate limited resources
(both human and material) to the prosecution of research efforts. The
scientific community has to offer a promising and practicable scheme for
the governing of itself. The setting of research goals always has some
practical or utilitarian considerations instead of those of pure search
for the truths about the natural world.
(3) Ethical Problems Regarding the Staffing of Research Activities
These problems relate to the allocation or even appointment of different

research duties of scientists in a collective research or institution: Which
role best fits a scientist, in just accord with his desire?
258 WU QIONGBING
(4) Ethical Problems Regarding Research Methods
These problems arise perhaps most acutely in biological, medical or

psychological experiments involving the use of experimental animals.
After the second world war, people were so shocked to find that
scientific experiments may be performed in such a terrible and brutal
way, when the Nazi war criminals were crossexamined in Nuremberg.
Scientific experiments must be carried out subject to be a series of strict
rules which apply independent of people's subjective will; however, once
scientists turn humans or animals into their actual objects, the arising
of ethical problems becomes unavoidable.
(5) Ethical Problems Regarding Standards of Proof
The problems arise in this way: At what juncture should scientific

evidence be reasonably regarded as strong enough to give warrant for
a conclusion, and how would the uncertainties of this conclusion be
presented? This relates to the motives and attitudes of scientists. The cold
fusion controversy is just such a case .
(6) Ethical Problems Regarding the Dissemination

of Research Finding
This kind of ethical problems is linked with plagiarism before the pub-
lishing of findings in the scientific community: How to deal properly
with the circulation of information from private communications.
(7) Ethical Problems Regarding the Evaluation of Research Findings
The problems are closely bound up with the boundary line between proper
science and pseudo-science. There is no doubt that the scientist has a
duty to protect both his own colleagues in other specialties and the lay
public against erroneous research findings, and has an obligation to
maintain the professional literature of his field at a high level of content
and quality. Can the scientist evaluate research findings in a purely objec-
tive, wholly rational and entirely open-minded way? This is quite
questionable. It is an important phenomenon in the historical develop-
ment of science: a scientist or the scientific community itself has often
opposed genuine scientific findings as being pseudo-scientific on the one
hand, and has overrated some strictly fraudulent and pseudo-scientific

findings on the other hand. One man 's interesting possibility may be
another man 's pseudo-science, and good ends do not justify question-
able means. We know, for instance, the nineteenth-century English
chemist J. J. Waterson; his groundbreaking papers on physical chem-
istry anticipating the development of thermodynamics by more than a
generation, were rejected by the referees of the Royal Society for pub-
lication in its Proceedings, with the comment (among others) that "the
paper is nothing but nonsense". As a result, Waterson's work lay for-
gotten in the archives of the Royal Society until rescued from oblivion
by Rayleigh some forty-five years later. For another example, the French
physicist Rene Blondlot allegedly discovered 'N-rays ' , which were
supposed to be something like X-rays. This curious finding won him a
prize from the French government. But the so-called "N-rays" proved
to be just a delusion. If the scientist evaluates a scientific finding with
some individual purpose or interest, the cases will become more intri-
cate, and I will not discuss such in detail here .
(8) Ethical Problems Regarding the Allocation of

Credit for Research Accomplishments
The allocation of credit for research accomplishments is particularly sen-

sitive. The distinguishing feature of modern science is the collectivization
and collaboration of research. Collaboration can immensely broaden a
person's scientific perspective and advance work far beyond what can
be accomplished alone. But it also can generate tensions between indi-
viduals and groups. Collaborative situations are far more complex now
than they were a generation ago. A notable feature of ' big science' is
the development of large-sized experimental instruments whose estab-
lishment and use cost dearly. The normal use of the big science apparatus
needs the cooperation of a large number of well-trained scientists who
play narrow and special roles respectively in a scientific project. For
example, every experiment in the field of high energy physics gener-
ally needs the collaboration of more than one hundred scientists who
are respectively responsible for the design and establishment of the instru-
ment, dealing with the information provided by computers, adjustment
of rays of light, supervision of the operation of the experiment, the
interpretation of data, etc. Finally, the experimental result is announced
by publication of a primary thesis, and each collaborator will seek indi-
260 WU QIONGBING
vidual 'acknowledgement' of his contribution to science. Here a delic ate

problem arises : how to allocate credit for research accomplishments in
a fair and reasonable way? The problem refers to the listing of so many
authors when the paper is published. Decisions about how credit is to
be allotted for these and many other contributions are far from easy
and require serious thought and discussion. Some research groups and
journals simply list authors alphabetically in order to avoid these knotty
decisions.
Another problem in respect to this is the apportioning of credit between
junior and senior researchers. The attribution of credit for a new idea
or discovery made by a professor and student, or by the student under
guidance of his supervisor, or by the student entirely on his own, can
not be viewed as a black and white situation. It is common (or even
conventional) that a mentor takes away the credit from a student who
takes part in the research. The following is just such a case.
Robert A. Millikan, a famous American physicist of the University
of Chicago, won the Nobel prize in 1923 for his experiments of 1910
in which he determined the electrical charge of the smallest component
of the atom, the electron. The basic idea to measure the charge of the
electron on droplets was proposed by Millikan. The original experi-
ment used water droplets, but they evaporated so quickly that observation
of their fall was possible for a few seconds only. At this stage of research,
a graduate student, Harvey Fletcher, advanced a breakthrough idea to
substitute oil droplets for water droplets. Betcher also made the discovery
in his experiment; Millikan, who was away from the laboratory that
day only saw Fletcher 's set-up the next day, and was very much excited.
From then on he continued working closely with Fletcher. They made
some improvements in the des ign of the appar atus and within six weeks
the discovery was announced publicly. Millikan published the paper in
1910, and although he gave credit to Betcher in the text of this paper,
he did not let him be a joint author on this first paper, which effec-
tively led to Millikan's being awarded the Nobel prize .
I have illustrated from the above that ethical problems crop up at
numerous points within the framework of scientific research, from the
initial inception of the work to ultimate reporting of its completed finding.
The course of producing scientific knowledge is morality-rich rather than
morality-free. But I do not mean that all and every point in scientific
activity is morality-laden. In traditional physics or chemical experiments,
when the scientist puts on his coat, he must act strictly by the experi-
IS NATURAL SCIE NCE FREE FROM MORALITY ? 261
mental rules; the course, we may say, is value free . In the case when
the experimental object is related to animals (including human) or the
experiment itself is directly bound up with the health and welfare of
human beings, its moral colour is striking.
Here I have just listed some main ethical problems scientists have
to face in their actual scientific activities, I have not given some general
or universal ethical rules for scientists to abide by, that is, offered answers
to these problem, for my article aims to prove that science is morality-
laden but not to propose a absolute ethical rule for science. I sincerely
agree with the view that the scientist should go about his work in an
objective and rational way, but ' what should be ' is one thing, and ' what
really is' is another one.
VI. ETHICAL DIMENSIONS IN SCIENCE AS KNOWLEDGE
As far as science as knowledge is concerned, we should deal with this

problem with caution. According to the neutral view, science aims at
the discovery of causes and regularities in the physical world indepen-
dent of us, and the facts and laws of science are valid irrespective of
the nation , race, politics, religion or class position of their discoverer.
The velocity of light is the same whoever carries out the experiment
which measures it. It is based on the dichotomy of fact and value;
scientific knowledge is expressed by factual judgements, that is, sen-
tences concerning ' what is' , and on the contrary, morality is expressed
by value (evaluative) judgements, namely , sentences concerning 'what
ought to be' . There is no logical leap from ' is' to 'ought', nor of course
from 'ought' to 'is' . That the world appears to us is irrelevant to the
way it really is. Then scientific knowledge must be value neutral.
Ju st as whether the scientist make value judgements in his scientific
activity is the key problem about whether science as an activity and insti-
tution is value free, I think, whether there are value judgements in
scientific knowledge is the key problem about whether science as a
system of knowledge is value neutral. The problem relates to how to
interpret the term 'truth', and touches on the everlasting issue between
realists and anti-realists. I would not now plunge into such an issue .
It seems that there is no logical problem for the neutralists to assert
the neutrality of scientific knowledge. I consent to the view that scien-
tific knowledge (perhaps truly and correctly) reflects nature and its course
of movement, its task being to describe and explain what happens in
262 WU QIONGBING
the natural world. However, the defect of the neutral view is, in my
opinion , that it fails to recognize the new development of modem science.
The distinguished feature of modem scientific research is its collec-
tivization and collaboration, and correspondingly, with the wide and deep
advancement of modem science, a notable feature of scientific knowl-
edge is the emergence of new disciplines, i.e., interdisciplines and
borderline disciplines, some of which bridge the natural sciences and
social sciences. Scientific research objects are not confined to the physical
world; science is increasingly turns man into a object of research. The
word 'science' now does not merely refer to the traditional physics, chem-
istry, astronomy and biology. I think that there are three kinds of research
objects of science : (1) the physical world (excluding human being); (2)
man as physiological body; (3) the relationship of the physical world
to man. I adopt the letters K" K2, K) that respectively stand for knowl-
edge about these three objects. There are two levels in scientific
knowledge , one which has been historically and fully confirmed by exper-
iments and universally accepted to be the 'correct' descriptions of nature .
I represent it as A; and there is one which is still in issue, represented
as B. My idea can be illustrated by the following figure:
B
A
Fig. I.
I think that K, is farthe st from ethical values . K, /\ A is believed to

truly describe and expla in what the physical world really is, and may
be value-free, that is, has no value judgement by the scientist. This part
embodies the natural laws and theories of traditional physics, chem -
istry, astronomy, etc. K, /\ B needs to be approached further. There are
many hypotheses or explanations or the same natural phenomenon, and
which one can be entitled to be called ' true' is still unknown, so I had
better keep silent as to it.
K2 and K) are open to ethical questions, especially K2 /\ Band
IS NATURAL SCIENCE FREE FROM MORALITY? 263
K3 /\ B. Scientific theory consists of concept, statement, and hypoth -

esis. In the body of K 2, there are few terms and statements that are
probably inextricably linked with values. In physiology, terms such as
'normal ', 'abnormal', and 'deviant' carry value implications. In biology,
the term 'adaptation' may also be value-laden; as it is often based on
the assumption that survival and propagation are the good results which
distinguish adequate from inadequate adaptation. In medicine, for
example, the simple statement ' Smoking is harmful to human health' pre-
supposes a general value context. Most of the knowledge statements in
K2, such as knowledge which 'truly' describe the internal properties of
the human body, are value free. In respect to K3, I think that the rela-
tionship of the human to the physical world can be distinguished by
two types, one non-ethical, the other ethical. The former consists, in fact ,
of the relational properties of the physical world which manifest them-
selves in the interaction of the human and the physical world. They are
beyond human control, but will not exist without human beings, such
as colour, taste, voice, etc. Knowledge about this type of non-ethical rela-
tionship is value neutral. The latter is the extension of ethical relations
of human beings. Traditional ethical relations exist only in the direct inter-
action of human beings; this type of ethical relation of human and
physical world is actual an indirect interaction of human beings, through
the medium of the physical world. These ethical relations can be
expre ssed in the following figures.
t
i
I
I
I
tnlAitioAAl etJ,.1C41 rtll/Jion \ '-./ \

e.thicJ rel«tim. Dj- 111M 4~J. rh"sia/ >AlDrlJ
Fig. 2.
It is often asserted that nature should certainly serve man , which

brings about many global problems: the overexploiting of nature, the
exhaustion of natural resources, the destruction of ecological balance,
264 WU QIONGBING
environmental pollution, etc. The blind exploitation of nature by man

leads to the serious punishment of man by nature. The traditional fit
and proper exploitation of nature now has to exhibit some ethical com-
plexion. Knowledge of this kind of ethical relationship, perhaps most
embodied in ecology, the science of environment, etc . may be value-
laden , that is, has some value judgements and implications which tell
us what to do or what must not be done something for survival, for the
health and welfare of human beings, including future generations.
Zhongshan University,
Guangzhou, China.
BIBLIOGRAPHY
I. Introduct ion Reading in the Philosophy of Science (1980) . Prometheus Books .

2. Loren R. Graham (1981). Between Science and Values, Columbia University Press .
New York.
3. Steven and Hilary Rose (1986) . 'The Myth of the Neutral ity of Science ', in Science
and Liberation , Black Rose Books, Montreal.
4. Gerard Radnitzky (1977) . 'Science and Values: The Cultural Importance of the Is/Ought
Distinct ion', in The Search for Absolute Values: Harmony Among the Sciences, Volume
II, International Cultural Found ation Press, New York .
5. J. Bronowski (1956) . Science and Human Values, Julian Messner , New York .
6. The Ethics of Science: Issues and Controversies (1989). English translation of the
revi sed Russian text, Progres s Publishe rs, Moscow .
YIN ZHENG KUN
TRUTH AND FICTION IN SCIENTI FIC THEORY
The deb ate bet ween scientific reali sm and anti-rea lis m deals mainly
with the problem whether the term s in scientific theory genuinely refer
or not, that is, wheth er the obje cts referred to by those terms genuinely
exis t in the objective world. In a se nse, this problem may be und er -
stood as " Are those objects real entities or fictions?"
Some scientific realists insist that the referent s of the term s such as
"e lectro n", "gene" are not artifici al fic tion s but real exis tence . They
regard the idea as a basic point of view of their truth theory. As L. Laudan
has pointed out,
a real ist would ne ver want to say that a the or y was approxima te ly true if its ce ntra l
theoretical terms fai led to refer. If there were no en tities similar to atoms, no atomic theo ry
co uld be approxi ma tely true; if there were no suba tomic part icles. no qua ntum theor y
of chemi stry cou ld be appro ximate ly true. In short, a necessary co nd ition - espe cia lly
for a scientific realist - for a theory being close to the truth is tha t its ce ntral exp lana-
tory term s genuine ly refe r. I
On the contra ry, anti-realists generally doubt the exis tence of the
referents of those term s. From the logi cal empiricism to the recent con-
structive empiricism of van Fraassen, all of them regard the observability
as the reaso n for the theoretical entity is existence , and the unobserv -
able objec ts referred to by the scientific theoret ical term s as "useful
fictions" or "convenient theoretical symb ols" which the cognitive subjec t
has to introdu ce to explain the empirica l phenomen on . As an empir i-
cis t, Qu ine regard s " the conceptua l sche me of sc ience as a tool ,
ult imately, for predict ing future experience in the light of past experi -
ence." He maintains that
Ph ysical objec ts are co nce ptually impor ted into the si tuat ion as convenie nt inter -
med iar ies - not by defi ni tion in term of e xpe rience, but simply as irreducible posit s
comparable. epistemo log ically, to the god s of Homer . . . the physical objec ts and the gods
differ on ly in degree and not in kind. Both sor ts of entities enter our conception only as
cultura l posits. Th e myth of phy sical objects is epis temolog ica lly superior to most in
that it has proved more efficacio us than other myths as a device for working a manage-
able structure into the flux of experience. '
265
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.) , Realism and Anti-Realism in the
266 YIN ZHENGKUN
Some post-modernists went so far as to make on attempt to dismantle

the line between truth and fiction. According to J. Derrida, there are
fictions in the central part of truth because the contradiction between truth
and fiction is deconstructed. Another post-modernist, Hilary Lawson,
claims. "All our truth are, in a sense fictions - they are stories we
choose to believe.'? But "although truth may be a fiction, it is the most
powerful fiction we have.?"
In order to clarify the controversies listed above, this paper attempts
to make further research about the following problems, in the scien-
tific theories, what is fiction? are there fictions or not? what is the
function of fiction? what is the relation between truth and fiction ?
1. REFERENCE AND FICTION
If a creative activity is the process in which people can produce a new

form in which something exists then fiction should also be one of the
creative activities of a human being's thinking (when it refers to a process
of thinking) In some fields such as in literature and art, fiction is even
an important creative activity. But by means of fiction , people only
construct some things or events existing not in the real world (or the
objective world) but in the ideas (man's mind and language). In other
words, the concepts and statements from fiction do not refer to the
actual objects and express the actual events. In general, there are fictions
in myths and novels. Characters such as Adam, Eve, Prometheus, Othello,
Don Quixote, Sherlock Holmes and the stories about them are all ficti-
tious. They were not like the historical characters and events that had
existed as the material entities in the real world . But in the history of
science there were also many concepts such as ' phlogiston' , 'caloric',
'aether ' and so on in the scientific theories. Now we know that they
do not genuinely refer in the real world , namely, they are fictitious .
However, the meaning of 'refer ' is equivocal. As Richard Rorty
says:
The term can mean either (a) a factual relation which holds between an expres sion and
some other portion of reality whether anybody knows it holds or not, or (b) a purely ' inten-
tional' relation which can hold between an expression and a nonexistent object. 5
He calls the former 'reference' and the latter 'talking about ' . He says
that we cannot 'refer to' Sherlock Holmes but we can 'talk about' him,
and similarly for phlogiston. 'Talking about' is a notion of common sense,
TRUTH AND FICTION IN SCIENTIFIC THEORY 267
and 'reference' is a term of philosophical art. We cannot refer to fictions .

Some of the scientific philosophers have remarked the difference too.
They call the former 'genuinely refer' so as to avoid confusing it with
the other. Evidently Rorty's use of the term 'refer' is clearer than the
other. I tend to accept his usage in this paper. But it make no differ-
ence whether we use ' refer ' or 'genuinely refer', if no one knows whether
the object expressed by a term actually exists (the material existence)
or not. In this case how can we decide it does not refer to fiction?
Therefore I consider that we may claim 'refer ' only when it means a
relation between a term and a matter that actually exists in the real world ;
and 'talking about' when it means the intentional relation between a term
and a matter that is considered existing, but in fact it may not exist at
all in the real world. Thus we can draw a clear distinction between 'refer'
and 'talking about', and avoid referring to fiction , when talking about
reality and fiction as well.
Nevertheless, the clear use of terms cannot solve the puzzle whether
the scientific theoretical terms refer or not. According to Quine's view
of 'Ontological commitment', when a person talks about a thing, he
has received an ontological judgement. Any scientific theory takes an
ontological position, or possesses an ontological premise that admits or
refuses the existence of something. But "ontological commitment" does
not tell us what there is. And as Quine says,
but in order to know what a given remark or doctrine, ours or someone else's says there
is; and this much is quite properly a problem involving language. But what there is is
another quest ion."
Hence, "ontological commitment" only commits "the ideal existence"

of the theoretical entities talked about in scientific theoretical terms,
but not " the material existence", namely, it does not guarantee that
the terms refer, or the entities have homologues in the real world. It
is probable that a certain theoretical entity is fiction or the relevant
theoretical terms refer to fiction before we have sufficient reason to deter-
mine the entity "materially existent".
Then how can we determine a scientific theoretical term refers?
The anti-realistic empiricists claim that we can do so only by means of
' observability '. Bas van Fraassen says:
Science aims to give us theories which are empirically adequate ; and acceptance of a theory
involves a belief only that it is empirically adequate . . . a theory is empirically adequate
exactly if what it says about the observable things and event s in this world is true.'
268 YIN ZHENGKUN
He especially emphasizes what is observable is what can be directly

observed by means of our naked organs of sense. If we can not measure
an entity (as a special micro-entity) without any apparatus, or say we
must use a cloud chamber or any other apparatus to measure the paths
of micro-particles, then the process can not be counted observing but
detecting by means of apparatus.
Van Fraassen's point of view really is rather limited. In fact, scien-
tists do not accept his view at work. As Dudley Shapere argues, when
citing example of the solar neutrino experiment, "although the central
core of the sun lies buried under 400,000 miles of dense, hot opaque
material, astrophysicists nevertheless universally speak of the experiment
as providing 'direct observation ' of that central core.:" In fact, they
only observed the effect produced by the neutrino emitted from the central
core of the sun in the apparatus. Most of the scientists tend to consider
that , if we can explicitly infer the particular effect produced by a theo-
retical entity, such as atom, electron, neutrino, etc ., under certain
experimental conditions from the theory, and really and truly observe that
effect under above conditions, especially when we can use the effect
to expand our cognition about any other entities, even to make some arti-
ficial products in practice (for example, we use the wave character of
electrons to make the electron-microscope and observe the structure of
a virus with the help of it), then it is reasonable to say the theoretical
entity is by no means a fiction , but actually exists in the real materi al
world, and the theoretical term talking about it has reference as well.
But if we cannot do so, then however much the theory seems coherent
and convincing, scientists would hardly accept it, and they doubt the exis-
tence of the entity involved in it. For instance, some of the quantum
physicists are suspicious of any 'hidden variables' models because there
are no observable effects related to these variables which, if ever
measured, could justify/ interdict their introduction. The ' hidden vari-
ables' may be fictitious, at least under existing conditions.
If after a series of changes in theory , especially the scientific revo-
lution, there is no better reasons to owe the observable effects to anything
else, then the reason that these entities producing the above effects
actually exist, i.e. their corresponding terms still refer, is more suffi-
cient and satisfying. For instance, with the continual change of theory,
such terms as 'atom ' , 'electron' could keep their reference in a new
one , though the meanings of the terms have undergone a great change,
because of the past known effects, and could be inferred from them in
the new theory. That may strengthen our faith in the existence of these
entities.
According to the above, whether a theoretical term refer or not is deter-
mine to a great degree not only in virtue of the empirical test but also
the evaluation of the theory system (or the conceptual scheme), even
in virtue of its whole background knowledge, What is fictitious in science
is not something given once and for all, but evolves along with scien-
tific knowledge. It is possible that a theoretical entity which had been
generally regarded as having actual existence for a long time would be
fictitious . For example, in the l780s, when J. Priestley said 'dephlogis-
ticated air ' and Lavoiser said ' oxygen', both of them almost depended
on the same experiment. 'Dephlogisticated air ' did ' refer ' in the light
of the phlogiston theory in fashion at that time, and was entirely observ-
able. But according to the oxygen theory , oxygen refer s and phlogiston
is fictitious. Although 'dephlogisticated air ' and oxygen in fact talk about
the same thing, we cannot think 'dephlogisticated air' has reference
because that means phlogiston also exists, which is contrary to the oxygen
theory accepted by us. Phlogiston theory was given up not only due to
the discovery of oxygen , but also, probably the main reason, due to the
consequence of phlogiston having negative weight , which contradicts the
mechanistic background knowledge.
Let 's see another example; in the 19th c. no one doubted the exist-
ence of aether as the aether theory was a commonly accepted background
knowledge in science. People thought that a lot of observable effects were
caused by aether. Maxwell hypothesized his electromagnetic wave in
aether. He regarded the wave as the transmission of aethereal vibra-
tion . H. Hertz confirmed the aether by demonstrating the existence of
radio waves. A. A. Michelson figured out a way to interact with aether.
He thought his experiment confirmed G. G. Stoke's aether drag theory,
which was not to deny the existence of aether at that time," and so on.
In fact, scientists regarded aether as fictitious only after they had accepted
Einstein's theory of relativity as the background knowledge, and all of
the above effects had the new explanation.
2. FICTION IN SCIENTIFIC THEOR Y
Many scientists unconsciously have faith in realism that the scientific

theory constructed or accepted by them should be the real description
or explanation of matters objectively existing, and all concepts in
270 YIN ZHENGKUN
scientific theory should have one-to-one correspondence with the objec-

tive world, instead of being fictions. This point of view is so-called
'metaphysical realism' (Putnam) or 'dogmatic realism' (Heisenberg on
Einstein's position). But when we open a book of the history of science,
we can find out with surprise that science attempts at any time to actually
reflect the present world facing us and is surely to have described and
explained the connection between/among the phenomena and their rules
in an increasingly extensive, deep, exact and efficient way. On the other
hand, there are fictional pictures of the world here and there in the history
of science. The scientists unceasingly discard the old fictions and con-
struct new ones. Even in modern science, we cannot get rid of the
components which the future scientists will perhaps regard as fictitious
and they may need them more and more , including the conjecture of
unknown fields, and idealized entities or states to reduce experimental
conditions, and mathematical treatment.
Any scientific theory emerges in the form of hypothesis. That is an
essential character of science as a creative enterprise. But the hypoth-
esis itself means an incomplete theoretical form, whose fundamental
concepts and ideas are to be tested. It is probable that hypothesis involves
some theoretical entities and laws which have not undergone experimental
test but only exist in the idea, temporarily. Thus, fictions are unavoid -
able in the hypothesis. When Newton established his theory, he claimed :
"Hypotheses non fingo", and made great efforts to establish his theory
on hard data, but his theoretical central core involves at least two fun-
damental concepts: absolute space and absolute time, which modern
scientists take for obvious fictions.
It is more important that scientists often have to postulate some unob-
servable theoretical entities to illustrate , explain or predict the observable
empirical phenomena. This is a conjectural and research process, so called
"Free creation of thinking" . It is reasonable to doubt whether they exist
or not, before we have sufficient evidences to confirm them. For example,
in modern physics scientists propose a number of new hypotheses
involving many new unconfirmed micro-particles so as to unite the
known natural fundamental forces. There are various kinds of these
micro-particles such as x-particle, super-electron, Higgs boson, etc .,
and there is a dazzling collection of names only for particles that make
up leptons and quarks on the subquark level: subquark, substraton,
prequark, preon, rishon, chromon, flavon, familon, alphon , beton , gleak,
omegon, even maon for commemorating the great Chinese Mao Zedong
who believed that matter can be divided infinitely. Who can assure us
none of them is a fiction?
Moreover, some theoretical entities whose existence had been gen-
erally accepted in a rather successful theory might ultimately be proved
to be fictions by new theory. Scientists predicted a new planet between
the sun and Mercury to explain the precession of Mercury's orbit
according to Newton's gravitation theory. Owing to the great success
of Newton 's theory in predicting new planets, such as Nepture (and
then they in fact discovered Pluto in 1930 in this manner), they had no
doubt of its existence though they had not found it for 85 years. They
did not know it is an unnecessary fiction until Einstein proposed his
general relativity. Another example is the magnetic monopole. There is
up to now no known reliable direct experimental evident for its existence.
Is it possibly a fiction? We need more and new experiments, perhaps a
new theory, to answer the question.
It is a misunderstanding to think that scientists make fictions as
the result of their boundless thinking, straying far from the reality as
scientific fancy novelists. But in fact most of known fictions in scien-
tific theories are proposed not due to their brave fancy but to, at least
it seems to me, scientists' rather conservative outlook. In 1917, De
Sitter proposed a simple cosmological model of an expanding universe,
but it was not accepted because it lacked sufficient evidence. Einstein,
almo st simultaneously, added a special cosmic term to his field equations
of general relativity, so as to maintain a picture of a static universe. Later
a few scientists discovered the cosmic term was an ent irely surplus
fiction. It seems that any scientific theory must be established on certain
background knowledge including background theory and known empir-
ical data. Scientists, even such great scientists as Newton and Einstein
with super-imaginative minds, are still often handicapped by people's
cognitive and practical levels at their times. During the 1920s, A.
Friedmann and G. Lemaitre discovered expanding universe models based
on the field equations of general relativity, but they attracted little atten-
tion until E. Hubble's observational evidence was generally accepted
in 1929.
There are many similar instances in the history of science. Although
chemical revolutionist Lavoisier overturned the phlogiston theory, he
could not entirely get rid of the restrictions of the old tradition. There
were still fictions in his oxygen theory. He not only retained some
alchemical symbols, but also put oxygen in the place of principle, as
272 YIN ZHENGKUN
the acid principle (Principle oxygine); he even maintained 'caloric'

without weight.
In addition to the above mentioned fictions, scientific theories need
some theoretical terms which have evidently no actual correspondents.
They discuss the theoretical 'entities' such as 'mass point', 'rigid body',
'point charge', 'ideal gas' and 'states' such as the thermally isolated
system, which only exist in some limiting cases and which could not
actually appear at all. Shapere calls these terms 'idealization terms' (or
idealization concepts) which clearly have fictional characters and are used
as tools in the theory, in order to make them different from the so-
called 'existence concepts' or 'entity concepts'. After studies of the three
cases about the use of 'rigid body ', 'point charge' and 'mass point', he
points out
There are often good scientific reasons for distinguishing between the way in which a
certain entity is asserted to be (or not to be) and the way in which it is treated (although,
again for scientific reasons, it could not really be that kind of thing) for the sake of
convenience in dealing with certain scientific problems . 10
According to our definition in this paper, 'idealization' is, in fact, a

kind of fiction and according to Shapere a fiction with sufficient scien-
tific reasons. Let us look at what are the reasons through Shapere 's
instance of 'point mass'. Shapere says:
. . . the rationale for supposing their employment to be both useful and possible is clear.
(I) There are certain problems to be solved-problems relating to the positions, veloc i-
ties, masses, and forces of bodies . (2) Mathematical techniques exist for dealing with
such problems if the masses are considered to be concentrated at geometrical points
(namely , the geometrical techniques of Newton and, for later scientists, the methods of
the calculus). (3) It is, as Newton showed, possible to treat spherically symmetrical
bodies as if their masses were concentrated at their centers (and, incidentally, Newton con-
ceived the elementary mass-particles to be spherically symmetrical). And, as for bodies
not spherically symmetrical, they could be considered in the same way, provided the
distances between their centers was large in comparison to their radii - a condition ful-
filled, happily, by the earth-moon system. More generally, bodies could be considered
as if their masses were concentrated at their centers of gravity . II
Therefore, a term such as 'mass point' does not refer to any entity
which actually exists in nature . And it even contradicts the other prin-
ciples of Newton's mechanics, because Newton explicitly claimed, "We
conclude the least particles of all bodies to be also all extended, and
hard and impenetrable, and endowed with their proper inertia,"!' but
not the geometrical point. In classical mechanics, an infinite gravitational
potential would result from the localization of gravitational mass in a

dimensionless point. Hence, such concepts as 'mass point' and simi -
larly ' rigid body ', 'point charge' are only artificial abstract tools or
reasonable fictions in order to enhance the exact degree of mathe-
matical calculation or to reduce the physical conditions in solving some
concrete scientific problems.
From the above argument we can see that scientific theory possesses
the characters of abstractive generality and conjectures. Science, as a
highly creative enterprise, can not avoid fictions in any time, but espe-
cially needs reasonable fictions, i.e. to use the abstractive, idealizative
concepts as tools to make our cognition purer and more profound.
3. TRUTH AND FICTION
To many people, fiction is opposite to truth, and equal to falsity or

falsehood. It seems that truth could not be established in a case in which
fiction appeared simultaneously, and truth should rule out fiction. This
is a point of view of metaphysical realism. In fact, fiction may be married
with falsity as well as allied with truth .
Fiction does not make up something out of nothing, and is not wild
imagination either. Fictions, both in literature and in science, are the
results of creative thinking aiming at some rational goals, based on certain
present models or empirical data, even on certain theories in science.
Fiction in science is the fiction 'Possessing scientific reason' . In litera-
ture, the more typical the character the more successful the fiction is.
From Hamlet to Holmes, from Don Quixote to Ah Q,I3 all of them
demon strate that. In science, fiction is one of the important ways in which
hypotheses should be postulated as well as one of the important means
of scientific research. It is because scientists presuppose some entities
and their attributes and interrelations (i.e. the laws of motion and inter-
action among these entities), which have not yet been confirmed in
practice, that scientists can be guided to make further experiments and
observations, hence the new discovery. Even if these imaginations are
indeed false in the final analysis, their anticipations are so successful that
we are often justified in accepting them, and make science progress.
Sometimes with the development of science, fiction, at present, may
yet be reality in the future. The history of the aether 's rising and falling
offers us a typical case .
The concept of 'aether' appeared almost as early as ' atom'. It had been
274 YIN ZHENGKUN
a mere philosophical fiction with a mysterious colour for a long time.

If the concept 'atom' expressed the ancient Greek's imagination of
material discreteness, then 'aether' expressed a view of material conti-
nuity. 'Aether' entered modem science as 'atom' did with the European
scientific renaissance in 17th century. Gilbert, Kepler, Harvey, Descartes
and others once introduced it into their own doctrines. But only after
Huygens, Hooke and others employed it in the wave theory of light
did 'aether' become a quite necessary concept. To people at that time,
any wave transmission must be helped by a certain medium vibrating .
Since light can go through the boundless intervening space, the space
must be full of the medium - aether in every place . They thought aether
existed. On the contrary, it seemed quite fantastic that a sort of wave
could transmit itself without any medium. Scientists proposed many
predictions that could be tested in virtue of aether, which developed
not only the wave theory of light, but also the electro-magnetic theory
between the 18th century and 19th century. For example, an influential
set of non-mechanical aethers, designed to mediate electrical interac-
tion, was proposed by several physicists, including J. Canton and G.
B. Beccaria, after 1750. Such a concept assisted H. C. Oersted in the
discovery of the magnetic effect of an electrical current. M. Faraday
also owed something to this tradition .
But Maxwell and Hertz used aether to advance the electromagnetic
theory of light, and on the other hand, their theory required that aether
had to be rare enough not to impede the planets and rigid enough to
transmit prodigiously rapid transverse vibrations. These conditions were
hardly consistent with other background knowledge For example, H.
A. Lorentz explicitly freed the aether from obedience to Newton 's third
law of motion, which placed aether in difficulty. Although aether no
longer plays the roles it did in the 19th century due to the theory of
relativity now, it remains in physical theory in various guises. There
are some similar ideas in Dirac's "vacuum" theory. Dirac said: "If one
reexamines the question in the light of present-day knowledge, one
finds that the aether is no longer ruled out by relativity, and good reasons
can now be advanced for postulating an aether.,,14
From the evolution of aether we can find out that a fictional meta-
physical concept could enter science under certain conditions, and it
was not only rationally accepted, but also advanced scientific theory in
a certain period, and people could obtain some true knowledge with
the help of the concept. It is just because of the development of the theory,
with the concept as its core, that it's fictional nature was gradually
revealed. People can treat it as a completely unnecessary fiction in a
new theory (such as relativity), and equate it to falsity . As a result neither
can we neglect its value during its long existence in history, nor can
we give up research concerning what it talks about in a new light even
while aether has lost its meaning in the new theory .
According to the structure of scientific theory, the more mature a
theory is, the closer to an axiomatized deductive system it will be. The
axiom, the central part of a theory, always consists of some concepts
and statements which are self-evident or cannot be directly tested, and
it must include some indirectly observable theoretical entities in a non-
phenomenalistic theory . We can indirectly test them only by means of
the observable effects which we assume to be produced by them. In other
words, as long as we discover some laws expressed with mathematical
relations between physical concepts and use them to correctly describe,
explain and predict some observable phenomena, we can consider this
is the result of the interaction between these entities assumed by us.
But we assume and name them in virtue of the experience within the
limits of the times. In the limits (including the background knowledge
and experimental technique), as long as the theory constructed by us is
logically self-consistent and not obviously contrary to the background
knowledge and, can offer, which is more important, satisfying explana-
tions and successful predictions of empirical phenomena, and even can
be used in practical production, we can not find out whether these entities
assumed in the theory are fictitious or not. For example, people could
use caloric theory in explaining the heat transfer phenomena very well.
If they were limited to the realm of heat transfer phenomena, people could
hardly find caloric to be fictitious with their background knowledge
between the 18th century and the early part of the 19th century. But
once knowledge breaks through, they can find out that the natures of
the entities assumed by us, and the names, are errors. And they can use
other entities which obtain their natures and names in new theories to
explain the above mentioned phenomena. At this time the entities past
named by us show their fictitious natures. But this does not mean that
we have no reason to insist on ontological realism, namely, insist that
aforementioned phenomena must be the result of the interaction between
certain unobservable entities. Nevertheless, although some 'entities' such
as phlogiston, caloric, have exposed their fictitious natures, this is a
change of our ideas, a new idea instead of the old one .
276 YIN ZHENGKUN
Any truth is not abstract, ultimate absolute truth , but a concrete truth .
Truth is not the theory which covers and contains everything, but the
correct theory which is defined and used in a certain limited realm and
under some concrete conditions. No matter what is used, the tested criteria
of truth theory are the correspondence theory, the coherence theory, the
pragmatic theory or the Marxist practice theory based on the corre-
spondence theory, etc . It can only test or define the concrete truth, or
relative truth or approximate truth. We have no way to rid the concrete
truth of what the later generations of scientists may regard as fictions.
Moreover, I claim that scientific truth should contain not only knowledge
of the material structure of entities, but also, more importantly, that of
the regularity of the interactions between entities which we may not know
yet. Even if someone discovers fiction in a theory so as to claim it erro-
neou s, he can still retain some theorem or laws of the repudiated theory
in the form of mathematical functional relations in the new one. In this
sense, we can say that the truth content passed on and on to a certain
degree.
Moreover, another kind of fiction such as 'mass point', 'rigid body'
obviously not only helps us to seek truth, but also is a component of truth
with profound epistemological significance.
Therefore if we had the idea that only scientific theory free from
fictions is truth, then neither could we have truth in the history of science,
nor could we affirm that the present theory is true . This plausible view
for maintaining the purity of truth is, indeed, the view of an extreme
relativism of denying truth. On the other hand, if one asserts that our
present theory has no fiction, he expresses a God 's Eye point of view,
with dogmatism.
Provided that science is a creative enterprise searching for the
unknown, it permits man's making fictions in it; As long as truth must
be realized by means of the form of scientific theory , truth and fiction
may coexist. But it is because scientific theory involves fictions that
it is falsifiable. No scientific theory can be the eternal, ultimate truth.
Huazhong University of Science and Technology,
Wuhan, China.
NOTES
I Laudan, L. (1984). Scien ce and Values , Univ . of California Press, Berkeley, pp .

120-121.
2 Quine, W. V. (1961). From a Logical Point of View (2nd edn), Harvard Univ. Press ,
Cambridge, p. 44.
) Law son, H. and Appignanesi, L. (eds) (1989) . Dismantling Truth, St. Martin's Press,
New York, pp. xxviii .
4 lbid ., p. 130.
5 Rorty , R. (1979). Philosophy and the Mirror of Nature, Princeton Univ . Press, Oxford,
p.289.
6 Ibid., p. 16.
7 Fraassen, B. van (1980). The Scientific Image , Oxford Univ . Press, Oxford, p. 12.
8 Shapere, D. (19 84) . Reason and the Search for Knowledge , Boston Studie s in the
Philosophy of Science, vol. 78, D. Reidel Publishing Company, Dordrecht, p. 342.
9 See Lakatos, I. (1978) . The Methodology ofScientific Research Programmes, Cambridge
Univ . Press, London, pp. 73-78.
10 Ibid., p. 365 .
11 Ibid ., pp. 364-365 .
12 Newton, I. (1946). Mathematical Principle s of Natural Philosophy (Principia), Univ ,
of California Pre ss, Berkeley, p. 899 .
1) Ah Q is the principal character in a famous novel of a Chinese modern author, Lu
Xun.
14 Dirac P. (1951). Is There an Aether?, Natur e 168, 906-907.
FAN DAINIAN
NIELS BOHR AND REALISM
Niels Bohr (1885-1962) was one of the greatest physicists of the 20th
century and a rival of Albert Einstein. As an atomic physicist, Niels Bohr
was concerned about the function of language as an unambiguous means
of communicating experience related to the atomic world , and about
the epistemological lessons arising from atomic physics. His main points
of view are as follows :
There is "a fundamental limitation in the classical physical ideas when
applied to atomic phenomena" (Bohr, 1934, p. 53).
The quantum postulate attributes to any atomic proce ss an essential discontinuity or
rather individuality, completely foreign to the classical theories and symbolized by Planck ' s
quantum of action (Bohr , 1934, p. 53)
From the very nature of the matter, we can only employ probability considerations
to predict the occurrence of the individual processes (Bohr, 1934, p. 109).
But Einstein disliked it and said: "I, at any rate, am convinced that
He is not playing at dice" (Einstein, 1926).
So, Bohr developed the notion of "complementarity" and wrote:
The very nature of the quantum theory thus forces us to regard the space-time coordin ation
and the cla im of cau sality, the union of which characterizes the classical theories as
complementary but exclu sive features of the description, symbolizing the idealization of
observat ion and definition respectively (Bohr , 1934, pp. 54-55).
But, in 1958, he revised his idea and said: "The very fact quantum
regularities exclude analysis on classical lines necessitate . .. in the
account of experience a logical distinction between measuring instru-
ments and atomic objects, which in principle prevents comprehensive
deterministic description." He stressed, in the wider frame of comple -
mentarity, he did not claim "any arbitrary renunciation of the ideal of
causality" (Bohr, 1963, p. 6).
He also expressed his notion of complementarity as follows :
We are apparently forced to choose between two mutually contradictory conception of
the propagation of light: one, the idea of light waves, the other, the corpuscular view of
the theory of light quanta, each conception expressing fundamental aspects of our expe-
rience (Bohr, 1934, p. 107).
279
R. S. Cohen , R. Hilpinen and Qiu Renzong (eds .), Realism and Anti -Reali sm in the
280 FAN DAINIAN
The problem of the nature of the const ituents of matter presents us with an analo-
gous situation. The individuality of the elementary electrical corpuscl es is forced upon
us by general evidence . Nevertheles s, recent experience, above all the discovery of the
selective reflection of electron from metal crystal s, requires the use of the wave theory
superposition principle . .. Just as in the case of light, we have consequently in the question
of the nature of matter , so far as we adhere to classical concepts, to face an inevitable
dilemma which has to be regarded as the very expression of experimental evidence. In
fact, here again we are not dealing with contradictory but with complementary pictures
of phenomena, which only together offer a natural generalization of the classical mode
of description (Bohr, 1934, p. 56).
Einstein did not like Bohr's idea of complementarity. He called it

"the Heisenberg and Bohr soothing philosophy or religion" (Einstein,
1928).
At the 5th and 6th Solvay Congresses held in 1927 and 1930 respect-
fully, the debate between Einstein and Bohr was developed. Einstein
attempted to negate the uncertainty principle and the notion of comple-
mentarity with the thought experiments of the double-slit and the photon
box, but both were refuted by Bohr (Bohr, 1958, pp. 32-58, Jammer,
1974, pp. 109-158).
In 1935, A. Einstein, B. Podolsky and N. Rosen wrote their paper 'Can
quantum-mechanical description of physical reality be considered
complete?' (Einstein, 1935) proposing the famous EPR argument, and
challenged Bohr 's interpretation of quantum mechanics once again.
Einstein and his collaborators started from the locality (or separability)
assumption and demonstrated the incompleteness of the quantum-
mechanical description of physical reality. On the contrary, if the
completeness of the quantum -mechanical description were insisted upon ,
then the locality assumption would be violated and special action at a
distance (or telepathy as Einstein called it) would be permitted.
Bohr replied quickly in the Physical Review with the same title
(Bohr, 1935). Based on the feature of wholeness (or individuality) typical
of proper quantum phenomena, in Bohr's view, the completeness of
quantum-mechanical description, challenged by Einstein and his col-
laborators, can be saved.
The EPR incompleteness argument was a big impulse to the revival
of hidden variable theories in quantum mechanics. The advocators of
local hidden variable theories denied the completeness of the quantum-
mechanical description of the state of micro-objects in space-time. In
1964, J. S. Bell, in his paper entitled 'On the Einstein-Podolsky-Rosen
paradox ' proposed Bell's famous inequality which showed that any
NIELS BOHR AND REALISM 281
local hidden variable theory cannot reproduce all statistical predictions

of quantum mechanics. From the early 1970's to the 1980's, about a
dozen experiments were carried out and the results obtained seemed to
confirm the predictions of quantum mechanics and not those of local
hidden variable theories.
From the early 1950 's, under the influence of the Soviet ideology,
philosophers and physici sts in mainland China believed that Bohr is an
idealist or a positiv ist. The situation has changed since the end of the
1970's. In recent years , during the debate about the subjectivity of cog -
nition , one scholar still believes that Niels Bohr is the main representative
of the international idealist trend of thought in physics owing to the
fact that Bohr had said: in quantum mechanics, " no sharp distinction
can be made between the behavior of the objects themselves and their
interaction with the measuring instruments" (Bohr, 1958, p. 61) and
"the new situation in physics has so forcibly reminded us of the old
truth that we are both onlookers and actors in the great drama of exis-
tence" (Bohr, 1934, p. 119, He, 1990) . He also thinks, "in respect to
the interpretation of quantum mechanics, Bohr is wiser than Einstein, but,
regretfully, he cannot get the materialist conclusion from the correct
physical interpretation (He, 1991). Another scholar believes Bohr is a
subjective idealist; and owing to the experimental tests of Bell's inequal -
ities in the 1970's which demonstrated that the Copenhagen Interpretation
is correct, that thereby materialism is falsified (Jin, 1988). These two
scholars, in spite of the fact that one adheres to materialism and the other
does not, both believe Bohr is an idealist or a positivist, believe that Bohr
claimed 'esse est percipi' and " the moon (or the electron) does not exist
when nobody looks at (observes) it" . And some others think Bohr is a
phenomenologist who recognizes only the phenomenal world and not the
atomic world .
But, according to Bohr's philosophical writings, I think Bohr is basi-
cally a realist, instead of being an idealist or positivist. The reason s are
as follows .
First , Bohr did not deny the reality of the external world including
atomic objects. He recognized their existence and their objectivity inde -
pendent of the subject.
For example, early in 1929, he wrote:
In physic s, where our problem consists in the coordination of our experience of the external
world (Bohr, 1934, p. I) .
Every doubt regarding the reality of atoms has been removed and as we have gained
282 FAN DAINIAN
a detailed knowledge even of the inner structure of atoms, we have been reminded in
an instructive manner of the natural limitation of our forms of perception (Bohr, 1934,
p. 103).
In 1958, 4 years before his death, he pointed out clearly:

It is also essential to remember that all unambiguous information concerning atomic objects
is derived from the permanent works - such as a spot of a photographic plate, caused
by the impact of an electron - left on the bodies which define the experimental condi -
tions .. . The description of atomic phenomena has in these respect a perfectly objective
character, in the sense that no explicit reference is made to any individual observer
(Bohr , 1963, p. 3).
According to an article written by the former Soviet academician M.

A. Markov, Bohr had said to Professor SHv clearly: the micro-world,
micro-objects, exist independent of our consciousness (Markov, 1990).
According to another Soviet physicist, Fock, in his recollection, during
a series of conversations on the philosophic significance of quantum
mechanics in the early 1957, Bohr told Fock he was not a positivist
and attempted simply to consider nature exactly as it is and he completely
recognized the objectivity of atoms and their properties (Graham, 1987,
pp. 337-338).
Therefore, imposing viewpoints such as esse est percipi, "the moon
(or the electron) does not exist when nobody looks at (or observers)
it" on Bohr, is groundless. And in my point of view the confirmation
of the Copenhagen interpretation of quantum mechanics does not falsify
materialism or realism either.
Second, Bohr never denied the distinction between object and subject.
He only opposes the mechanical conception of nature which fixed the
object-subject separation between the object and the measuring instru-
ment. He only claimed the different placing of such a separation.
For example , Bohr wrote:
While , in the mechanical conception of nature, the subject-object distinction was fixed,
room is provided for a wider description through the recognition that the consequent
use of our concepts requires different plac ing of such a separation (Bohr, 1958, pp. 9 I-92).
He also wrote:
In view of the influence of the mechanical conception of nature on philosophical thinking ,
it is understandable that one has sometimes seen in the notion of complementarity a
reference to the subjective observer, incompatible with the objectivity of scientific descrip-
tion . Of course, in every field of experience we must retain a sharp distinction between
the observer and the content of observations, . . . In quantum physics, as we have seen,
an account of the functioning of the measuring instruments is indispensable to the defi-
nition of phenomena and we must, so-to-speak, distinguish between subject and object
in such a way that each single case secures the unambiguous application of the elemen-
tary physical concepts used in the description. Far from containing any mysticism foreign
to the spirit of science, the notion of complementarity points to the logical conditions
for description and comprehension of experience in atomic physics (Bohr, 1958, pp .
90-91).
Bohr's notion of complementarity indeed claimed:

no sharp distinction can be made between the behavior of the [atomic] objects them-
selves and their interaction with the measuring instruments (Bohr, 1958, p. 61).
But, he pointed out clearly :

The notion of complementarity does in no way involve a departure from our position as
detached observers of nature, but must be regarded as the logical expression of our
situation as regards objective description in this field of experience (Bohr, 1958, p. 74).
Here, he placed the object-subject separation between the detached

observer and the atomic phenomenon which was produced by the inter-
action between the atomic object and measuring instruments.
Bohr expressed his viewpoint more clearly in his letter to Pauli on
March 2, 1955.
A phrase like 'detached observer' . .. used in connection with the phrase 'objective descrip-
tion' . . . had to me a very definite meaning . In all unambiguous accounts it is indeed a
primary demand that the separation between the observing subject and the object content
of communication (i.e., the phenomenon to be described) is clearly defined and agreed
upon . .. this condition is indispensable in all scientific knowledge .. . It appears that
what we have really learned in physics is how 10 eliminate subjective elements in the
account of experience, and it is rather this recognition which in turn offers guidance as
regards objective description in other fields of science, To my mind this situation is
well described by the phrase 'detached observer' (Folse, 1955, p. 215).
Therefore, Bohr used the phrase 'detached observer' and clearly dis-
tinguished the object and subject, meanwhile, noticing the dependence
of atomic phenomena upon the measuring instruments, or the condition
of observation. There is no trend of idealism or positivism. What Bohr
had done was to opposed the mechanical conception of nature which
improperly extended the subject-object separation in classical physics
into atomic physics.
Indeed, according to Bohr's viewpoint, in the account of conscious
life, psychical experiences, mental activities, the subject-object separa-
tion, is more flexible and the sharp distinction between subject and object
is impossible. As Bohr wrote:
284 FAN DAINIAN
For describing our mental activity, we require , on one hand, an objectively given
content to be placed to a perceiving subject, while, on the other hand, . . . no sharp
separation between object and subject can be maintained, since the perceiving subject
also belongs to our mental content. . . . Indeed, strictly speaking, the conscious analysis
of any concept stands in a relation of exclusion to its immediate application. The neces -
sity of taking recourse to a complementary . .. mode of description is perhaps most familiar
to us from psychological problems (Bohr, 1934, p. 96).
In introspection it is clearly impossible to distinguish sharply between the phenomena
themselves and their conscious perception (Bohr, 1958, p. 27) .
It must be emphasized that the distinction between subject and object, necessary for
unambiguous description, is retained in the way that in every communication containing
a reference to ourselves we, so to speak, introduce a new subject which does not appear
as part of the content of the communication. It need hardly be stressed that it is just
this freedom of choosing the subject-object distinction which provides room for the
multifariousness of conscious phenomena and the richness of human life (Bohr, 1958,
p. 101).
It is evident, however, that all search for an ultimate subject is at variance with the
aim of objective description, which demands the contraposition of subject and object (Bohr,
1963, p. 14).
But, Bohr still insisted:

Of course, in objective physical description no reference is made to the observing subject,
while in speaking of conscious experience we say 'we think' or 'we feel' (Bohr, 1963,
p.28).
To the question whether complementarity has introduced a subjec-

tive element in physics or not, Bohr answered in his last published
paper very succinctly:
This development has essentially clarified the conditions for an objective account in atomic
physics, involving the elimination of all subjective judgement (Bohr , 1963, p. 24).
As to Bohr's statement that: "We are both onlookers and actors in

the great drama of existence" (Bohr, 1934, p. 119), he just meant that
the role of the subjective practice of human beings in changing the objec-
tive world should not be neglected. What he opposed was also the
mechanical conception of nature .
Third, Bohr is not a phenomenologist who recognized only the phe-
nomenal world and denied the atomic world .
In quantum physics, Bohr believed, the interaction between atomic
objects and apparatus
forms an inseparable part of a phenomenon. Accordingly, the unambiguous account of
proper quantum phenomena must, in principle, include a description of all relevant features
of the experimental arrangement (Bohr, 1963, p. 4).
Bohr really claimed that only these atomic phenomena (i.e ., the phe-
nomena produced by the interaction between atomic objects and
measuring instruments) can be visualized, can be accounted by the clas-
sical terms . But he also claimed, the atomic object can be described by
the mathematical formalism of quantum mechanics (Psi-function). This
is a kind of theoretical abstraction. He also insisted we use the concepts
of radiation in free space and free material particles having no interac-
tion with any apparatus.
And, Bohr wrote:
On the whole, it would scarcely seem justifiable , in the case of the interaction problem,
to demand a visualization by mean s of an ordinary space-time picture. In fact, all our
knowledge concerning the internal properties of atoms is derived from experiments on
their reactions or collision reactions, such that the interpretation of exper imental facts ulti-
mately depend s on the abstraction s of radiation in free space, and free material particles.
Hence, our whole space-time view of physical phenomen a, as well as the definit ion of
energy and momentum, depend s ultimately upon these abstractions (Bohr, 1934, p. 77).
Concerning these atomic objects, according to quantum mechanics,

we cannot visualize them to form a space-time coordination, but can make
probabilistic predictions of the possibilities of the phenomenal appear-
ances of atomic objects under certain measuring instruments (such as
corpuscular or wave pictures, space-time coordination or energy-
momentum descriptions [claim of deterministic causality]). In view of
these , Bohr did not confine himself to the domain of phenomena, he
recogni zed theoretical abstraction, the quantum description of the atomic
objects.
In 1953, Max Born had written to Bohr that he "dislikes thoroughly"
the instrumentalist standpoint of those physicists and philosophers who
hold that the goal of science is a purely observational description in which
one does not inquire what there is behind the phenomena." And Bohr
"agrees entirely" with Born's position that the mathematical symbols
of the formalism which allow us to predict the phenomenal appearances
of an atomic object express what we know about this reality behind the
phenomena (Folse, 1985, pp. 247-249).
Therefore, the Einstein-Bohr debate is not, as someone thought, a
debate between a realist and an idealist, positivist, or phenomenologist,
but a debate between a classical, mechanical , deterministic realist with
a modern, sophisticated, dialectical realist, a debate between two scholars
with different epistemic or conceptual, framework within the realism.
Someone thought that Einstein upheld realism and disliked the quantum
286 FAN DAINIAN
mechanical description, while Bohr upheld quantum mechanics and aban-

doned realism. Actually, Bohr upheld realism and quantum mechanics,
and insisted on wholeness (or individuality) instead of the separability
of the quantum systems.
In 1937, during Bohr's visit to China, Professor Shu Xingbei of the
Zhejiang University bluntly asked Bohr's own opinion of the Einstein-
Bohr debate. Bohr answered candidly that he was right and Einstein
was not (Wang, 1985). From 1970 to the early 1980's, Bohr's view-
point was supported by most experiments testing the Bell inequalities.
Just as Leon Rosenfeld said,
Untrammelled by formal schooling, guided only by the sure intuition of the investigator
of nature, Bohr rediscovered the dialectical process of cognition which had so long been
obscured by unilateral development of epistemology on the basis of Aristotelian logic
and Platonic idealism (Rosenfeld, 1979, p. 522).
The epistemic framework of complementarity is a new form of realism

with some elements of dialectics and someone think it may have some
similarities with some ancient Chinese philosophical doctrines.
A leader of the School of Names, Gong-sun Long (c320B.C.-
c250B.C.) expressed his doctrine of the separation of hardness and white-
ness, stated that: "There is a hard, white stone. It is hard when you
touch it with your hands. It is white when you look at it with your eyes .
Whiteness cannot be felt by hands, while hardness cannot be felt by eyes.
So, whiteness and hardness can be separated. The properties of matter
that can be separated are hidden in the matter. Whether we can discover
it or not, depends upon which sense organ we are using." Someone
think this doctrine and the Taoist Tao of Yin and Yang seem similar to
Bohr's notion of complementarity.
However, hardness and whiteness of a stone, Yin and Yang, coexist
and can be observed at the same time , while the two aspects of com-
plementarity do not. This is the essential distinction between Bohr's
notion of complementarity and these two ancient Chinese doctrines.
Zhongguancun,
Beijing, 100080,
China.
BIBLIOGRAPHY
Bohr , Niel s (1934, 1987). Atomic Theory and the Description of Nature, Ox Bow Press,
Woodbridge, Conn.
Bohr, Niels (1935). ' Can Quantum-Mechanical Description of Physical Reality be
Considered Complete?' , Physical Review 48, pp. 696-702.
Bohr , Niels (1955). Letter from Niels Bohr to Wolfgang Pauli , 2 March, 1955, BSC, NBA.
Bohr, Niels (1958). Atomic Physics and Human Knowledge, John Wiley, N.Y .
Bohr, Niels (1963). Essay s (1958-1962) on Atom ic Physics and Human Knowledge,
Richard Clam, Bungay, Suffolk.
E instein, Albert (1926). Letter to Max Born, 4 December, 1926, The Born-Einstein Letters,
Walker, N.Y ., p. 91.
Einstein, Albert (1928). Einstein to Schroedinger, 31 Mai, 1928, Schroedinger. Planck.
Einstein. Lorentz. Briefe zur Wellenmechanik, Springer, W ien, 1963, p. 29.
Einstein, Albert (1935). 'Can Quantum-Mechanical Description of Physical Reality be
Considered Complete?', Physical Review 47, pp. 777-780.
Einstein, Albert (1949). ' Remarks to the Essays Appearing in this Cooperative Volume ',
Albert Einstein, Philosopher-Scientist , Harper & Brothers, N.Y ., p. 669.
Folse, Henry J. (1985). The Philosoph y of Niels Bohr - The Framework of Com-
plementar ity, North-Holland, Amsterdam.
Graham, Loren R. (1987). Science. Philosophy and Human Behavior in the Soviet Union,
Columbia University Press , N.Y.
He, Zuoxiu (1990). Zhongguo Shehui Kexue 2, pp. 69-83.
He, Zuoxiu (1991). Studies in Dialectic of Nature 7, 3, pp. 28-38.
Jammer, Max (1974). The Philosophy of Quantum Mechanics, Wiley , N.Y.
Jin , Guantao (1988). My Philosophical Exploration, People's Press, Shanghai.
Markov, M. A. (1990). 'G lazami Ochevidcha, Priroda 5, p. 100.
Rosenfeld, Leon (1979) . Selected Papers ofLeon Rosenfeld (Ed. R. S. Cohen and J. Stachel
(Boston Studies Phil. Sci. vol. 21), Reidel, Dordrecht.
Wang , Gan-chang (1985). .A Profound Friendship and an Unforgettable Meeting' , Impact
of Science on Society 35, I, pp. 55-61.
HE NRY 1. FOLSE
THE BOHR-EINSTEIN DEBATE AND THE

PHILOSOPHERS' DEBATE OVER REALISM
VERSUS ANTI-REALISM
Although debates surrounding the interpretation of quantum theory are

generally thought to have some bearing on the que stion of reali sm , the
exact nature of the connection ha s remained controversial. Thus it is
common to find the anti-realist arguing that the succe ss of quantum theory
has dealt the death blow to realism, while simultaneously some reali sts
herald the quantum revolution as demanding a reconstruction of the
conceptual framework of physics in order to attain a fuller understanding
of nature 's structur e. At the center of thi s debate on physical reality
and the qu antum description of nature is the monumental encounter
between Bohr and Einstein, certainly by any standard one of the most
significant intellectual confrontations in the history of science. Yet the
preci se relation of this particular debate to the philosopher 's discus sion
of reali sm remains not surprisingly as obscure and ill-understood as the
relation ship between quantum theory and reali sm generally.
In recent years many philosophers have complained that the opposing
views in the traditional formulation of the issue of reali sm versus anti-
realism are ill drawn. This suspicion is confirmed by the fact that the
attempt to apply the philosophers' preconceived sch em a to the debate
between Bohr and Einstein on the description of microsystems in
quantum mechanics distorts that debate. Thi s pap er attempts to 'turn
the que stion around' by taking what really was at issue in the discus-
sion between Bohr and Einstein and applying it to the philosophers'
deb ate over reali sm as a way of illuminating and reformulating what
really is at issue.
I. THE R E ALISM DEBATE AND THE BOHR -EINSTEIN D EBATE
The que stion of realism was present at the birth of modern science in
the Copernican revolution. In his famous anonymous "Introduction to
de Revolutionibus" Andreas Osiander strikes three distinguishable chords
in favor of his anti-realist reading of the "hypotheses" of Copernicus's
work. First is the familiar " epistemological" theme which disengages
289
R. S. Cohen, R. Hilpin en and Qiu Ren zong (eds.), Reali sm and Anti-Realism in the
290 HENRY J . FOLSE
"truth or probability" from acceptance.' Second, there is the equally

familiar "ontological" theme that theoretical constructions are compu-
tational fictions and not descriptions of real processes." Finally there is
the somewhat less discussed "axiological" theme that the correct goal
of science must be only "to present a correct basis for calculation" and
not to "persuade anyone it is so" by constructing pictures or models of
"reality. ,,3 In subsequent episodes throughout the history of science
these anti-realist themes have recurred, generally with some selective
emphasis on epistemological or ontological issues.
On one level the debate between Einstein and Bohr does reflect these
familiar themes. After all Bohr 'accepted' quantum theory and Einstein
'rejected' it. Isn't it the epistemic question of the acceptability of quantum
theory which was at stake? Bohr is easily cast as arguing an anti-realist
line for the acceptability of the quantum description on its ability to
save the phenomena. He did indeed argue that we must abandon the
attempt to visualize atomic processes and rest content with an abstract
' symbolic ' formalism for predicting the probability distribution for a
range of possible experimental outcomes. At the same time, Einstein is
easily portrayed as defend ing the realist demand for a description which
would represent the quantum mechanical object in terms of properties
possessed apart from any observation of them.
Moreover the ontological theme of the realism debate surfaces when
the disputants disagree over what physical 'reality' can be accorded to
the properties of the systems whose behavior quantum mechanics purports
to describe. Again Bohr appears on the anti-realist side for he rules out
any attempt to describe atomic systems apart from the circumstances
in which they are said to be 'observed ' . Indeed Bohr does insist that
the interpretation of experimental phenomena as measurements of the
values of various properties of observed objects makes use of stories
that picture trajectories of particles moving through space or wave dis-
turbances propagating across a field , but Bohr denies these pictures are
pictures of observer independent microsystems and their real behavior
in space and time. Einstein appears as the classical realist precisely
because he wants to accord an "element of physical reality" to just
those unobserved properties to which Bohr denies reality.
Philosophers debating realism speak about these issues in terms of
' metaphysical truth' versus mere 'pragmatic utility' or in terms of 'real'
existing entities versus purely constructive 'fictions'. This way of posing
the issues arises from the philosopher's tendency to see the scientists'
DEBATE OVER REALISM VERSUS ANTI-REALISM 291
acceptance of theories in their ability to lawfully connect groups of empir-

ical phenomena. However, in providing such lawful connections an
abstract theory can be interpreted by a model which may enable one to
tell a 'story' about how empirical phenomena are produced by the
behavior of unobservable' theoretical' entities, their properties, and inter-
actions. Thus the issue of realism is transformed into the issue of the
truth of these stories and the reality of the players in them. On the one
hand, for the anti-realist philosopher the criteria for theory acceptance
warrant asserting as true only those statements which can be confirmed
by empirical phenomena. Whatever role the stories might have in theory
acceptance is purely pragmatic and has nothing to do with their possible
or probable truth in the metaphysical sense of describing the properties
of an objective reality . The realist philosopher, on the other hand, con-
siders the criteria for theory acceptance as providing rational warrant
for holding that at least some theoretical statements do truly refer to
an objective world which possesses the properties we predicate of it inde-
pendently of its relationship to empirical phenomena.
However, when we inspect more closely the way in which Bohr and
Einstein debated the issues of "theory acceptance" and "physical reality",
what we find is that they did not speak about these issues in this typi-
cally philosophical way." We find absolutely no direct reference to the
'truth' of theoretical claims, nor would it be helpful to try to 'trans-
late' their debate into a question over the truth conditions for theoretical
statements, for like most scientific debates 'truth' is a parameter irrele-
vant to acceptability. Moreover, philosophers who speak of the realism
debate as concerned with the ontological question of the existence of
"theoretical entities", are surely not concerned with the same issue that
arises in the opposition between Bohr and Einstein, both of whom were
clearly convinced realists with respect to the independently real existence
of atomic systems.' These facts suggest that if we attempt to stage this
confrontation on the philosopher's battle plain of realism versus anti-
realism we will inevitably misconstrue the problem with which it deals.
Therefore, here I propose not to translate the Bohr-Einstein debate into
the preconceived terms of the philosophers' debate, but to reflect on what
Bohr and Einstein actually debated as a way of formulating a question
which philosophers concerned with realism might find of interest. So
we tum now to consider what was in fact actually maintained by each
disputant.
292 HENRY 1. FOLSE
2. RECENT CHANGES IN THE PERCEPTION

OF BOHR AND EINSTEIN
The traditional perception of the Bohr Einstein debate presents Bohr as

a phenomenalist who banishes the objective description of 'physical
reality ' from physics and preaches the instrumentalist message that the
goal of physics is to "save the phenomena". Furthermore, the basis of
Einstein's objection is often taken to be the indeterminism of quantum
mechanics. Nevertheless, if we ignore this anecdotal history of the text-
books and follow the more recent research on Bohr and Einstein, it
becomes obvious that Bohr is miscast when made to play the typical
philosophical anti-realist's role and that, at least by the time of EPR,
determinism was not the basis of Einstein's objection.
In previous publications I have argued that Bohr should be termed a
'realist' because his arguments are based upon what he considered to
be a discovery about the real nature of microsystems, namely that they
had to be described as changing their classical mechanical states dis -
continuously." His objection against predicating the classical empirical
properties to objects independently of their observation in experimental
phenomena is derived from this physical discovery, not from any instru-
mentalist line that the criteria of theory acceptance warrants accepting
only what can be empirically verified as true . Because Bohr speaks of
different experiments as revealing complementary phenomenal appear-
ances of the same physical system, the framework of complementarity
requires we accord some transphenomenal reality to the atomic systems
whose complementary phenomenal appearances the theory predicts so
well.
As the tense debate s between Bohr and Heisenberg in 1927 reveal, the
opposing views were the young Heisenberg 's truly anti-realist hope to
dispense with classical models and rest content with a predictive for-
malism versus Bohr's view that understanding the experimental
phenomena which confirm the formalism required using the classical
descriptive models to ' interpret' these phenomena as observations of
the properties of the systems thereby described." The lawful connec-
tions between phenomena are made through the use of particle and
wave stories that are expressed in terms of the properties to which the
classical concepts refer and which provide the only way we have for inter-
preting the phenomena as observations in which one determines the
measured value of such properties. Bohr tells us we cannot read these
stories 'realistically' as visualizable pictures of transphenomenal objects.

The physical discovery that microsystems change state discontinuously
makes it "ambiguous" to attribute such properties to objects apart from
their phenomenal manifestations in observational interactions.
Nevertheless, Bohr nowhere denies reality to the microsystem as the
same entity which appears in a 'preparation' phenomenon and a 'detec-
tion ' phenomenon and is the basis of the story by which we understand
(not merely predict) the lawful connection between these phenomena.
Moreover he must presuppose the independent reality of the microsystem
when he speaks about that which could appear phenomenally as mani-
festing either complementary wave-like or particle-like properties,
depending on the nature of the apparatus with which it interacts to
create the experimental phenomenon described as an "observation"
of such properties." While this argument for calling Bohr a 'realist' is
my own, I would venture to claim that other recent interpretations of
Bohr at least agree to the extent that it is scholarly anachronistic to
characterize Bohr's outlook as a rejection of the concept of physical
reality.
New research on Einstein has also changed the traditional percep-
tion of his rejection of quantum theory. The work of Don Howard and
Arthur Fine has shown that at least by the time of EPR, the source of
Einstein's discontent with the quantum description had little to do with
determinism and a lot to do with what is now called "separability"."
The assumption of the principle of separability, that space-like sepa-
rated systems can always be assigned separate mechanical states, was for
Einstein an ontological presupposition so necessary for individuating
distinct systems that he could not conceive abandoning it. Assuming
separability, the EPR argument leads to the conclusion that physical
reality contains elements which quantum theory, on Bohr's interpreta-
tion, prohibits attributing to the system, thus justifying rejection of
quantum theory's 'completeness '. But as Bell's work has shown us,
separability all by itself has empirical consequences which differ from
quantum theory, and here the evidence seems to tell against separability
in favor of the quantum description. 10
Thus in the perception of more recent research , the difference between
Bohr and Einstein consists in this: for Einstein, an ontological presup-
position, essential for individuation, constrains what can be accepted
as a 'complete' description of physical systems, leading to his rejec-
tion of quantum mechanics as complete. For Bohr, the discontinuous
294 HENRY J. FOLSE
change of state of a microsystem in an interaction implies that predicating

the classical state defining properties to an object independently of obser-
vation is 'ambiguous' and hence cannot be the basis of an objective
description. For Einstein ontological constraints determine the accept-
ability of empirically based theory; while for Bohr, the conditions
governing the well-defined predication of empirically defined proper-
ties constrain permissible ontological presuppositions regarding the
objects to which such properties are predicated.
3. THE AXIOLOGY OF INQUIRY AND THE

BOHR EINSTEIN DEBATE
The Bohr Einstein debate may be seen as concerned with the axiology
of atomic physics. II Though their discussion prima facie concerned the
proper description of the familiar gedankenexperiments, their different
views reflected a deep division over whether or not the aim of atomic
physics had been achieved, expressed in terms of the question "Can
Quantum Mechanical Description of Physical Reality be Considered
Complete?" Their opposing answers were based on a disagreement over
the necessary conditions for predicating properties to the systems
described by atomic physics. We can understand why this disagreement
arose by considering how the physicist's philosophy of nature, i.e.,
one's view of the relationship between ontology and empirical science,
affects the choice of goals and values. Though they finally agreed that
the theory was 'successful' in attaining the goal of empirical adequacy,
their different views in the philosophy of nature led them to disagree over
its final acceptability.
Here I want to alter our perception of the debate by stressing that
they held in common the typically realist goal of securing 'harmony',
'unity' , and 'consistency' in the conceptual framework physicists employ
for describing phenomena. An anti-realist needs pursue no goal other
than empirical adequacy. Both of our disputants of course valued pre-
dictive success, but for neither was this value paramount. Einstein's
confidence in general relativity apart from its successful predictions is
well known, and Bohr's opposition to Heisenberg reveals he accorded
much higher value to understanding the phenomena than merely getting
the numbers right. Bohr and Einstein were both involved in the philos-
ophy of nature, and accepted the typically realist goals of understanding,
as opposed to merely predicting, phenomena. In order to secure the
goal of a harmonious understanding of natural phenomena both were

prepared to reweave the conceptual nets by which the natural philoso-
pher working in the empirical tradition of physics hopes to capture nature.
Indeed if Bohr and Einstein were as far apart in their goals as are the
realist and the anti-realist, it becomes astonishing that Bohr insists on
arguing that the conceptual revolution required in quantum physics is
analogous to the one Einstein precipitated with relativity."
Perhaps more remarkable than the fact that both physicists held values
that typify a realist axiology is the fact that both men also shared another
goal which typically motivates the anti-realist. In seeking to fashion
new perspectives for the description of nature Bohr and Einstein both
followed the mathematician's drive to reduce fundamental assumptions
to the simplest set by means of pruning claims to the truth of excess meta-
physical assumptions. For Einstein this was the presupposition of absolute
space-time as the locus of the ether; for Bohr it was the assumption
that an independent physical reality can be described in terms of prop-
erties which can be predicated unambiguously only of systems observed
in physical interactions.
Consequently, while it is certainly correct to say that each physicist
held to a different standard for a complete description, and hence had
different goals for atomic physics, it is misleading to say that one's
goals were 'realist' goals and the other's those of an 'anti-realist'. A more
revealing description would point to their different positions regarding
the relation of ontology to empirical knowledge. The primacy Einstein
accorded separability was a result of his conviction that only in this
way could physics individuate and accord separate identity to the physical
systems it describes. Thus in Einstein 's philosophy of nature, ontolog -
ical commitments constrain the goals of empirical knowledge , and failure
to attain those goals reflects negatively on the final acceptability of any
theory which aspires to empirical knowledge. For Bohr the value
accorded accepting an empirically adequate theory capable of describing
all possible phenomena forces even our most deeply held ontological
commitments to submit to revision ." But while empirical adequacy may
well be the ultimate anti-realist value, Bohr does not take the pursuit
of this aim to imply abandoning all ontological commitments. Instead,
for him it clearly implies the need to revise ontological commitments
in order to accord with what attaining an empirically adequate theory
requires us to assume about the phenomena, namely that they arise in
the interactions between observing instruments and independently real,
296 HENR Y J . FOLSE
physical microsystems which possess some identity between their phe-

nomenal appearances.
However, even though complementarity must make reference to the
microsystem apart from its phenomenal appearances, Bohr fails to explain
how we are to individuate such systems after abandoning separability.
Indeed, while Bohr certainly wants to accord reality to atomic systems
apart from our observations of them, his reluctance to engage in onto-
logical reconstruction delimits an unfinished sector of the framework
of complementarity. Einstein's values make his ontological commitments
clear, but he is left without any acceptable theory, even though he has
to admit reluctantly that an empirically adequate one is at hand . Bohr can
lay claim to an acceptable theory, but the ontological implications of com-
plementarity have yet to be made clear.
The Bohr Einstein debate arises from the need of physics to attain two
goals: one is the acceptability of a theory which, as it appears, is in accord
with all possible empirical evidence; the other is the demand for a
description which is 'complete' against a standard posed by prior onto-
logical presuppositions. Different physicists may prioritize the value of
attaining these goals differently, and thus be said to adhere to different
axiologies, but this is not the same as debating realist versus anti-realist
axiologies, or at least not the way philosophers have debated it. How
are we to decide, when these goals conflict, which is to take priority,
and which is to be refashioned? What reason do physicists give for the
preferences they display? Here perhaps philosophers can find an issue
which could be discussed more profitably than the way in which realism
and anti-realism traditionally have been counterposed.
Loyola University ,
New Orleans ,
U.S.A .
NOTES
I Osiander warns the reader that ".. . it is not necessary that his [Copernicus's] hypotheses
be true , nor even probable. This alone suffices that they provide a computation that
tallies with the observations" (Gingerich (1975), p. 30 I).
2 Osiander addresses specifically those " ignorant of geometry and optics" who might
"perchance" think the "epicycle of Venus as likely" . He also actually argues that other
observations than those the epicycle is conjectured to predict rule out the reality of any
such motion. Ibid., pp. 301-304.
J Osiander makes it clear that anyone who mistakenly "takes models - with their own
purpose - to stand for reality . . . leaves the discipline more ignorant than before he entered
it". Ibid., p. 304.
4 When reading Bohr 's account of these discussions, one cannot help feeling that this
bears little or no resemblance to the philosopher's debate over realism; cf. Bohr (1949) .
5 No one doubts Einstein 's realism, but many seem to forget that Bohr clearly proclaimed
that "every doubt regarding the reality of atoms has been removed", and that "we have
gained a detailed knowledge of the inner structure of atoms", includ ing not only the
behavior of aggregates of atoms, but also of single atoms (Bohr (1958) , p. 103). For further
analysis of this theme see Folse (1985), pp. 224-227; and (1987), pp. 161-167.
6 See Folse (1985), (l986a), (I 986b), (1987), and (I 989b).
7 Heisenberg (1949) , pp. 103-107; and (1963), pp. 13-17 .
s As he does , for example, at Bohr (1958), p. 26.
9 Howard (1985), and Fine (1986), pp. 26-39. I have discussed the connection between
separability and Bohr 's response to EPR in Folse (1988) and (1989a) .
10 This theme appears in the papers of Howard , Wessels, Shimony , and Teller in Cushing
and McMullin (eds.) (1989) .
II I have analyzed the debate from this point of view in Folse (1990) , for which I am
much indebted to work in the NEH Summer Seminar conducted by Larry Laudan in
1989; see Laudan (1984) for his views on the axiology of inquiry.
12 One might object that this is only a strategic move Bohr makes ad hominem for the
sake of convincing Einstein; however, Bohr strikes this theme from the beginning of
his arguments for complementarity, well before Einstein's opposition could have been con-
sidered ' entrenched'. The fact that Bohr clearly believed that his reasoning in many
ways paralleled Einstein 's made his failure to convert Einstein all the more frustrating,
and this was felt as especially frustrating just because they held many goals in common .
IJ Bohr makes his values obvious when, commenting on their disagreement, he notes
that " In my opinion, there could be no other way to deem a logically cons istent for-
mal ism as inadequate than by demonstrating the departure of its consequences from
experience or by proving that its predictions did not exhaust the possibilities of obser-
vation , and Einstein 's argumentation could be directed to neither of these ends ". Bohr
(1949) , p. 229.
BIBLIOGRAPHY
Bohr, Niels (1935) . ' Can Quantum Mechanical Description of Physical Reality Be
Considered Complete ? ', Physical Review 38, pp. 696-702.
Bohr, Niels (1949) . ' Discussion with Einstein on Epistemological Problems in Atomic
Physics ' , in P. A. Schilpp (ed.), Albert Einstein: Philosopher-Scientist, Open Court,
LaSalle, Illinois , pp. 199-241.
Bohr, Niels (1958). Atomic Physics and Human Knowledge, J. Wiley , New York.
Cushing , James T. and McMullin, Eman (eds.) (1989). Philosophi cal Consequences of
Quantum Theory, University of Notre Dame Press, Notre Dame.
298 HENRY J. FOLSE
Fine, Arthur (1986). The Shaky Game: Einstein, Realism and the Quantum Theory, The
University of Chicago Press , Chicago.
Folse , Henry (1985). The Philosophy ofNiels Bohr: The Framework of Complementarity,
North Holland Physics Publishing, Amsterdam.
Folse, Henry (1986a). 'Niels Bohr , Complementarity , and Realism', in A. Fine and P.
Machamer (eds.), PSA 1986: Proceedings of the Biennial Meeting of the Philosophy
of Science Association, Vol. I, Philosophy of Science Association, East Lansing,
Michigan, pp. 96-104.
Folse, Henry (1986b) . ' Complementarity and Scientific Realism' , in P. Weingartner and
G. Dom (eds.), Foundations ofPhysics, Holder-Pichler-Tempsky, Vienna, pp. 93-10 I.
Folse, Henry (1987). 'Realism and the Quantum Revolution' , in Abstracts of the 8th
International Congress ofLogic, Methodology, and Philosophy ofScience, Vol. 4, Part
I, Institute of Philosophy of the Academy of Sciences of the USSR , Moscow, pp.
199-200.
Folse , Henry (1988) . 'Niels Bohr's Concept of Reality ', in P. J. Lahti and P. Mittelstaedt
(eds.), Proceedings of the Symposium on the Foundations of Modern Physics: The
Copenhagen Interpretation 60 Years After the Como Lectu re - Joensuu, Finland,
August 6-8, 1987, World Scientific Publishing , Singapore, pp. 161-179.
Folse, Henry (I 989a). 'Bohr on Bell' , in Cushing and McMullin (eds.) (1989), pp. 254-271.
Folse , Henry (I 989b) . ' Complementarity and Space-Time Descr iption', in Menas Kafatos
(ed .), Bell' s Theorem, Quantum Theory and Conceptions of the Universe, Kluwer
Academic Publishers, Dordrecht , pp. 251-259.
Folse , Henry (1990). 'Laudan 's Model of Axiological Change and the Bohr-Einstein
Debate ', in Fine, A. et al. (eds .), PSA 90, Vol. I , The Philosophy of Science
Association, East Lansing , pp. 77-88.
Folse , Henry (1992). ' Complementarity and our Knowledge of Nature ', in M. Carvallo
(ed.), Nature, Cognition, and System, Vol. 2, Kluwer Academic Publishers, Dordrecht,
pp.51-66.
Gingerich, Owen (ed.) (1975). The Nature ofScientific Discovery, Smithsonian Institution
Press, Washington, D.C.
Heisenberg, Werner (1949). ' Quantum Theory and Its Interpretation' , in S. Rozental
(ed .), Niels Bohr: His Life and Work as Seen by his Friends and Colleagues, North-
Holland , Amsterdam , pp. 94-108.
Heisenberg, Werner (1963). Interview with Thomas Kuhn, Munich, 25 February , 1963,
in Archive for the History of Quantum Physics, Tape 52a.
Howard, Don (1985). ' Einstein on Locality and Separability', Studies in History and
Philosophy of Science 16, 171-201 .
Laudan, Larry (1984) . Science and Values, University of California Press, Berkeley .
HE ZUOXIU
ON THE EINSTEIN, PODOLSKY

AND ROSEN PARADOX AND THE
RELEVANT PHILOSOPHICAL PROBLEMS
I. INTRODUCTION
The controversy on " Can the Quantum Mechanical Description of

Physical Reality be Considered Complete?" was everlasting for almost
sixty years [1,2]. A clarification was made by J. Bell [3], who showed
that any local, deterministic theory leads to a relation called the Bell's
inequality, whereas the quantum mechanics violates the inequality. From
1972-1982, different kinds of experiments performed by experimental-
ists, gave results that indicated the violation of the Bell's inequality
and were in agreement with quantum mechanics [4]. That the Bell
inequality is violated now is generally accepted .
In 1979, B. d'Espagnat gave a detailed analysis of the philosophical
problems pertinent to this topic. Finally, he drew the conclusion, the
doctrine that the world is made up of objects whose existence is inde-
pendent of human consciousness turns out to be in conflict with quantum
mechanics and with facts established by experiment [5]. Similarly N.
D. Mermin gave an alternative representation of this point of view, "we
now know that the moon is demonstrably not there when nobody looks"
[6].
Are their conclusions wrong or correct ?
II. D 'ESPAGNAT 'S ARGUMENTS
B. d'Espagnat's arguments can be expressed by the following sketch ,

which is taken from "The Quantum Theory and Reality" [5].
He showed that local realistic theories originate from three premises :
"One is realism, the doctrine that regularities in observed phenomena
are caused by some physical reality whose existence is independent of
human observers. The second premise holds that inductive inference is
a valid mode of reasoning and can be applied freely, so that legiti-
mate conclusions can be drawn from consistent observations. The third
premise is called Einstein separability or Einstein locality and it states
299
300 H E Z U O X IU
?
I REALISM f
I
?
:> I INDUCTION f
Ul
W
s
Ul
W
c:
Cl.
i ? ") f SEPARAB ILITY
Ul
w ,
I i
LOCAL REALISTIC
THEORIES
I
I
2
0 I QUANTUM MECHANICS I
!!! I
;::
I I
! J
l
Ul
z
0
BELL INEQUALITY ;:: I BELL INEQUALITY
o
I
I
IS VALID C IS SOMETIMES INVALID
w
a:
Cl. I I
t ~
EXPERIMENTAL
TEST
RESULT
Fig. 1.
ON THE EINSTEIN, PODOLSKY AND ROSEN PARADOX 301
that no influence of any kind can propagate faster than the speed of
light" [5].
In short, these premises can be identified as: (1) realism, (2) the free
use of induction and (3) separability.
Since the logical consequence of the local realistic theory is the Bell
inequality, and there is strong experimental evidence that the inequality
is violated, thus , at least one of the three premises of local realistic
theories must be false.
d' Espagnat argued, that (l) "in a physics experiment, the separa-
bility assumption expressed the intuitively reasonable idea that the spin
components of one proton have no influence over these of the other
proton, provided the two particles are far apart , . . . this assumption
must now be regarded as highly questionable". (2) If an unbiased and
large amount of samples were tested, "the confidence of these asser-
tions approaches certainty as the size of the sample increases". Hence,
he leads to the conclusion: the "quantum mechanics curiously disagrees
with the doctrine that the world is independent of mind".
III. D'ESPAGNAT 'S LOOPHOLE
However, if we examine the proof of Bell's inequality more carefully,

the assumption of realism really is one of the premises of a local real -
istic theory, but this premise is only a special form of realism, the
deterministic realism, i.e. the existence of a hidden parameter. So that
the violation of Bell's inequality can not be regarded as a violation of
realism in general, e.g . a general statement, such as "disagreeing with
the doctrine that the world is independent of mind "!
IV . REALISM IS ONE OF THE PREMISES OF

QUANTUM MECHANICS
In this history of philosophy, there were different definitions about

realism. What we shall adopt, is
Realism = the existence of world is independent of mind .
In order to show that the realism is one of the premises of the quantum
mechanics also, we shall analyze the following example, the spin cor-
relation of a pair of electrons or protons, in detail.
(i) Assume we have some composite system, composed of the elec-
trons or protons A and B, with total angular momentum equal to zero.
302 HE ZUOXIU
(ii) The spin correlation function E(a , b) is defined as

(1) E(a,b) == 4Au • Bb,
where Au and B; are the spin component of particles A and B along the
direction of unit vectors and b. a
(iii) According to quantum mechanics, one easily shows, that
= (O+IOA . a ° bIO+),
°
(2) E(a, b) 8 '
in which 0A and 8 ' are the Pauli spin operators of the particules A
and B respectively, and the wave function 10+) is
1
(3) 10+) = .""2 ('V+l(A)'IU(B) - 'V_l(A)'V+l(B»
v~ 2 2 2 2
and the 'V's are the spin wave functions.

(iv) Direct calculation shows
(4) E(a,b) = -a . b
However, J. Bell started from local and deterministic theory, i.e., exis-
tence of some hidden parameter, he obtained the relation
(5) IE(a. b) - E(a , 2)1 ~ 1 + E(b, 2)
in somewhat different form from quantum mechanics.
(i) If one lets a = b = 2 = n = unit vector, then
(6) E(n,n) ~ -1,
which is compatible with quantum mechanics.
a
(ii) If the angles lie between and b. band 2 is 60°, and angle lies
a
between and 2 is 120°, then according to quantum mechanics, one
has E(a, 2) = -t,
E(a,b) = E(b, 2) = t,
substitute them to the Bell's
inequality, one has
(7)
I +-I
-I
2
I
2
~ I - -I ::::) I
2
~ -I !?
2
Now we trace back to the procedure of the derivation of the spin
correlation. One can find that the spin correlation E(a, b) = -a· b is essen-
tially originated from the special form of spin wave function
1
(8) 10+) = '2 ('V+l(A)'V_l(B) - 'V_l(A)'V+l(B»
v~ 2 2 2 2
ON THE EINSTEIN , PODOLSKY AND ROSEN PARADOX 303
where the spin component +~ of particle A is always combined with

the spin component -t
of the particle B and vice versa, i.e. the (+t) +
(-t) = (-t) + (+t) == 0, because of law of conservation of angular
momentum.
All these are objective, independent on the human mind, it could exist,
even far before the appearance of the mankind.
V . THE ROLE OF THE APPARATUS
It should be clarified, whether the spin correlation is arised or origi-

nated from the interaction between the particles and apparatus or not.
This is related to the roles of apparatus in the theory of quantum
mechanics.
As the theorem of measurement shows: the physical quantity L of
certain particle system is equal to
(9) L = I 1e;l2 t;
where L is the corresponding operator of the physical quantity, L, is the

eigenvalue of the operator L, with the eigen function "'i(X), i.e.,
(10) L"'i(X) = Li"'i(X),
the c, is the coefficient of expansion of the wave packet ",(x) in the

basis of the eigenfunction "'i(X), i.e.
(II) ",(x) = I Ci"'i(X),
The value of L; can be measured and obtained by some apparatus, while

the Ic;1 2 is the transition probability, i.e, the wave packet collapse to
", ,(X) . Ictl 2

2
(12) ",(x) ~ ~2(X) with th.e transition Ic21
{ : probability as :
"'n(x) Icn/2,
which can be measured by experiments also. This theorem of measure-
ment clearly answers: even if the apparatus interferes with the wave
packet of the particles system, this theorem still guarantees how the
real value of physical quantity L of the original quantum system could
be found!
304 HE ZUOXIU
It seems difficult to understand what is the essentiality of the collapse

or the reduction of the wave packet. Actually, the collapse of the wave
packet is a general phenomenon occurring in any statistical measurement.
For example, a "die" will collapse to certain number, ego the "red 4",
when the "die" hits on the table, while before this acting, all the number
1, 2, 3 . . . , 6 is indefinite, each of them has a probability of 1/6. Although
the table disturbs the original state of the "die" , it does not prevent the
objective of studying the probability of the appearance of the "red 4" .
When we apply the same theorem to the process of the spin correla-
tion measurement, the wave packet of this composite system
1
10+) = '2 ('l'+I(A)'lU(B) - 'lU(A)'l'+I(B))
v~ 2 2 2 2
will collapse to
'l'+I(A)'l'_!(B)
2 2 (Ar=~
with the transition
(13) =>
{ probability as
'l'_!(A)'l'+!(B) ...L)2 = 12
(-{'i.
2 2
The peculiarity of the collapse of the wave function of this composite

system, is that the eigenfunctions should be taken as the products of
'l'+!(A)'l'
2
I(B) and 'l'- 1I(A)'l'+I(B).
-2 1
This proce ss of measurement is like a special "die" in the following
form .
Fig. 2.
with the following law of conservation. i.e.

(14) Arabic No + Chinese No. = constant = 7
ON THE EINSTEIN, PODOLSKY AND ROSEN PARADOX 305
VI. ON THE MECHANISM OF THE COLLAPSE

OF THE WAVE PACKET
Of course, the collapse of the wave function in quantum mechanics is

different from the classical "dice", since the former is always accom-
panied with "the disappearance of the interference". This had been
explained, for instance, by a successful theory of measurement, devel-
oped by A, Daneri, A. Loinger and G. M. Prosperi in 1962 [7], The
essential of this theory is that the detectors are made from certain
thermodynamical quasi-stable systems, the collapse of the wave func-
tions occurs during the interaction between the particles and the
apparatus. It can be shown that the inference terms are multiplied by
some numerical factors practically zero, while the non-inference terms
are multiplied by a factor 1, both of which originate from the variables,
which describe the apparatus.
That is the role the apparatus plays, i.e., the confirmation of the spin
correlation of the quantum mechanical system, a reflection of the physical
reality which is independent of the human consciousness.
VII. CONCLUSIONS
(1) Quantum mechanics is compatible which objective realism i.e ., the

world is made up of objects, which are independent of human con-
sciousness.
(2) Deterministic realism is not supported by the experiments done
in quantum mechanics. The realism of quantum mechanics should be
interpreted as local and stochastical or statistical realism.
Zhongguancun, lluangzhuang,
Beijing, China.
BIBLIOGRAPHY
I. Einstein , A., Podolsky, B. and Rosen, N. (1935). Phys. Rev. 47, p. 777.
2. Bohr, N. (1935) . Phys. Rev. 48, p. 696.
3. Bell, J. (1964). Physics 1, p. 195; (1966) . Rev. Mod. Phys. 28, p. 447.
4. For example, Kasday, L., Ulman, J. and Wu, C. S. (1975) . Nuovo Cimento Lett. 825,
p.633.
5. d'Espagnat, B. (1979). Scientific American 241(5), p. 128.
6. Mermin, N. D. (1981). J. Phi/os. 7, p. 397.
7. Daneri, A., Loinger, A. and Prosperi, G. M. (1962) . Nucl. Phys. 23, p. 297.
HONG DINGGUO
ON THE NEUTRAL STATUS OF QM IN

THE DISPUTE OF REALISM VS . ANTI-REALISM
There is a widely spread impression that reali sm has been swept away
from the reign of physical science once and for all since quantum
mechanics came into being. This implies that QM provides a decisive
support to anti-realism and is not compatible with any reali sm, whichever
form it takes. However, such an op inion is unable to hold its position.
In fact, QM is neither a menace to realism (or anti-realism) nor a decisive
support to anti-realism (or realism). That is to say: the status of QM in
the dispute of realism vs. anti-realism is neutral.
ON THE ONTIC REALITY AND THE EMPIRICAL REALITY
The word 'reality' frequently appears in philosophy literature. Its

primitive meaning is 'existing being'. Owing to its various implications
in different contexts, this word is in fact of multiple meanings. To make
the discu ssion proceed clearly, we clarify two different notions about
the ontic reality and the empirical reality.
The ontic reality is the reality which is external and independent of
human being s, having strong objectivity. One of its essential character-
istics is its transcendentality. Naturally, those philo sophers who think that
the only source of knowledge is experience will hold that the ontic reality
is unrecognizable, and so deny that the notion of ontic reality pos sesses
any epistemological significance. In contrast, holders of reali sm (refer-
ring to onti c realism) think that the notion of the ontic reality possesses
an epistem ological implication. They insist that although the ontic reality
is beyond reach of experience, it is within reach of mind. Human being s
can explore ontic reality by means of intuition and insight produced by
the free creation of the mind, and grasp it in the form of scientific
concepts. As Einstein said [I], "physics is an attempt conceptually to
grasp reality as it is thought independently of its being observed." No
doubt, the exploration and grasp of ontic reality can be convincing only
when it is in concord with experience. In fact , the notion of on tic reality
is continually sublimated and deepened in the interactions between the
free creation of mind and the verification of experience. Trying to explore
307
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.) , Realism and Anti-Reali sm in the
Philosophy of'Sc ience, 307-316.
© 1996 K/uwer Academic Publishers.
308 HONG DINGGUO
the ontic reality is in the nature of mind. Since the ontic reality transcends
the sum of all experience and knowledge of mankind at any time, such
a process will never end .
The empirical reality is the reality which depends on subjects and
cannot be separated from experience so as to have the weak objectivity
characterized by intersubjectivity. Emp iricism (i.e. empirical realism)
insists that science (and ordinary knowledge as well) is indissolubly
linked with human experience, so that the task of science is but to
describe the phenomena constructed by the collective experience of
human beings .
Obviously, empirical realism contains a thick atmosphere of instru-
mentalism (or operationalism). However, d'Espagnat correctly points out
that there is a difference between empiricism and strict instrumentalism.
To offer an image, empiricism is a boat that is sailing between ontic
realism and pure instrumentalism, and completes the instrumentalism
with ideas about empirical reality as natural as possible.
We suggest an open-minded philosophy under which the ontic reality
and empirical reality are two complementary notions existing side by side
with their own implications and functions respectively. On the one hand ,
we confess that intersubjectivity is the foundation of all empirical
sciences, and the notion of empirical reality is of paramount impor-
tance to practical sciences. On the other hand, we believe that the ontic
reality potentially dominates the activities of mankind and the progress
of sciences. Although the ontic reality is always hidden, its various
surface structures can be perceived by human beings in the form of empir-
ical reality. And its strong temptation for human beings is the eternal
motivation of the development of sciences.
THE MINIMAL STRUCTURE OF QM AND ITS EXTENSION
Any complete physical theory is a system with four dimensions: its

mathematical formalism, the operational definitions of its basic concepts,
its interpretation, and its unfolding [2]. The first two dimensions con-
stitute the minimal structure of a physical theory. It is a formal but
operational algorithmic system, which simply and effectively describes
the relevant physical phenomena. The last two dimensions constitute a
philosophical extension to the first structure to manifest any physical
meaning it may have. Such an extension is a self-contained system of
metaphysical ideas which is compatible with the first structure.
ON THE NEUTRAL STATUS OF QM 309
What on earth is the minimal structure of QM? That is the quantum

algorithmic system centered around the concept of probability amplitude.
In Bohr's words [3], "the whole formalism (i.e. QM) should be regarded
as an instrument, by which the predictions with definite or statistical
information could be deduced under certain experimental conditions
described in classical terms."
The minimal structure of QM has the following features: (1) it has
been deprived of all physical meanings, being a complete theory only
in the operational sense ; (2) it has so far been one of the most suc-
cessful theories in relating experimental information; (3) the phenomena
it applies to, all possess the characteristic of the irreducible wholeness
of experimental context; (4) it can apply only to the ensembles of
quantum measurements described in classical terms.
Instrumentalism insists that human knowledge is but the set of all
the effective 'recipes ' . In the view of instrumentalists, the minimal
structure of QM is all of QM. However, most physicists are not willing
to follow instrumentalism. In a sense, even the proponents of the
Copenhagen School oscillated between a de facto instrumentalism and
a declared but undeveloped realism. It is inconceivable for a physicist
to have a formalized physical theory without any physical meaning.
However, to explore an interpretation of QM is to explore a philosoph-
ical extension of it. How is such a metaphysical business relevant to
physical science? What is the value of such a business? Let us examine
these questions from two sides.
From the passive side, we notice that metaphysics is inevitable, and
the point is how to take a correct position. On this Bohm said [4]:
We don 't know the nature of reality ultimately. What is the point of describing it or talking
about it? Well, one view is to say there is no point, and a lot of modern philosophers
and scientists say we shouldn't do it. And that 's a reasonable approach up to a point.
But everybody has tacit metaphysical ideas even though he has explicitly disavowed them.
And therefore ultimately that way of doing it is only going to put you under control of
whatever metaphysical ideas you happen to have adopted as presuppositions, perhaps in
early childhood. So, I think it is valuable to explore your metaphysical ideas, to question
them , and to propos e new metaphysical ideas, and so on.
From the active side , we notice that metaphysics is necessary. This

point is suggested by the so-called Godel theorem about the incom-
pleteness of a formal system in mathematics. The theorem tells us that
in any formal mathematical system some true propositions can be con-
structed, which cannot be proved nor negated within the framework of
310 HONG DINGGUO
the system. That is to say: the truth of such propositions must be judged
by means outside the formal system. No doubt, Godel's theorem can
be applied to formalized physical theories . In other words, such a for-
malized physical theory contains some true propositions which cannot be
proved or negated, and the truth of such propositions should be judged
by means and notions outside the formalized physical theory. Therefore,
to complete the meaning of a formalized physical theory, it is neces-
sary to make some philosophical extension of it. For creative theoretical
work, such a philosophical extension is indispensible. The only point
is to make such an extens ion reasonable and consistent.
Moreover, any formalized physical theory results from the creation
and evolution of some physical insight, while the creation of any new
physical insight and the format ion of any new physical notions is not
logical and algorithmic but meta-physical and non-algorithmic. In this
sense, the minimal structure of QM can be thought of as the result of
sublimation and crystallization of the metaphysical ideas which have
nourished the growth of QM.
The above analysis indicates that the birth of QM is by no means to
deny meta-physics but put more severe requirements to the possible forms
of the concrete notions of metaphysics.
THE FIRST TYPE OF PHILOSOPHICAL EXTENSION OF QM
Therefore, when we talk about the philosophical extension of QM, we

speak of a consistent system of metaphysical ideas which cannot be
deduced from nor refuted by the minimal structure of QM. If such a
system has the notion of empirical reality as its essential content, we
call it the first type of extension of QM . According to D 'Espagnat's
analysis [5], the notion of empirical reality includes the following ideas:
(1) At any time , the center of mass of macroscopic bodies has the
property of locality.
(2) A property of a system must be defined operationally but counter-
factually.
(3) The time evolution of the properties of a system in an ensemble is
determined by the computation rules of a theory.
(4) It is impossible to influence the past.
(5) A theory of empirical reality is obtained by postulating (a) that
replacing very large times by infinite times and/or very large particle
numbers by infinite numbers is a valid abstraction; and (b) that on
the other hand the possibility of measuring observables exceeding

a certain degree of complexity is to be considered as nonexistent,
even in matters of principle, even though this non-measurability does
not follow from the theory and even the only way we have for making
it compatible with QM seems to be to ascribe it to some basic inap-
titude of men.
It is not difficult to find that the first four ideas come from the notion
of reality in common sense and the essential requisite of the minimal
structure of a physical theory; while the last is the product of efforts to
coordinate the first four ideas with the minimal structure of QM . To
elucidate the last point , let us look at some important progress in the field
of fundamental researches of QM.
At first, we must say that the conventional (or orthodox) theory of QM
has been a somewhat philosophical extension of QM because it contains
some concepts implying physical meanings such as 'particle', ' wave ' ,
'complementarity principle', ' uncertainty principle', etc. However, such
extension is unsuccessful since if it does not regress to strict instumen-
talism , various conceptual difficulties cannot be exempted. A prominent
example is that it cannot explain the collapse of the wave function in a
single quantum mechanical measurement. Many quantum physicists have
claimed that many conceptual troubles in the conventional QM may be
avoided by resorting to the more elaborate theory known as algebraic
QM . This means that by extending QM in the algebraic framework ,
and endowing the observable algebra with a primitive status, it may be
possible to speak of both classical reality and quantum reality in a
united epistemological framework.
Here, we would like only to mention some works of signpost signif-
icance. First, S. Machida and M. Namiki were able to prove that [6]
the disappearance of the cross terms in the measurement process (cor-
responding to the collapse of wave function) results in the macroscopic
nature of instruments. They used the continuous direct sum of Hilbert
spaces H, I labeled by the continuous different number of particles n, E
(no, An) to describe the instrument involved, when no is the mean of
particle numbers of the instrument. In their theory, observables are
described by self-adjoint operators defined within each HnI., and the fol-
lowing limits were taken : no ~ 00, 6.n ~ O. This means that the continual
superselection rule is an essential component of their theory.
Next, H. Araki has further proved that [7] (1) A continual super-
selection rule can result in the collapse of wave packets in the separa-
312 HONG DINGGUO
tion procedure of QM in the infinite time limit. (The so-called separa-

tion procedure differs from a measurement procedure in that it doesn 't
bring instruments into different states.) (2) In the case of a discrete selec-
tion (or no selection at all), it is impossible for the collapse of wave
packets to take place. (3) Even for the case of a continual superselec-
tion rule, the collapse of wave packets and the measurement procedure
must be carried out in two separate steps (i.e, step by step).
Finally, based on the above results, especially on the work of Araki
and K. Hepp [8], C. M. Lockhart and B. Misra proposed an interesting
algebraic quantum measurement theory not long before [9]. In this theory,
when the instrument dynamics is complex enough, the states of 'system
plus instrument' will spontaneously evolve from a pure state to a mixture
(after an infinite time) via an irreversible process. Thus, a well-known
essential prerequisite in quantum measurement problems had been log-
ically deduced.
From the above brief introduction, we could see that it is based on
his investigation of the above theoretical progress that D'Espagnat has
designed his fifth idea. It is reasonable to say that such an idea is an
embodyment of an empirical realism compatible with QM. Thus, it is not
strange that he called it the axiom of empirical reality.
THE SECOND TYPE OF PHILOSOPHICAL EXTENSION OF QM
The second type of philosophical extension of QM is a system of QM

that includes the notion of ontic reality. It is goal pursued by those
quantum physicists who persist in their ontological research. What we
are mainly concerned with here is the total new program of quantum
mechanical ontology developed in the gap between the ontologies of
Einstein and Bohr.
This research line was at first proposed by L. de Broglie in the name
of 'the pilot-wave theory' [10], and was given up soon because it was
attacked intensively and had no support then. Quite later, Bohm proposed
similar ideas independently [11], introducing particle variables (local
beables) into non-relativistic dynamics and endowing the wave function
with the meaning of a realistic field (global beables), so as to transfer
a non-determinant QM into a determinant QM without losing experi-
mental information. Moreover, in his reformation of QM, all quantum
paradoxes are resolved in a consistent way, with all quantum processes
being explained 'intuitively' . Since Bohm has cleverly answered all
reasonable criticisms, his theory stands and thus not only makes de
Broglie return to his original position but also attracts J. S. Bell to join
their efforts; and Bell later proposed his famous and elegant 'Beables
for quantum field theory ' [12]. Although there are many differences
between the individuals, we call their common thought BBB theory in
short.
In BBB theory, 'observer ' is no longer at a fundamental level.
Observables are replaced by beables. The so-called beables are referred
to those quantities corresponding to realistic elements (i.e. existed
beings). Their existence is not dependent on 'observation' . On the
contrary, observing instrument, observation behaviour, and even observers
all come from beables. Naturally, the crossed influence between the
state of global beables and the behaviours of local beables are determined
by the equations of QM . The totally new ontic relationship between
the whole and parts appears running through all quantum mechanical
problems (both the measurement problems and non-measurement ones).
This is a fundamental characteristic which distinguishes the quantum
ontology from the classical one. It is simply called the wholeness of
quantum reality, or the unseparability of the quantum system, or the
EPR non-local correlation of quantum system. However, as long as
the influence of global beables of QM to local beables is far less than
the effects of classical interactions, a quantum mechanical system
is dissolved, and quantum reality is thus retrograded to a classical one
which can be described by classical physics. So, in this type of exten-
sion model of QM, a smooth transference between the quantum reality
and the classical reality can be realized, without a non-transcendent gap
between them.
People may be anxious that the causality paradox may be brought
by the confrontation between the non-local correlation of quantum
mechanical ontology and the principle of special relativity. You should
not worry about that for two reasons. First, the non-local correlations
of QM are very fragile. Once one tries to draw information from some
location of a quantum system, the correlation is dissolved immediately.
So, one can never utilize it to transfer a distant signal simultaneously.
Second, the quantum measurements carried out so far are only of statis-
tical significance. The principle of special relativity thus possesses only
stati stical significance. That is to say: the non-locality refused by experi-
ments is the one contradicted by the principle of special relativity in
the statistical sense but not one in a presumably basic ontic process.
314 HONG DINGGUO
Here it is probably suitable to mention the ontological interpretation

of the Dirac equation, proposed by Bohm and B. J. Hiley recently [13].
In their ontological model, the equations satisfied by global beables
(i.e. the Dirac equation and the continuity equation) is Lorentz covariant,
while the guidance equation of local beables (i.e. velocity equation of
particles) is of a non-locality characteristic. That is to say: no matter how
far the distance between the particles, the displacement of any particle
may generally mean the corresponding instant variations of the veloci -
ties of all the particles . However, this does not mean that the ontological
interpretation of QM cannot be extended toward the situation of rela-
tivity, since the theoretical statistical prediction relies only on the
covariance of the equation of global beables. So, it may be suggested that
there exists a preference-based reference frame , in which the non-locality
correlations are simultaneous, and there exists a non-covariant basic
stochastic process, which makes whichever distribution of the initial
velocities tend dramatically to the equilibrium distribution determined by
globe beables. To all practical purposes , the equilibrium probability dis-
tribution is the sole distribution. The statistical results for all existent
quantum mechanical experiments computed by means of the Bohm-Hiley
theory is therefore covariant, as the results computed in the conven-
tional QM. It is obvious that at the moment there is no experiment to
prove their theory wrong.
In fact, Bohm and Hiley have brought us to the vicinity of a new
physics. The statistical process on the sub-relativity level proposed by
them cannot be revealed in any quantum mechan ical experiment. This
implies that it is sub-QM as well. The existent quantum theory and
relativity come only from the statistical results of experiments. They
are not sensitive to the details of basic ontological models, just as the
laws of thermodynamics are not sensitive to the basic atomic models. So,
in this period, we have a wide range of probabilities within which to
choose various models of sub-relativity and sub-QM .
Last, we would like to mention the relationship between the
algebraic QM and the extension of QM. It is remarkable that H. Primas
has made a concise and systematic research study of it [14]. He empha-
sized that the early Hilbert space formalism of QM is limited to locally
compact phase space, so as to rule out the proper descriptions of gen-
uinely irreversible stochastic processes, spontaneous symmetry breaking,
phase transition and the classical behaviour of quantum world. It is there-
fore unsuitable to take it as a universally valid theory. However, the struc-

ture of algebraic QM contains all physical systems, not only finite
systems (with a local compact phase space), but also infinite systems
(whose phase space is not locally compact). Infinite systems have infi-
nitely many physically non-equivalent W*-representations which account
for the stupendous complexity of observable phenomena in nature. For
example, it may be proved that an open quantum system generally expe-
riences symmetry breakings, and possesses classical observables, while
contextual classical observables are not the function of intrinsic observ-
abies but the product of the intrinsic algebra combined with a contextual
topology.
Basing on the algebraic QM, the philosophical extension of QM
constructed by Primas is a consistent system of idea s with the following
distinguished statements: (1) A strictly closed physical system without
any concept of an observer is called an endosystem. (2) If the endosystem
is divided into an observing and an observed part, we speak of an
exophysical description. (3) The world of observers with their com-
munication tools is called an exosystem. (4) The empirical inter-
pretation of a physical theory refer to our knowledge of the properties
or modes of reactions of the system "as we perceive them". (5) The
ontic interpretation of a physical theory refers to the nature of 'object
itself'. It is irrelevant with respect to our knowing it or not, and it is
independent of any perturbation by observing actions. In the view of
Primas, the ontic interpretation of QM refers only to a fictitious theo -
retically immanent reality, and not to the ultimate reality. But, under
the working hypothesis that QM is a universally valid theory , an ontic
interpretation allows us a consistent way of speaking as if we would refer
to reality .
CONCLUSION
In sum, we might construct various consistent extended structures of QM

on the minimal structure of QM with different physical commitments,
which constitute an interpretation spectrum of QM in accord with a series
of different attitudes toward reality, from extreme anti-realism to onto-
logical realism. The work of extending QM is not only a philosophical
undertaking but also a physical one . It is an unfinished business. We hope
that more philosophers familiar with theoretical physics and more physi-
316 HONG DlNGGUO
cists with philosophical quality will take care of, and be engaged in, such
a fundamental project concerning the future of physical science and
that of modem philosophy as well.
Department of Physics,
Hunan Nonmal University,
Changsha, China.
BIBLIOGRAPHY
1. Einstein. A. (1937). Mein Weltbild. English version by C. Seelig. Ideas and Opinions.
Dell Publishing Co.• New York. p. 292.
2. Hong Dingguo (1988). The Structures of Physical Theories and Their Unfolding •
Science Press. Beijin.
3. Bohr. N. (1949). Dialectica 2. pp. 321-329.
4. Bohm, D. (1985). 'A Talk on the Value of Metaphysical Discussion', in D. Facctor
(ed.), Unfolding Meaning - A Weekend of Dialogue with David Bohm ,Foundation
House Publication.
5. D'Espagnat , B. (1989). Reality and the Physicist. Cambridge Uni, Press. pp. 232-253 .
6. Machida, S. and Namiki. M. (1980). Prog. Theor. Phys. 63. pp. 1457. 1833.
7. Araki. H. (1980). Prog. Theor. Phys. 64. p. 719.
8. Hepp, K. (1970). Helv. Physica Acta 45. p. 237.
9. Lockhart, C. M. and Misra. B. (1987). Physica A 136A. p. 47.
to. Broglie, L. de (1927). Journal de Physique 5. p. 225.
11. Bohm, D. (1952). Phys. Rev. 85. pp. 166. 180.
12. Bell, J. S. (1987). ' Beables in Quantum Field Theory'. in B. J. Hiley and F. D.
Peats (eds.), Quantum Implicat ions . Routledge and Kegan Paul, London.
13. Bohm, D. and Hiley, B. J. (1989). Phys. Rep. 173(3). pp. 93-122 .
14. Primas, H. (1991). 'Realism and Quantum Mechanics', preprint.
JIN WULUN
DOUBLE STRUCTURE OF TAO REALITY
The legitimacy of interpreting scientific theory in a realistic way has been

increasingly accepted by physicists and philosophers of science. For cen-
turies the results of relevant investigations in natural sciences have
made it possible to use the hypothetico-deductive method fruitfully.
This tremendously strengthens the rationality of, and the belief, in real-
istic interpretations.
Heisenberg (1958) used potentia to denote the independent reality.
Shimony (1978, 1986) and Stapp (1979, 1985) followed Heisenberg,
exploring further the realistic interpretation of the quantum formalism;
Bell (1964) talked about physical reality by 'Beable' in quantum field
theory; Bohm (1957, 1980) used the concepts 'quantum potential' and
'the enfolding and exfolding of implicate and explicate order ' to describe
the wholeness of reality ; d'Espagnat (1983) put forward 'veiled reality'
to describe reality independent of all observers. And beside these there
are other interpretations, such as the many-worlds interpretation of
Everett and others. These inquiries are fruitful, but still there is much
room for a more detailed account of reality.
In this paper I'll try to discuss the structure and properties of reality
with Tao and Yin-Yang in ancient Chinese Philosophy, and to explain
some quantum paradoxes. In doing so, I prefer d'Espagnat's opinion:
to solve the difficulties a realist faces we shall establish that "a link
with some philosophical or cultural tradition is necessary". I
THE THEORY OF TAO-REALITY
Tao is the ontology of the universe and also the laws of the universe.
As an ontology, just as Professor Jin Yue-lin says: "Tao, as the most lofty
concept, as the most fundamentally primal power, is not absolutely empty.
Tao must be real .. . , Tao can be spoken both integrally and separately.'?
First, Tao is real, so we can call it 'Tao-reality', and therefore we
can call the theory about it the 'theory of Tao-reality' .
Second, integrally speaking, it can be called Tao. Tao is "the begin-
ning of heaven and earth", "the mother os Wanwu (ten thousand things)".
317
R. S. Cohen. R. Hi/pinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
318 JIN WULUN
It is the unity of having (1f) and no having (JG), [being and non-
being, nothing] these two spring from the same source but differ in name.
Separately speaking, it can be divided into "the eternal Tao" and "not
the eternal Tao", that is "having and no having".
Third, the unity formed by having and no having is nonobservable.
It "differs from the observable concrete things": "look, it cannot be
seen - it is beyond form. Listen, it cannot be heard - it is beyond sound.
Grasp, it cannot be held - it is intangible. These three are indefinable;
Therefore, they are joined in One." (Tao Te Ching, p. 14). However, it
is not absolutely empty. On the contrary, "Tao is elusive and intangible.
Oh, it is intangible and elusive, and yet within is image, Oh, it is elusive
and intangible, and yet within is form ." (Tao Te Ching,p. 21) It is
"standing alone and unchanging, even present and in motion." (Tao Te
Ching, p. 25)
Tao-reality has having and no having as two components, therefore
we can say it has a double structure. Having and no having are two
existential states of Tao in the cyclic movement. No having is the
start-point and the end-point, whereas having is the middle-point or
extreme-point in this cyclic movement. Therefore, although no having
and having are different, they are to be used to refer to Tao. They are
two indispensable aspects. Tao on the one side and having and no having
on the other side have relations as follow.'
e
no having
having
Fig. I. The relation between Tao and having and no having .
Here Tao is the ultimate reality. It is the united body of having and
no having. It is the primal source of WanWu. Everything emerges from
it, and returns to it, but Tao itself is eternal existence. WanWu have the
variant principles, but Tao comprises the principle of WanWu. The unity
and mutual interrelation of all things and events, the experiences of all
phenomena in the world are manifestations of this basic oneness. All
things are seen as interdependent and inseparable parts of this cosmic
whole; and as different manifestations of the same ultimate reality.
DOUBLE STRUCTURE OF TAO REALITY 319
From this , we can know that Tao-reality has many more implica-
tions than Heisenberg's potentia. Heisenberg 's potentia is simply
tendencies or possibilities of Being. It has no internal structure, and it
only expresses a process from potentia to actualization by measure-
ment. Tao-reality has a double structure. It is an unity of having and
no having, but potentia is only equivalent to ' having' in the Tao-reality;
however 'having' cannot exist without 'no having'. More importantly,
Tao-reality cannot be directly measured, only WanWu generated from
Tao-reality can be measured. That is to say, only the observable things
generated from Tao-reality can interact with a measurement. Tao-reality
and its emergence are all outside our consciousness; to use d'Espagnat's
words, Tao-reality is a far reality," far realism is the theory about it.
THE THEORY OF TAO-BECOMING
Tao is real, it has a double structure.

Tao can be born and destroyed. It has the mechanism of double-
directions movement.
Tao is changeable. All things arise from it. (The Tao begot one, one
begot two, Two begot three , and three begot WanWu) This begetting is
natural, there is no artificial force in it, and it involves only the force
of Yin and Yang, as Harvard University astronomer Professor David
Layzer says:
In Chinese Philosophy, it recalls the doctrine of Yin and Yang; Yang, the integrative force,
corre sponds to the growth of order ; Yin, the disruptive force , to the entropic decay of
order.'
Becoming is an important feature of the unity of having and no having.

Becoming is the unity of having and no having . Becoming has gener-
ating and destroying, and the generating has the beginning. According
to Hegel, when a thing is at its beginning and has not yet become actu-
ality, it is not a mere nothing but already contains something, or existence,
of it. Beginning itself is becoming, but beginning also contains the
meaning of development.
In fact, a thing or an object that the objective probability expresses
is the representation of the unity of having and no having. There is
nothing in appearance, but there is the having in fact. Not only the having
is real , but also no having is real. Having and no having are inter-
changeable, " something for nothing"." There is the ' becoming ' and
320 JIN WULUN
'beginning' implicated in it. For example, consider radioactive material:

before it spontaneously emitted a particle a radioactive atom has neither
happened nor not happened. For the emitted particle, it will be in the
unitary state of having and no having . No having means no radiation;
Having means a tendency state of radiation. When this state will be
destroyed, the particle will come out of it. This emerging particle is
exactly observable, though the unity of having and no having here is
nonobservable.
The vacuum fluctuation is the best example of Tao-becoming. Vacuum
is certainly not absolute emptiness. The vacuum state of a unity of having
and no having will spontaneously break the symmetry or lose the
symmetry. The symmetry breaking requires a new field, the Higgs field,
and the associated particles are also called Higgs. We now know that
the Higgs field is not zero. So " the universe and everything in it may
be no more and no less than one of those vacuum fluctuations that
allow collections of particles to burst forth out of nothing, live for a while,
and then be reabsorbed into the vacuurn'" The vacuum symmetry-
breaking, the vacuum phase changing, generate elementary matter and
fundamental interactions. "A closed universe has zero energy overall , and
it is not so difficult to make something with zero energy overall out of
a vacuum fluctuation."?
Only by this time , WanWu, namely observable things , can emerge
and the empirical phenomena can appear, and then we can observe,
measure and grasp them. This suggests that the process of becoming from
potentiality (the unity of having and no having) to actualization (empir-
ical phenomena) is very complex.
An important feature of this generative process is an integrative
genesis. For example, the generation of a particle is not isolated from
the original potential state which existed in it, but the result of the
integrity of all information in the net of the unity. It is not like a water-
melon seed in the watermelon which is peeled off, but rather like a young
shoot growing from the watermelon seed. This is an integrative genesis.
The view of integrative genesis emphasizes that the explication of
becoming or changing would be made in terms of 'coming into being ',
' passing away' or ' transformation' . This view can be called the genesis
theory. It is plainly contrary to the constructive theory which is dominant
in western thought, for the latter claims that the changing of things is
the ' combining ' and 'separating' of the unchangable elements, and the
method is the method of analytical reconstruction.
As we known, the Feynman diagram represents the creation and

disappearance of particles and their interactions. For example, a photon
traveling through space and time spontaneously creates an electron/
positron pair, and the positron meets another electron and disappears,
and the like. However, the constructive theory cannot interpret these
phenomena; only the genesis theory can interpret them.
The genesis theory specifically emphasizes that this generation is
not a mechanistic separation, but an integrative genesis.
To see measurement from this view, we can say that measurement
concerns only generative things, not Tao-reality itself. Reality is inde-
pendent of measurement. This is the view of reality we need to defend.
This view of reality differs from that of metaphysical reality. And the
main difference consists in that the latter takes measured things not as
generated ones, but as originally real existence there. Observation and
measurement simply pick out the measured objects from the universe,
and the mea surement does not act on the rest of the universe, only on
the measured object. The view of Tao-reality emphasizes that the
measured object is a generative object of the unity of having and no
having. This generative thing is the product of the integrative genesis
of Tao-reality. The measurement in turn affects the whole reality.
The mechanism of the generation and interaction have been governed
by Yin-Yang force . The doctrine of Yin-Yang tells us that the unity of
having and no having is a combination of Yin and Yang which are unsep-
arable. The generative thing s are loaded with both Yin and Yang. This
is becau se "WanWu carry Yin and embrace Yang. They achieve harmony
by combining these force" (Tao Te Ching, p. 42).
To speak with emphasis, for measurement the generative things only
appear as Yin force, and the measurement represents Yang force .
Therefore, the relation between the measured object and the measuring
device is a relation of Yin-Yang complement. This situation is consis-
tent with Bohr's framework of complementarity. To see this change of
Yin-Yang roles, we might as well see the change of Yin-Yang roles in
the relation between father and son, male and female. For a single male
has both Yin and Yang, but for his wife, he has only a Yang force . The
relations between WanWu in the universe are all like this . However Yin
and Yang ever combine two into one, but separable.
Starting with this view, we can explain the EPR correlation, non-
separability and nonlocality.
When the system AB is separated into A and B, far from each other,
322 lIN WULUN
anyone of them under measurement will influence the other though there
is not any physical interaction between them. The measured object, say
A, will soon be transformed into a Yin force , while the measuring device
appears as the Yang force as we said above; when A is transformed
into Yin, then B is transformed into Yang. That is to say, if we measure
A, then A appears as Yin and B as Yang; and vice versa (see figure 2).
When observation or measurement begins, the measured object becomes
a mediator linking reality with the measuring device . Since Yin-Yang
is always co-ex istent, it integrates reality, observable objects and the
device into an undivided whole. Thus, A and B are a whole connected
by Yin-Yang, even if they are separated from each other and have no
physical interactions. We can say that Yin-Yang's becoming is exactly
the very essence of the EPR correlation mystery; that is the essence of
the microparticle's entanglement.
8 Yang
)1n
o
g
Ya6:2
..
measunng device
\,
~easured object
( J Yin.Yang
Fig. 2. Interpretation of EPR correlation with the Doctrine of Yin-Yang.
We can hold that the correlation-at-distance linked together by

Yin-Yang force in the result of measurement is characterized as non-
separability, nonlocality, world's wholeness, and general connection. It
is characterized as 'passion-at-a -distance ', as Professor Shimony says."
It is a kind of overall holism .
The double-slit experiment can also be interpreted with the doctrine
of Yin- Yang. According to this doctrine, the combination of Yin and Yang
produces everything in the world, which are composed of Yin and Yang.
Nothing can exist unless neither Yin nor Yang is absent. But two
force, Yang and Yang or Yin and Yin, are in conflict, and they cannot
co-exist within a single thing. This feature of Yin and Yang leads to
showing the special pictures on the screen of the single-slit and double-
slit experiments.
When combining Yin and Yang into one, particles are produced, so
a peak appear on the screen. However, Yang and Yang would repel one
another, and Yin and Yin would repel one another too, just as two like
electric charges repel one another. Therefore, the wave valleys appear
on the screen when Yang and Yang or Yin and Yin repel one another. See
figure 3.
~I wavepeak
A
~I dOObt
slit ~y;"-y,".
Yin-Yinor
---+1
Yang-Yang
B . wave valley
I
Fig. 3. Interpretation of Double-slit experiment with the Doctrine of Yin and Yang.
Some of other quantum phenomena can be also interpreted with the

Doctrine of Yin and Yang. Since we have limited space we cannot offer
a detailed discussion.
The Tao-becoming is a holistic process . To clarify this process further
we need to make further inquiry for holism , including a search of the
physical mechanism of Yin-Yang interaction. In the 5th section of this
paper we will discuss this problem.
THE THEORY OF CONSTRUCTION
Scientists construct scientific theories to explain the observational char-

acteristics of the physical world. They often construct hidden-structured
models of the objects studied. This structure is used for a causal account
324 JIN WULUN
of observable phenomena, and the theoretical models provide approxi-

mate explanations of phenomena.
Scientific realism argues that to a certain extent the theoretical struc-
ture is an insight into the structure of the world or of reality, but the
scientific realist does not maintain that the constructed theoretical struc-
ture must be the true structure which we live in.
Generally speaking, this procedure is roughly this: scientists observe
sensory phenomena, from which they apply their cognitive structure
(including background theory) to advance a set of basic concepts or
basic principles as the theoretical foundation from which they deduce
a sequence of conclusions with the aid of logical deductive procedures.
These conclusions can be compared with sensory phenomena, and are
required to be compared with other well-known theories as well as the
new observable phenomena. In the course of comparing, they will be con-
tinuously revised, substantiated and increasingly completed.
This construction is not, as a Chinese saying goes, to "draw calabash
according to the original example", because there is no original calabash.
The construction is rather like to "draw a tiger in the light of a cat".
The tiger drawn consists simply in a eat's form and a painter's imagi-
nation . As a result the picture drawn is often like a dog rather than a tiger.
This constructive procedure is called the 'method of analytical recon-
struction' ; in other words, it attempts to represent the whole structure ,
or the whole of reality by constructing local phenomena. There are
some apt metaphors in China : 'looking at the sky from the bottom of a
well' or 'a blind person groping for the elephant' . Although a part of
the sky is real and true, and a part of an elephant is also real and true,
yet one part differs entirely from other parts. Every part falls under
conditions and circumstances. Maybe they are really incommensurable
in Kuhn 's sense .
THE DOCTRINE OF EMBEDDING
Our present world picture is the result of constructing and embedding,

just as a beautiful painting drawn by a famous painter. It is a product
which has been made bit by bit, just as a painter works out the plot of
a picture .
The scientific world picture is a bit similar. The scientists first con-
struct a theory from local phenomena and then embed it within the
general appearance of a picture; as Einstein says, "the real difficulty

lies in the fact that physics is a kind of metaphysics; physics describes
'reality'. But we do not know what 'reality' is; we know it only by means
of the physical descriptionl'" That is to say, the reality we know is only
the result of models constructed by scientific theories that are hope-
fully embedded in the nature .
The atomic pictures we know, from Democritus's atom to Dalton's
atom to Bohr's atom, are all the products that scientists constructed.
The electron was also constructed in this way. Therefore, atom, electron,
elementary particle, and quarks too are all products constructed and
then embedded in nature by scientists. Their existential features are
closely linked to scientific theories; as the physicist Heinrich Hertz
said, "from the view of humankind, light is a kind of electromagnetic
wave." That is to say, "light" and "electromagnetic wave" both are some-
thing seen from the view of humankind. They are the descriptions and
interpretations for natural phenomena in terms of language.
Physicist D. Bohm made this clearer. After having thought deeply,
he said, "I suddenly realized that everything is all subjective." I am
sure that the world picture provided by science has artificial proper-
ties. Of course, experience provides limiting constraints upon this world
picture. But construction and embedding are not arbitrary, they must
be subjected to many constraints so as to satisfy the weak objectivity
in science. To make an embedding is to assign theoretical models con-
structed according to the objectivity principle in nature, to Tao-reality.
In the words of Professor Jin Yue-lin, this process is "to receive from
the given and to return to the given.?'?
However, the reality picture constructed by science is greatly different
from picture of nature itself. It is in view of this that various difficul-
ties arise for convergent realism just as Larry Laudan criticized; II And
it is also in this situation that the transformation of incommensurable par-
adigms and the Gestalt shifts of world views take place, as Thomas Kuhn
shows clearly."
HOLISTIC METHOD FOR SEARCHING TAO-REALITY
Reality is a whole . It is correlated by Yin and Yang. Observable things

spring from it with the aid of Yin and Yang. Observable things are only
its generating products. Tao-reality itself cannot be measured.
326 JIN WULUN
Reality (or the world picture) as revealed by science is a constructed

and embedded picture . Its method of revealing reality is the method of
analytic reconstruction. For this method , "there continue to be influen -
tial dissenting voices, even among physicists."!' The limitations of this
method are as follows :
(1) The description of a whole would be achieved by knowing its parts
plus a mechanical assembling method. If we take a man as example, then
this constructed man is a robot, not a living man.
(2) Its philosophical basis is reductionism. According to reductionism,
a whole is simply composed of its parts. One can achieve a completely
adequate understanding of the behavior of the whole by analyzing
it into its component parts and discovering the laws according to
which these operate . This so-called "simplistic principles of the universe"
as I. Prigogine wrote, is a myth which has been challenged by new
sciences.
(3) It devotes itself to revealing the analysis of static structure but it
neglects to search for the connections, processes, and becoming. It notes
that the whole is constructed of its proper parts which relate differently
from the sum of the various parts; rather they have a "supervening
relation" as Paul Teller says."
The method of analytic reconstruction has met refutation by holists.
The holists claim that some individuals in the domain in question have
properties that are not determined by the properties of their parts . "Now
it is clear that historical or individual properties of a whole cannot be
expected to supervene on a basis of properties and relations of their
parts which fail to include their historical or individual properties and
relat ions". 15
There is also a similar view in Chinese philosophy, that is the so-called
Yin Tuo Luo (m~'i?;~) net. According to this view, every intersecting
point in this net has a bright pearl which illuminates other pearls and
is also illuminated by others." All relations in the net are based on the
nonrelational properties of every bright pearl in itself (a single bright
pearl's illumination). This, however, is not the dominant view of Tao-
reality in Chinese philosophy.
We have distinguished the holistic method of Tao-reality and the
method of the analytic reconstruction. The differences between Tao-
reality and Heisenberg's theory of Potentia or Everett's many-worlds
interpretation can be listed as following.
Comparive list of variou s views of reality .
Heisenberg Potentialities actualisation by measurement

Shimony, Stapp
Everett Many-worlds real world by measurement
Wh itehead Potentialities actualisation by value
Bohm Implicate order explicate order by soma-significance

I .. . I
d 'Espagnat Veiled I empmcia phenomenal world
reality I reality by observation
I
our model no having I observable phenomenal world
@ I things by observation
:ii I
having ~I generated ~
I
The theory of Tao-reality maintains the holistic method of integra-

tive genesis . It especially emphasizes the following ideas:
(l) Tao-reality is existence independent of our consciousness. It
is a really independent reality. This reality is an indivisible whole. All
sensible phenomena are generated from this independent reality.
Humankind itself is the product of natural evolution, according to
Darwin's theory of Evolution; Humankind's consciousness is also a trans-
forming process, from no having too. The independent reality doe s
precede consciousness. Admitting the independent reality, such concepts
as subject-object interaction, or the unity and separation of subject-object,
might be significant concepts in epistemology.
(2) Tao-reality goes beyond the space-time framework. It cannot be
described by existing scientific concepts; by WanWu are in the space-
time framework, and can be described by the scientific concepts.
(3) The processes of construction and embedding show that our current
picture of the universe is obtained through generalization of scientific
knowledge. We have this picture thanks to the work that scientists do
in understanding, describing and expressing scientific achievements. This
picture itself is only a part of the real world . The independent reality
is large than and precedes the picture constructed by the sciences.
(4) WanWu generated from Tao-reality are not mechanically cut off,
328 JIN WULUN
but organically and integratively generated from it. Everything in it

carries information of the whole universe. Therefore, the method for
understanding and grasping reality cannot be limited to analytic recon-
struction and induction. We should turn to the holistic method of
integrative genesis, and transform our method from static analysis to
dynamic integration.
The theory of Tao-reality can be also called the genesis theory; its core
is integrative genesis . The mechanism of the process of this integrative
genesis remains to be explored further .
Chinese Academy of Social Sciences,
Beijing, China.
NOTES
I d'Espagnat, B. (1983). In Search of Reality, Springer, N.Y., p. 97.

2 Jin Yue-Iin (1987). On Tao, The Commercial Press, p. 17.
3 Wang Bo (1991). 'Exploring the Relation between Tao and Having , No Having in
Chinese Philosophy ', Philosophical Research, No.8, pp. 44.
4 d'E spagnant, B. (1990) . In Search of Reality , p. 95. He said there , "The vision s of
the world of Buddha, of the Tao, ... are examples of far reality".
5 Layzar, David (1990). Cosmogenesis. Oxford University Press, Preface for the Chinese
Edition.
6 Gribbin, John (1984). In Search of Schriidinger' s Car - Quantum Physics and Reality,
Bantam Books, pp. 194-202, p. 271.
7 Ibid.
8 Cushing, James T. and McMullin , Ernan (ed.) (1989). ' Reflections on Bell' s Theorem ' ,
Philosophical Consequences of Quantum Theory, University of Notre Dame Press,

p.77.
9 Einstein to Schrodinger, 19 June 1935.
10 Jin Yue-Iin (1986). Theory of Knowledge , The Commercial Press, p. 493.
11 Laudan, Larry (1984). Science and Values, Chap. 5, University of California Press.
12 Kuhn, Thomas S. (1970). Structure of Scientific Revolutions, University of Chicago
Press, Second Edition.
13 Healey, Richard A. (1991) . 'Holism and Nonseparability ', Jou rnal of Philosophy
88, No.8, August, p. 398.
14 Teller, Paul (1986). ' Relational Holism and Quantum Mechanics ', Brit. J. Phil. Sci.
37, pp. 71-81. 'Relativity, Relational Holism, and the Bell Inequalities', in Cushing and
E. McMullin (eds.)
15 Healey, Richard A. Holism and Nonseparability , pp. 401-402.
16 Xu Zhou-yun (1991) (i-tf$'E), Chinese Culture and World Culture, Chinese Edition,
GuiZhou People 's Press, p. 67.
LARS-GORAN JOHANSSON
REALISM AND WAVE-PARTICLE DUALITY'
Quantum mechanics is a touchstone for realism: everyone claiming

to be a realist faces the challenge of giving a realist interpretation of
quantum mechanics. There are two features of quantum mechanics which
trouble the realist: one is wave-particle duality, the other is quantum
theory of measurement. This paper discusses wave-particle duality only.
The extent to which the ideas put forward will highlight the measure-
ment problem is discussed elsewhere.'
Realism, as opposed to idealism, holds that the reality is indepen-
dent of human minds . This independence consists of three aspects: (l)
real entities exists whether or not there are minds in the world, (2) these
real entities are external to human mind s, and (3) real entities exist objec-
tively, i.e. independent of anyone 's opinions.
These views imply that any scientific theory candidate for being true
must describe its subject matter in conformity with these demands. In
other words, if a description of a real entity contains a (implicit or
explicit) reference to a human perspective or point of view, any such
reference must be possible to get rid of. A description which contains
irreducible relations to human minds is not acceptable for a realist.
Einstein, for example, felt the force of this demand when developing
his relativity theory. The demand for a theory free of human references
implies that the fundamental equations must be invariant under changes
of coordinate systems and relativity theory meets this demand.
The dual nature of the objects in quantum mechanics appears to be
a violation of these realistic demands, because it seems as if quantum
objects show wave or particle properties depending on which experiment
we perform. This dualism is by itself not contradicting realism, if strictly
physical conditions which tell us when the objects are waves and when
they are particles can be given. However, this seems impossible: one
and the same physical experiment shows the wave or the particle char-
acter of the fundamental entities depending on which question we are
interested in. Consider for example the well-known two-slit experiment:
we can choose photons, electrons, neutrons or whatever when performing
this experiment. The result of the experiment is an intensity distribu-
329
R. S. Cohen , R. Hilpinen and Qiu Renzong (eds.), Realism and Anti -Realism in the
330 LARS-GORAN JOHANSSON
tion on a photographic film which reflects the wave character of the

objects having passed the double slit. By measuring (l) the distance
between the interference fringes, (2) the distance between the photo-
graphic film and the double slit and (3) the distance between the two
slits, we can calculate the wave-length of the particles using the well-
known formula nA = d sin c. This calculation takes for granted that
each object has passed through both slits in the screen . In other words,
when calculating the wave-length we assume that each object is at two
places simultaneously, viz. when passing through the screen with the
double-slit. (Young, who first made this type of experiment with visible
light, was thus convinced that light must be waves and not corpuscles
as Newton had said, and the experiment was considered the crucial
experiment in favour of the wave theory of light.) On the other hand,
each particle had a definite position when hitting the photographic film,
because each particle had made a tiny spot on the film. The position is
determined with an accuracy determined by the mean diameter of the
grains in the film emulsion. This accuracy, i.e., the mean diameter of
the grains of the emulsion is much smaller that the distance between
the two slits in the screen. But if the object really was a particle when
hitting the photographic film, how then is it possible for these objects
to show an intensity distribution corresponding to waves spread out
over the entire target region? The conclusion generally drawn from
this and other experiments is that quantum objects show wave or
particle properties depending on what questions we pose to nature and
not depending on any objectively existing circumstances. This is not
an acceptable stance for a realist; consequently, every realist faces the
problem of accounting for wave-particle duality in a way which is both
empirically adequate and philosophically acceptable. At the very least
we need a criterion which in an observer-independent way tells us when
an object is a wave and when it is a particle. Here is my proposal :
EVERY OBJECT IS A WAVE PHENOMENON DURING MOTION

AND A PARTICLE WHEN TAKING PART IN IRREVERSIBLE
INTERACTION WITH OTHER OBJECTS
The criterion should be understood so as to imply that if no irreversible

interaction occurs, the object is moving and thus a wave.
This criterion uses three vague concepts, viz. (l) wave, (2) particle
and (3) irreversible interaction.
REALISM AND WAVE-PARTICLE DUALITY 331
1. Waves. The natural language word 'wave' has many connotations.

In this context it is appropriate to focus attention on two properties
common to all sorts of waves : they obey the superposition principle
and they have a non-negligible extension in space . The last point means
that no wave can be said to be in one point.
2. Particle . The word 'particle ' hardly belongs to natural language,
it is a theoretical concept in classical mechanics. One often thinks of a
particle as some small object, but that is not correct. A planet is in
some contexts a particle. The core of the particle concept is that we
can disregard the extension of the object called a 'particle'. Furthermore,
particles do not follow the superposition principle because that would
mean taking into account the inner const itution and thus its extension.
It is obvious that no object can be considered to be a particle and a
wave simultaneously.
3. Irreversibility. Irreversibility is a conceptual problem in physics
generally: how can it be the case that truly irreversible processes exist
when all fundamental equations of motion are reversible? Without further
argument I will take it for granted that irreversibility enters physics in
systems made up of a great number of (at least partly) independent
objects. A monoatomic gas (at a pressure higher than the air pressure)
in a container is a simple example. Once the container is opened most
of the gas leaves the container. The reverse process, i.e., a gas dis-
tributed in a big volume assembling into a small part of that volume,
never occurs and molecul ar kinetics expl ains this as a statistical effect.
The reverse process is not logically conflicting with the fundamental laws
of motion , but it is extremely improbable. But considering only a few
gas atoms, it not at all unlikely that they will be found inside the con-
tainer long after the container has been opened. The irreversibility thus
enters at the macroscopic level as an effect of the motions of many
independent parts of the macroscopic system .
I assume that this line of reasoning is correct even in the framework
of quantum theory. Hence, the phrase 'irreversible interaction with other
objects' refers to interaction with objects big enough to have the irre-
versibility property, i.e., in which the probability of a reversal of state
change can be neglected.
A measurement on a quantum object is undoubtedly an irreversible
interaction, but is should be observed that I am not supposing that mea-
surement interactions are the only irreversible processes. For example,
in many experiments only a fraction of the prepared objects is measured,
the rest being lost outside experimental control. All these objects will
sooner and later hit some macroscopic object, for example some irrele-
vant part of the experimental equipment. These collisions are not
measurements but nevertheless irreversible interactions. There is reason
to believe that even those interactions can be irreversible because the
hit part of laboratory can undergo an irreversible state change .
(The term ' measurement' here refers only to those measurements
where the wave function collapses, but those cases are the problematic
ones for the realist.)
Thi s analysis describes the wave-particle duality as a mind-indepen-
dent property of quantum objects, and so far the analysis does not conflict
with any empirical evidence.
How then can this criterion be used as an explanation of the two-
slit experiment? The answer is: every object is a wave from the emission
from its source to its collision with some big object. When passing the
screen the object passes both slits and does not interact irreversibly
with the screen: the object neither exchanges energy nor momentum with
the double -slit screen . Behind the screen the wave is separated into two
parts whose total momentum is the same as it was before the screen.
These two parts interfere with each other and the intensity distribution
of the total wave shows an interference pattern. When hitting the pho-
tographic screen the wave collapses and gives away its kinetic energy
to one of the grains. This collap se is a sudden disappearance of the wave;
it ' condenses' into one small region. Which one? That is impossible to
predict, the place of the collapse being an truly indeterministic event.
If we simply assume that the probability for collapse at a certain point
equals the intensity (the squared amplitude) of the wave, the interpre-
tation fits observed data .
I think this interpretation resolves the conflict between realism and
wave-particle duality, but it could be argued that that is not enough:
the sudden collapse of the wave when hitting a macroscopic object is still
a complete mystery. A philosopher could reply that the explanation of
the wave collapse is now a task for the physicist, not the philosopher. I
think this is a proper defence, because the collapse of the wave is a
completely physical process without relevance for the realism debate.
However, the wave collapse can in fact be given some sort of explana-
tion without making any assumptions which are not already implicit in
quantum mechanics.
When assuming that real objects are wave-like objects except in irre-
versible interactions we would like to know why these wave-like entities

show particle behaviour in some circumstances only. The answer is:
the particle behaviour follows from quantization of action?
Although it belongs to the very foundation of quantum theory, quan-
tization of action is not a very well understood concept. Quantization
does not mean that physical quantities generally are made up of multi-
ples of least possible portions of every quantity, such as the elementary
charge e. It is rather the case that each interaction process is one
indivisible whole: when two objects approach each other either one
indivisible quantity of energy is exchanged or nothing at all happens.
This is the meaning of quantization of action and it is the guiding idea
behind the choice of quantum mechanical operators to represent mea-
surable quantities. Bohr was of the same opinion:
.. . its (quantum theory) essence may be expressed in the so-called quantum postulate,
which attributes to any atomic process an essential discontinuity, or rather individuality,
completely foreign to classical theories and symbolized by Planck's quantum of action:
It is to be noted that quantum theory does not contain any general

rule by which we can find the operator representing an observable.
Specifically, it is generally not true that if the observable A is a function
of the observables B, C, . . . , i.e., A = f (B, C ...), a similar func-
tional relationship A = f(B, C ..•) exists between the corresponding
operators A, B, C, . . .5
Now, I conjecture that quantization of action is the root of the particle
character of waves in interaction. In order to illustrate the argument,
the following thought experiment is suitable. Imagine a possible world
in which Planck's constant h has a macroscopic value, I Js say, instead
of its actual world value 6.6.10- 34 Js. Let us further assume that we
perform a wave experiment in a water tank, akin to those made in
elementary physics courses in wave mechanics. A dipper in the form
of a straight line touches the water surface and water waves expand on
the water surface from the dipper. Let us further assume that at the
opposite wall of the water tank we have placed a row of sensors capable
of registering the incoming waves. Taking into account quantization of
action, the interaction between the dipper and the water must be quan-
tized. That implies that each wave is an indivisible portion of energy.
When this energy portion approaches the row of sensors, at most one
sensor can take up this energy. Even if the energy spacing in each sensor
is small enough to allow for a distribution of the energy to all sensors
334 LARS-OORAN JOHANSSON
• i i i i i i i '0
o
o
(]
(]
(]
[J
(]
D
(]
n
~! ·n
...... \ I / ,/ ~ .f'~ -
""~/
.a
wave detector
+
fronts row
Fig. I.
hit by the wave, this will not occur, because that would contradict quan-
tization of action. Let us assume that the only observable results of this
experiment are the records of the sensors. This record has the property
that it never occurs that two sensors simultaneously are triggered. After
a sufficiently large number of registrations the distribution of these
registrations among the sensors maps the intensity distribution of the
waves approaching the sensor row. But it seems as if the sensor row
has been hit by a number of particles. There is nothing in the records
telling us that waves were propagated on the water surface.
One could complain that this explanation does not show how the
collapse of the wave happens: we have not been given any new insight
into the details of the process. My answer to this argument is that this
complaint confuses philosophical and physical explanation; asking for
a detailed account of the collapse of the wave during interaction with
certain other objects is to ask for more physics, it is not the philosopher 's
task. Moreover, the indivisibility of the energy exchange implies that it
is impossible to describe further details. Hence it seems that a deeper
physical explanation is not to be had . This is the limit where physical

explanation must stop . But still I think this account has substantial
explanatory value, because we have a deeper understanding of quantum
phenomena. In my view, the salient feature which provides better under-
standing is that we no longer have to accept collapse of the wave function
as a brute fact, but it is shown to be a consequence of a previously
accepted trait of nature. Hence, the number of independently accepted
statements in the theory is reduced. That this is a form of explanation
is now a common view which is elaborated by Friedman" and somewhat
differently by Kircher.' Kitcher dubbed this form of explanation 'unifi-
cation' , and that term is now widely accepted, I think.
Schrodinger once proposed a wave interpretation of quantum
mechanics, claiming that the intensity 'P*(x, y, z)'P(x, y, z) of the wave
function is a literal description of the object in question. In the case of
the electron, the intensity of the wave function describes the density
distribution of the charge of a single electron governed by the stated wave
function . However, Schrodinger met severe criticism and was convinced
that his ideas could not be correct. The objections raised against his wave
interpretation," could be summarized in three points:
(l) Wave functions are complex whereas real waves must be real entities.
(2) The wave function for a multi-particle system is multi-dimensional;
the number of dimensions of the wave function is three times the
number of particles in the object system. How should we interpret
this wave function in physical three-dimensional space?
(3) The wave interpretation must take for granted that waves collapse
during interaction with other objects. What is the mechanism and
cause of this collapse?
1. The first problem can be solved if the real object is identified
with the intensity of the wave function because this quantity is always
real. Admittedly, the intensity, i.e., the squared amplitude, is not a wave
function so one could criticize this stance by saying that real objects
are not waves after all. However, the intensity has a number of features
in common with real macroscopic wave phenomena and it is quite
reasonable to extend the everyday word 'wave' to cover this type of
objects. Specifically, the intensity of a wave packet made up of a number
of monochromatic waves is a rather wave-looking phenomenon, and a
wave packet is a much more plausible model of a real object than a mono-
chromatic wave.
2. The second difficulty, the multi-dimensionality of many-particle
wave functions raises an interpretative problem only in EPR-type exper-

iments, i.e., when objects partake in non-local interaction. In order to
see this , consider first two non-interacting objects, a and b. The total
wave function for this pair is 'VIOl = cjla + cjlb' i.e., a sum of two independent
wave functions, one for each particle. We can ascribe to both objects a
definite state independently of the other. The situation is not funda-
mentally different from a system of classical particles in this respect.
However, if the particles are connected to each other non-locally, the
situation is different. The well-known wave function for a spin zero
pair,
'VIOl = ffI (cjla(+)cjlb(-) + cjla(-)cjlb(+))
shows this clearly. It is not possible to ascribe a definite state to each

particle independently of the other. So in this case the objection is valid.
However, this is no objection specifically against the wave interpreta-
tion . The scientific community has settled on the view that non-local
correlations are with us no matter what interpretation of quantum
mechanics we choose. Non-locality seems to be an irreducible fact of the
world and no philosophical interpretation can explain away this feature.
The multi-dimensionality of the wave functions thus merely tells us, in
the case of a correlated pair of particles, that the two waves are depen-
dent on each other. It does not make up an argument against viewing
objects as waves.
3. The objection that no explanation is provided for the wave collapse
can be given an answer analogous to the answer to the second objec-
tion. No one else has provided a physical explanation of the mechanism
of the collapse, so the wave interpretation fares no worse than its com-
petitors. And if you accept that the proper form of explanation is
unification, an explanation is in fact given . In any case , the advantage
of the wave interpretation is that the projection postulate is superfluous.
The collapse follows from the two assumptions made above , viz. , that
quantum objects are waves and that exchange of conserved quantities
is an indivisible process .
Using this wave interpretation we can resolve not only a number of
problems connected with wave-particle duality but some other problems
as well. One such problem is the interpretation of probability formulas
in quantum mechanics. The standard interpretation, due to Born, goes:
P(x, y, z) = W*(x , y, z)W(x, y, z) is the probability that we will detect

the object at the point (x, y, z) when performing a measurement.
It is well-known that this formulation cannot consistently be translated
so as to say that P(x, y, z) is the probability for the object to be in the
point (x, y, z) irrespective of any measurements. That this latter more
committing interpretation is impossible is a great obstacle for realism
because this impossibility forces us to say that the objects have no definite
properties before any measurements are made. However, this obstacle for
realism can be removed if we adopt the wave interpretation. A wave with
a non-negligible spatial distribution cannot reasonably be said to have
a precise position. It is natural to propose that the probability P(x , y, z)
= W*(x, y, z)W(x, y, z) is simply the part of the wave being in this
point. If we further assume that the probability for interaction at a certain
point equals the intensity of the real wave at that point, we have solved
the interpretative problem. Thus, P(x , y, z) = W*(x, y, z)W(x, y, z) is
not the probability for the entire object to be in (x, y, z), but the propen-
sity for interaction, i.e. the propensity for exchange of energy at this point.
This latter formulation fulfils realistic demands and if we assume that
the proportion of the wave at a certain point equals the probability for
interaction with measuring devices at this point, the interpretation is
empirically equivalent to the Born interpretation.
But, it could be argued, using propensities in the explanation is a
case of obscurum per obscurus. Propensities, and dispositions in general,
are as much in need of clarification as are quantum probabilities. So
we need a further analysis. The standard analysis of a disposition is in
terms of inner constitution. Fragility, for example, is explained in terms
of the strength of chemical bonding. Could we find an analogous expla-
nation of the concept of propensity? Lawrence Sklar has argued that a
general explanation of propensity is not to be had, because the propen-
sity and hence the probability for an event depends on which class the
event belongs to and there seems to be no objective way of classifying
events. The propensity for an event depends on what we happen to
know about the conditions, not on which conditions obtain. This argument
we must accept, I think, but what if we know everything there is to
know about the conditions and the future development still is indeter-
minate? What more can one say than there is a certain propensity for a
change in a specified direction? Nothing, I think, and the request for
an analysis of the concept of propensity when applied to irreducibly inde-
terministic events is to ask for too much. Hence, propensities could be

acceptable in an analysis of indeterministic quantum phenomena.
Lawrence Sklar hinted at the same conclusion. He rejected propensi-
ties as a general analysis of probability, but admitted that in the case
of quantum mechanics, propensities might be acceptable because quantum
events are irreducibly indeterministic."
In conclusion: two important problems for a realistic interpretation
of quantum mechanics can be solved by assuming that quantum objects
are waves which exchange conserved quantities in discrete steps.
Stockholm University,
Sweden.
NOTES
I The ideas presented in this paper are part of my doctoral thesis Understanding Quantum
Mechanics ; a Realist Interpretation without Hidden Variables, Almquvist & Wiksell
International, Stockholm, 1992.
2 Op. cit.
3 The ~oncept of action was introduced by d' Alembert. It denotes the time integral of
energy [Edt.
4 Cf . Bohr, N. (1934). Atomic Theory and the Description of Nature , Cambridge
University Pres s, Cambridge, p. 53.
5 Cf. Cohen , L. (1973) . 'Joint Prob ability Distributions in Quantum Theory ' , in C. A.
Hooker (ed .), Contemporary Research in the Foundations and Philosophy of Quantum
Theory, Re idel, Dordrecht.
6 Friedman, M. (1974) . 'Explanation and Scientific Under standing ' , J. Phil. 71, no. I,
January 17.
7 Kitcher, Philip (1981) . ' Explanatory Unifi cat ion ', Phil. Sci. 48 , pp. 507-531.
K For an overview of this discussion see Jammer, M. (1974) . The Philosophy of Quantum
Mechanics, John Wiley & Sons , New York , pp. 31-33.

9 Sklar, L. (1970) . ' Is Probability a Dispositional Property? ', J. Phil. 67(11), June I I,
pp. 355-366.
LARS-GORAN JOHANSSON
VAN FRAASSEN'S CONSTRUCTIVE EMPIRICISM

- A CRITIQUE
Van Fraassen's Constructive Empiricism rests on the distinction between

believing a theory to be true and believing it to be empirically adequate.
This distinction in turn rests on an assumption which seems doubtful
in the light of Goodman's paradox.
According to van Fraassen, I to believe a theory to be empirically
adequate is to believe that the theory has a model, a part of which is
isomorphic with all appearances. An appearance is a structure which
can be described in a measurement or observation report. Van Fraassen
takes pain to point out that this applies to all appearances, whether
actually observed or not. In other words: a theory is empirically adequate
if and only if a part of this theory is isomorphic with an infinite set of
observation statements.
That means that to believe a theory to be empirically adequate is to
use induction: having made a number of observations and measure-
ments and found that a theory accounts for all observed phenomena,
we might conclude that the theory is empirically adequate. Thus we
are assuming that the theory also in the future will be in accord with
our observations. Hence, van Fraassen claims that induction from
observed to unobserved but observable phenomena sometimes can be
justified.
On the other hand he claims that we are never justified in inferring
the truth of a theory. To believe a theory to be true means to believe
that it correctly represents the world and such a belief can never be
justified according to van Fraassen. His reason for this rejection is under-
determination. If any amount of evidence is insufficient to distinguish
between a number of incompatible theories, then none of the alterna-
tives is more trustworthy than the rest. But then we must admit that
theoretical statements cannot justifiably be believed to be true . This
does not mean that they are void of meaning. Van Fraassen's position
does not imply instrumentalism, he merely claims that we must stay
forever agnostic about theoretical statements.
Obviously, van Fraassen's position rests on the distinction between
observational and theoretical statements and the main criticism against
339
R. S. Cohen , R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-Realism in the
him has focused on this issue. He has not been able to tell us exactly
where to draw the distinction between theoretical and observational state-
ments, i.e., the crit ics have argued he has exactly the same problem as
the old pos itivists. His answer to this criticism could be summarized as
follows:
1. Constructive empiricism does not face the same problem as logical
empiricism because there is no semantic difference between theoretical
and observational statements according to constructive empiricism. Both
observational and theoretical statements are meaningful and have
truth-values, whereas in logical empiricism theoretical statements lack
truth-value. This implies that for the constructive empiricist it is accept-
able to have a somewhat vague distinction observationalltheoretical. It
suffices if there are some clear cases on both sides.'
2. It is not the philosopher's task to draw a general distinction between
observational and theoretical statements. Such a distinction is the business
for scientists, because it is an empirical question, a question about us qua
organisms in the world.'
3. The alternative of not using the distinction observational/theoret-
ical is either to believe every accepted theory to be completely true, or
to be a sceptic. Rejecting both these extreme positions, something in
between must be found and if we want to say more about a scientific
theory than it fits the facts so far, it is reasonable to say that it will
continue to do so, i.e., that it is empirically adequate."
It seems to me that these remarks make up a rather good defense of
using a somewhat vague distinction observational/theoretical.
However, I believe that there are other grounds for disbelieving van
Fraassen's distinction. I want to recall Goodman's discussion about induc-
tion in which he introduced the famous example of the predicate grue,
which "applies to all things examined before t just in case they are
green but to other things just in case they are blue".' It is clear that
Goodman means that the time t lies in the future, hence it is plain that
all hitherto available evidence is equally good evidence for emeralds
being green as well as being grue. So simple induction cannot distinguish
between projectible and non-projectible predicates and we can not
distinguish between law-like and accidental generalisations."
Goodman's conclusion was that the induction problem cannot be
solved using any merely syntactical criterion. In other words, we cannot
solve the induction problem by only considering the formal relations
between the proposed general hypothesis and its evidence, i.e ., its
VAN FRAASSEN 'S CONSTRUCTIVE EMPIRICISM 341
instances. Goodman proposed the concept of entrenchment as additional

vehicle: among a number of rival hypotheses with equal evidence we
shall choose the most entrenched one . This is admittedly not very clear.
What sort of criteria do we use when deciding out of a number of rival
hypotheses which one is the most entrenched?
In the present context there is a crucial question: 'is it possible to
distinguish between rival hypotheses without using statements which van
Fraassen would count as theoretical statements?' In other worlds, can
we solve the induction problem solely within the observational realm?
It seems impossible. Singular observational statements cannot do the job:
either they are instances of all the rival hypotheses or they are irrele-
vant. Neither is it possible to rely on another general hypothesis about
observable states of affairs: that move only triggers the ensuing question:
how do we know that this new general hypothesis is a law-like gener-
alisation and not merely an accidental generalisation? It is clear that
we still have the same problem. These two alternatives, singular obser-
vational statements or generalisations about observable states of affairs,
outrun our possibilities of finding distinguishing evidence in the domain
of statements which according to van Fraassen can be known to be true.
Hence, we must use theoretical statements when choosing the better
entrenched hypothesis among a number of altern atives.
Now, the question is 'c an van Fraassen use this way out of the induc-
tion problem?' No he cannot! Theoretical statements can according to
constructive empiricism never be known to be true. But then they cannot
be used as evidence for or against any hypothesis. Using an unsettled
statement in an inductive inference to a statement which we claim
to be true (a statement about a future observable event) is plainly
irrational.
The conclusion to be drawn is that van Fraassen cannot accept
Goodman's conclusion that we must go outside the realm of singular
observational statements when making reliable inductive generalisations.
Van Fraassen's constructive empiricism in fact presupposes exactly what
is denied by Goodman and that throws serious doubt on van Fraassen's
distinction between believing a theory to be true and believing it to be
empirically adequate.
Of course, van Fraassen could reject Goodman's conclusion by pro-
viding a method for distinguishing between law -like and accidental
gen eralisations without using theoretical statements. What he cannot
do is to take for granted that this can be done. But that is precisely
342 LARS -GORAN JOHANSSON
what he does when he makes the distinction between believing a theory

to be true and believing it to be empirically adequate.
Stockholm University,
Sweden.
NOTES
I Cf. Fraassen , Bas van (1980). The Scientific Image, Clarendon Press, Oxford, p. 12.
2 Ibid., p. 16.
3 Ibid., pp. 57-58.
4 Cf. Fraassen, B. van (1985) . 'Empiricism in the Philosophy of Science ', p. 254, in
Churchland and Hooker (eds.), Images of Science , The University of Chicago Press,
Chicago.
5 Cf. Goodman, N. (1972). 'The New Riddle of Induction ', p. 318 in his Problems and
Projects, Bobbs-Merrill Inc., Indianapolis & New York.
6 Ibid., p. 388.
LILI
ON COMPLEMENTARITY REALITY
Many interpretations have been given to quantum mechanics.

Unfortunately, no one of them is completely successful and satisfac-
tory. The most influential interpretation is the 'orthodox Copenhagen
interpretation', which has been criticized as a positivistic and vague
theory, and which always concludes with anti-realism. The realistic
'hidden variables theory', recently of greater and greater influence,
faces the problems of nonlocality and unobservability, and it does not
have more content than the orthodox theory . What I want to argue is
that we should and can establish a realistic foundation for the anti-
realistic Copenhagen interpretation - that is a foundation of comple-
mentarity reality to provide a realistic and reasonable interpretation,
and to avoid the problems in interpreting quantum mechanics.
I. PROBLEMS IN THE COMPLEMENTARITY INTERPRETATION
Why do people feel that complementarity is difficult to understand?

The enthusiastic advocate of complementarity, von Weizsacker, gets a
definitely negative answer from Bohr after his most careful study of it.
The famous opponent of complementarity, Einstein, complains that :
"despite much effort which I have expended on it, I have been unable
to achieve a sharp formulation of Bohr's principle of complementarity.'"
And in the current argument between realism and anti-realism, why is
Bohr considered as a realist by some people and as an anti-realist by
others? I think the reason is that there are some hidden contradictions
in the complementarity principle, which seem to fit Bohr's aphorisms
that "the opposite of a deep truth is also a deep truth"; "truth and clarity
are complementary.'? These contradictions give to complementarity a
sense of vagueness, mystery and inconsistency.
1. About the Reality of Atomic Systems
Even though Bohr still uses the term 'atomic object', emphasizes the
objectivity of the description of atomic phenomena, and tries to exclude
343
R. S. Cohen , R. Hilpinen and Qiu Renzong (eds .), Realism and Anti-Realism in the
344 LILI
the subject from the phenomena,' in fact, he does not consistently keep
the reality of the atomic system in the complementarity interpretation.
In fact, conversely, when he explains atomic phenomena by his com-
plementaryprinciple, he denies the independent reality of the atomic
system. In his Como Lecture, he writes:
now the quantum postulate implies that any observation of the atomic phenomena will
involve an interaction with the agency of observation not to be neglected. Accordingly,
an independent reality in the ordinary physical sense can neither be ascribed to the
phenomena nor to the agencies of observation."
Obviously, he denies the independent reality of the atomic object and

holds that the reality of the atomic object cannot be independent of obser-
vation. He notes that
the finite interaction . . . entails . . . the necessity of ... a radical revision of our altitude
towards the problem of physical reality,' of course there is in a case like that just con-
sidered no question of a mechanical disturbance of the system under investigation during
the last critical stage of the measuring procedure . .. there is essentially the question of
an influence on the very conditions which define the possible types of predictions regarding
the future behaviour of the system, since these conditions constitute an inherent element
of the description of any phenomena to which the term ' physical reality' can be properly
attached."
He denies the assumption of E-P-R that there is an independent reality

of the object without disturbance by observation, and proposes that the
term 'physical reality' must be revised so that it include not only the
instrument which interacts with the object, but also the circumstances
of the object, and these influence the conditions for predictions about
the object. The dependent reality of the object results in the positivist
interpretation of quantum mechanics. Bohr remarks
there is no quantum world . There is only an abstract quantum physical description/
quantum mechanical formalism which represents a purely symbolic scheme . . . ;8 In
this theory . .. in every stage of the computation only directly observable quantities
enter .. . the new quantum mechanics does not deal with a space-time description of
the motion of atomic particles;" the properties of atoms are always obtained by observing
their reactions under collisions or under the influence of radiation; 10 radiation in free space
as well as isolated material particles are abstractions, their properties on the quantum theory
being definable and observable only through their interaction with other systems. II
These quotations show that Bohr denies that quantum theory describes
the independent reality of the object and asserts that the properties of
the atomic object can exist and be described and observed only under
ON COMPLEMENTARITY REALITY 345
observation. He does not assume the independent reality of the atomic

object and only admits the reality of the object with its circumstantial
content under observation.
The denial of the independent reality of the object leads to the fol-
lowing problems in Bohr's interpretations: (A) If there is not an
independent reality of the object, how can he say that the quantum theory
describes the phenomena of the atomic object? What is the meaning of
the "atomic object"and how can he explain the law which the quantum
theory describes if the reality of the atomic object depends on the various
and changeable circumstances? (B) If there is no independent reality
of the object, what do the instruments which are used to observe the
atomic phenomena act on, and how are the atomic phenomena created?
Are the phenomena created by the instruments alone which have the
ordinary meaning for their reality, and how can the object get its prop-
erties under observations? (C) If the reality of the object depends on
our observations, the existence of the world would depend on our obser-
vations. But our practice in an experiment without an observer, and
the existence of the world before our existence, both contradict this
conclusion.
In fact, mutually exclusive phenomena under different observations
reflect the fact that showing the properties of the object depends on its
circumstances such as provided by the instruments but also properties
both of the object and of the instrument are independent of any observer
and his observation. Otherwise, there must be the problems mentioned
above in the interpretation.
2. About' the Reality of Complementarity

In the article 'Quantum Physics and Philosophy', Bohr states that the
complementarity description is similar to relativity theory in keeping
objectivity and avoiding subjectivity. In his 'Discussions with Einstein
on Epistemological Problems in Atomic Physics', he notes that
the tenn 'complementarity,' suited to embrace the char acteristic features of the individ-
uality of quantum phenomena . . . evidence obtained under different experimental
conditions cannot be comprehended within a single picture but must be regarded as
complementary in the sense that only the totality of the phenomena exhausts the possible
information about the object. 12
From these words, it seems he believes that the complementarity

represents an objective attribute of the phenomena of the atomic objects,
346 LILI
and it seems that it can be seen as representing the reality of the objects.
But in fact he does not admit the reality of complementarity. He points
out that the concept of complementarity represents ".. . the objective
existence of phenomena and is related to the method of observing
them."!' That means that complementarity does not represent an objec-
tive attribute of the phenomena. It represents the dependence of
phenomena on observation of them. It does not derive from the mutually
exclusive properties of the reality of the objects. It derives from the
different observations which create mutually exclusive phenomena of the
objects. This meaning is also shown in these words of Bohr: "... the
study of the complementary phenomena demands mutually exclusive
experimental arrangements."!" This says that the complementary phe-
nomena result from the mutually exclusive experimental arrangements,
not from the reality of the objects. But how do the instruments create
mutually exclusive phenomena? He explains:
According to the quantum theory, just the impossibility of neglecting the interaction
with the agency of measurement means that every observation introduces a new uncon-
trollable element . . . the measurement of the positional co-ordinates of a particle is
accompanied not only by a finite change in the dynamical variables, but also the fixation
of its position means that a complete rupture in the causal description of its .. . momentum
always implies a gap in the knowledge of its spatial propagation."
That is to say, the interaction with the instruments destroys the foun-
dation of a consistent description of the object and causes a gap in our
knowledge of the object, which results in the complementarity of our
knowledge about the objects. This idea is expressed in another style in
the following words of Bohr:
the occurrence of individual effects [permits] neither a choice of ' nature ' nor a choice
of ' observer', that only can be accounted as we are dealing with individual phenomena
and that our possibilities of handling the measuring instruments allow us only to make
a choice between the different complementary types of phenomena we want to study . 16
In other words, the complementarity description does not describe

the choice of nature; it represents our limited ability in handling the mea-
suring instruments. That is why he always relates the complementarity
to the failure of our forms of perception. For example, he declares,
a close connection exists between the failure of our forms of perception, which is founded
on the impossibility of a strict separation of phenomena and means of observation, and
the general limits of man's capacity to create concepts, which have their roots in our
differentiation between subject and object. J7
ON COMPL EMENTARITY REALITY 347
These words show that complementarity, which indicates the impossi-

bility of a strict separation of phenomena from means of observation,
establishes the failure of our forms of perception, and has a close con-
nection with the impossibility to differentiate subject and object. This
results in the general limits of man 's capacity to create concepts. In short,
complementarity has a close relationship with the impossibility to dif-
ferentiate subject and object. So it does not represent the reality of the
objects.
His denying of the reality of complementarity leads to the following
problems: (A) If complementarity results from the mutually exclusive
instruments, why do different instruments react on the same object
according to mutually exclusive styles? Why are the different instruments,
which have the same physical reality in ordinary common sense usage,
mutually exclusive? (In fact, they are mutually exclusive just because
they show mutually exclusive phenomena of the atomic subjects.) (B)
If complementarity shows the limitation of our perception and the
capacity to create concepts, the reality of the independent world would
be beyond our epistemic ability, and it would be unknowable. This con-
clusion contradicts the scientific object in its capacity to recognize the
reality of nature .
Actually, both the objective existence of mutually exclusive phe-
nomena of the objects, together with successful complementary
description of the object prove the reality of the complementarity. The
complementarity is the reality of the object.
3. About the Reality of the Instruments

On the one hand, Bohr does not consider the instrument as indepen-
dent reality when he says that, because of the interaction between the
instruments and the objects, "an independent reality in the ordinary
physical sense can neither be ascribed to the phenomena nor to the
agencies of the observation.':" which means the instruments have no inde-
pendent reality, just as is said of the objects. So in Bohr's analysis of
the experiments in "Discussions with Einstein on Epistemological
Problems in Atomic Physics", the momentum of the instrument is also
uncertain when its position is certain. But if the instrument is also depen-
dent and uncertain, how can it interact with the object and how can we
get definite phenomena from it? On the other hand , in order to get the
complementary phenomena and descriptions, he considers the instruments
348 LILI
as certain and an independent reality in the ordinary physical sense. He

claims that in the analysis of the quantum phenomena, the essentially
new character is the fundamental differentiation between the instru-
ments and the objects, which is the result of the necessity that we must
use classical concepts to explain the function of the instrument and we
exclude any consideration about the quantum of action . 19
The instrument is described as having double realities: independent
and certain reality in the classical physical sense and dependent and
uncertain reality in the quantum physical sense. How can it play double
roles at the same time? And if the instruments are considered as inde-
pendent reality and can be described by classical concepts, the atomic
objects would belong to a different world from the instruments. This con-
clusion contradicts the consistency principle of science.
About the influence of the instrument on the object, Bohr again
provides two explanations. On the one hand, it is considered as physical
interaction, in which there is exchange of energy and momentum. On
the other, in these analysis of the E-P-R experiment, it is considered as
an influence upon the conditions of the types of predictions of the objects
and there is no mechanical disturbance in the process (as note 6 notes) .
What the actual influence of the instrument on the object is remains
unclear. In reality, in order to keep scientific con sistency, we must
consider the instruments to belong to the same world of the objects. Both
of them must be considered as independent reality but their properties
are uncertain according to quantum theory. When they exist in a certain
relationship, such as when the instrument is used to measure a certain
property of the object, both of them show a certain property and certain
phenomena. What is special is that their showing of the properties depend
on the relationship between them , but not the mysterious influence of the
dual-roles -instruments.
4. About the Reality of Physical Concepts in Quantum Theory

We ask whether the physical concepts in quantum theory represent any
reality and what reality do they represent? Bohr declares that the clas -
sical theory cannot describe atomic objects: "the quantum theory is
characterized by the acknowledgement of a fundamental limitation in the
classical physical ideas when applied to atomic phenomena.t'i? the essen-
tial discontinuity or rather individuality in any atomic process is
"completely foreign to the classical theories.'?' Correspondingly, the
classical concepts are also considered limited in describing atomic

phenomena, which is shown in these words of Bohr:
the fundamental contrast between the quantum of action and the classical concepts is imme-
diately apparent from the simple formulae . . . ET = II. = h.22 The Limitation in the classical
concepts expressed through relation (2) (lHl'iE = l'ixl'iI, = l'iYl'iIy = l'iZl'iI, = h), is, besides,
closely connected with the limited validity of classical mechanics."
But he contends that only the classical concepts can be used to defi-
nitely explain the result of the experiment. He asserts:
however far the phenomena transcend the scope of classical physical explanation, the
account of all evidence must be expressed in classical terms."
Even though the classical concepts are limited in describing atomic

phenomena, the phenomena must be expressed in classical terms. How
can we do that? His analysis:
on one hand, the definition of the state of a physical system , as ordinarily understood,
claims the elimination of all external disturbances. But in that case, according to the
quantum postulate, any observation will be impossible, and, above all , the concepts of
space and time lose their immediate sense. On the other hand, if in order to make obser-
vation possible we permit certain interactions with suitable agencies of measurement,
not belonging to the system, an unambiguous definition of the state of the system is
naturally no longer possible, and there can be no question of causality in the ordinary sense
of the WOrd.25
According to Bohr, the definition of state claims the elimination of all

observation including the observation of the state. The definition of
state and the observation of it are mutually exclusive. If so, what is the
meaning of the 'observables' in quantum mechanics and how can we
get a definite description of the state of the object from an experiment
on it? According to Bohr's words above, the definition of state excludes
observations of the state and the space and time as well as the meaning
or definition of space and time. But in the analysis of the ideal experi-
ments in 'Discussion with Einstein on Epistemology Problems in Atomic
Physics', the observation of the energy and the momentum make it
possible to get the definite description of the energy and the momentum,
which implies the definite definition of them, and bring uncontrollable
changes of time and space making both the definitions and the obser-
vations of them impossible. And the observation of space and time
make it possible to describe the space and the time and bring uncon-
trollable changes of momentum and energy, making both the definitions
and the observations of them impossible. The observation of the state
350 LILI
both excludes and infers the definite definit ion of it. The relations among
the observations and definitions of the energy, the momentum, the space
and the time of the object are ambiguous and entangled. How these
concepts can properly be used in quantum mechanics to become the
' observables' is unclear.
Bohr declares:
an adequate tool for a complementary way of description is offered precisely by the
quantum -mechanical formalism which represents a purely symbolic scheme permitting
only predictions. on lines of the correspondence principle . as to result s obtainable under
conditions specified by means of classical concepts."
This means the quantum theory is merely a symbolic scheme which

can only predict the result under conditions specified by means of clas-
sical concepts. But how can the conditions be specified by means of
classical concepts? Bohr claims:
In the quantum-mechanical description our freedom of constructing and handling the
exper imental arrangement finds its proper expres sion in the possibility of choosing the
classically defined parameters entering in any proper application of the formal ism."
This means that the proper use of classically defined parameters is con-
ditioned by the proper constructing and handling of the experimental
arrangement. But how can the proper experimental arrangement make
the limited classical parameters suitable to the atomic phenomena if
the observation in atomic phenomena is essentially different from clas-
sical observation? Bohr does not give a clear explanation. He just declares
ambiguously:
the very nature of the quantum theory thus forces us to regard the space-time co-ordination
and the claim of causality. the union of which characterizes the classical theories. as
complementary but exclusive features of the description. symbolizing the idealisation of
observation and definition respectively . . . In the description of atom phenomena. the
quantum postulate present us with the task of developing a complementarity theory the
consistency of which can be judged only by weighing the possibilities of definition and
observation."
That is to say, the space-time co-ordination and the claim of causality are
exclusive. We get definite space-time co-ordination only under the con-
dition of the impossibility of the claim of causality. The limitation of
classical concepts is overcome by using the two exclusive descriptions
without combining them in one picture. The complementary descrip-
tions can describe the atomic phenomena but the complementarity
does not come from the reality of the objects, it comes from the limi-
tation of the classical concepts. The classical concepts can represent

the reality of the atomic phenomena through the method of com-
plementarity.
As a matter of fact, the de Broglie relation and the uncertainty
relations show that the concepts of momentum p, position q, energy
E and time t in quantum theory are no longer the classical concepts,
because the classical p.q.E. t have not these special relationship. They
have the classical meaning when the instrument and the objects are in
certain relationships and they show their proper p.E. or q.t. So the
physical parameters in quantum theory represent the reality of the atomic
objects.
In brief, the presumption of the uncontrollable, interaction between
the object and the instrument which is the core of the complementarity
interpretation, is bound to the denial of the independent reality of the
object and the instrument, and the limitation of the classical concepts,
which cannot be observed and defined simultaneously. But in order to
get definite phenomena and a finite description of them, the indepen-
dent reality of the instrument and the proper application of the classical
concepts under the certain using of the instrument must be introduced
in the interpretation. The interaction between the object and the instru-
ment actually presupposes the independent reality of the object and the
instrument. These contradictions are all caused by the presumption of the
uncontrollable interaction between the object and the instrument. In
reality, no uncontrollable interaction between the object and the instru-
ment happens. There are only the different and certain relationships
between them, which determine the different properties and the dif-
ferent phenomena that they show.
II. THE VIEW OF COMPLEMENTARITY REALITY AND

ITS INTERPRETATION OF QUANTUM THEORY
The view of complementarity reality can be concluded from the analysis

in Part I: both the object and the instrument are independent comple-
mentarity realities which contain mutually exclusive properties (which
are described by uncertainty relations); each of the properties is uncer-
tain when the object has no certain relationship with its circumstantial
content. When the object is in a certain relationship with its circumstance
the object changes to show the certain property which corresponds to
the certain relationship at the same time, the circumstance such as the
352 LILI
measuring instrument also changes to show the same kind of property.

Because of the quantum state of the object and the process of the change,
the objects statistically show certain values of the property under certain
relationship with its circumstance.
I. The Reasonableness of the View of Complementarity Reality

(A) The view of complementarity reality can resolve the problems in
the complementarity interpretation. (a) in this view, both the object and
instrument are independent complementarity reality, which can avoid
the problem that the reality of the object depends on observation and that
the instrument belong to a different world from the object. The depen-
dence of the atomic phenomena on the observation does not mean the
reality of the object and the instrument depend on the observation. It
means that the certain presentation of reality depends on the certain
relationship between the object and the instrument. (b) According to
this view, the mutually exclusive phenomena directly derive from the
complementarity reality of the object, which avoids the problem of
attributing the complementary phenomena to different instruments and
the limitation of man's epistemic ability, (c) In the light of this view,
the influence between the object and the instrument is mutual. Their
certain relationship determines their certain properties, which avoids
the problem of the unclear and different interactions between the instru-
ments and the objects. (d) On the basis of this view, quantum mechanics
describes the complementarity reality of the object. The uncertain descrip-
tion of the state of the object without measurement, and the certain
description of the state of it under measurement in quantum mechanics,
reflect the uncertain state of the complementarity reality of the object
without a certain relationship with its circumstance and a certain state
of it under certain relationship. It avoids the problem of using the limited
classical concepts to describe the atomic phenomena, in which the
concepts have no clear meaning and do not describe the reality of the
object but merely describe the phenomena through the complementary
method .
(B) The complementarity reality is consistent with facts, theory, and
practice in the atomic field. The complementarity reality is shown by
the objective complementary phenomena of the atomic objects. The fact
that our view of complementarity reality can reasonably interpret quantum
mechanics (as this paper shows later) indicates that the complemen-
tarity reality is described by the successful quantum theory . In practice
the processes of the atomic objects (without observers) obey the quantum
theory which shows the independent reality of the objects and the objec-
tivity of the quantum theory. In fact, in our ordinary common sense field ,
there are also some remarkable phenomena which show the comple-
mentarity reality. For example, the metaphor that Bohr always quotes,
that we are both actor and audience in the stage of life, describes our
complementary properties. When we are on the stage, relating to the
audience, we are actors, but we also can suddenly become the audience
when watch the actors on the stage. When there is no stage, or all
people are on the stage, our properties are uncertain. We have the
mutually exclusive properties of actor and audience, our property is
certain when we have a certain relationship with our circumstance. We
change from uncertain property to a certain property, or from one certain
property to another one, when our relationship with our circumstances
changes. But the ordinary object always show a certain property and it
is difficulty to observe its complementary properties, because an ordinary
object always exists in a certain relationship with its circumstances, which
are the ordinary world in which we live. Therefore the complemen-
tarity reality is general, not suitable only to the microfield. The
complementarity does not derive from our limitated ability to observe
and recognize the atomic objects.
(C) The view of complementarity reality can keep the consistence
between ontology and epistemology. It assumes that what we describe
and observe are what exist.
(D) The view of complementarity reality is consistent with the
scientific belief which is the foundation of all scientific researches. The
scientific belief is definitely described by Einstein as "the belief in an
external world independent of the perceiving subject is the basis of all
natural science.?" It is also expressed by Bohr, when he notes that the
description of atomic phenomena is absolutely objective - it does not
involve any observer.3D The scientific belief that there must be reality
outside of us and independent of our observation, and that our theories
objectively describe it, is retained in quantum mechanics by the view
of complementarity reality.
2. The Reasonable Interpretation of Quantum Mechanics

in the View of Complementarity Reality
(A) About the Wave Function. According to the view of complementarity

reality, the '\jf function describes the uncertain state of the complemen-
354 LILI
tarity reality of the object. Because the complementarity reality contains

mutually exclusive properties, its state is uncertain and each of the
properties is uncertain, but the range of the properties obey the uncer-
tainty relations. The properties of the energy E and the momentum P, and
the properties of the position q and time t, are the essential mutually
exclusive properties of the object. The object cannot be described by
space-time description of its causal motion. The properties of E and P
and the properties of q and t are not certain at the same time. The 'If
function reflects the attributes of the uncertainty and mutual exclusion
of the complementarity reality of the object. The Schrodinger equation
(H'If = ih <hv/dt) reflects the law of complementarity reality. The 'If(K)2
describes the probability of the change of the object to a certain k value
of a certain k property under a certain k relationship with its situation.
For example, when the position of the object is measured, the object
has a certain position relationship with its circumstance (the instru-
ment), and 1'If(x)1 2 describes the probability of the transition of the object
to a certain positional value x. When the result of the measurement of
the po sition of the object is An, the state of the object is one of the
eigen-functions <1m (An is the eigenvalue of position, <l>n is the eigen-
function of position).
(B) About the Observables. In the view of complementarity reality,
the concept ' observables ' should be changed to the mutually exclusive
properties of the object, because the exi stence of these properties is inde-
pendent of observations, even though the certain appearance of the
properties depends on observations. The certain extents of the appear-
ances of the mutually exclusive properties are determined by the
uncertainty relations [~p~q ~ (-h), ~EM ~ (-h)], which reflect the mutu-
ally exclusive relations between the extents of the appearances of the
complementary properties of the object by the reciprocal relations. The
conventional axiom about the observables in quantum mechanics is:
the probability of observables (A, B, . ..)
if observed at time t
being observed to be (a, b, ...)
is
II{a, b, ... qlt)f
where q denotes additional quantum numbers which together
with the eigenvalues (a, b, . . .)
form a complete set.
should be replaced by
the probability of the properties of complementarity reality

(A, B, .. .).
under the (A, B, . . .) relationship with its circumstance at
time t
being (a, b. .. .)
is
II(a, b, . .. qlt>1 2
where q denotes additional quantum numbers which together
with the eigenvalues (a,' b, . . .)
form a complete set.
(C) About Measurement. In the light of the view of complementarity

reality, the measurement does not mean an uncontrollable interaction
between the object and the instrument; it means the changes of both
the instrument and the object to show certain properties under certain
relationship between them in the measurement. Because both the object
and the instrument are the independent complementarity reality, the
appearances of their properties are uncertain before the measurement.
When they are given in the measurement, their certain relationship deter-
mines their changes to show the same certain property. When the
appearances of their properties are certain, the situation in the mea-
surement is similar to that in the classical measurement, i.e. the
instrument reflects the property of the object. Then the classical concept
which describes the property in the measurement can reasonably be used,
even though at the same time the other concept (which is mutually exclu-
sive to this concept) can not be used since the object does not show
the other property under this particular certain relationship of the object
with its circumstance. The wave packet reduction in quantum mechanics
reflects the change of the object from an uncertain state to a certain value
of a certain property under the certain relationship with its circum-
stance.
(D) About the E-P-R Paradox. On the basis of the view of comple-
mentarity reality, the E-P-R paradox reflects the relative attribute of
the complementarity reality, which shows a certain property under a
certain relationship with its circumstance. In the E-P-R experiment, when
object A interacts with object B, there is a relationship between them.
When they do not interact to each other, the relationship still exists.
356 LILI
This relationship determines that both of them show the same property
of their complementarity reality. When object A is measured and changes
to show its certain momentum. P A (or position qA)' the object B also
changes to show its certain momentum P B (or position qB) because the
object A and the instrument acting on object A, which are the circum-
stance of object B, determine the certain appearances of the properties
of object B. Therefore, the E-P-R paradox does not mean we get the com-
plementary descriptions of object B simultaneously without disturbing it.
It just shows the radical relativity of the appearances of the properties
of the complementarity reality of the object to its circumstance.
Hebei Academy of Social Sciences,
Shijiazhuang, China.
NOTES
1 Jammer, M. (1974) . The Philosophy of Quantum Mechan ics , John Wile y & sons , pp.
90-91.
2 Bell, J. S. (1987). Speakable and Unspeakable in Quantum Mechanics , Camb. Univ .
Pre ss, p. 190.
3 Bohr, N. (1958) . Quantum Physics and Philosoph y - Causality and Complementarity.
4 Bohr, N. (1934). Atomic Theory and the Description of Nature , Camb. Univ . Press,
p.54.
5 Schilpp, P. A. (ed .) (1949) . Albert Einstein: Philosophyer-Scientist, Tudar Pub . Co.,
p.233.
6 Ibid., p. 234.
7 Jammer, M . The Philosophy of Quantum Mechanics, p. 204.
s Schilpp, P. A. (ed .), ibid., p. 210.
9 Bohr, N., ibid., p. 48.
10 Bohr, N., ibid., p. 95.
11 Bohr, N., ibid., pp. 56-57.
12 Schilpp , P. A. (ed.), ibid., pp. 209-210.
13 Bohr, N. (1958) . Quantum Physics and Philsophy - Causality and Complementarity.
14 Schilpp, P. A. (ed .), ibid., p. 21 I.
15 Bohr, N., Atomic Theory and the Description of Natur e, p. 68.
16 Schilpp, P. A. (ed .), ibid., p. 223.
17 Bohr, N., ibid., p. 96.
IS Bohr, N., ibid., p. 54.
19 Bohr, N. (1958) . Quantum physics and philosophy - Causality and Complementarity .
20 Bohr, N., Atomic Theory and the Description of Nature, p. 53.
21 Bohr, N., ibid., p. 53.
22 Bohr , N., ibid., p. 57.
23 Bohr, N., ibid., p. 60.

24 Schilpp , P. A. (ed.), ibid., p. 209.
25 Bohr, N., ibid., p. 54.
26 Schilpp , P. A. (ed.), ibid.. pp. 210--211.
27 Schilpp, P. A. (ed.), ibid.• pp. 230.
2K Bohr, N., ibid.• pp. 54-55.
29 Schilpp, P. A. (ed.), ibid.. p. 248.
30 Bohr, N. (1958) . Quantum Physics and Philosophy - Causality and Complementarity.
31 Jammer, M. (1966) . The Conceptual Development of Quantum Mechanics, McGraw-
Hill Book Comp o
32 Cohen, R. S. and Stachel, J. (ed .) (1979). Selected Paper of Leon Rosenfeld, D.
Reidel Pub. Comp., 1979.
33 Bohrn, David (1954) . Quantum Theory, Prentice-Hall Inc., New York.
34 Ge Ge (1985) . Niel's Bohr - His Life, Research. and Ideas, Shanghai People 's Pub
House.
35 Zhou, Shi-xun (ed.) (1979). A Course of Quantum Mechanics, People's Educational
Pub. House, Beijing .
36 Lu, He-Fu (1984). Textual Interpretation of the Quantum Theory of the Copenhagen
School, Fudan Univ. Pub. House, Shanghai.
37 Guo, Gui-chun (1991). Contemporary Scientific Realisms. Scientific Pub . House,
Beijing.
38 Wu, Da-You (1984) . Quantum Mechanics, Scientific Pub . House, Beijing .
LUO JIACHANG AND HU XINHE
RELATIONAL REALISM ON REFORM OF

THE VIEW OF PHYSICAL REALITY AND
ITS LOGICAL MANIFESTATION
The interpretation of quantum mechanics (QM) is one of the most con-

troversial issues of 20th century science and philosophy of science. As
some scientists and philosophers mentioned, many problems in it, such
as wave collapse and the measurement problem, wave-particle duality
and non-v isualizable reality, the EPR argument and correlation-at-a-
distance, etc., concerned both physical theory and philosophical concepts.
Many quarrels came from the divergences of the disputers' philosoph-
ical views, especially of their view of reality. Then, a key question for
interpretation of QM is whether we should and can hold on to the clas-
sical view of reality after such a radical change from classical physics
(CP) to QM, and if not, how should we modify it? By examining the
classical view of reality, the conceptual features of QM, and some affir-
mative interpretations of QM, we try to give a somewhat new view of
reality, its logical form, and some applications of it. We hope that our
tentative answer to this question may be helpful to provide a new per-
spective in understanding physical reality.
I. THE CLASSICAL VIEW OF REALITY
The classical view of reality could be called " the view of substance
(orientities)", or "the view of the primary quality". The main ideas are:
The objects of our recognition, especi ally in scientific research, are the
entities in the world. These entities and their attributes are realistic and
objective and by these terms the classical view stresses their indepen-
dence and invariance. An entity is both qualitatively and quantitatively
definite, independent of human observation, existent in itself, as well
as independent of other entities or environment, existent separably or
locally.
But as a fact even found in ancient times, we know that some physical
phenomena, like the colour of light, sound, smell and contact feelings,
etc., are related to human senses as well as entities themselves. To explain
these phenomena, a strategy had been developed in traditionally phi-
359
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.) , Realism and Anti -Reali sm in the
360 LUO JIACHANG AND HU XINHE
losophy, from the atomist and sceptical school of the ancient world
via the Middle Ages to Kepler, GaIiIeo and to its climax with Locke.
The strategy was to make a clear distinction between that in the world
which is absolute, objective, immutable, mathematical, and that which
is relative, subjective, fluctuating and sensed. The former used to be
called primary qualities which could be attributed to substance as its
intrinsic, essential attributes, while the latter were called secondary
qualities, declared to be effects of the action of primary qualities on
human sense organs . So, the secondary qualities are binary (or plural)
functions, neither reducible to the substantial entities, i.e., the primary
qualities completely, nor ascribable to the state of human senses in and
of themselves, namely
(I) y = f(x, x2 ' •• •)
where yare the secondary qualities, x, are entities and their essential
attributes, X2 are the states of human senses. Then, the y's are relative
manifestations, or projections of XI or X2' or relations between XI and
x2• But when the y's are the primary qualities.
(2) y = g(x,)
they have nothing to do, in their exi stence , with surrounding variables.
That is the very ideal of the view of an [objective] entity . Therefore it
was based on the objectivity and invariance of the primary quality, called
also "the view of primary qualities".
The classical view of reality is cons istent with concepts and experi-
ments of CP as well as our daily experiences. In fact, some basic
constants, like mass , extension, time interval and some basic character-
istics of matter, like particle and wave , in CP were often used as examples
of primary qualities, while the view of entity [or substantial thing] con-
stituted a part of the conceptual structure of CPo In Lakatos's terms, we
can analyse the ' hard core' of a physical research program into com-
ponents: a theoretical one and a metaphysical one. Referring to the
former, we mean those basic postulates and laws of a theory which
constitute the physical foundation of a research program. In CP, for
example, Newton 's three laws of motion and the law of universal
gravitation are such components which are applicable to all the mechan-
ical phenomena of objects. The latter refers to the basic conceptual
framework of the world-picture of phys icists, consisting of sediments
accumulated by physical and philosophical thinking from time to time ,
RELATIONAL R EALISM 361
and serving as criteria both in construction and evaluation of theories,

in a sense called 'physical ontology ' . Causality, continuity of motion,
concepts of space and time, the view of reality, are all cases of it and
its example in CP is just such a mechanism that all the various phenomena
in the world could be explained in terms of continuous motion of the
particles with definite, objective properties in absolute space and time,
according to exact causality. In principle, we have no reason to hold
on to either of these two sorts of components during a scientific revo-
lution, although the latter is more resistant to variation. In practice, we
can give the field form of matter in Maxwell's theory and the view of
space and time in Einstein's relativity as historical cases of such reforms
of metaphysical components.
2 . THE BASIC FEATURES OF QM
With the development of physics, the ideals of the view of [substan-

tial] entities seems to have gone to ruin . From the theory of relativity,
it follows that the geometric-kinematic properties of position, time
and velocity (as well as basic constants in CP like length, size , duration
and mass), previously regarded as invariant features, depend upon
reference frames. They have different values in different frames but all
of the values are real and equivalent. So, like the secondary qualities,
the primary qualities are also not intrinsic and invariant but relative
to, or manifested in, the relations with observation. Concerning the
primary qualities, physical phenomena are binary functions of both XI
and x2•
(3) Y = g(x\, x2)
where now x2 are the reference frames. There are still some philosophers,
especially in China, who want to insist on the view of entities proper
by giving other interpretations to the theory of relativity. One depends
on the distinction between the effects of kinematics (due to velocity of
motion) and dynamics (from acceleration), and argued that most effects
of relativity belong to the former which are not real but apparent. So,
measurement values = proper value + kinematical effects, in which proper
value is measured in the rest frame . But this interpretation seems to
contradict the basic idea of relativity that all measurement values are real
and all frames have equal right. In other words , it seems to come back
to Lorentz's idea of 'local times' in some extent.
362 i.uo JIACHANG AND HU XINHE
Another interpretation tries to maintain objectivity of the primary

qualities by appealing to invariant variables. With this aim, it defines
physical reality in 4-dimension space in which the interval is invariant
to Lorentz transformation. But it seems that to retain the definition of
reality as invariant, you have to pay a very high price: giving up the
reality of all physical properties in the 3-dimensional world, defining a
set of new primary qualities in the 4-dimensional world which might
be grasped only in ideas, not practices, and getting the 'block world'
which is abstract, static, not evolving, all determined including all causal
chains from the past to the future. Such a world picture will cause a
lot of difficulties.
Let us go back to modern physics. We find that much more serious
and fundamental clashes come from QM in which the whole frame-
work of concepts seems to contradict the conceptual view of entities.
In fact, QM, in contrast to CP, is an essentially new research program
with a hard core whose theoretical and metaphysical components are both
radically different from classical ones . So we shouldn't hold on at any
price to an invariant concept of reality, but rather adjust it to be coherent
with the whole conceptual structure . From the basic postulates of the QM
formalism, including postulates of state -vectors, operators and eigen-
values, probability of prediction, superposition, Schrodinger equation,
etc., to its inferences and interpretations such as the measurement
problem, uncertainty relations, wave-particle duality, we can find a
far-reaching change, not only in the content of physical knowledge, but
also in the fundamental concepts in terms of which such knowledge is
to be expressed. The three basic features of these concepts are (Bohm,
Chapter 8): (i) Discontinuity of the quantum process, i.e., replacement
of the notion of continuous trajectory by that of indivisible transitions;
(ii) Probability or indeterminacy of quantum laws, i.e., replacement of
complete determinism by that of causality as a statistical trend; (iii) Wave-
particle duality of quantum objects, manifested in different experimental
conditions, i.e., replacement of the assumption that the world can be
analysed correctly into distinct parts, each having some fixed 'intrinsic'
nature (for instance, wave or particle), by the idea that the nature of
quantum objects depends on the external experimental conditions. That
means a quantum phenomenon, even the world, is an indivisible whole
in which parts appear as abstractions or approximations, valid only in the
classical limit.
The quantum feature, namely the indivisibility of the quantum, implies
RELA nON AL REALIS M 363
the indivisibility of both the quantum of action and the quantum process.
From the former, we cannot draw a clear line of demarcation between
quantum object and measuring instrument; in principle, we cannot recog-
nize the state of an object as thing in itself, but rather the integral
phenomena as the results of interaction. From the latter, we can find a
way to a probabilistic feature, i.e., deterministic laws in CP which were
based on the principle of continuity are invalid in QM, and we can only
give probabilistic predictions concerning changes of state.
So, causality becomes statistical trend, physical properties become the
probabilistic spectra of eigen-values, and even the quantum objects,
described completely by wave functions, become probabilistic. Just as
was revealed in the wave-particle duality, quantum objects are unique
in the sense that they are essentially different from those in CPo It is
not completely identical with either a particle or a wave, nor both a
particle and a wave, but instead it is something potentially capable of
developing either one of these aspects of its behavior at the expense of
the other. Which of these opposing potentialities will actually be realized
in a given case depends as much on the nature of the system with which
the quantum object interacts as on the object itself. Then, wave property
and particle property, which were thought to be utterly different and
absolutely opposite forms of matter in the classical framework, become
relative manifestations of the same system under different experimental
conditions in the quantum formalism.
Obviously, the above features of QM clash not only with the basic
principles and world picture of CP, but also with the classical view of
reality, because they reveal the quantum integrality, the probability and
potential possib ility of quantum objects, while denying the separability,
locality, and definite, invariant properties, both qualitative and quanti-
tative, of quantum objects. The alternative ways are: holding on to the
view of entities, you must at least deny some of above features, as
Einstein did, or give them compatible interpretation, such as local hidden
variable theory; admitting and insisting on these features, it is neces-
sary to develop a view of quantum reality which is consistent with
them.
3. THREE VIEWS OF THE QUANTUM REALITY
Due to the aim of this paper, we will discuss only three views of reality
in QM whose common features may be found after our sketch of them.
3.1. Complementary Realism or Phenomenal Realism
This is Bohr's view of reality. At first, the significance of the principle

of complementarity is mainly in epistemology, in revealing the com-
plementary limitation of the classical concepts (which are suitable to
continuous processes) when applied to atomic phenomenon . Because "the
quantum postulate implies that any observation of atomic phenomena will
involve an interaction with the agency of observation not to be neglected"
(Bohr, 1934, p. 54) and not to be analyzed, so two basic concepts of
CP, the definition of state, and of observation or measurement, are incom-
patible. But as long as classical concepts are used , both of them are
necessary for the complete description of quantum phenomena . Therefore ,
clear definition and precise observation (also space-time description
and causal description) are mutually exclusive and complementary. From
this reply to EPR, in which he claimed the ambiguity involved in not dis-
tinguishing the system as an observed phenomenal object from the system
as an independent reality led to the invalidity of Einstein's argument,
Bohr began to pay great attention to ontological meanings of comple-
mentarily (see Folse, pp. 154-161). He emphasized "it should be clear
that the whole situation in atomic physics deprives of all meaning such
inherent attributes as idealizations of CP [i.e., 'particle ' and 'wave'
pictures] would ascribe to the object" (1937, pp. 292-293). He used
the word 'phenomena' to refer now not to independent reality but to
the whole observational interaction, "to reserve the word ' phenomena'
for the comprehension of the effects observed under given experimental
conditions" (Bohr, 1939, p. 24). The "rational account of a phenom-
enon . .. must be taken to involve a complete description of the
experimental arrangement as well as observed results" (Bohr, 1947, pp.
59-60). From this notion, we can get different phenomena, namely dif-
ferent observational interactions under different experimental conditions,
and also get the complementarity of different phenomena as a logical
consequence of the complementarity of experimental conditions.
We can thus draw the outline of Bohr 's view of quantum reality:
quantum objects cannot be described by classical concepts except through
their interactions with observing instruments, the attributes of quantum
object are real only in a context involving the experimental arrangements.
These interactions and real quantum attributes are what we called 'phe-
nomena '. It is the key word in Bohr's view of reality, at least since
1939. The characteristics of quantum phenomena are:
RELA nON AL REALISM 365
a. They are indivisible.

b. They are real.
c. They are complementary.
So all the information about the same object x. can be obtained by a
set of phenomena yri)
(4) {y(i)} = {F(x l , x20 )) }
where xz<i) refer to experimental arrangements. So the conclusion is
that Bohr 's phenomenal realism is characterized by the relations between
quantum objects and experimental conditions, by stressing "These con-
ditions constitute an inherent element of any phenomenon to which the
term 'physical reality' can be properly attached" (Bohr, 1935, p. 700) .
3.2. Potential Realism and Latent Realism

The first is what we call Heisenberg's view of reality. As also a main
figure of the Copenhagen School, Heisenberg surely had many ideas
which were identical or similar to Bohr's, but also he had his own unique
ideas . As an eminent physicist, Heisenberg clearly recognized "that the
results of modem physics do touch such fundamental concepts as reality,
space and time, the confrontation may lead to entirely new developments
which cannot yet be foreseen." "It is in quantum theory that the most
fundamental changes with respect to the concept of reality have taken
place" (Heisenberg, p. 28). In the analysing the relations between
'Language and Reality in Modem Physics ' (Ibid., chap. x), Heisenberg
in fact distinguished two sorts of language (mathematical and ordinary)
and two levels of object (reality, and facts as the results of measurement);
his main ideas may be illustrated thus:
Language - - - - - - - - - - - - - - - - reality
mathematical - - - - - - - - - - - - - - - reality
1 qua~-i---actual
ordin~- - - - - - - ~ fact
In CP, each language is applicable to both level of reality, but in QM,
"we cannot speak about the atom in ordinary language. The analysis
can now be carried further in two entirely different ways. We can either
ask which language concerning the atoms has actually developed among
the physicists in the thirty years that have elapsed since the formula-
tion of QM. Or we can describe the attempts for defining a precise

scientific language that corresponds to the mathematical scheme" (Ibid .,
p. 179). The second way seems to choose resolution of the contradic-
tion between ordinary language and quantum objects by replacing the
former with a different precise language which follows definite logical
patterns in complete conformity with the mathematical scheme of
quantum theory. Some progress has been ach ieved in this direction
(quantum logic), but a regression is always involved because "we have
to speak at least about our eventual modification of logic in the natural
language" (Ibid., p. 181). The ontology that underlies this view is 'co-
existent states' or 'coexistent potentialities'. The first way is the same
as what Bohr had done in introducing complementarity, namely to use
an ambiguous language to talk about the atomic world. Thus atoms and
their attributes, when they appear in such a scheme, only acquire the same
statistical significance as 'temperature of the atoms '. That is a modified
concept of reality. "One might perhaps call it an objective tendency or
possibility, a 'potentia' in the sense of Aristotelian philosophy. In fact,
I believe that the language actually used by physicists when they speak
about atom events produces in their minds similar notions as the concept
'potentia' . So the physicists have gradually become accustomed to
con sidering the electronic orbits, etc ., not as reality but rather as a
kind of 'potentia'" (Ibid. , p. 181). Thus, according to Heisenberg,
quantum reality, when it appears inevitably in the framework of clas-
sical concepts and ordinary language, will be described as statistical
expectations, objective tendency, or 'potentia'. Only after interacting with
measuring instruments, does it become as certain as in classical state,
namely a well defined fact. Reality tends from ' potential' to 'actual'
during a measurement.
Similarly, Margenau held a view of reality which we call 'latent'
realism. He distinguished properties in QM to be attributed to objects
conditionally (latent attributes) with those in CP attributed uncondi-
tionally (possessed attributes) (H. R. Holcomb III). "In QM the problem
of reality seems to center in the meaning of states. Here it became clear
for the first time , and not without a shock to the physicist, that even
physical state, previously regarded as a nice and tidy collection of pos -
sessed observables geared together and running like wheels in a clock,
is in fact a composite of latent observables loosely coupled to nature
by rules of correspondence of a statistical sort" (Margenau, p. 452). Only
a measurement of an observable projects it from latent to actual. So
RELATIONAL REALISM 367
both Heisenberg and Margenau described quantum reality as poten-

tiality which may be actualized in measurements.
3.3. Dispositional Realism

Similar to our 3.2, this describes quantum properties as dispositions of
objects which manifest themselves as observable actions under certain
conditions because of common groundings (R. Harre, 1985, 1991).
Together with conditions, or more precisely in our case, with apparatus,
as one of us argued (see Xinhe Hu , 1991), a dispositional account of
physical properties can give us a satisfactory resolution to some problems
associated with it. It can be sketched by the formula:
(5) Propensity + Apparatus ~ Physical Properties
among which propensity is a probabilistic & potential micro-world.
Taking the left side as a whole and giving the apparatus the meaning
of ' open to development' , we can get a self-sufficient theory to answer
such problems as regre ssions in both directions, namely ungrounded basic
dispositions and an unobservable and unmaniputable cosmos, etc. But
it seems we still can say nothing about disposition in itself. What we
can say is just that it is discovered and conformed statistically on the
experimental setup. We cannot talk about its physical content and give
it an objective description except as it is bonded to experimental arrange-
ment. But taking disposition in that sense, we know that even thinkers
of the 17th or 18th century like Descartes, Hobbes and Boyle, viewed
powers as not being inherent in matter; that means 'powers' are not
ascribable to bodies in and of themselves. Rather, powers are mani-
fested only when bodies are in relation to specifiable extrinsic circum-
stances. So it projects the common features of all these interpretations
as follows:
(l) Distinguish two levels of reality: independent and phenomenal.
The former can only be described by mathematical formalism as a
catalogue of potential interactions or as a propensity to produce various
possible results on measuring instruments, while the latter, as actual
results of interaction, as realized possibilities, can give us all the infor-
mation we can get by different instruments.
(2) Emphasize the role of instruments in the concept of quantum
reality.
(3) Project the relational character of quantum reality. Described in
368 LUO JlACHANG AND HU XINHE
quantum phenomena, the attributes of quantum objects are manifested

as results of interactions, as conditional properties, and as relations
between objects and instruments. In a word, it seems relations take a
much important place in our description of Nature, one which was once
occupied by properties. What we are describing is not the isolated prop-
erties of objects but their extrinsic relations with circumstances. This
is very thesis of this paper : relational realism .
4. RELATIONAL REALISM
Relational realism emphasizes the relational characters of reality. It

regards reality as potential possibility, provided and emergent in a set
of rel ational structures, and aims to explore these relations, and the
relations between them.
The emphasis on the relational characters and relational structures
of reality is not a fresh point of view in the history of both philosophy
and science. As early as in his Parmenides, Plato had discussed the
relational structure of existence and constructed entities as combina-
tions of relations, although those relations were abstract, prior and about
ideas (see Sternfeld, R., Chapter IV, V, VI).
Following Plato , Leibniz also developed an ontology of relations and
the ontological structure of relational facts (see Hector-Neri Castaneda).
Leibniz thought there are universal relations between the elements of
the universe. Every individual, whether person or thing, in some way
reflects everything that is happening in the universe, each from its own
point of view.
We can cite many more historical examples (see Jammer [1966]
'Concluding Remarks'), but the most impressive progress of relational
realism comes from modern physics. Besides the three interpretations
of QM described above , we can find the most complete and convincing
statements from the two well known books by Max Jammer. In the first
one, he indicated those statements of relational realism (though not with
such a title) "found their full corroboration in modern physics. At first ,
the theory of relativity revealed that the geometric-kinematic proper-
ties of position, time, and velocity (as well as length, size, duration,
and mass), previously regarded as objective features, depend upon the
frame of reference. Subsequently, QM showed in addition that these prop-
erties are relative also to the means of observation" (Ibid., p. 404). And
in the second, he analysed " the relational conception of quantum states"

(Jammer, 1974, pp. 197-211). Its main points were :
(I) Bohr "found the solution of the emerging difficulties in the
adoption of a relational and holistic conception of the state of a physical
system" (Ibid., p. 199). "In other word s, the descriptions of the state of
a system, rather than being restricted to the particle (or system of
particles) under observation, expresses a relation between the particle
(or particles) and all the measuring devices involved" (Ibid., pp.
197-198).
(2) " Because every measurement involves both the entity to be
observed and the measuring experimental setup, no attributes of physical
reality can be ascribed to the former alone. In particular, a particle,
even after having ceased to interact with another particle, can by no
means be considered as an independent seat of ' physically real attributes '
and the EPR thought-experiment loses its paradoxical character" (Ibid.,
p. 198).
(3) "One may also regard the connection between Bohr 's relational
conception of state and Einstein 's theory of relativity from the general
point of view that in the historical development of physics attributes were
gradually replaced by relations" (Ibid., p. 201). The finite size of the
velocity of light removed the discrepancies originating in the theory of
relativity and led to the choice of different frames of reference, and the
relational conception of quantum states reconciles the contradictions in
QM . In this sense , Fock said : "A QM description of an object by means
of a wave function corresponds to the relativity requirement with respect
to the means of observation. Thi s extends the concept of relat ivity with
respect to the reference system familiar in CP" (Ibid., p. 202).
Another more recent example is Paul Teller's relational holism. By
this term, he means "the claim that objects which in at least some cir-
cumstances we can identify as separate individuals have inherent
relations, that is, relations which do not supervene on the non-relational
properties of the distinct individuals" (Teller, p. 73). Applying this idea
to QM, he thought that inherent relations are important to interpreting
QM, because all these cases involving superpositions manifest the
inherent relations. According to this doctrine , we can see the EPR
argument as revealing inherent relations of two particle systems and
see the key to the measurement problem as the relations which hold
between measuring devices and measured objects.
370 LUG JIACHANG AND HU XINHE
We have outlined relational realism and given it a sufficient argument

in the micro-recognition field . Now, let us elaborate our relational realism
further as following:
(1) Reality is relational, it exists in a set of relations which are essen-
tially inseparable from it. The general patterns of existence may be
expressed as
(6) y = I(x. , X2' ••• x.)
where y are physical phenomena, as recognized properties, x. are physical
objects, or realities, and its essential attributes (in some cases), X2, • • •
x, are different relational parameters, but in most cases, only one para -
meter is the principal one which is determined by our perspectives. So,
its simplified form is
(7) y = I(x., x2)
As we have known from QM, when the coupling of x. and X 2 is not

neglectable and reducible, we must admit that relations are realities, phe-
nomena are realities, the quantum objects can be represented by a
conceptual reconstruction of all concerned complementary quantum phe-
nomena, by a set of inverse functions like
(8) x, ---7 {g(y, x2) }
According to this view of reality, quantum properties are not intrinsic ,
but relational. As the results of interaction, their reality depends on the
measuring set up, and even relates to the type of measuring setup of
the next time. Similarly, quantum states, as Jammer analyzed, are rela-
tional. They should be interpreted as the catalogue of potential
interactions with all concerned measuring setups, as the probabilistic
spectra in these relations, as the relativity requirement with respect to the
means of observation; the relationalization, relativization and projec-
tivization of reality are the very features of quantum reality . In fact, in
the macro-field, a physical object is also in relations with its surrounding,
as a part of a much larger system . Only in such a larger system including
instrument and observer, could the object get its phenomenological
manifestation and become an object of measurement and theoretical
investigation. In a word, reality always gets its peculiar ontological and
epistemological positions by its relations with its surroundings.
(2) Relations are real, the reality of relations is precisely in its mate-
riality and its objectivity. We have emphasized the relational character
of reality, but no matter whether those relations are between moving body
and reference system in the theory of relativity, or those between micro-
object and measuring instrument in QM, they are all realistic relations
between material systems. Although reference systems and measuring
instruments are chosen, even made by human beings, as soon as they
are chosen, or made-up, the results of experiment are definite or prob-
abilistically predictable. Similarly, the relativity we have emphasized
is the relativity of object to different material systems like frames of
reference, measuring instruments, etc . But in each special case, namely
each special material relation, the information provided to all scientists
is equivalent, objective; there is nothing related to their individual sub-
jective characteristics such as personality, will etc.
(3) Relation is prior to relata in many cases . "One can give a strong
sense to the idea that a relation does not 'reduce' to non-relational prop-
erties by saying that a relation does not supervene upon the non-relational
properties of its relata" (P. Teller, p. 71). Systematic properties of sub-
systems in system theory, quantum properties in QM, and some
high-energy particles produced by cyclotrons, are all examples of such
relata. So contrary to the trend of going deeper and deeper to find the
most 'elementary' particles and their intrinsic fundamental properties, we
find at least some such properties, and even particles, are related or come
from their relations with other material systems. A famous sutra of
Chinese Buddhism says: Things have no inherent attributes, which
emerge from their relations. We think at least we can say some of them
emerge from their relations with others.
(4) Relations and attributes are mutually transformable in different
perspectives. In the general pattern of existence, when we define the
different sorts of relational parameters x2, we face different phenom-
enal worlds, and for each sort of x2, when we give it different values,
we get different phenomena. Then we find not only that attributes become
relations, but also relations become attributes when X2 is defined. So,
when x 2 is a measuring setup in QM, we face the objective world of
quantum properties which may be called the zero-th qualities. When X2
is the frame of reference, we face the physical world of the primary qual-
ities . In this way, we can also confine x2 to the state of human senses
(in the sense of normal people), and then face the common-sense world
of the secondary qualities, even confine X2 to the common aesthetic (or
ethical) experience and then face the aesthetic (or ethical) world of the
third qualities, although for the latter we see the narrower and the weaker
nature of their objectivity, and the stronger of their relativity. In so doing,

the view of relational realism could embody and reconcile the views
of reality at different levels and fields, and manifest a theory of multi-
worlds, as a much more fruitful and broader view of reality .
5. RELATIONAL LOGIC
Relational realism came from the reform of the view of physical reality .
In order to give it a much rigorous and general expression, we must enter
into the logico-linguistic field in order to develop the relational logic
which is suitable to relational realism .
Different from the subject-predicate (s-p) logic which was set up by
Aristotle on the base of substance-attribute realism, relational logic
stresses the feature of centering its attention on the relations between
objects. It restricts or replaces the classical judgement formula's is p'
by a formula 'aRb ', adds a relational operator in the standard system
of propositions to serve as a necessary condition for getting the truth
value of a proposition. So corresponding to the existence pattern of
relational realism, a sentence is meaningful in the syntax structure of rela-
tional logic only in form of
y = f(x, r) or Pr(p)
where r is relational operator. That will lead to a series of important
conclusions.
Obviously, s-p logic is suitable to substantial realism which was
centered upon entities, but in such a logical frame of monadic predi-
cates combined by the principle of bivalence, substantial realism
will go to its opposite - antirealism in many cases. When some p
(red) couldn 't be attributed to s (the flower) itself (attributes couldn't
be attributed), or some contradictions arise like s is both p and not p
(the flower is one-meter high and not one-meter high), we have to deprive
s of these p, reduce them to human subjective perceptions and go
to the anti-realistic conclusions of p. In this way, we had deprived
the physical realities of the secondary qualities, we should also deprive
them of the primary qualities, and quantum properties because most of
them are also relational and could not be attributed to realities them-
selves; they are also secondary qualities as many philosophers have
indicated. In the end , we would be metaphysical realists with a bare,
RELA nON AL REALISM 373
abstract concept of reality but anti-realists concerning most of the

physical properties.
Contrary to the traditional logic, by maintaining the description of
physical reality by binary (multivariate) predicates, relational logic is
compatible with relational realism through their common symbolic
expression of binary (multivariate) functions . So, it provides a possibility
to overcome metaphysical realism and avoid anti-realism. For example,
'x is one-meter length' in the theory of relativity and ' x is particle' in
QM are meaningless according to classical logic, but meaningful in
relational logic because relational operator r was introduced as an essen-
tial element in its syntax structure. Namely, ' length ' , 'particle property'
are not monadic predicate but relational predicate, are functions of an
ordered couple (x, r), each of them is separately a projection of a 4-dimen -
sion invariable to a reference system or of a quantum object to a
measuring setup . Corresponding to such a change of ontological category ,
the syntax structure of the concerned judgments is correct when intro-
ducing r, i.e., ' in the reference system r ' or 'on the setup r ' . That is
the meaning of the symbolic expression Pr(p), namely "the proposition
P is tenable in the relation r". Strictly speaking, the generating of a propo -
sition and the getting of its definite truth value depend on the indication
of the value of r.
As we stressed the relational character of relational realism contrasts
to classical reali sm, the essential feature of relational logic is to intro -
duce a relational operator in the standard system of propositions.
The relational operator is similar to the tense operator in tense logic ,
but it has much broad implications. It can express the relational items
like 'to whom', ' when' , 'where', ' what situation' and 'how ', etc. , and
it includes the tense, spatial , contextual factors , etc.
By introducing r, combined with the principle of bivalence, we can
accept that contradictory statements are true at the same time (both's
is p' and 's is not p' are true for different r, e.g . that ' the flower is red'
is true for a normal retina and that ' the flower is not red ' is true for an
abnormal retina) and thus avoid the tendency of antirealism (p is sub-
jective), the tendency of dogmatism (only's is p ' is true and reflects
the real features of thing), and the tendency of so-called dialectic logic
(s is both p and not p). From the formula of phenomenal structure, we
can see, neglecting the parameter r (or X 2 •• .) is the cause of difficul-
ties in the above three tendencies or views.
374 t.uo JIACHANG AND HU XINHE
By introducing r, a variation of the view of truth is necessary contrary

to such a abstract view of truth as the classical correspondence theory,
simply regarding a proposition as the seat of the truth; generally, we
should regard truth as a property of a function of the triple of a speaker,
conditions of speaking and a sentence. In this way, the concept of truth
could be used in natural language.
By introducing r, we can go to a much broader view of reality. On
the one hand, because as soon as the relations are defined, the sentence
has truth value, and truth value is determined by the objective charac-
ters of the system f(x., Xl' • •• ), irrelevant to personal will and
characteristic, so in the first place, relational realism is a sort of realism
(p can be attributed to s). But on the other hand, instead of a metaphysical
external realism, it is an internal realism. This is so because without
the participation of Xl (e.g., retina) on the ontological sense, according
to relational realism, not only the p (red) could not emerge , but also it
could be said there is no such a P (e.g., colour). Colour as a property
of the system is internal in the system f(x., Xl ' • • •) and will change
with the Xl ••• Because relational logic provides us a new method of
regarding all those properties affected by evaluating the subject (such
as ' cool' , 'warm ', 'red', 'yellow', 'good', 'evil', 'beautiful', ' ugly' ) not
as monadic predicates but as relational predicates through introducing
new variables relating aspects which are concerns of the judging subject,
it provides us with a possibility of describing them and the relations
between them objectively too. For example, ' x is good ' should be
expressed completely as 'for r, x is good' and thereby receive its truth
value . Such a general pattern of sentence 'for r, s is p' implies that a
sentence of evaluation can become a conditional sentence of a state-
ment, namely a descriptions of facts (of cause in reverse, a so-called
statement sentence is always a description of fact in relation r, and it
contains the evaluation). In this way, ' the flower is beautiful', ' the flower
is red' , 'the flower is one-meter high' are all objective statements when
confined to the relational operator, with the degree of objectivity of
each proposition determined by the r concerned. It is helpful for our
understanding the objective significances of aesthetic and ethical judge-
ments and provides us with a logical foundation for constructing a much
broader, richer and more colourful view of reality.
6. ON SOME PROBLEMS
6.1. On the Paradox of the Grue Emerald

The puzzle of the grue emerald is that the same set of observations, if
made prior to t, supports two incompatible generalizations (' All emeralds
are green ' and 'All emeralds are grue') and leads to two incompatible
predictions ('The emerald is green after t' and 'The emerald is blue
after t') (Goodman, N.). In the view of relational logic, the puzzle comes
from the fact that two sorts of language are combined in describing the
same world, that the r which constricts the range in which each language
is suitable is neglected. Let us provide a formalized analysis. As we have
said, colour is a relational predicate generally (especially in the scien-
tific world) and has an expression f(x, r). In the present situation, r is
t, then c = f(x, t), For the commonsense world, colour is a constant
and becomes a monadic predicate, c = f(x , t) = f(x) = G(green). But
for a special possible world expressed by 'Grue' language,
C = f(x, t) = {G when t < 2000 A.D.

B when t > 2000 A.D.
Therefore, two sorts of language, corresponding to two possible worlds,

are confined separately by two values of t, and the combined usage of
them will cause contradictions . It might be much popularly accessible
and vivid if we replace t with v (velocity), for according to the Doppler
effect, colour is a function of velocity, and we will see a differently
colourful world-picture in different moving systems, just as the length
of a rod is different in different frames and cannot be combined.
Up to now, most resolutions to this paradox held that only one world
(the commonsense world) exi sts . That led inevitable to the choice of
different sort of language in describing the world and only gave us such
pragmatic criteria as ' habit' , 'entrenchment' in the choice of predicates.
But in the view of relational logic, each language corresponds to a
possible world to which we should not refuse its reality, in principle.
Then we can give a much formal rule to allot the predicates and the
validity of inductive references.
6.2. Relational Realism v. Metaphysical Realism
Obviously, relational realism is a challenge to the metaphysical realism

represented by Kripke. In the view of Kripke, the bearer of a name,
i.e., the individual and its essential properties, are identical in all possible
worlds. Identity will be definite forever as soon as it is defined. The
essence of the individual is sole [unique], necessary and absolute. This
metaphysical view of identity and essence is the cornerstone of Kripke's
metaphysical realism (see Kripke).
But in the view of relational realism, things (as relata) and their
essence are defined by special relations. They would be changeable in
certain conditions, with the variation of these relations. The objective
connections and relations between objects and surroundings include the
relation between human practice and object, to form a context of recog -
nition. So the definition of things often reflects human practical action
and its end, the 'essential properties ' described in the ' definition' could
not be absolute properties separated completely from certain relations
and conditions, and neither could they be pure ontological, metaphys-
ical necessary properties which are completely irrelevant to a human
being's existence and mode of understanding. On the contrary, because
of the manifold patterns of a thing connected with its surroundings, and
the manifold ways of a human being's practice and recogn ition, it should
be possible to reveal the different ' essential properties' of the thing in
different 'definitions'.
In contrast to essentialism, relational realism shows the complexity
of classifying and naming of things. There is a relativity not only in
the differentiating of thing, varieties, and essential properties, but also,
to some extent, in the problem whether they exist or not, related to certain
language communities, criteria and modes of understanding. We can give
definite answers to questions such as 'what is a object?', 'What essen-
tial properties does it have? ' only after all these conditions are defined.
In our linguistic practice , the relativity of objects and essential prop-
erties is shown by the relativity of references , which means that all objects
referred in the variety of language communities are concrete and describ-
able. It is this describability that provides those predicates, which have
the same extension and different intension in the given world (often a
certain scientific world), some conditions by which to manifest their
different intensions in other possible worlds.
Again, different from Kripke's admitting only one world, relational
RELA nON AL REALIS M 377
realism holds a position of reali stic comprehension of (relationally

relevant) possible worlds, i.e., admits of the existence of many possible
worlds with very different properties and features. Such a view of multi-
worlds not only finds its supports from the advances in physics and
cosmology, but also may rationally satisfy modern people's request for
a much richer and more colourful view of reality in their practical and
spiritual lives.
6.3. Oriental Thought as Background of Relational Realism
Relational realism has been based on modern science, mathematics and

logic, but it has a fundamental connection with traditional Oriental
thought.
Contrary to Western substance thinking and logic , in the Orient, espe-
cially in China, a mode of thinking centered on the relativity of things
originated even in ancient times. At the origin of Chinese philosophy, the
Yi-k ing (Book of Changes) had distinctly manifested such a feature.
According to it, every phenomenon could be interpreted by the combi-
nations of a pair of elements, Yin and Yang which are negativity and
positivity respectively (symbolized as - and + respectively). All phe-
nomena possess negative and positive aspects and all judgements have
negative and positive values. But it should be stressed that both of them
are neither substances nor inherent essences of things, but jointly ways
for expressing some relations between things. For example, when such
a pair of concepts applies to relation s between male and female, a man
is Yang to his wife (in husband-wife situation) and he is Yin to his mother
(in mother-son situation). That means Yin and Yang are not monadic
but binary or relational predicates. They are constrained by the rela-
tional operators, such a relativity of Yin and Yang could be expressed
as following ,
[y] Yin [a, vl [z] Yang [a, z]
namely 'thing a is Yin relative to thing Y and Yang relative to thing
Z '. It is easy to see that Y and Z are two values of relational opera-
tors. Of course , such embryos of ancient relational logic were clarified
through the view of modern logic, while in the history of more
than two thousand years, and even for many current Chinese scholars,
philosophers were unaware of the roles of relational operators, and
the relativity of things were expressed as 'being both Yin and Yang'.
Later, relational thinking was drowned entirely by so-called dialectical

thinking.
It is an easy passage from the relativity of Yin and Yang to a rela-
tivity of 'existence' and ' nothing ' , and everything. That means we cannot
talk about 'existence' or 'nothing' in isolation. They are mutually pro-
moting and restricting, complementary and interexclusive. Reality is
always relative to a certain relation and condition. Such a relative view
of reality in Chinese philosophy was expressed in Chinese Buddhism
as 'impermanence ', 'endless genesis', namely everything is transient,
there is neither permanent, invariant self-nature, nor an independent,
isolated thing-in-itself. Everything depends on other things, generates,
and emerges in certain conditions and relations. Quite the contrary,
from Parmenides to Bradley, the Western absolute view of reality pursued
the ultimate reality which is behind common sense, behind variation,
behind phenomena and particulars. That is the reality before God's-eye.
The relational view of reality underlines both the objectivity of reality
and the relational limitation of reality, namely the multiplicity of reality.
It holds onto grasping reality by different abstract methods, and also holds
that phenomena are realities, processes are realities, reality is in the unity
of human practice and its objects.
Inst itute of Philosophy,

Chinese Academy of Social Sciences ,
Beijing, China.
BIBLIOGRAPHY
Bohm, D. (1951). Quantum Theory, Prentice-Hall.

Bohr, N. (1934). Atomic Theory and the Description of Nature, Cambridge University
Press.
Bohr, N. (1935). 'Can the Quantum Mechanical Description of Reality be Considered
Complete?' Physical Review 48.
Bohr, N. (1937). 'Causality and Complementarity', Phil. Sci. 4, pp. 292-293.
Bohr, N. (1947) . 'Newton 's Principle and Modern Atomic Mechanics' , Newton
Tercentenary Celebrations, Cambridge University Press, pp. 59-60.
Castaneda, H. N. (1982). 'Leibniz and Plato 's Phaedo Theory of Relations and Predictions ' ,
in M. Hooker (ed.), Leibniz: Critical and Interpretive Essays, Manchester University
Press.
Folse , H. J. (1988). The Philosophy of Niels Bohr, North-Holland.
Goodman, N. (1955). Fact, Fiction and Forecast, Harvard University Press.
Harre, R. (1986). Varieties of Realism, Basic Blackwell.
Harre, R. (1991). Lectures in the Sino-British Summer School of Philosophy .

Heisenberg, W. (1970). Physics and Philosophy; The Revolution in Modern Science, Harper
& Row .
Holcomb III, H. R. (1986). 'Latency Versus Complementarily: Margenau and Bohr on
QM' , Brit. J. Phil. Sci. 38, pp. 193-120 .
Hu, X. (1988, 1991). In N. Bunnin (ed.), Best Examination Essays, Sino-British Summer
School of Philosophy .
Jammer, M. (1966). The Conceptual Development of QM, McGraw-HilI.
Jammer, M. (1974). The Philosophy ofQM, John Wiley .
Kripke, S. (1980). Naming and Necessity, Basil Blackwell.
Margenau, H. (1950). The Nature of Physical Reality, McGraw-HilI.
Sternfeld , R. and Zyskind, H. (1987). Meaning , Relat ion and Existence in Plato ' s
Parmenides: The Logic of Relational Realism, Peter Lang .
Teller, P. (1986). ' Relational Holism and QM ', Brit. J . Phil. Sci. 37, pp. 71-81.
MAO SHIYING
A REALISTIC INTERPRETATION
OF QUANTUM MECHANICS
I. INTRODUCTION
The contemporary debate between scientific realism and anti-realism

prominently presents the question whether or not a realistic interpreta-
tion of quantum mechanics (q.m., afterwards) can be given adequ ately
when it is reflected in concrete scientific theories. Scientific realism
generally demands that a scientific theory aims to be a true or approxi-
mately true description of entities, properties, structures, processes, etc.
that exist in an objective world independently of our observation and con-
sciousness. All classical physical theorie s satisfy the demand. However,
the Copenhagen interpretation, as the orthodox interpretation of q. m.,
is usually regarded as an instrumentalist interpretation with an episte-
mological meaning alone. It avoids answering the ontological questions
such as whether the quantum theory really describes the microworld. Just
as Bohr wrote : "The entire formalism is to be considered as a tool for
deriving predictions, of definite or statistical character, as regards infor-
mation obtainable under experimental condition described in classical
term s" [I]. It is obvious that this interpretation is in accordance with
the anti-realist doctrine of Van Fraassen, which holds that "science aims
to give us theories which are empirically adequate, and acceptance of
a theory involves a belief only that it is empirically adequate" [2]. The
primary reason for the Copenhagen school to take the instrumentalist
position instead of the realistic position, I think, lies in the fact that in
a quantum measurement the measuring device or the observer can no
longer be separable from the observed object, as it could in classical
measurement; the measuring device and the object system constitute
an inseparable unified whole together, the entire experimental situation
is already of an inseparable wholeness or nonseparability. It is according
to the concept of wholeness that Bohr refuted the EPR argument against
the completeness of q. m. [3]. In fact, the experimental results of Aspect
et at. about the Bell-type inequalities in recent years have undoubtedly
verified that the microreality possesses nonseparability, i.e. a microsystem
cannot be divided into two local parts which are independent of each
381
R. S. Cohen , R. Hilpin en and Qiu Renzong (eds.), Realism and Anti-Realism in the
382 MAO SHIYING
other in space, the principle of separability, that is based on classical

measurements and used as a basic principle of the realistic interpreta-
tion of classical physics, no longer holds for the quantum theory.
However, the presence of nonseparability does not certainly shows that
scientific realism cannot hold any more . The basic demands of scien-
tific realism such as objectivity, causality, and completeness, should be
still upheld, or perfected further on the basis of new scientific dis-
coveries, but they should not be abandoned simply, otherwise the
scientific enterprise has no way to resist the aggression of subjectivism.
In fact, it is due to one-sided depending on the nonseparability, that the
Copenhagen interpretation cannot give a realistic interpretation to the
phenomenon of the 'wavepacket reduction' of measurements, and only
considers it as a sudden change of knowledge of the observer. Thus
this viewpoint cannot guarantee objectivity independent of any observers,
causality and completeness of the theoretical statements, and leaves a
backdoor for the subjectivist interpretations of E. Wigner et ai. Actually,
scientific realism is not certain to be in conflict with nonseparability.
As I shall show in the following, it is not impossible to construct a
realistic interpretation of q. m. which is compatible with nonsepara-
bility. The key to the settlement of the question is to construct a new type
of correspondence theory of truth on the basis of quantum measure-
ment. In what follows I shall first give an analysis of the quantum
measurement.
2. THE QUANTUM MEASUREMENT IS THE UNITY OF

NON SEPARABILITY AND MACROINTERV ALITY
Because of the presence of nonseparability of the entire experimental

situation, the measuring device (including the observer as a recorder) and
the object system constitute an inseparable composite system together,
and in the case of a measurement the disturbance which is imposed on
the object system by the measuring device, can no longer be neglected
or compensated as it could in a classical measurement in principle, and
various measurement results, such as prepared quantum states, the phe-
nomena of wavelike or particlelike, and etc. , are formed in the measuring
interaction between object and device. Therefore, the measuring device
already has the ontological function of acting as a macroscopic physical
system or macroscopic surroundings to participate in the objective world,
and the measurement process should be in principle seen to form a
A REALISTIC INTERPRETATION OF QUANTUM MECHANICS 383
physical process, and the measuring interaction should be considered

as an objective physical interaction. E. Wigner holds that in the quantum
measurement the consciousness of the observer plays a central role , but
this point of view cannot explain how the universe existed and evolved
before man was born.
Besides nonseparability, I think, the quantum measurement still pos-
sesses another universal character: macrointervality, which is what the
macroscopic spatio-ternporal conditions of the measuring device con-
stituting a kind of relatively independent region show in the measurement.
The conclusion ·is drawn by contrasting the theoretical description with
the empirical facts about a microobject. The nonclassical phenomena
appear in quantum measurements, in which the theoretical description
cannot in principle correspond to the empirical fact directly and wholly.
Firstly, the phenomenon of the wavepacket reduction shows that during
a measurement the superposition of states of an object transforms into
one term from the superposition, and the other eigenstates in the theo -
retical description do not appear. Secondly, the theoretical description
of the wave-particle dualism of an object cannot correspond to the phe-
nomenon of the mutual exclusion of wavelike property and particlelike
property directly and simultaneously. The nonclassical phenomena men-
tioned above actually show that every observable phenomenon of an
object is limited within some relatively independent space-time interval.
Due to the space-time region of an observable phenomenon being deter-
mined by a macroscopic measuring device, the interval is a macroscopic
one . The phenomenon of wavepacket reduction is actually thought to
occur when the measuring device possesses a macrointerval which has
the same spatial size as some term from the superposition of states,
thus resulting in the so-called wavepacket reduction. A deta iled analysis
about this will be given later. The phenomenon of simultaneously mutual
exclusion of wavelike property and particlelike property in measure-
ment shows the two properties have a different macrointerval respectively,
and want us to use two kinds of different devices to measure respectively.
The relation between the macro interval in an experiment and the spatio-
temporal region of a microobject in the theoretical description, is one
between part and whole. Because the objects of classical measurements
are macroscopic physical systems, the macrointervality of measuring
devices could not show up in classical measurements. It can only be
manifest in the measuring interaction between device and microsystem.
The macrointervality of measuring devices determines a device in a
384 MAO SHIYING
quantum measurement to possess its nonclassical peculiar epistemo-

logical function of selecting and manifesting some physical property of
a microsystem which is latent relative to our sense organs, where the
function of selection can be classified as two types:
(1) Active selection: determination of wavelike or particlelike prop-
erties of a microsystem within some macrointerval. In the measurement
we can estimate the measurement result in advance. Due to the com-
plementarity principle, the two selections are mutually exclusive and
complementary.
(2) Stochastic selection: determination of some eigenvalue of an
observable of a microsystem. Since the eigenstates are generally count-
less, and the macrointerval of the device only accommodates one
eigenstate in a measurement, we cannot estimate the result of each
measurement in advance, thus the selection is wholly passive and
random.
A selecting act of a measuring device for some physical property is
also a manifesting act in the meantime a measuring device as a carrier
of physical information, can fetch and manifest some piece of latent
physical information of a microsystem in an active or a stochastic
selecting way within a macrointerval. Moreover, there exists the poten -
tializing (or concealing) act in the meantime i.e , of the manifesting
act, i.e., the other pieces of physical information of the microsystem
simultaneously lose the possibility of appearing within the interval.
For instance, the wavelike property and the particlelike property of a
microsystem are manifested respectively through measuring devices. And
the mutation-like measurement actually shows that in a measurement one
and only one eigenstate is able to be manifested by a device, and all
the other eigenstates are potentialized.
The macrointervality of the measuring device actually constitutes
the mechanism of macrointervalized action, which is able to make the
information of a microsystem manifest partly and potentialized partly.
The quantum measurement is actually a physical process in which the
microsystem as an indivisible whole manifests some interval part of itself
corresponding to a localized observable phenomenon; meanwhile the
rest of the whole is potentialized. The empirical structures coming from
empirical data are actually the limited macroscopic localized manifes-
tations of the physical structure of the world in measurements. The
macrointervalized action generally results in the uncontrollable loss of
a certain amount of energy and negative entropy, so the measuring
A REALISTIC INTERPRET A TION OF QUANTUM MECHANICS 385
process, by and large, is an irreversible process, just as a die moving

in the air, due to the constraint of a table plane, finally only manifests
one of six numbers at rest.
3. A NEW TYPE OF CORRESPONDENCE THEORY OF TRUTH
The very core of scientific realism is the correspondence theory of truth .

It claims that there exists a direct one-to-one correspondence relation-
ship between correct scientific theory and reality. Whether or not a
scientific theory is true lies in whether the correspondence relation can
be tested by experiment. In fact , the correspondence relation is one
relating a theoretical model, empirical facts and the objective world,
and is constructed on the basis of measurements. The traditional corre-
spondence theory of truth, I think, is actually constructed on the basis
of classical measurements subject to Einstein's principle of separability.
This is to say, during the measurement the device or the observer can
be in principle separable from the macrosystem, and be spatially isolated
from it; thus the measurement outcomes basically are purely objective
records about the object, and a theoretical model can directly corre-
spond to the objective world. The separability of classical measurements
actually reflects the separability of the macroworld. The traditional cor-
respondence theory of truth actually only holds good for classical theories
about the macroworld, i.e., whether a classical theoretical model is true
lies in whether we can in principle verify the isomorphic relations
between the empirical structures constructed from empirical data and
all the substructures of the structure of the theoretical model on the
basis of classical measurements (Historical tests are of course always
limited and incomplete). Because an object of study of classical physics
is the same macroscopic physical system as a measuring device, they
have the common macroscopic property, so the physical structure of
the macroworld that can be represented by the structure of a certain
theoretical model, can in principle manifest wholly as various observable
phenomena through various experimental mean s, and be directly con-
firmed by all observers, i.e. intersubjectively. There are in principle no
unobservable quantities and substructures in classical physics, which
cannot be forever examined by macroscopic experiments. Historically,
the concepts of absolute space and time had no empirical support for a
long time, but they were finally falsified only by experiments.
However, as to the quantum theory, due to the nonseparability of the
386 MAO SHIYING
experimental context, we cannot already talk about the observed object

under the condition of getting rid of the measuring device or the observer;
moreover, the macrointervality of the device makes it impossible for
us to obtain the whole information of the microobject, therefore the
traditional correspondence theory of truth does not hold for the quantum
theory . It cannot show the direct correspondence relations between
quantum theory and objective world under the condition of crossing
experiment. The failure of the traditional correspondence theory provides
us with two options : either give up realism and take up instrumentalism,
or reconstruct it. The Copenhagen school takes the first, de Broglie, Bohm
et al. take the second . I favor the second one, and think that a new type
of correspondence theory of truth - the quantum correspondence theory
- which is compatible with nonseparability, should be constructed on
the basis of the quantum measurement. First I favor D. Dieks ' point of
view : whether a theoretical model of q.m. is empirically successful,
consists in whether a number of substructures of the theoretical struc-
ture can be isomorphic to empirical structures constructed from empirical
data [4].
But the key to the question lies in how to treat the unobservables of
the theory. As indispensable elements of a theoretical model, unob-
servables come from the free invention of scientists on the basis of
experience. Whether unobservables are of reality, is already a focus of
the realism -antirealism debate . For realists , it is logical coherence and
the unity of theoretic al structure itself that ensure the relative reliability
of the causal relation between unobservables and observables, and make
unobservables be probably of reality. Historically, the success of a lot
of predictions of classical physics verifies this point of view. Anti-realists
completely deny the relative reliability of abductive inference from
observables to unobservables, thus completely deny the possibility of the
existence of unobservables . Clearly, the point of view of thinking in terms
of absolutes restricts the powers of interpretation and prediction of a
theory, and makes it unacceptable to many working scientists. But there
is a unresolvable fundamental problem for realists . Since there actually
are various theoretical models that possess common empirical substruc-
tures corresponding to observable phenomena, e.g., quantum mechanics
have various theoretical models, such as the many-worlds interpreta-
tion, the hidden-variable theory, etc . that agree with the orthodox
interpretation in experience, each contains its own ontological commit-
ment, and possesses some unobservables, then which theoretical model
is closer to the real structure of the objective world on earth? Of course,

this cannot be judged only according to the criterion of empirical
adequacy. It is necessary to synthesize according to other criteria, such
as logical consistency, simplicity, usefulness, economy, elegance, etc .
meanwhile. Perhaps we cannot obtain a unique correct answer forever.
This paper takes the conventional quantum theory as the standard model,
and tries to construct its realistic interpretation, because according to
various different criteria, the prevalent theoretical model is easiest to
accept rationally and universally.
As to the unobservables of q. m., I think, they are unobservable in
principle. This is different from the case of the unobservables in classical
physics, which can be in principle examined sooner or later by experi-
ments. Firstly, the uncertainty relations show that some pair of canonical
conjugates in the theoretical description cannot simultaneously have
the two empirical counterparts. That is to say, relatively to a canonical
conjugate being observed, another conjugate is uncertain, namely an
unobservable. We can call them 'relative unobservables'. Secondly, the
phenomenon of the wavepacket reduction shows that during a mea-
surement one and only one outcome is actual, and all the other eigenstates
(which are possibly countless) in the theoretical description do not
emerge. These disappeared eigenstates can be called 'absolute unob-
servables' . According to the tremendous success of q. m. in experience,
the logical consistency and the unity of the theoretical structure itself,
I think, these unobservables have real existence. Because of the macro-
intervalized action of devices (see Sec. 2), the theoretical observable
corresponds to the manifested part of an object in a measurement, there-
fore the relevant unobservable should be thought to correspond to the
potentialized part of the object beyond the macrointerval in the mea-
surement. The unobservable actually is of the non-macroscopic reality
(concrete analyses about this will be given afterwards). Whether the
magnitude of an unobservable will change after a measurement, needs
to be concretely analyzed in terms of the quantum theory. Therefore,
investigating whether the quantum theory corresponds to the physical
reality, is to investigate not only whether the observables correspond
to the manifested part of the world, but also whether the unobservables
can represent the potentialized part of the world in terms of relatively
reasonable inference. In this way, Einstein's demand of completeness:
'Every element of the physical reality must have a counterpart in the
physical theory' [5], is concretely ensured. After passing the investiga-
388 MAO SHIYING
tion mentioned above, we can get a realistic interpretation of q. m.: the

mathematical structure of q. m. corresponds to the physical structure
of the world, and is of the same inseparable wholeness transcending
the macrointerval as the latter. The empirical substructure of the quantum
theory corresponds to the manifested localized part, and the relevant
unobservable part of the theory corresponds to the potentialized part of
the physical structure of the world in measurement. According to the
point of view, as I will show in the following, the uncertainty relations
and the mutation-like measurement are given a realistic interpretation
respectively.
4 . THE UNCERTAINTY RELATIONS REFLECT THE DUAL

STRUCTURES OF SOME MICROSCOPIC PHYSICAL QUANTITIES
Heisenberg's uncertainty relations are usually regarded as mathemat-

ical expressions of Bohr's complementarity principle. They show that the
canonical conjugate physical quantities such as momentum and space,
energy and time, due to the constraint of Planck's constant, can no more
be measured simultaneously and accurately as they could in classical
measurements. An uncertainty relation as a substructure of the quantum
theory, actually reflects an internal relation between two canonical con-
jugates, in which two sides are independent of each other and mutually
restrictive, but no longer absolutely independent of each other as in
classical physics . I call the relation the 'dual structure'. It is a kind of
physical structure independent of observation by man. The conjugates
in the dual structure are relative unobservable, and can be measured
respectively in experiments. The measurement outcome about a dual
structure is actually only the concrete manifestation of some canonical
conjugate within a macrointerval, meanwhile another conjugate is poten-
tialized, i.e. loses the possibility of accurately manifesting itself in the
macrointerval. The dual structure makes the space-time coordination
and the claim of causality (generally referring to the law of conserva-
tion of momentum and energy) be mutually exclusive and complementary,
but not freely combined as in classical physics [6]. The dual structure
also shows that the wavelike property (represented by position, time or
frequency, etc.) and the particlelike property (represented by momentum,
energy, etc.) are mutually exclusive and complementary. Therefore, we
are no longer able to construct an intuitive picture of the dual struc-
ture, and can only construct an abstract composite picture about it - a
'picture of abstraction' about the microworld - in complementary mea -

suring and describing ways. Bohr 's complementarity principle should
be regarded as not only a description of some observable phenomenon,
but also a description of a dual structure with an ontological status.
5. THE REALISTIC INTERPRETATION OF THE

MUTATION-LIKE QUANTUM MEASUREMENT
The key to the construction of a realistic interpretation of q. m. lies in

the treatment of the mutation-like measurement. The conventional treat-
ment about this is to consider the measurement as a process of the
'wavepacket reduction', in which the microsystem transforms from a
superposition of states to a certain term from the superposition. Suppose
the initial state of the object system is I'V) = :E C, l'Vn)' where l'Vn) is
an eigenvector from the Hilbert space spanned by the complete orthog-
onal set of eigenvectors {1'Vn)} of some observable to be measured.
Suppose a position of the pointer of the measuring device is represented
by a statevector IAn) from the space spanned by the complete orthog-
onal set of vectors {IAn)}, and the zero position of the pointer before
the measurement is labeled I Ao)' Before the measurement there is no
interaction between the two systems, two systems possess together an
initial total statevector I'V) IAo)' After the measurement begins, two
systems affect each other, the initial statevector at last transforms into
the final total statevector l'Vm) IAm), where l'Vm) refers to the final state
of the object system. The whole process of measurement can be repre-
sented as:
(1) I Cnl'Vn) lAo) ---7 l'Vm) lAm)
not
(2) I Cnl'Vn) lAo) ---7 I Cnl'Vn) IAn)'
According to the linearity of the evolution of a statevector with time,

the measurement process should be (2). However, the actual final
outcome of the measurement is not the superposition of the right-side
of (2), but a single term of its component terms. Where do the other
eigenstates go? This constitutes the largest obstacle for a realistic
interpretation of q. m. . Historically, von Neumann generalized the phe-
390 MAO SHIYING
nomenon as the famous ' projection postulate' [6], which supposes that
during the measurement the initial statevector is 'projected' onto one
eigenvector of the Hilbert space. It becomes one of the basic postulates
of the conventional quantum theory. However, it is after all an addi-
tional postulate coming from outside the quantum theory, and is unable
to give the measurement a genuine realistic interpretation in terms of
the intervally valid quantum theory itself. As is well-known, the postu-
late results in the violation of the causal cont inuous evolution of
the statevector governed by a Schrodinger equation, breaks down the
symmetry of time evolution, makes the theoretical description fail to
be in a one-to-one correspondence with the empirical fact rationally,
and leaves a back door for the subjective interpretations of E. Wigner
et al.
According to scientific realism , the quantum theory should be able
to really reflect such a physical process as the measurement. Only a
theory can determine what we are able to observe. Because the theoretical
description does not show that the statevector reduction will occur, if
we still hold that the quantum theory is correct, then we should not
hold that the statevector really ' reduces' . Moreover, there is no time scale
for automatic reduct ion of the state vector, which is compatible with all
observations [7]. Besides, the "statevector reduction" is said to be relative
to the apparatus. Since the device and the object already constitute an
inseparable composite system, we should analyze from the perspective
of the framework of the composite system . In this way, the eigenstates
"disappeared in measurement" cannot be logically considered to have
disappeared. They should unchangedly exist beyond the macrointerval of
the device and inside the composite system. The so-called statevector
reduction, I think, actually shows that there is the mechanism of macro-
intervalized action made by the measuring device. This is to say,
the measuring device constitutes the macrointervalized action to the
statevector of the microsystem, which makes a certain eigenstate manifest
as an actual macrostate, i.e., leaves a permanent macroscopic localized
record in a measurement instrument that can be confirmed by all
observers. Meanwhile, it destroys the coherence of physical states, and
makes the other eigenstates be potentialized beyond the macro interval.
For instance, in the Stern-Gerlach experiment about a spin I12 atom,
the total spin about the z-axis has ±I12 two values corresponding respec-
tively to two opposite oriented statevectors. A mark emerging in the upper
or lower side of the detecting screen along the z-axis, reflects a mani-
festing act of the device to some eigenstate of spin of the atom, mean-
while another eigenstate of the spin is potentialized beyond the
macroscopic screen. Suppose the macrointervalized action is represented
by the operator I, which is equivalent to the projection operator in the
mathematical status .
Therefore, the measurement process can be represented as follows:
(3) 11'1') -7 I Cnl'l'n)p + Cml'l'm).

n,om
where 'p' refers to potentialized states relative to the measuring device,

and 'a' represents the manifested actual state alone. It is obvious that
the operator I can definitely illustrate the overall measurement process.
In contrast to the projection operator, it has really the realistic meaning.
When a microsystem manifests one state of its superposition of states
through a measuring device, the measuring device itself manifests one
state, i.e, the pointer points to some position corresponding to the
eigenvalue of the manifested state of the object. Since the device and
the object already constitute an inseparable whole, the other states
representing the other positions of the pointer should be considered to
be potentialized out of the macrointerval represented by the actual
position of the pointer meanwhile, they correspond to the eigenvalues
of the potentialized states of the object respectively. Therefore, the overall
measurement process can be seen as a process in which the microsystem
and the measuring device constitute a composite system, and two
sets of statevectors are finally unified as a set of composite statevec-
tors governed by a Schrodinger equation. The measurement process can
still be represented as follows :
(4) I Cnl'I'n} IAn}

n
-7 I
nern
c, l'I'n)p IAn)p + CmI'I'm). lAm}.·
The right side of (4) is the state of the composite system (supposing
it is represented by 'P), which has been the superposition of states con-
sisting of a manifested state and potentialized states simultaneously,
and in which the macrointervality of the measuring device has been
reflected . In this way, the theoretical description corresponds wholly to
the measurement process, only corresponds to the empirical fact partly.
According to (4), the probability of the actual eigenvalue is:
(5) Ic mf = 1('I'mAmllIJ)12/1('PllIJ)12.
392 MAO SHIYING
From this expression it is thus clear that the probability of the

actual eigenvalue actually means the squared modulus of the relative
proportion of the manifested macrostate of the composite system, in
the superposition of all states as a whole with the weight ('II). It is
already verified by the statistics of a number of experimental results in
experience.
In this way, the quantum measurement is given a realistic interpre-
tation: the quantum measurement is a process of transformation of a
physical structure, in which the structure of the Hilbert space of a
microsystem and the spatial structure of the adequate coordinates of a
device constitute the structure of a composite Hilbert space through the
mutual coupling where the macrointervalized action is included, and there
is no time scale . Relative to the measuring device, the measurement is
an irreversible physical process in which the spatial structure of the
microsystem manifests itself as the limited macroscopic localized empir-
ical structure, and which is of the statistical law about the measurement
outcome determined by a statevector subject to a Schrodinger type of
equation. The quantum measurement is just like a process in which a
two-dimensional living thing measures a die moving in the air with a
table-like device with the function of manifesting-number. When the
die falls down on the 'table plane ', the device measures one number,
but the other numbers of the die as the component elements of the spatial
structure correspond to the other coordinates of the device potentializedly.
The probability of emergence of some number is one sixth. Th is can
be derived by the distributive relations of the numbers of the die in the
spatial structure, and be verified by a lot of observational results. The
quantum measurement actually shows the gap between the limited
marcoscopic measuring power of man and the abstract power of cogni -
tion of man about the limitless microworld in terms of the quantum
theory . The interpretation restores the symmetry of time evolution, and
ensures that the measurement is still governed by a Schrodinger type
of equation after going through the macrointervalized action.
In summary, this paper proposes a consistent realistic interpretation of
q. m. on the condition of the conventional formalism of q. m. . In contrast
to the many-worlds interpretation of Everett [8], which supposes that
all eigenstates correspond to countless different worlds, along with their
measuring devices and observers respectively, the present interpreta-
tion holds that there is only one actual world, one device and one
observer, but permits the presence of countless relatively potentialized
property worlds corresponding to the potentialized eigenstates in the

measurement respectively. Because the countless eigenstates of the object
belong to some macroscopic observable operator, which corresponds to
a macroscopic observable in classical physics, these eigenstates should
correspond to one apparatus and one observer together, so I think , the
postulate of the many-worlds interpretation is not adequately reason-
able. Bohm 's hidden-variable theory [9] agrees with the orthodox
interpretation in experience, but its additional concepts and structure of
determinism, due to satisfying the demand of nonseparability, maybe
are forever unable to be tested by localized experiments, and appear
like a surplus burden in practical application. The present interpreta-
tion keeps the conventional formalism of q. m. invariant, and of course
agrees with the Copenhagen interpretation empirically. Moreover, it
satisfies the demand of simplicity, and is logically consistent. Not only
does it keep the epistemological meaning of the formalism of q. m.,
but it endows the formalism with an ontological meaning. In this sense,
so to speak, the present interpretation is a realistic 'Copenhagen' inter-
pretation.
Institute of Technology & Society ,

Northeastern University,
Shenyang,
China.
REFERENCE
I. Bohr, N. (1948). ' On the Notion of Causality and Complementarity', Dialectics 2,
pp.312-319.
2. Fraassen , B. C. van (1980). The Scientifi c Image, Oxford University Press, Oxford,
p.I2.
3. Bohr, N. (1935). Phys. Rev. 48, pp. 695-702.
4. Dieks, D. (1989). ' Quantum Mechanics Without the Projection Postulate and Its Realistic
Interpretation', Found. Phys. 19, p. 1402, p. 1416.
5. Einstein, A., Podolsky, B. and Rosen, N. (1935). Phys. Rev. 47, pp. 777-780.
6. Jammer, Max (1986). The Conceptual Development of Quantum Mechanics , McGraw-
Hill Book Company, p. 351, p. 371.
7. Squires, Euan (1986). The Mystery of the Quantum World, Adam Hilger Ltd, Bristol
and Boston, p. 53.
8. DeWitt and Graham (1973). The Many-Worlds Interpretation of Quantum Mechanics,
Princeton University Press, Princeton.
9. Bohm, D. (1952) . ' A Suggested Interpretation of the Quantum Theory in Terms of
'Hidden' Variables' , Phys. Rev. 85, pp. 166-179, 180-193 .
MORTON L. SCHAGRIN
RUMFORD 'S EXPERIMENTS ON

THE MATERIALITY OF LIGHT 1
More Heat Than Light
Rumford's experiments on the heat generated during the boring of

cannons, from which he argued for the "inexhaustible" creation of heat
by friction , were cited and used by James Prescott Joule many years later
to calculate roughly a mechanical equivalent of heat.? I shall suggest
in what follows that the fact that this work was so cited and used by a
later work that itself provided the reference point for even later research
is a key factor in the growth of science.
In contrast to this use of Rumford 's earlier work on heat, his researches
on the materiality of light' made little impression on his contempo-
raries, and are far less known today than his work on heat. We recall
that Rumford's interests always turned to practical improvements in
living. His work on heat was not purely theoretical: he designed fire-
places with improved efficiencies; he devised a coffee percolator; he
explored ways to improve cooked field rations for soldiers. In his studies
regarding light he also devised improved lighting arrangements; the use
of frosted glass in lamps rather than clear glass is due to Rumford. To
assist him in measuring the amount of light produced by different sources,
Rumford constructed a sensitive photometer based on the earlier models
employed by Pierre Bouguer (1729) and Johann Heinrich Lambert
(1760).4
As were the earlier ones, Rumford's photometer was based on two
simple principles:
(1) The intensity of light diminishes as the square of the distance
travelled.
(2) If two sources illuminating a single body each cast a shadow of
the same observable darkness, then the intensities of the light from
each source ar the same.
The instrument (Fig. I) had a white screen upon which the shadows
fell, an opaque object for casting a show,' and two tracks leading away
from the screen at an acute angle to each other bisected by a perpen-
dicular to the screen. These tracks were calibrated by "degrees" marked
off in proportion to the square of the distance. One track was twelve
395
396 MORTON L. SCHAGRIN
s,
b
- -- - - - - - - ------~ c - --
d
S1
Rumford's photometer
By permission from: W. J. Sparrow, Count Rumford of woburn , Mass . (1964)
Fig. I.
RUMFORD'S EXPERIMENTS 397
feet and the other was twenty feet long. The procedure employed a
standard source for comparison, say a candle, placed on one arm at,
say, 10°. Another source was then moved up and down the second arm
until the shadows were equally dense. For instance, if the second source
were to cast an equally dense shadow at the position marked 400°, then
the test light would be forty times as intense as the standard candle.
The experiment proceeded by using a standard wax candle at 100°
on one arm of the photometer. Next an Argand lamp, cleaned and
trimmed, was placed successively at 100° intervals on the other arm. The
wick was adjusted so that the shadows were equally deep. Thus the
lamp shone with intensities of 100° to 900° in 100° steps. The amount
of oil consumed in thirty minutes at each station was carefully measured.
(The lamp was weighed before and after the thirty minute period, with
the difference in weights recorded.) After some arcane juggling of figures
in order to get a basis for comparison, Rumford arrived at the following
table, showing how much intensity of light can be attributed to the con-
sumption of 100 "units" of oil per hour at the different intensities of
flame:
Intensity of the lamp Light furnished per hour per
100 parts of oi I
1000 48 0
200 74
300 98
400 112
500 121
600 138
600 149
800 155
900 160
The first row tells us that if the Argand lamp were to burn for one
hour at 100° of brilliance, then 100 "units" of oil would furnish 48° of
the brilliance of the light. And in the last row, if the lamp were to bum
for one hour at a brilliance of 900°, then 100 units of oil would contribute
160° of the brilliance.
Rumford remarks:
On comparing the results of these nine experiments, it appears that the quantities of
light furnished were very far from being in a constant ratio to the quantities of oil
consumed, as they would doubtless have been, were light one of the chemical products
of combustion,
In the terminology of Bogen and Woodward," Rumford is claiming

there is no phenomenon where one should be expected: the data should
show a relation between the light given off the oil consumed. The data
in this case are presented in the table above. Yet, pace Bogen and
Woodward, those are not Rumford's observations. What I shall call
Rumford's personal experiences were such things as the intensity of
shadows, the pointer of a balance coinciding with the null mark on a
scale, the position of a carriage on a track, etc . His observation reports
used such notions as the brilliance of an Argand lamp, the weight of
oil consumed, etc. These observation reports were further refined to
produce the table of data, which according to Rumford does not reveal
the expected phenomenon: the constant ratio of light produced and oil
consumed.
It is difficult for me to understand how anyone who has conducted
an actual experiment relevant to a genuine hypothesis could have
accepted such positivistic accounts of scientific data in terms of obser-
vation reports such as "Red here now", with the additional claim that
theories are confirmed or refuted by such subjective experiences.
But what about the key experiment described above? Is there no
relation between the amount of oil consumed? at given lamp intensities
and the amount of light released ? And if there is not, what was Rumford
entitled to conclude?
Fir st of all , if one assumes that the amount of light produced per
unit of oil consumed depends on the brightness adjustment of the wick,
that is, the "chemical" reaction is more efficient with more intense flames ,
and that the amount of light given off asymptotically approaches a
maximum, we might conclude that the amount of light produced depends
on the brilliance of the flame as a logarithmic growth curve with a
maximum." It does not seem implausible that the reaction releasing light
becomes more complete as the flame bums more intensely. A plot of such
an equation with actual data noted is as follows (Fig. 2).
Yet suppose my ad hoc hypothesis is not confirmed by further obser-
vations at less intense or more intense flames? Must we then embrace
Rumford's conclusion? Not at all! Because the experiment deals only
with visible light. Recall that W. Herschel had discovered infra -red light
in 1800 and Ritter had found ultra-violet rays in 1801. Since more
(invisible) light might be present in his experiments, Rumford cannot
(logically) refute a theory of light as a material ingredient of chemical
compounds by using observations on visible light alone .
RUMFORD 'S EXPERIMENTS 399
JO(! ....--- -- - - - - -- - - - -- - - -- 1'

1lIO •
180
170 -
l Ull - ~ _--- £1.-- --'. . ~_.- -
~ 150 • ~ -- ~
_ ._- 1=-1
o
H lJ -
J:
<,
..... ,- / '
"'8
1.10
•.-1
/9
...oo /
~~ ~I
/ [=:1
,/
on
....,
....
: 6
§ gO
//
o
o
.....
80
70 "
ii"
<, ,I
~ j E/
Q)
U
c:
....-e
~I sn ~ ,.:
~~ ] :,.:/
CJ
o ¥-~---,-- "-~-_y_-_r-_r- , I I I ,
o 1. ~ B 10
Br ill ianc e o f l arr.?
-kx
y M(l- e )
-0.003x
y 170 (1- e )
Fig. 2.
Regardless of the reasons for its neglect, the fact remains that the
paper was neglected, and so remains in the archives of scientific litera-
ture . But Rumford's experimental results never became a part of the
corpus of science (scientific knowledge, scientific truths). The contrast
with the cannon-boring paper is striking. There Rumford subscribed
to a(n untenable) vibratory theory of heat (not random motion of
particles) . There Rumford's theory was incapable of handling radiant heat

phenomena. Yet Rumford's work there is enshrined in the corpus of
science, even though his thoughts were quite at odds with the finally
accepted theory of heat. Here Rumford's experiments on light are con-
ceptually more compatible with the emerging theories of light. Here
Rumford makes substantial practical suggestions. Yet unfortunately no
one cites Rumford's work on light, and thus it never becomes a part of
science.
PART II
A practitioner of science may cite an earlier work in order to give

credence to his own current efforts. Numerous sociologists beginning
with Merton himself have remarked on this feature of citations.
I call attention to a reverse effect: the cited work gains in scientific
stature because someone can build upon (an interpretation of) it:
One function of reference is therefore to act as a device which establishes the authority
on which the author's argument is founded . At the same time. in supporting his argument
by referring to a particular knowledge claim, the author implies that he accepts the validity
of the claim; in short, a citation used to justify an argument suggests that the author
recognizes the cited claim as a contribution to knowledge.' (emphasis added)
But more than this

A public expression of approv al is given to a paper only when it is cited by some further
paper, which uses its findings as evidence to support a new knowledge claim ."
Yet, so far, we have no closure. There is a cited paper and a citing

paper that approves of the first . I suggest that what is needed is a third
paper that claims to build on the second citing paper. That is, a research
finding becomes an item of scientific knowledge only if it is cited by a
report that is itself cited in tum by another report. Of course, mere citation
is insufficient: it must be favorable citation, and the new work must
use the cited information in a constructive manner.
My "double citation" proposal as a necessary condition for acceptance
into science is similar to a recursive definition except that there is no
base case and it is forward looking rather than backward based. I coin
the new term procursive to characterize the necessary condition:
A published item x is in the corpus of science only if x is favorably cited by publica-
tion y and y is favorably cited by publication z.
RUMFORD'S EXPERIMENTS 401
An important consequence of the proposed procursive condition should

be noted. Suppose the status of x is secured by y , and the status of y is
secured by z. but Z is explicitly refuted or simply not cited by future work
in science. The whole sequence of items is threatened with "unzipping".
That is, y's status is now threatened and a fortiori x's. But it usually
will be the case that some items in the sequence will be connected to
other items in other sequences. Such multiply-connected items will
remain in the corpus of science. Other supposed results that cannot find
new connections, that do not, that i~, provide a basis for new work, are
indeed rejected from science. One ~ s reminded of spurious phenomena
such as N-rays, polywater, or Bohm plasma-diffusion. A small flurry
of mutually supporting experimental results began to penetrate science,
but eventually all of this work was extirpated because it led nowhere;
it couldn't be built upon .
Sometimes when a line of investigation becomes halted through refu -
tation or neglect, and when all the items of a sequence consequently
are threatened, a new idea or hypothesis may lead to a reconstruction
of the linkages. A new sequence of results can emerge, but with some
items missing. The "lost" items may have to wait a while before they can
be used again in some viable sequence of results. Moreover, items may
be permanently lost; not all items find their way into a future recon-
struction. And while they are so lost they are not in the corpus of science.
They are not taught to novices; they are not favorably mentioned in
current research. Notice that these reports have not been refuted in any
strong sense. They just are no longer important or interesting or accepted .
These remark s may help us to understand the phenomenon, emphasized
by Kuhn and others, of non-cumulative paradigm shifts .
The above remarks also show why the argument for Realism based
on the so-called "inference to the best explanation" for why science is
successful is simply the Argument from Des ign in modern clothing.
Again D. Hull makes the point succinctly:
All scientists are trying to understand the world in which they live, but only a small
percentage succeed. Science as a process looks so directional because of the retrospec-
tive bias that we bring to it. It can be made to look globally maximising by careful
editing. Most research fails or leads nowhere . Of the research that finds its way into
print, most has no appreciabl e effect. Once we weed out all the failures and false steps,
science looks very efficient, goal-directed, and globally maximising . II
Rejecting the design argument for the existence of God does not make
one an atheist. Likewi se, rejecting the inference from the fact that science
enables us successfully to control and predict features of our world to

the conclusion that the un cognized world is as we represent it to be
does not force one to be an anti-realist. The fact that our currently
accepted representations are successful is due to the fact that we do
not accept representations unless they are successful, and can be built
upon - the survival of the fittest. Yet, just as the presence of natural
evil suggests to some the atheistic conclusion, so too do the past failures
of previously successful representations (e.g., caloric theory, phlogiston,
aether) suggest to some that the elements of our currently successful
representations are not necessarily similar to elements of an uncognized
world."
SUNY at Fredonia, NY,
U.S.A.
NOTES
I An earlier, and more complete, version of this paper will appear in Synthese
99(1) (1994) 111-121 .
2 Phil. Trans. 140 (1850) . Joule mentions Rumford in his earlie r papers, but this is the
first time he uses Rumford 's reported observations.
) 'On the Light Manifested in Combustion ' , Phil. Trans. (1812)
4 A brief descript ion of these photometers can be found in A. Wolf, A History of Science.
Technology . & Philosophy in the 18th Century, Vol. I (Harper Torchbook, 1961, New
York) pp. 167-170. According to W. J. Sparrow, Count Rumford of Woburn . Mass.
(Thomas Y. Crowell , 1964, New York), Rumford deve loped his instrument in ignorance
of the earlier work. Since the principles upon which all of them were constructed are so
simple , Rumford may have learned about the earlier models through conversations. The
point is not important here and will not be pursued.
5 Rumford actually had two opaque objects so that the shadows from the two sources
would be next to each other on the screen . This variation increased the sensitiv ity of
the judgment of equal darkness, but did not affect the point of the photometer.
6 James Bogen and James Woodward , ' Saving the Phenomena ' , The Philosophical Review
XCVII 3 (July 1988) 303-352. Jim Woodward , ' Data and Phenomena ' , Synthe se 79 (1989)
393-472.
7 It is not possible to estimate the error s of measurement in these experiments, since
Rumford did not mention the weights of the Argand lamps nor how much oil was used .
Because he determined the oil consumed by subtracting the weight of the lamp after
burning for a half hour from the weight of the full lamp at the beginning , it is clear that
the error in the difference of two large numbers can be quite substantial.
S My thanks to S. Brown , J. McKenna, and D. Jelski, colleagues at SUNY at Fredonia,
for helping me with a little curve fitting experiment.
RUMFORD 'S EXPERIMENTS 403
9 G. Nigel Gilbert, 'The Transformation of Research Findings into Scientific Knowl-

edge', Soc. Stud. of Science 6 (1976) 281-306, see especially p. 287. This extremely
important paper came to my attention after the first draft of the current essay, well after
my earlier attempts in 1987 to persuade the Staff Seminar at Fredonia that my insight
had merit.
10 Ibid., p. 294.
11 D. L. Hull, Science as Process (Unive rsity of Chicago Press , 1988, Chicago) p. 474.
12 I must confess that I am equally at a loss as to how to establish (prove) that God
does not exist as I am to prove that the uncognized world is not as we currently repre-
sent it to be. And why would I want to prove either result?
RICHARD H . SCHLAGEL
RESOLVING THE REALIST-ANTIREALIST DILEMMA
The realist-antirealist dilemma consists in deciding whether the results

of current research in the physical sciences (such as genetics, neuro-
physiology, and particle physics) justify belief in the approximate truth
of the respective theories and in the conditionaL or contextuaL reality
of the physical structures designated by the theories, or whether the
fallible and transitional nature of scientific explanations preclude
believing in anything other than the observabLe results of experimental
inquiry. For realists, accepting the confirmed predictions deduced from
theories while denying reality to the insensible, underlying mechanisms
and interactions from which the predictions are derived makes scien-
tific inquiry miraculous or magical.
Antirealists, on the other hand, following a long tradition from
Hipparchus and Osiander to Mach and van Fraassen according to which
astronomical systems and physical theories were derided as mathe-
matical fictions, argue that anomalies in quantum mechanics (such as
the wave-particle duality and the fuzzy status of the electron), impasses
in the general theory of relativity (pertaining to the continuity or dis-
continuity of space -time), as well as the continuous revision of theories,
preclude theoretical concepts from designating any actually existing
physical entities, states, or processes. As Bas van Fraassen states:
It is the intensional correlate of the conceptual framework through which I perceive and
conceive the world . But our conceptual framework changes, hence the intensional cor-
relate of our conceptual framework changes - but the real world is the same world . I
From this he concludes that "to accept a theory is . .. to believe that

it is empirically adequate - that what the theory says about what is
observabLe . . . is true"." Because theories are ultimately verified by
observable data observation constitutes the limits of what can be known,
conforming to the ancient dictum of "saving the phenomena" while
refusing to believe in the reality of the explanatory mechanisms.
Antirealist philosophers like van Fraassen are the Osianders of contem-
porary physics.'
Arthur Fine, who describes himself as a "nonrealist", states that realists
405
406 RICHARD H. SCHLAGEL
mistakenly make "the question-begging move from explanatory efficacy

to the truth of the explanatory hypothesis"," and erroneously "suppose
that . . . confirmation is a mark of an approximately correct ontology
and approximately true laws".' He even denies that scientific inquiry
offers any evidence of "a [progressive] change where we learn more accu-
rately about the same things"? In my opinion these antirealist assertions
are so implausible given the meticulous competitive confirmations in
physics (for example, the joint confirmation of the J/psi particle by Burton
Richter's group at SLACK and Ting's group at Brookhaven), the frequent
theoretical convergences among the sciences (as in molecular biology),
and the remarkable technological inventions and applications (such as
the radio telescope, tunnelling microscope, and current use of com-
puters in simulating physical experiments), that we simply can reject
the antirealist hom of the dilemma.'
It requires an implausible credulity to believe that all the remark-
able consequences of current experimental inquiry , purported to be based
on discovered entities and properties, are actually derived from ficti-
tious theories and nonex istent entities: e.g. , the magnetic acceleration
of electrons and protons in particle accelerators, genetic engineering
and the genome project based on the atomic-molecular structure of the
DNA-RNA, high speed transit utilizing electromagnetic fields, and the
role of anions in superconductivity. Rather than imposing arbitrary a
priori limitations on the possibility or actuality of scientific discov-
eries, what is required is a conception of science that preserves and
accounts for its achievements while also acknowledging the idealized,
approximate, and fallible nature of scientific theories .
From the beginning of modem science with Kepler 's discovery of
the three planetary laws , Galileo's experimental confirmation of the
correct law of free fall, and Newton's derivation of the universal law
of gravitation from Kepler's third law, it was apparent to the investiga-
tors that their laws applied to idealized states and approximate conditions
where the extraneous factors had been abstracted and eliminated. This
does not mean , however , that these laws are false but that they are true
within the limiting conditions specified. In fact, as science progre sses we
discover why previous laws only held approximately, as Newton 's law
of gravitation demonstrated with respect to Kepler 's astronomical laws
and Galileo 's gravitational law, and as Einstein's general relativity theory
showed with respect to the anomalous shift in the perihelion of Mercury ,
unexplainable on Newton's law.
RESOL VING THE REALIST -ANTIREALIST DILEMMA 407
While it may be dramatic to declare "how the laws of physics lie",8

false laws are difficult to reconcile with the precision of scientific pre-
dictions (as with the spacecraft Voyager that flew billions of miles over
a period of eight years and arrived within seconds of the predicted
time), or with the construction of extremely complex apparatus (as in
the design of the magnets required to accelerate the electrons or protons
in the previously funded but now cancelled superconductor-supercollider),
or the exact prediction of particles with specific properties (as the omega
minus). As Kepler stated: " 'It is a most absurd fiction . . . that the phe-
nomena of nature can be explained by false causes ' " . 9 While not all
the properties and interactions of every model used in science are con-
sidered real, as Cartwright affirm s, this does not mean that the laws of
science on which confirmed theories are based are false.
In addition to idealized states and approximate conditions, scientific
inquiry utilizes abstract concepts such as mass, force, instantaneous
velocity, and inerti a, as well as isotopic spin, symmetry, charm and color,
and quarks. In so far as these concepts are precisely defined in terms
of experimentally detectable properties and magnitudes - and that is what
mainly differentiates twentieth century science from past science - they
have exact meanings and designations within the theoretical framework
despite their abstractness and remoteness from ordinary experience.
Furthermore, because natural processes are understood as functional
relations among constants and variables as expressed in mathematical
equations (F = MA, E = MC 2, e = hv, etc.), it is the formalism that
conveys our most precise representation of the underlying physical
structures and interactions producing phenomena. What anchors this
formalism in the real world is the experimental determination of the
magnitudes of the theoretical structures and their properties, with the
discovery of constants (like the invariant velocity of light and Planck's
constant of action) attesting to some encounter with an inherent feature
of physical reality. It is hardly credible that these constants, any more
than Mendeleev's periodic chart, are merely phenomenal uniformities
with no underlying physical significance or causes.
The reason experimental inquiry is necessary is that ordinary occur-
rences are not self-existent or self-explanatory, but require explication
in terms of deeper causal structures and processes or inherent mecha-
nisms. Moreover, because sensory observation discloses the world relative
to those sense modalities, the use of optical and radio telescopes , spec -
troscopy, refractive, electron, and tunnelling microscopes, cloud or bubble
chambers, particle accelerators and supercolliders, etc ., have been

invented to augment or supplement our sensory knowledge. Just as the
qualitative macroworld appears as it does due to the nature of the human
organism, as Kant maintained, so the detected states and properties
of the microworld essential for scientific theorizing are manifested
experimentally, as Bohr claimed.
Like all knowledge, our access to and understanding of deeper domains
of physical reality depend upon the type, power, and precision of the
instruments used in probing nature as interpreted by the conceptual frame-
work and mathematical formalism." To claim that only the world as
displayed within the limitations of human sensory observation is worthy
of belief is to succumb to a naive anthropomorphism. Even the distinc-
tion between what is a "theoretical construct" as opposed to "empirical
phenomena" constantly shifts with improvements in technology: e.g., the
prediction of Neptune based on Newton's laws to account for the irreg -
ular motion of Uranus was later telescopically confirmed; the molecular
structure of DNA-RNA can now be observed with the tunnelling micro-
scope; and even the lattice structure of atoms comprising the surface
of metals can be 'seen' using the latest microscopic technology aided
by computer enhancement. The later observable confirmation due to
improved technology of what at one time were purely theoretical entities
should reinforce the belief that theories can represent actual, if insen-
sible, structures.
So rather than concluding that none of our theories has any veri simili-
tude or referential validity beyond what is optically observable and
tactually detectable, it should be recognized that physical reality itself
has induced scientists to develop improved investigative techniques and
forced them to assimilate the striking, unexpected results obtained -
typified in the remark of the American physicist I. Rabi, "Who ordered
this? ", when the muon was correctly identified. II
This brief justification of a realist interpretation of science is based
on actual scientific achievements. However, in the realist-antirealist
debate two theoretical arguments opposed to realism are often introduced
which, if shown to be mistaken, can remove much of the opposition to
realism. The first, related to Quine's thesis of "the underdetermination
of theories by empirical evidence"," claims that because the empirical
support for theoretical explanations is almost always inconclusive it is
possible that alternative theories could account for the same phenomena.
So if it is possible that any present theory could be replaced by another
theory that would serve just as well, why should we believe that any
one theory is true? This is a variation of the argument Pope Urban VIII
stated to Galileo which he repeated in The Two Chief World Systems ;13
namely, that however persuasive the empirical evidence might be it
should be remembered that God in his omniscience and omnipotence
could have caused things to occur by many means inconceivable to
man beyond those they appear to have . The argument has a certain
heuristic value (as in Galileo's case where he was mistaken about the
cause of the tides), in that it acknowledges the fallibility of scientific
inquiry and cautions against premature acceptance of a theory, but
does it warrant the conclusion that we never are justified in believing
that any theory is true (within certain domains and relative to certain
approximations)? - for example, the kinetic-molecular interpretation of
gases, electrical conduction, the explanation of chemical reactions in
terms of molecular structures and bonds, the chemical-electrical inter-
pretation of synaptic discharges and neurotransmitters, etc.
What weight should be given to an argument based on a theoretical
possibility as compared to actual historical precedents ? For if we consider
modern scientific controversies, invariably there were two or three com-
peting theories, not an indefinite number of possible contenders. While
during the proto stage of scientific development (as in Presocratic phi-
losophy) a wide diversity of explanations was possible, as scientific
inquiry advanced with its methodology becoming more selective and
its theories more refined, the possible interpretations became more
restricted. Consider the classic example of the Copernican revolution
where there were just three competing theories, the Aristotelian-
Ptolemaic-geocentric view, the Copernican-Keplerian-heliocentric sys-
tem, and the Tychonic compromise. Would anyone claim that the sub-
sequent evidence provided by Gal ileo's new telescopic observations
and Newton's universal law of gravitation (finally "resolving the problem
of the planets") did not justify astronomers in believing that the helio-
centric theory of the planetary system was true? Should Galileo, Kepler,
and Newton have remained skeptical as to the truth of heliocentrism
because another explanation theoretically was possible?
Or consider another classic example, the "hypotheses" offered to
explain light in terms of part icles or corpuscles and pulses or waves?
Based on the evidence of the sharp outline of shadows and the refrac-
tion of light in denser mediums (which he attributed to attraction of
the corpuscles), Newton concluded that the corpuscular hypothesis was
410 RICHARD H . SCHLAGEL
correct, even though he was aware that other phenomena such as color
films and diffraction seemed to support the pulsation or wave hypoth-
esis of Hooke and Huygens. Newton's prestige, along with the success
of atomism in chemical explanations, was such that the corpuscular
hypothesis prevailed until the early nineteenth century when experiments
by Young and Fresnel demonstrated that diffraction patterns supported
the wave theory of light, and Foucault proved that light is retarded in
denser mediums contrary to Newton's corpuscular prediction . As a result,
the wave theory was accepted until Einstein's explanation of the photo-
electric effect reintroduced the conception of discrete quanta of light
as photons, which was confirmed by the scattering experiments of Arthur
Compton. Now light is believed to have both particle and wave char-
acteristics, as detected under different experimental conditions.
While this brief history suggests that it would have been imprudent
earlier to have decided conclusively in terms of either particles or waves,
it also shows that the experimental evidence constrained the possibili-
ties of explanation. This was true of other controversies, such as the fluid
or particle interpretation of electrical conduction, the caloric versus the
kinetic theory of heat, Stahl's phlogiston versus Lavoisier's oxygen expla-
nation of combustion, Darwin's natural selection conception of evolution
versus Lamarck' s inherited characteristics, the Golgi reticulum versus the
Cajal neuronal theory, or Linus Pauling's protein versus Watson's and
Crick's nucleic acid interpretation of genes, etc. In each case the evidence,
along with the theoretical constraints based on accepted theories, decided
the issue, despite the possibility of other unknown explan ations!
At the frontiers of science (as today in quantum mechanics) there is
considerable uncertainty as to what is the correct interpretation, but this
was always true. When reading the works of Galileo and Newton 14 one
is struck by the fact that they were very doubtful as to whether suffi-
cient evidence ever would be forthcoming to decide whether the universe
was spherically finite or indefinitely extended (Galileo) or to explain
the nature of gravity and of attractive and repulsive forces (Newton).
While it is true that some of these problems are still with us, many others
have been settled. Thus the mere theoretical possibility of alternative
theories should not be an impediment to believing a theory once the
evidence is compelling, as it is in many instances. It surely is an unwar-
ranted inference from the fact that it is always possible that another theory
could be found to explain the same phenomena to the conclusion that
no scientific theory , regardless of the evidence, can be considered true.
RESOLVING THE REALIST - A NTI R E A LI ST DILEMMA 411
Newton himself was aware of the deficiency of this argument as indi-

cated in the following quotation, although the point of his argument
was to oppose the introduction of merely hypothetical explanations.
If anyone offers conje ctures about the truth of things from the mere possibility of
hypotheses, I do not see how any thing certain can be determined in any science ; for it
is always possible to contrive hypotheses, one after another. . . . Wherefore I judged
that one should abstain from considering hypotheses as from a fallacious argument, and
that the force of their opposition must be removed, that one may arrive at a maturer
and more general explanation. IS
The second objection pertains to theory change, reference, and incom-

mensurability. Ever since Kuhn's influential monograph on paradigm
shifts in science," and Feyerabend's "anarchistic" thesis that "anything
goes "!" in scientific inquiry, it has been argued that scientific revolu-
tions are incommensurable because the concepts in the opposing theories
refer to different entities. Because the competing theories describe dif-
ferent universes they cannot be compared, nor can one be the result of
a transformation of the other. Revision and development of concepts
occurs within paradigms, but paradigms themselves are so self-contained
and disparate that they cannot be compared, only rejected and replaced.
While one can find some instances illustrating this, as when the cor-
puscular-kinetic, mechanistic universe was developed by Boyle and
Newton to replace the outmoded substance-attribute, organismic cos-
mology of Aristotle, or when the theory of oxygen replaced that of
phlogiston, it is extremely rare and does not characterize most scien-
tific developments, even revolutionary ones. This notion of scientific
revolutions as involving rationally incomparable theories is based, I
believe, on a mistaken interpretation of conceptual frameworks and how
they are revised.
A world-view is contained in a theoretical framework consisting of
a system of basic concepts and a network of various semantic implica-
tions (plus the syntactic structure) based on the meanings of the concepts
as stated by the sentences composing the framework. A scientific world-
view consists of the conceptual system derived from the accepted
methodology. These conceptual-linguistic frameworks comprise the
background reference system used in communicating or exchanging
views, as Whorf and Quine pointed out. Without any commonality
in the background framework no meaningful exchange is possible.
Moreover, it is these entailment connections among the meanings of
scientific concepts and laws that make possible scientific reasoning, infer-
ence, and prediction.
Intellectual development, whether personal or historical, consists of
a reorganization of the conceptual framework by revising the meaning
of older concepts and introducing new ones, which in turn modifies the
implicative or entailment connections composing the strands in the
network, and by adjustments in the weights given to the evidential
significance of the statements. This is why theory change, even if
revolutionary, is never wholesale. Some of the concepts and/or their
meanings are modified while new ones are added disrupting the implica-
tive relations. Whereas in the order framework the inferential links tended
to be consistent or coherent, modifications introduce incongruities and
anomalies that require reinterpretation of the meanings as well as the
rejection of older, and the creation of newer, concepts. But if all the
underlying concepts and their entailment relations were changed, the
system would be completely unintelligible. As Paul Thagard states: "Thus
even revolutionary conceptual change occurs against a background of
concepts that have relative stability". 18
This is why ordinary language plays such a central role in any dis-
course - it provides the broadest and most familiar background by means
of which new concepts can be explicated and defined. Bohr made the
same claim for classical physics, asserting that discourse in quantum
mechanics presupposes the description of laboratory equipment and
experimental phenomena in the language of Newtonian physics. There
must be some underlying conceptual-symbolic system that remains rela-
tively stable and intelligible if one is to talk sensibly of paradoxes and
anomal ies.
This can be illustrated best in terms of actual historical develop-
ments , such as the Copernican Revolution, the favorite example of Kuhn
and Feyerabend. Certainly there was considerable ambiguity, misun-
derstanding, and irrationality underlying the opposing arguments in the
early seventeenth century. But even granted this, the transformation of
the traditional geocentric universe into the heliocentric system was a
gradual process, as Kuhn himself argued in his earlier book." Copernicus
left the traditional universe essentially as it was, arguing that a simpler
more interconnected explanation of planetary motion could be obtained,
and artificial computing devices such as the equant eliminated, if one
described the orbital revolution of the planets with the sun rather
than the earth at the center - as Aristarchus had suggested nearly two
RESOLVING THE REALIST-ANTIREALIST DILEMMA 413
millennia earlier. This interchange of the sun and the earth had its jarring
conceptual implications affecting the meanings of ' terrestrial' and 'celes-
tial bodies ', since the stationary earth would revolve among the celestial
bodies and the rising and setting sun would be stationary and centrally
located in the terrestrial world . In addition, terrestrial objects would no
longer fall to the center of the universe but to the center of the earth.
The motion of the earth introduced further problems: if the earth had
an orbital revolution why was no stellar parallax observed and why are
not clouds and birds left behind as the earth revolves? If the earth
rotates on its axis diurnally causing the apparent rising and setting of
the sun , why are not objects thrown from the earth's surface, why do
objects propelled straight upward return to the same location, and why
do cannon balls fired in opposite directions with the same force land
an equal distance away (since the earth should have rotated further from
one and towards the other during their flights)? These objections are
indicative of disrupted semantic or entailment relations based on the
changed meaning of the concepts induced by the heliocentric conception,
yet the general framework of the older system remained intact: e.g .,
the assumption of a finite spherical universe with a center and periphery
bounded by the fixed stars, the inferior and superior planets revolving
around the center with a circular uniform motion, etc. Most of the dif-
ficult ies had to do with reconciling the earth to its new position as a
revolving, rotating planet.
It was Kepler who, in order to describe precisely the orbital revolu-
tion of the planets based on a sun-centered system and Tycho 's
astronomical observations of Mars, challenged the ancient simplifying
assumptions that the planets as celestial bodies must have perfect circular
and uniform motions. By discarding these assumptions he was able for
the first time in history to formulate precise laws depicting and predicting
planetary motions along with their relative orbital distances from the sun.
Then Galileo introduced new astronomical data derived from his tele-
scopic observations that undermined the ancient distinction between the
celestial and terrestrial worlds: the observation of mountains and craters
on the moon resembling those on the earth , spots revolving on the surface
of the sun suggesting clouds, satellites revolving around Jupiter resem-
bling the earth and its satellite the moon , and the evidence of the phases
of Venus supporting the predictions based on the heliocentric system
and refuting those derived from the geocentric view.
In addition, Galileo skilfully devised arguments rebutting the objec-
tions opposing the heliocentric system based on terrestrial implications

and the Aristotelian distinction between circular and rectilinear motion
as definitive of celestial and terrestrial bodies. These dialectical argu-
ments demonstrate irrefutably that for Galileo the Aristotelian-geocentric
and Copernican-heliocentric frameworks were not incommensurable
because the arguments presuppose a thorough understanding of both
systems (not surprising since he taught the Aristotelian system at the
University of Pisa before committing himself to heliocentrism).
So gradually the theoretical framework containing the classical geo-
centric cosmology became unravelled and a revised network of concepts
woven into a new astronomical system. But this was a gradual process
of change that took nearly a century from the publication of Copernicus's
De Revolutionibus to the eventual confirmation and acceptance of the
heliocentric system . Yet even this radical transformation involved some
continuing stable designations regardless of the shifts in meaning: e.g .,
despite the change in significance of ' celestial bodies' the term 'planets '
still designated six of the seven traditional bodies; regardless of the
altered position of the sun and the earth the terms referred to the same
entities; although elliptical, the planets still revolved in orbits, etc.
Attributing the inability to change frameworks to being incapable of
shifting conceptually from one paradigm to another because they are
incommensurable overlooks the ancillary role of additional beliefs. Even
when one clearly understood the nature of and rationale for the two
systems there were all kinds of reasons for resisting the change beyond
the astronomical evidence: the displacement of man from the center of
the universe, the challenge to Catholic doctrine and authority, the refu-
tation of scriptural passages claiming the sun moved, the threat to
university professors teaching the Aristotelian system , and the disruption
of common sense and religious convictions.
I think a similar analysis would show that Einstein 's relativity theory,
despite the radical change in meaning of the basic concepts, was not
incommensurable with Newtonian science if that is construed as not being
able to compare rationally the two systems or understand the reasons
for the introduction of Einstein 's theory, especially in connection with
the null results of the Michelson-Morley experiments. Einstein did not
describe his relativity theory as incommensurable with Newton's cos-
mological system, but as reducing to it in the limiting case where the
velocity of light is infinite or instantaneous, as it is for all practical
purposes on the earth. Similarly, Bohr described quantum mechanics
as reducing to classical mechanics when the magnitude of interactions

is large relative to Planck 's constant.
The main flaw in the interpretation of theory change as incommen-
surable, rather than transitional, is that it does not recognize that the
intensional or connotative meanings of words can change while their
denotations or designations remain the same. This misunderstanding
underlies Fine's assertion quoted earlier that the history of science offers
no evidence of "a change where we learn more accurately about the same
things". But can it sensibly be maintained that when we talk about the
rings of Venus we are not referring to the same planet as did Copernicus,
that when we describe the heart as a pump producing the circulation of
the blood we are not designating the same organ meant by Aristotle or
Galen , that when we refer to neurons in the brain we are not denoting
the same organ as did Descartes, or that the reference to atoms is not
continuous with that of Leucippus and Democritus, Epicurus, Lucretius,
Dalton, Rutherford, and Bohr?
If learning more about the inner structures and properties of atoms ,
cells , neurons, and planets precluded referring to the same entities, then
this would have to be true of our increasing knowledge of ordinary
objects as well. The term 'whale' which once meant a fish could not
be used to designate the same aquatic creature when it was learned that
whales are mammals. We could not claim bread consists of carbohydrates,
vitamins, and minerals, that the composition of water is H20, or that
genes are actually chromosomes which in turn are DNA-RNA struc-
tures because the terms' bread', ' water ' and' gene' would have different
referents from those of their ingredients. But denotation can remain fixed
while the intensional cluster of properties change . When the earth became
a planet, it was still the earth that was designated.
Unless this is recognized we would have to admit that our knowl-
edge of the physical universe had not advanced beyond that of Aristotle
and Newton, or that our conception of the atom had not progressed
beyond Democritus or Dalton, or that our understanding of how the
nervous system functions is no better than Galen and Descartes because
each investigator would have referred to a different physical phenom-
enon. And if "explanatory efficacy is not a sign of the partial truth of
explanatory hypotheses" and "predictive confirmation not a mark of an
approximately correct ontology and true laws", as Fine claims, then
why are not concepts of crystalline spheres, the five Platonic solids, phlo-
giston or genes just as effective in scientific explanations as gravity,
oxygen, Wand Z particles, and the DNA-RNA? As Sheldon Glashow

claims regarding the Grand Unified Theory (GUTs), " 'I'm convinced the
theory is sound .. . the essential framework of GUTs simply explains
too many things for it to be fundamentally wrong' "20 - the basic claim
also of realists. Since it is experimental inquiries as interpreted by these
theories that account for the achievements of science, an adequate con-
ception of science must include an explanation of how this occurs. The
basic choice , then, is between accepting an interpretation of science (as
antirealists or nonrealists do) that precludes actual scientific discov-
eries, approximately true explanations, and continuous progress, or
accepting the obvious fact of the latter and formulating an interpreta-
tion of scientific inquiry that accounts for it. After all, what is actual must
be possible.
The George Washington University,
U.S.A.
NOTES
1 Fraassen, Bas van (1980). The Scientific Ima ge, Clarendon Press, Oxford , p. 81.
2 Ibid., p. 18.
3 Cf. Schlagel, Richard H. (1988). ' "Critical Study": "Experimental Real ism: A Critique
of Bas van Fraassen's " Constructive Empiricism?", in Review of Metaphysics 41, June
1988, pp. 789-814.
4 Fine, Arthur (1986). The Shaky Game: Einstein Realism and the Quantum Reality ,
The University of Chicago Pres s, Chicago, p. 120.
5 Ibid., p. 117.
6 Ibid., p. 130. Brackets added .
7 Cf. Schlagel, Richard H. (1991). '''Critical Notice ": " Fine ' s ' Shaky Game' (And Why
NOA is No Ark for Science)"', Philosophy of Science 58 , June 1991, pp. 307-323.
8 Cf. Cartwright, Nancy (1983). How the Laws of Physics Lie, Oxford University Press,
Oxford, p. 152.
9 Kepler, Johannes (1963). Astronomia Nova, Prefatory Malter . Quoted from Koestler,
Arthur, The Sleep Walkers, Universal Library, New York, p. 169.
10 Cf. Schlagel, Richard H. (1985). Contextual Realism : A Meta-physical Framework
for Modern Science, Paragon House , New York, Ch. IX.

II Cf. Ne'eman, Yuval & Kirsch, Yoram (1986). The Particle Hunters, Cambridge
University Press, Cambridge, p. 86.

12 Quine, W. V. O. (1970). 'On the Reasons for Indeterminacy of Translation ' , The
Journal of Philosophy 67 , pp . 178-179.

13 Galilei, Galileo (1970). Dialogue Concerning the Two Chief World Systems, trans .
by Stillman Drake, University of California, Press, Berkeley, p. 464.
14 This is based on my research for Volume II of From Myth to the Modern Mind ,
Copernicus through Quantum Mechanics .
15 Letter to Oldenburg, Secretary of the Royal Society, in reply to Pardies ' objections
to Newton' s theory of colors. Quoted from Sir Isaac Newton (1962). Principia, Vol. 2,
"The System of the World", Motte's trans., revised by Cajori, University of California
Press, Berkeley , p. 677.
16 Cf. Kuhn, Thoma s S. (1962). The Structure of Scientific Revolutions , University of
Chicago Press, Chicago.
17 Cf. Feyerabend, Paul (1978). Against Method, Verso, London, p. 23.
18 Thagard , Paul (1992). Conceptual Revolutions , Princeton University Press, Princeton,
p.56.
19 Cf. Kuhn, Thomas S. (1957). The Copernican Revolution , Harvard University Press,
Cambridge, pp. 180-184. In my opinion his earlier analy sis is more correct than the
later one.
20 Quoted by Crease , Robert P. and Mann, Charles C. ' How the Universe Works' , The
Atlanti c Monthly , August 1984, p. 92.
ZHOU JI HONG
THE MATHEMATICIZED PRACTICAL ATTITUDE
Arthur Fine suggested 'The Natural Ontological Attitude', which accept

the results of scientific investigation as ' true' , on a par with more homely
truths. He believed that both realists and antirealists all accept scien-
tific truth at this 'core position' . The natural ontological attitude (acronym
NOA) is just the core position itself and all by itself, neither realist nor
antirealist but mediates between the two, and it is for commonsense epis-
temology. Considering that realism adds an outer direction to NOA
(that is, the external world and the correspondence relation of approxi-
mate truth), and antirealism (typically) adds an inner direction (that is
human-oriented reductions of truth, or concepts, or explanation), Arthur
Fine tried to conciliate the controversy between realist and antirealist
through NOA: 'NOA suggests that the legitimate features of these addi-
tions are already contained in the presumed equal status of everyday
truths with scientific ones, and in our accepting them both as truths.
No other additions are legitimate and none are required.' I Thus, the
controversy between realist and antirealist seems to be nonsignificant,
and the puzzling problems about modern physics and the challenge they
give to philosophy all will be covered up by the empirical view of
NOA.
In my opinion, NOA is too near a naive emp iricism to deal with the
problems about modern scientific realism. Although NOA correctly
realizes the fundamental defects in realism and antirealism, it does not
offer a better means to draw an alternative scheme but recedes to a
nonactive position, evidently not enough for our debate. So I put forth
a 'Mathematicized Practical Attitude' (acronym MPA) to supplement and
modify NOA . In this essay I want to analyse scientific theory in detail
and elaborate MPA, which argues that scientific theory actually consist
of a mathematicized structure - it is an operational framework - from
an inner-view, while scientific explanation as an abductive inference
just reflects a subjective perspective. Finally, MPA suggests that, from
a point of view of an operating level, science provides a mathernati-
cized operational system for getting knowledge of what the world is
like and for controlling nature, and thus antirealism is approximate; on
419
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.), Realism and Anti-R ealism in the
420 ZHOU 11 HONG
the other hand, from a point of view of an onlooking level, science is

the only way through which we can touch the world beyond our senses,
and the operational framework which scientific theory provides can reveal
an image about the world. This image's rightness is obvious within a
related scientific theoretical system and a scientific practical system,
and thus a special realism is the very reasonable attitude. According to
MPA, realism and antirealism seem to be incommensurable and are
complementary at different levels .
The Empiricist and Popperian views of science take scientific laws and
theories to be statements of the form; for all X, if all X has the property
F, then X has the property G. The first step is conceptually to delin-
eate a universe of X's having property F, and then it is to be empirically
determined whether such X's also have property G. A number of these
X's are thus to be observed and found either to have or not to have
this property.' But this kind of view, based on a Deductive-Model, was
rejected by later philosophers after the criticism of Kuhn and Feyerabend
who offer a Gestalt-Model as the substitute. From the point of view of
a psychologist and historian, Kuhn used the term ' incommensurable '
in order to express what is perhaps the same particular insights concerning
the nature of the relation between certain scientific theories.' So scien-
tific theories are not taken to be statements corresponding with a truth
value , but are instead likened to individual empirical concepts or pred -
icates which are intended to apply only to certain phenomena. Thus
theories are conceived of as being intended to apply to certain states
of affairs, and to be such that they may be judged to be more or less
successful in their application."
The progress of scientific philosophy of Post-Kuhnian and post-
Feyerabendian [analysis] shows that restricting attention to syntactical
at the expense of semantic problems is a lopsided view, which does not
conform to the history of science (particularly the history of physics).
Although Kuhn's explanation is on a correct path by which scientific
theory is regarded as a correlated and historical unit, it has so strong a
psychological tint that it is apt to introduce people to an indefinite and
imprecise state.
Considering that these views are not all satisfactory I put forward
an alternative view about scientific theory as the basic idea of MPA. In
THE MATHEMATICIZED PRACTICAL ATTITUDE 421
my opinion, science is rather a mathematicized practice than a theory-

system which would exhibit a more and more precisely natural picture,
or than a subjective paradigm which mixes together many non-scien-
tific factors . This practice is neither merely a magnificent organizing tool
nor just corresponds with realism itself, but provides us with an opera-
tional framework ; my core position depends upon the practical notion
rather than on a commonsense epistemology. The difference between
MPA and NOA is that the former always considers scientific theory from
the angle of view of practice.
MPA holds that the basic components of scientific theory are math-
ematical deductive systems, dimensional systems and their frames of
reference. Many basic concepts of theory are operationally defined. In
formulations of laws, these concepts appear as corresponding parame-
ters with certain dimensions. By means of these basic components and
parameters which have specific space-time features, scientific theory
mathematically defines and expresses the properties and dispositions
of analysed objects or phenomena, and further makes it possible for them
to go into an experimental operational process in order to investigate
and test their structures or relations, and so on. Thus the consistency
of intersubjects based on this operational process can offer a stable and
reasonable way for scientific communication and development. If we
do not rigidly adhere to linguistic analysis of scientific theory, but take
MPA's practical viewpoint, we should agree that scientific theory consists
of an especially abstract element such as model and hypothesis which
cannot simply be reduced or separated into isolated statements.
Before the twentieth century, many scientists and philosophers
believed that there were some laws of nature laying hidden in the night,
and scientists could discover them. Now, however a law is conceived
as the reason which accounts for uniformity in nature, not the mere
uniformity or regularity itself.' MPA suggests that a law is just a working
hypothesis for scientific practice and a temporary achievement of
theoretical explanation. A theory might include many elements such as
postulates, axioms, principles and theoretical terms , but its core content
is mathematical formula and physical (or chemical or other special) para-
meters and working hypothesi s. The theory 's axioms, principles, and
basic postulates determine its operational structure together with its exper-
imental laws and its core content.
Of course, models embedded in scientific theory would provide an
explanation of mathematical systems, but they are not enduring and can
422 ZHOU 11 HONG
only playa role of enlightenment. The really core part of a theory lies
in its operational framework which is determined by some basic elements
of the theory. For example, as to the mechanical law: F = m, 11vi11t. While
in appearance this law seems to explain the cause of an object's motion
and its regularity, the formula itself gives us nothing else than an instru-
ment by which we can measure interrelated parameters, examine relations
between objects or motions and predict the result of a given mechan-
ical process. Here all these terms (F, m, v, t) are operable in experiment
although they may be attached with unnecessary principles or metaphors.
In short, it is an operational framework within which people can exper-
iment and forecast to understand the relations between elements and
dispositions of an object.
In the case of classical physics, no postulate contains the term 'mea-
surement' and in whatever expressions the term may appear, it is always
eliminable in favor of other theoretical terms. Now in empirical science,
a theoretical statement must be confirmed either directly or ind irectly
by empirical evidence." According to MPA's opinion, a scientist always
brings experimental data into a correlative methematicized connection or
context, the cluster of laws, or a specific model, and harmonizes them .
The viability of theory depends rather on the fact that it indicates the
mathematicized angle of view, and the operational program of thus
observable phenomena can be treated by mathematical means and exper-
iments, more than that giving a kind of plausibly speculative hypothesis.
For scientists, theory is the means which they design to solve those
problems left unresolved. For instance, force to Newton was never
an obscure qualitative action, as the sympathies and antipathies of
Renaissance Naturalism had been. He set it in a precisely mechanical
context in which force was measured by the quantity of motion it could
generate.' Within the operational framework and mathematical system of
theory, the parameters, terms and theoretical entities would take a defined
denotation which is operational directly or indirectly in experiments.
When going into the operational framework and mathematical system,
the property and disposition of object can be described, tested and pre-
dicted during the mathematicized practice.
II
From the surface, we can only see that scientific theory seems consti-
tuted of principles, and explanatory model(s); we cannot see its core
content. It is obvious that scientific theory does not directly describe mea-
surement processes and operational structures, but the mathematical
system, dimension system and basic frame of reference which theory
contains will jointly indicate the operational framework. This in-formed
framework also exists in the explanational model, the inductive hypoth-
esis. Even if the theoretical parts, more precisely, the inferential and
explanatory parts are denied, the in-formed framework will remain and
still be the base of the development of theory, as it is stable in its
operational style and experimental data/phenomena. The operational
framework makes it possible for scientists to go further to construct
mathematical laws, universal principles, and to 'abduct' a working
hypothesis and a theoretical explanation.
However, MPA is different from instrumentalism, for it does not regard
scientific theory only as a useful means, but argues that theory can exhibit
a scientific picture of the world through its operational framework .
From scientific theory's axioms, deductive - hypothesis system, and other
theoretical components we can infer some universal conclusions. Based
on these conclusions people can explain and comprehend matter's struc-
ture and motion, phenomena's cause and result. In other words , people
can understand what the world is and why it so happens, relying on
abductive inference within the operational framework. In short, scientific
theory is a cognizing apparatus of mathematicized practice: it constructs
mathematical models and an operating program by which it would reflect
a picture about nature. So MPA not only treats scientific theory as a
kind of means which is a practical medium but also treats it as an
ideological mediation between subject and object.
From this opinion of MPA, I maintain that theory is the ' eye ' of human
cognition, that we can realize matters underlying the natural phenomena
only by this ' eye ' . This is by means of scientific theory and the human
eye in order to reflect the abstractly rational image and the perceptual
image which human being can 'see' in the world . The rational image
of the world goes far beyond the commonsense image , for it is a kind
of abstract structure which often renews and replaces, and it not only
comes from abductive inference and conjecture based on scientific laws
and theories but also comes from subjective impressions based on the
explanations of such theories.
Of course, this scientific image is not as stable as a sensory image
which depends on human sensory experience, while also not as trust-
worthy for it is the result of mental activity as to what originates from
424 ZHOU JI HONG
body-activity. However, what can we know about this world except by

our sense organs? Scientists rely on a large number of tests to confirm
their hypotheses and to examine the efficiency of the hypotheses, and
thereby the abductive or inductive inferences drawn from those theories
would be acceptable. Considering that scientific theory and language
always go along within its operational framework, so scientific terms
would correspond with specific denotations, and in the terms of practice,
the well-confirmed theory would be ' acceptable truth' within the oper-
ational framework. Here, the acceptable truth is neither homely truth
nor pragmatic criterion; it takes scientific practice as its standard and also
relies on its theoretical framework . According to MPA, this kind of
truth would be nonsense if beyond the theoretical operational framework,
and the framework itself transcends commonsense and the problem about
truth. Meanwhile, the acceptable truth does not use an unobservable, tran-
scendental criterion as approximate truth does .
In addition, when some correlated theories and experiments progress,
the angle, scale and degree of cognizance of the previous operational
framework will change correspondently. MPA argues that scientific
revolution is not only a change of mind-model and the conclusion of law;
it would accompany change in theoretical statement, particularly the shift
of the operational framework - including space-time framework, dimen-
sional system, and operational-model. As MPA sees it, from Newtonian
mechanics to Einstein's theory of relativity, what had changed was indi-
cated not only by the shift of theoretical conclusions, but more important,
was a change from a three-dimensional operational space-framework
to a four-dimensional space-time-framework. In terms of history of
science, here the narrowly operational framework would be broken and
an obscure hypothesis would be modified, for more elaborate mathe-
maticized practice after a long-term scientific development. Although
during a special period the result of mathematicized practice could not
be regarded as absolute truth, we can tentatively accept the image under
theory's operational framework as reliable.
III
By now, it is apparent that the distinctive feature of MPA, which sepa-

rates it from NOA and other similar views, is MPA refusal to analyse
scientific theory (or terms) from the angle of a linguistic or a semantic
view. Rather, MPA recognizes science at the point of practice and offers
a new idea about truth.
Inside scientific practice, that is, at the operating level, we cannot
regard scientific theory as a literally true story about the world and
should realize that it is just constructed by scientists, grounding it on
basic mathematicized elements which correspond with certain operations.
As a subjective structure, it has no necessary relation with the natural
world; realists cannot provide full reasons in favor of themselves, so anti-
realism is more applicable. On the other hand, outside scientific practice,
that is, at an external, onlooking level, scientific theory is the only way
with which we can aim to understand the natural world, and it can be
treated as 'acceptable truth' when its results are well-confirmed and its
mathematical model and operational framework are empirically adequate.
MPA does not oppose Tarski's definition of truth, but it believes that,
as an anthropocentric practice, science can acquire a kind of accept-
able truth. Based on this notion of truth, scientific theory can be true and,
then, realism's attitude is unavoidable.
So realism and antirealism are incommensurable and complemen-
tary at different levels in MPA. In fact, we can find that realism and
antirealism cannot always offer abundant reasons in defence of their
opinions, and their strong attitudes have obvious shortcomings which are
difficult to accept; in fact NOA had analysed them clearly. Here I suggest
MPA to go a step further. Certainly, MPA cannot completely solve the
divergences between realism and antirealism, but it helps to provide a
unified thought for us to clear up some confusion in the realist and
antirealist controversy.
Center for Science and Society,

Beijing University,
China.
NOTES
I See Fine, Arthur (1984). 'The Natural Ontological Altitude ', in Jarrett Leplin (ed.),
Scientific Realism, University of California Press, California , pp. 97-101.
2 Dilworth , Craid (1986). Scientific Progress, D. Reidel Publishing Company, Dordrecht,
p. 77.
3 Kuhn, T. S. (1962). The Structure ofScientific Revolutions . University of Chicago Press,
Chicago, p. 102.
426 ZHOU JI HONG
4 Ibid., p. 147.
5 Fraassen, Bas van (1989). Laws and Symmetry, Oxford, Clarendon Press, p. 22.
6 Cf. Cohen, R. S. and Wartosky, M. W. (eds) (1984). 'Physical Sciences and History
of Physics', Boston Studies in the Philosophy of Science 82, p. 148.
7 Westfall, Richard S. The Construction ofModern Science. Cambridge University Press,
London, p. 143.
USKALI MAKI
SCIENTIFIC REALISM AND

SOME PECULIARITIES OF ECONOMICS*
I. INTRODUCTION
I suggest that the issue of realism about science should be contextual-

ized in terms of the peculiarities of particular disciplines and kinds of
theories. Instead of any absolute and universal assertions for or against
scientific realism we end up with a sort of relativization of realism.
This amounts to a defence of concrete and local as against abstract and
global philosophy of science. This suggestion is supported by using the
case of economics as evidence.
Without pretending that there obtains a single standard formulation
of scientific realism, let us nevertheless list a few general statements char-
acteristically attached to this doctrine: Scientific theories depict a realm
of entities which (i) lie beyond the boundaries of the ordinary realm of
commonsense entities and (ii) exist independently of the human mind .
Ideally, theories succeed in (iii) referring to those entities and (iv)
describing them correctly, approximately at least. Finally, (v) the facts
stated in (iii) and (iv) explain the uncontroversial technological success
of science. According to the inference to the best explanation, this is con-
strued as an argument for scientific realism. Let (i) to (v) con stitute
one standard statement of scientific realism.
It will be shown that the ideas (i) to (v) become problematic
if ' science ' is replaced by 'economics' and ' scientific theories' by
' economic theories' . It is argued that these ideas, created with the char-
acteristics of physics in mind, are at odds with what has to be required
of a scientific realism capable of accommodating economics. Several
themes of relevance emerge.
First, and perhaps least harmfully, there is the non-philosophical usage
of 'realism' amongst economists. They use the term to designate a set
of properties attributable to economic theories and their constituent
parts in phrases such as "the realism of the assumption that business firms
maximize their profits". Here, ' realism' does not designate a philo-
sophical doctrine. In many cases, this usage of 'realism' is not even
427
428 USKALI MAKI
conceptually related to realism understood as a family of philosophical

doctrines.
Secondly, the fundamental thesis of ontological realism needs a careful
adjustment if it is to apply in the case of economics. The most abstract
formulation of ontological realism is the thesis that ' X exists '. Different
more specific versions can be gained by specifying both 'X' and 'exists' .
The concepts of mind-independent existence and mind-external existence,
characteristic of many formulations of scientific realism, are not suffi -
cient to accommodate the ontology of economic theories, since these
theories typically invoke mental and social entities.
Third, whereas in physical sciences the issue of realism has typi-
cally been formulated in terms of the existence of nonobservational
postulated theoretical entities (electrons, photons, electromagnetic fields,
etc.), economics, for the most part, deals with observational or com-
monsense entities (business firms, households, goods, money, prices, and
folk psychological entities such as aims and beliefs) . Economic theories
accomplish various rival modifications and rearrangements among such
"folk" entities and their attributed properties. Relative to the situation
in the physical sciences, this implies two things . First , ontological and
referential realism are not or should not be major issues regarding the
basic entities of economics. Second, some sort of commonsense realism
becomes particularly relevant in economics.
Fourth, while popular arguments for scientific realism in the case of
physical sciences involves appeals to technological success and experi-
mental manipulability, it is very problematic to use these arguments in
the case of economics. Questions of existence and truth in economics are
not typically settled by invoking manipulab ility or success . In most cases,
the existence of economic entities seems to be pre-experimentally unprob-
lematic. Also, economic theories do not seem to be the sort of things
of which we could relatively unambiguously say that they have or have
not had technological success. No wonder then that judgements based
on intuitive plausibility or compatibility with commonsense experience
play such a prominent role in economics.
2 . REALISM AND REALISTICNESS
The first characteristic of economics to point out is one of terminolog-

ical usage . Indeed, the common usage of 'realism' itself in economics
does not fit with the conceptual substance of scientific realism. On the
SCIENTIFIC REALISM 429
one hand, this is a trivial point, but on the other, it is significant since
only by making it are we able to see the crucial issues clearly.
Economists talk about realism on a daily basis. They argue about
"the realism" of this or that theory or model or assumption, such as those
of profit maximization, stable preferences and technology, full infor-
mation on the part of agents, perfect competition, full employment
equilibrium, benevolent state. Some economists - such as Herbert Simon
- hold the view that the "realism" of such assumptions is a major issue
and that more "realism" is better than less. Some others - such as Milton
Friedman - think that the "realism" of assumptions does not matter a
bit if only they help construe theories or models that provide good pre-
dictions or possess other pragmatic virtues.
Philosophers have a long tradition of talking about realism. They
discuss the problem of the existence of universals, of natural necessity,
of the mental, of subatomic particles, of moral values . They also argue
about how language relate s to the world, of reference and truth. Those
who espouse realism may hold the view that either properties, moral
values, the theoretical entities of physics or all of them may really exist ;
and that language can be used to refer to such realities and to formu-
late true statements about them. The anti-realists reject such views.
Most of the time, these two discourses, by economists and by philoso-
phers, do not overlap. Indeed , when economists talk about the "realism"
of their theories and their constituents, they do not talk about realism
in the philosophers' sense at all. Using the term economists charac-
terize various properties of economic representations. But realism as a
philosophical doctrine is not a property of economic theories and their
constituent parts. Thus, using the term economists must mean some-
thing else . Perhaps due to the poverty of the Engl ish language, there
has not been available a separate expression for this something. My
suggestion is that when using the term 'realism ' economists are talking
about what may be called realisticness. Realisticness and unrealistic-
ness are properties attributable to economic representations. Given this
stipulation, ' realism' can be reserved for denoting a philosophical
doctrine or many such doctrines (Maki, 1989).1
The introduction of the distinction between realism and realisticness
has two significant consequences. First, it facilitates the recognition
that neither of these categories is unambiguous, that there are a number
of relevant kinds of both. Second, it helps us understand that within
economics there are widely discussed kinds of realisticness (and unre-
430 USKALI MAKI
alisticness) which are not conceptually connected to realism at all (such

as testability, comprehensiveness, plausibility, practical relevance). There
are others, however, which are so connected, such as referentiality
("referential realisticness") and truth ("veristic realisticness"), suitably
understood. This means that, as against a misguided habit, the dispute
between those who favor more "realism" in economic theory and those
who are against it cannot always be construed as a simple controversy
between the realists and the instrumentalists. For instance, the dis-
agreement between Simon and Friedman does not seem to be a standard
case of a realist quarrelling with an instrumentalist. In what follows
the focus will be on those sorts of realisticness that are conceptually con-
nected to the realism issue.
3. ONTOLOGICAL AND SEMANTIC REALISM
The ontological issue of existence is the place to begin when dealing with
the questions of realism. The most general definition of ontological
realism goes as follows:
Def 1. Ontological realism is the thesis that X exists or that Xs
exist.
This formulation permits a vast number of alternative specific versions
of ontological realism. Here X is a variable which has to be given
appropriate values. 'Exist' is also in need of specification. Its presumed
extension and intension can be specified separately. I will make a sug-
gestion to this effect in the next section.
The next natural step in our clarification of realism is to face
the semantic issue of truth and its prerequisites, reference and repre-
sentation:
Def 2. Semantic realism is the thesis that (not any, but many) lin-
guistic expressions refer to entities that exist (referential
realism) and represent their features (representational realism)
and are either true or false about them partly by virtue of what
they are like , i.e., by virtue of the way the world is (veristic
realism).
The respective semantic forms of (un)realisticness are the ones that
are connected to semantic realism. They are referential, representa-
tional and veristic (un)realisticness, attributable to some linguistic
expressions. A referentially (un)realistic expres sion is one which refers

(does not refer) to entities that exist. A representationally (un)realistic
expression is one which represents (does not represent) features that
existing referents possess. A veristically (un)realistic expression is one
which is true (false) about its referents partly by virtue of what they
are like.
There seems to be an interesting difference between economics and
physics regarding their respective focus on the various aspects of the
issue of realism . In physics, existence and reference constitute a major
issue, whereas in the context of economics, existence and reference would
appear to be relatively unproblematic and, where the issue of semantic
realisticness is posed at all, the emphasis seems to be on truth. While
the existence of quarks and black holes is construed as problematic in
the philosophy of physics, the existence of business firms and house-
holds, preferences, goods and their prices, appears as more or less
unproblematic. The issue in economics is rather geared to, for instance,
the representation of the actors as having rational expectations and tran-
sitive preferences and taking the single price of a homogenous good as
given. In Section 5 below I attempt to give reasons for this difference
of emphasis between physics and economics. Before that, we have to
address the ambiguity of 'existence' .
4. CONCEPTS OF EXISTENCE
It is customary to define the ontological side of realism as "the thesis

that there is a domain of mind-independent existence" (Rescher, 1987:
p. 1). It is then natural to suggest that knowledge of this domain is
acquired by natural science. The objects of astronomy, physics and chem-
istry such as galaxies, positrons and molecules, might indeed exist
independently of the human mind. A realism qualified for accommodating
economics cannot build upon such a concept of existence. Economics
clearly deals with objects which do not exist independently of the human
mind; it even deals with the contents of the human mind itself. A realism
about economics has to employ a different concept of existence. '
I defined ontological realism as the thesis that X exists or that Xs exist.
The extension of the predicate 'exist(s)' consists of the set of objects
to which the predicate can be truthfully applied. Partial specifications
of the extension of 'exisus)' are partial specifications of the set of appro-
priate values of X. The set of candidate members in that extension include
432 USKALI MAKI
sorts of objects such as universals and particulars, essences and appear-

ances, material and mental entities, observables and unobservables, all
of which are relevant for studying social and economic theories from a
realist point of view. The issue of ontological realism in economics and
other social sciences involves the question of whether there are objects
of those sorts in the social world.
Partial specifications of the intension or meaning of 'exists ' are
provided in terms of three specifications in the order of descending
strength (Miiki, 1990b: p. 294).
Def 3. Independent existence. X exists independently iff it exists
independently of the human mind. (Only material entities may
exist independently.)
This concept of existence leads to a materialist or physicalist realism,
since it leads to the exclusion from the realm of real existents things such
as mental states and processes (e.g., preferences and expectations) and
anything the existence of which is dependent on the mental (e.g., social
institutions such as money and the market) . Only material entities (such
as neurophysiological entities as well as photons, plants, and planets)
may exist independently. Rescher 's version of realism subscribes to this
concept.
Def. 4. External existence. X exists externally iff it exists externally
to or independently of any individual human mind. (Both
material and social entities may exist externally.)
Again , this (more or less Durkheimian) notion excludes mental entities
from the set of existing items. Contrary to what is implied by ontolog-
ical realism employing the concept of independent existence, here social
entities gain a membership in the set of real existents. Social entities
can variously be characterized in terms of social rules, collective patterns
of behavior, unintended consequences of individual action mediated
through a social mechanism , or the like. The set of such entities relevant
to various streams of economics comprises the market, money, debt,
property rights, classes, corporations, etc. Social entities in this sense
depend on individual human minds for their existence (in this they are
unlike material entities), but exist externally to them (just as material
entities do). I here ignore attempts to do without irreducible social entities
and thereby without the concept of external existence. Whether or not
those attempts are deemed successful does not affect my main argument,
which hinges upon the concept of objective existence and the centrality
of mental entities in much if economics.
Def. 5. Objective existence. X exists objectively relative to a given
representation iff it exists unconstituted by that particular
representation. (Both material, social, and mental entities may
exist objectively.)
In accordance with this specification, something exists objectively only
relative to particular representations of that something. Objective exis-
tence relative to representation R is denied those entities whose existence
depends on R or is constituted by R. Along with material entities (such
as the molecular structure of gold) and social entities (e.g ., money, the
market mechanism), also mental entities (such as preferences, purposes,
and expectations) may exist objectively. This is because mental entities
represented from a third-person point of view are not constituted by
such representations even though they may be constituted by other (first-
person) representations. The preferences of a consumer are not brought
to existence by being represented by an economic theoretician.
It used to be a popular view in the nineteenth century, from Nassau
Senior and J. S. Mill onwards, to regard economic theory as based on
two kinds of premises, material and mental, such as the law of dimin-
ishing returns and the desire for wealth respectively. The objects of
material premises may exist independently, but the referents of the mental
premises may exist objectively at most.
This means that a realism about economics and a realism about physics
do not have a single shared concept of existence. A realism about
economics needs the notion of objective existence and finds that of mind-
independent existence too restrictive. A realism about physics is able
to combine the two .
5 . SCIENTIFIC THEORIES AND FOLK VIEWS
In commonsense experience, we confront our physical surroundings as

being composed of hard rocks and green trees and drinkable water and
other such visible and tangible macroscopic objects. The science of
physics hypothesizes an entirely different realm of objects, composed
of configurations of subatomic particles, gravitational and electromag-
netic fields, black holes, curved space-time. This theoretical realm of
scientific objects radically transcends the boundaries of ordinary com-
434 USKALI MAKI
monsense experience. The issue of scientific realism about physics is one

of the reality of the scientific realm and the truth of statements about
it.
In economics, the situation seems to be drastically different. The
constituents of the worlds hypothesized in economic theories appear to
be more or less the same as those inhabiting the realm of our ordinary
conceptualized experience. No radical gap seems to prevail between
the scientific realm and the ordinary realm. Economic theories speak
about objects that are confronted in our ordinary experience about
economic matters - business firms and households, their aims and expec-
tations, money and goods and their prices , land and labor and capital,
wages and profits and taxes. Therefore, the existence of the objects of
the scientific realm should not be a major issue in economics. The
referential realisticness of the fundamental elements of economic theories
is more often than not beyond doubt: since the terms of economic theories
seem to refer to entities with which economists and others are familiar
on the basis of ordinary experience, the referential status of these terms
should not be taken as suspect.
The relevant issue lies elsewhere, namely in how the ordinary objects
are represented in economic theories. Even if the referents of ordinary
experience and economic theories are pretty much the same, the
ways in which they are represented often dramatically diverge. There
are two main ways in which the theoretical representations seem to
dev iate from commonsense representations. I call them modification
and rearrangement.
To see the point more clearly, let me postulate the existence of three
simplified and hypothetical levels of thought, characteristic of an imag-
inarily pure Western culture. First, there are general folk views of "man"
and society at large. They comprise folk psychology, among other
elements.' These folk views are widely and flexibly applicable. Second,
there is what may be called folk economics. It is a modification and appli -
cation of the general folk views to economic matters. Third, there is
scientific economics. Scientific economics accomplishes a modification
and rearrangement of the ordinary objects of folk economics." We can
now imagine a quasi-historical tale which delivers a message of the
systematic relationships between the three frameworks.
The general folk views depict human life and social phenomena in
terms of men and women, families and nations, castes and classes, desires
and aims, hopes and fears, values and norms, beliefs and expectations,
deliberation and routine, producing and consuming, working and eating,

tools and what you make with them, love and hate, duty and honor,
kinship and partnership, gifts and saleables, right and wrong , freedom
and subordination, etc. Elements such as these constitute the "on tic fur-
niture" of the general folk views of "man" and society . Folk economics
modifies this furniture by eliminating some of the elements and focusing
on some others. For instance, gender is excluded in favor of "man",
the generalized economic actor. Emotions are excluded in favor of
rational deliberation. Making love is excluded in favor of making money.
Gifts are excluded in favor of saleable goods. Kinship is excluded in
favor of partnership. And so on.
Scientific economics goes further in modifying the elements of the
ordinary realm. A fairly trivial aspect of this modification is the refor-
mulation of the linguistic expressions used for the elements in the
ordinary realm in terms of technical vocabulary (such as characterizing
buying decisions in terms of 'indifferent curves' and 'budget constraints';
and expres sing the idea that a change in the price of a good has a variable
impact on how much more or less people then wish to buy it as 'the price
elasticity of demand') and in terms of formal languages (such as replacing
'price' by 'p'; and ' price elasticity of demand ' by 'dX;dp;'). Modifications
of substance include selection (such as including case -by-case means-
ends deliberation and excluding routines in the picture of economic
agents; and focusing on trade between two countries and forgetting about
the rest of the world) ; abstraction (such as talking about labor power
in general instead of you and me; and about the representative firm
instead of particular business firms) ; idealization (such as the assump -
tions of perfect information; zero transaction costs; perfect substitutability
between goods and factors of production; and firms as price takers
rather than price makers); exaggeration (such as the assumption of
maximization; and that of transitive preferences); projection (such as
the notion of expected permanent income, equal to the actual measured
income of an individual minus a transitory part); aggregation (such as
summing up unemployed people to yield the rate of unemployment);
and their various combinations.
There are a number of ways to perform these operations in the con-
struction of economic theories, but none of them accomplishes a major
departure from the ontic furniture of the ordinary realm. No new kinds
of entities or properties are introduced. Only the ordinary entities and
properties are modified (see Maki, 1992a).
436 USKALI MAKI
Take the example of the concept of business firm in standard neo-

classical general equilibrium theory. We may take it to refer to real
business firms even though it represents these firms by attribut ing to them
properties that they are not ordinarily believed to possess: they are all
alike to relevant dimensions; they are devoid of internal structure, they
are functionally uniform; they are cognitively omniscient, they are fully
informed about the market conditions; they are impotent regarding price
setting, they take prices as given in the market; they maximize their
profits by using refined marginalist calculations; they are not engaged
in active entrepreneurial rivalry but rather react passively to changes
in cost and demand conditions. Some commentators are led by such attri-
butions to the conclusion that this version of neoclassical theory does not
refer to real business firms. My point is that folk economics and neo-
classical economics have real business firms as their shared referent even
though they represent these firms differently.
Or take the folk psychological notion of preference. The preferences
of consumers are represented by the axioms of standard neoclassical
theory as complete (the consumer has a preference order over any two
bundles of goods), reflexive (any bundle is at least as good as itself), and
transitive (if the consumer prefers bundle x to bundle y and y to z, then
she prefers x to z). It would be a mistake to conclude that if consumers
do not have preferences with these characteristics, they do not have
preferences at all. There may be other reasons to doubt the reality of pref-
erences (along with the rest of the folk psychological realm), but, say,
the intransitivity of preferences should not be one such reason. The
axioms of consumer theory may refer to real entities irrespective of
how these entities are represented.
I am suggesting that, ontologically speaking, a major part of economics
is a modified form of economic folk views . The implication is that the
scientific status of modified folk economics cannot be articulated simi-
larly to that of theoretical physics.While in physics the ontic furniture
of the theoretical realm represents a radical departure from that of the
ordinary realm , in economics it is a modification of the furniture of the
ordinary realm .' As regards reali sm, the implication is that, concerning
the individual referents of theories, ontological scientific realism may
fit physics, while economics seems to require some sort of ontological
commonsense realism. At this level, no major abyss is opened between
folk economic and scientific economics.
I suggest that there is a more significant difference though . The key

to understanding this difference between folk economics and scientific
economics is connected not so much to the modifications of the folk
objects, but rather to their arrangement. This is where entire theories
or theory-like systems of ideas enter the picture. I mention such semi-
theoretical formations, since it is not clear that folk economics is a theory
in any genuine sense (cf. the situation regarding folk psychology). In any
case, while the modifications of the folk objects are accomplished by
the individual concepts and statements of a theory or a theory-like for-
mation, their arrangement is a matter of a theory as a whole .
To begin with, we can postulate a form of economic thought that
can be treated as having some of the typical characteristics of folk
views in other fields as well. It is based on the idea that the functioning
of national and global economies can be considered analogically to the
functioning of household or business management. To introduce another
analogy, while folk psychology is akin to Aristotelean psychology and
folk physics is similar to Aristolelean physics, folk economics could
be considered as akin to Aristotelean oikonomike. Recall that oikonomike
was understood as the study of household management. Folk economics
involves essentially the idea that the economy is managed, that is, that
economic phenomena are typically intentionally brought about by people
in charge: it is the government in power or the big business which is
taken as respon sible for what takes place in the economy. Mercantilist
and cameralist streams of economic thought may be taken to exemplify
this belief.
It is only when questioning this belief that we enter the realm of
scientific economics as understood since Adam Smith or so. Notions such
as "the invisible hand" and the "spontaneous order" or the "character
masks" and "the natural law of motion" of the capitali st economy exem-
plify this questioning. This idea is brought out by Ullman-Margalit (1978,
pp. 268-270) who points out that an invisible-hand explanation of social
phenomena is not the most "natural" or "plausible" account, because it
is "an explanation in terms of intentional design [which] will naturally
suggest itself". This may be reformulated by saying that an intentional-
design explanation is a kind of commonsense account for social patterns
and institutions. Let us formulate the two perspectives as follows ([FE]
stands for ' folk economics', [SE] for 'scientific economics') (see Maki ,
1990a: p. 337) .
438 USKALI MAKI
[FE] Economic phenomena and institutions are results of inten-

tional design by human individuals and their groups.
[SE] Economic phenomena and institutions are unintended con-
sequences of human actions mediated by a social mechanism.
We may say that [FE] and [SE] are two principles of social constitu-
tion . In any economic doctrine, one of them typically predominates. I am
suggesting that in moving from the predominance of [FE] to the pre-
dominance of [SE] no mere modification of folk objects but their major
rearrangement takes place. The ontic furniture of folk economics remains
pretty much the same, but the elements have been reorganized. This
means that the nature of the causal relation between human action and
economic phenomena is depicted differently in the two views. In folk
economics, the outcomes are intended results, while in scientific eco-
nomics, the outcomes are unintended consequences.
There are a number of alternative theoretical specifications of such
rearrangement, ranging from Adam Smith's depiction of the market as
an arena of relatively harmonious coordination of self-interested
economic activities through the invisible hand down to Karl Marx's
presumed picture of the downfall of capitalism due to the tendency
of the rate of profit to fall , unintended by any single capitalist. Such
alternative further arrangements within [SE] involve rival isolations
of causally relevant objects and features from those that are sup-
posed to be causally less potent or relevant, using idealizations of
various kinds, including the ceteris paribus clau se. Any given re-
arrangement presupposes certain modifications concerning the elements.
Different rearrangements presuppose different modifications of the
elements. For example, if the modifications do not serve to exclude entre-
preneurship, ignorance and learning on the part of economic actors, the
resulting rearrangement is different from one based on the assumption
of perfect information and the exclusion of entrepreneurship (see Maki,
1992b) .
It is now easy to see why the issue of realism in economics is centered
on representation and truth, while existence and reference should be
regarded as relatively unproblematic. There is a potential existential
and referential issue , though, related to the social mechanism posited
in [SE] . But the issue is genuine only under non -individualist construals
of the social mechanism. For instance, we may then ask whether the
metaphor of the invisible hand refers to an existing mechanism which
is not a folk entity. Even in this case, however, the issue can be dealt
with as that of the truth of the metaphor.
For the most part, the issue is that of the truth (or otherwise benefi-
cence) of the many representations involved in scientific economics. The
modifications of the folk objects typically involve statements that
are obviously false (such as the assumptions of perfect information
or transitive preferences or zero tran saction costs). Their falsehood,
however, may at least sometimes serve the purpose of pursuing true
specifications of rearrangement [SE]. Since there are alternative sets of
modifications and rival specific rearrangements, the issue of true rep-
resentation plays a significant role, at least potentially (see Maki, 1994).
It does so on two fronts, as it were. The modification of individual folk
objects involves the issue of semantic commonsense realism, i.e., the
issue of whether the representations of the folk objects are usefully true
or false. The rearrangement of those objects in the form of entire theories
involves the issue of semantic scientific realism, Le., the issue of whether
this or that suggested rearrangement succeeds in truthfully representing
the detailed mechanism of social constitution.
6. EXPERIMENTAL MANIPULABILITY AND

TECHNOLOGICAL SUCCESS
Let us next consider two arguments which relate the correctness of

scientific realism to the presumed technological success of science . The
first is the argument from experimental manipulability, found in the
work of Ian Hacking and others . The second is the popular no-miracle
argument for realism about science. Both arguments are problematic in
the context of the physical sciences themselves, but I will ignore this
here. I just point out why they are particularly problematic in the case
of economics.
Hacking 's experimental realism is a realism about the ontology of exis-
tence. He argues that we can become convinced about the correctness
of realism by referring to certain experimental practices in science. In
particular, we can be convinced of the existence of subatomic particles
if we are capable of manipulating them. Experimental manipulability
of an entity is, for Hacking, an indication of the existence of that entity.
The example of electrons summarizes the idea : "We are completely
convinced of the reality of electrons when we regularly set out to build
- and often enough succeed in building - new kinds of device that use
440 USKALI MAKI
various well-understood causal properties of electrons to interfere in other

more hypothetical parts of nature ." (Hacking, 1983: p. 265).
Again, economics looks different. It seems obvious that no experi-
mental manipulability by economists qua economists is required for them
or others to become convinced of the existence of business firms or prices
or price expectations and other folk entities. Economists have other
sources for most of their existential beliefs. The argument from exper-
imental manipulability does not seem to be needed for an ontological
realism about economics.
Let us then move on to the second idea of realism which builds upon
the idea of technological success. A popular argument for scientific
realism is the no-miracle argument. This argument takes it as an uncon-
troversial empirical fact about science that scientific theories are
successful in the prediction and control of phenomena. Scientific realism
is then suggested as an explanation of this fact. It would be a miracle,
the argument goes, if scientific theories were not true about the really
existing entities and yet would be technologically successful. Scientific
realism is offered as a hypothesis which would make the success of
science an expla inable fact, a non-miracle.
Let us take a closer look at the argument, by using Richard Boyd's
formulation, as paraphrased by Hilary Putnam. Realism as an empir-
ical hypothesis is spelled out by Boyd in term s of two principles: (1)
"Terms in a mature science typically refer" (i.e., are referentially real-
istic). (2) "The laws of a theory belonging to a mature science are
typically approximately true " (i.e., veristically realistic) . Boyd then
suggests that "scientists act as they do because they believe (1) and (2)
and that their strategy works because (1) and (2) are true" (Putnam,
1975-76: p. 179). The notion of the "strategy" of scientists "working"
obviously means that scientists attain success in prediction and control.
Simply put, the non-miracle argument is this:
[RSS] (1) Terms in a mature science are typically referentially

realistic.
(2) The laws of a theory belonging to a mature science are
typically approximately true (i.e., veristically realistic).
(3) Science is successful in the prediction and control of

phenomena.
Of course, [RSS] (R for ' realism' and S's for 'success' and 'science')
is not a complete argument. For instance, premises concerning the beliefs,
aims and behavior of scientists are missing. The important point is
being delivered though. Realism, in the form of the two premises, is
provided as an explanation of the technological success of science. In
accordance to the inference to the best explanation, we are invited to
conclude that realism about science is true. Since (3) is true, in all like-
lihood (l) and (2) are true as well.
Applied to the case of economics, we get the following enthymemic
argument [RSE] (E standing for ' economics' ):
[RSE] (l') Terms in economic theories are typically referentially
realistic.
(2') The laws of an economic theory are typically approxi-
mately true (i.e., veristically realistic).
(3') Economics is successful in the prediction and control

of phenomena.
There are several problems with [RSE] and using it to argue for realism
in economics. First, [RSE] substitutes 'economics' for 'mature science',
which is problematic in itself. It is not clear what 'maturity ' signifies
here, but, quite clearly, economics does not count as a mature science
in the same way and degree as physics, notwithstanding its celebrated
status as the queen of the social sciences.
Secondly, it is doubtful if the explanandum of [RSE] is there at all,
waiting for being explained. In any case, it cannot be maintained that (3')
represents an undoubted fact (cf. Hands, 1991, p. 93) . It is no doubt
true that much of post -Keynes (i.e. post-1936) economics has pursued
technological success, often with great confidence. Policy interventions
have been designed based on quantified estimations of the effects of alter-
native policy measures using multiple regression and other econometric
techniques. Sometimes the se measures have been implemented in
economic policy, and sometimes they have appeared to work to some
degree of satisfaction. More often than not, economists' policy proposals
have not been implemented, or have been implemented only partially.
Often enough, the outcomes have not been those intended.
No wonder then that, whatever their reasons are, most people seem
to believe that economics has not achieved technological success of
442 USKALI MAKI
anything like the same kind and scale as physics (it is possible to take
this as one indication of the immaturity of economics). Among the ways
of grounding this belief are the following two. It may be maintained that,
as a historical fact, economics does not have a comparatively strong
predictive and technological record. In addition, or alternatively, it may
be argued that, due to the nature of its subject matter, economics is not
capable of attaining technological success which would even distantly
approximate that in physics. The tradition of Austrian economics and
a version of rational expectations macroeconomics exemplify this
argument. At the core of these versions of the argument are the ideas that
the subject matter of economics is peculiarly complex and that economics
is confronted with the problem of reflexivity."
The mere existence of such arguments indicates that the success of
economics is not an uncontroversial fact. The status of the explanandum
of [RSE] is suspect. Therefore, it is problematic to propose explaining
technological success by appealing to semantic realisticness.
Thirdly, the inference to the inverse direction is problematic, too. This
is because technological success does not seem to be needed to estab-
lish the semantic realisticness of economic theories (nor is technological
failure needed to infer to unrealisticness of theory). Regarding (I ') , econ-
omists are fairly convinced of the referential realisticness of their basic
terms and of the existence of their referents independently of whether the
theories containing them are technologically successful. This is because
those referents are folk objects, as argued above. Business firms and
households, their beliefs and aims, money, goods and their prices, etc .
are not theoretical entities in the same way as electrons and quarks are.
This also appears to support the old experiential tradition in econo-
mists' thinking about the character of their theories, going back to Nassau
Senior and J. S. Mill. In 1888, John Elliot Cairnes compared the situa-
tion in physics and economics in this way :
[Tlhe laws of gravitation and of motion [. . . ,1 the undulatory theory of light, the theory
of the molecular constitution of matter, the doctrine of vis inertia? - all alike elude direct
observation, and are only known to us through their physical effects. [. . .] The econo-
mis t starts with a knowledge of ult imate causes. He is already, at the outset of his
enterprise, in the position which the physicist only attains after ages of laborious research
(Cairnes, 1888 [1875]: pp. 84, 87).
Decades later, after major transformations in the substance of economic

theory, Lionel Robbins expressed the same idea: The postulates of eco-
nomics are not such that "the existence of [their] counterpart in reality
admits of extensive dispute once their nature is fully realised" (Robbins,

1935: p. 79) .
In Economics, [. .. J, the ultimate constituents of our fundamental generalisations are known
to us by immediate acquaintance. In the natural science s they are known only inferentially.
There is much less reason to doubt the counterpart in reality of the assumption of indi-
vidual preferences than that of the assumption of the electron." (Ibid., p. 105).
Thus, the justification of (I') would not seem to require technological

success. Regarding (2') , economists also seem to form their judgements
about the veristic realisticness or truth of many of the "law" statements
of their theories independently of technological success. It seems to be
rather their own everyday experience or direct statistical or other empir-
ical evidence that counts, not the controversial technological success.
Economists' beliefs about the truth or falsity of, say, the assumption of
profit maximization or the expected utility hypothesis or the " law" of
diminishing returns or the Jevonian "law" of indifference (there exists
only one price for a good in a market) or the "law" of the tendency
of the falling rate of profit, are not primarily based on technological
considerations.
There is a natural reason for this. Economists systematically practise
what may be called the method of theoretical isolation whereby a system
of factors is closed from influences from the rest of the universe, this
closure being accomplished by assumption. The theoretical method of
isolation involves an imitat ion of a limited portion of the logical struc -
ture of experimental research capable of material isolations (see Maki,
1992a). Besides the disanalogy between theoretical and material isola-
tion, there is a further limitation to the analogy: the feedback from
aposteriori experimental ev idence to theory is weak or missing in eco-
nomics. At least part of the reason for this is that all or most of possible
technological success in economics takes place in the conditions of
open systems. Therefore, such success hardly provides entirely reliable
evidence for or against economic theories which depict relations in
theoretically isolated closed systems. This turns the focus on the direct
plausibility of the basic statements of theories (see also Hausman, 1992).
In Boyd's argument, it is assumed that scientists believe in (2). In eco-
nomics, it often appears to be the case that economists do not believe
in (2') . Even more , many economists appear to believe in the falsehood
of many "law" statements in economics. On the other hand, many econ-
omists do believe that at least some such statements in their theories
444 USKALI MAKI
are close to the truth. Robbins provides a radical version of this belief:
"We do not need controlled experiments to establish their validity; they
are so much the stuff of our everyday experience that they have only
to be stated to be recognised as obvious." (Robbins, 1935: p. 79). The
important thing is that none of these beliefs is usually based on con-
siderations of technological success .
To conclude, an admission to the technological success argument can
be made. I have suggested that assessments of the semantic realistic-
ness of a given economic theory T are not typically based on the
perceived technological success of T. However, this does not preclude
the possibility that it may be based on pretheoretical, that is, pre- T
success. In a sense, folk economics is based on the practical experi-
ence, the successes and failures, of economic agents in their everyday
business operations, whether as "housewives" or "businessmen". If it
is the case that T shares with folk econom ics most of its referents, then
whatever technological success folk economics may have, serves as a
practical foundation for the ontological commitments of T. In this way,
pretheoretical success might play the role of grounding existential belief
within scientific economics. Yet, it is not technological success in a more
demanding sense and it is not the success of an economic theory itself
that is being used as a criterion of existence.
7. CONCLUSION
The above discussion suggests several interrelated points. Ad economics:

Economics is not on a par with standard examples of physical sciences
on a number of important dimensions . In other words, the suggested argu-
ments illuminate some of the peculiar characteristics of economics, and
remind philosophers of science of the fact that science is not uniform.
Scientific realism has served as the framework for bringing this point
home.
Ad realism: The paper suggests a relativization of scientific realism
by using some of the specific features of economics as supporting
evidence. It argues for a relativity of realism to kinds of scientific dis-
ciplines and theories. Scientific realism already appears in a number of
versions, but this multiplicity is not based on the recognition of the
multiplicity of types of disciplines and theories .
Ad economics and realism: Standard versions of scientific realism
are tailored to suit the characteristics of physics. If we accept the premise
that economics is not the kind of intellectual endeavour that would

abhor scientific realism tout court, economists have to reject the standard
versions. In order to check whether scientific realism is a viable descrip-
tive philosophy of economics, we need to develop a version which is able
to accommodate the peculiarities of economics without diverging from
some fundamental minimum tenets of realism. Whether this is possible
remains to be seen. Some ingredients for such a solution have been
provided above.
There is an alternative conclusion. It is to fix a version of scientific
realism and to deny any discipline not satisfying the tenets of that version
the status of science. For instance, the incompatibility of the character-
istics of economics with tenets (i) to (v) of our standard version could
be taken as an indication that economics is not a science. The unifor-
mity of science would then be guaranteed by dropping some disciplines
from the set of the sciences. The task would not be to adjust our for-
mulation of scientific realism but rather to develop an entirely different
economics. This option is also worth examination.
Academy of Finland.
NOTES
* Earlier versions have been presented at the Annual Conference of the History of
Economics Society, Fairfax, 30 May-2 June, 1992, and at the conference on Realism
and Anti-Realism in Science, Beijing, 15-17 June, 1992, and to audiences at the World
Institute for Development Economics Research of the United Nations University and at
the Hong Kong Institute of Economic Science . Thanks to the four audiences for lively dis-
cussions. Special thanks for helpful comments go to Siddiq Osmani, Nancy Wulwick,
Victor Mok and, in particular, Alex Viskovatoff.
I The inaccurate usage of ' realism' (and even 'reality') is not restricted to the eco-
nomics profession. We can find a physicist talking about "the realism of physical theories"
(Franklin , 1984, p. 381) and " the reality of theories" (ibid., p. 389), and a philosopher
discussing the " reality of assumptions" in economics (Musgrave, 1981). Yet, the scale
of such a usage seems to be of a completely different order in the case of economics.
2 It turns out that economist Tony Lawson's suggestion for a definition of ' realism' is
not quite apt for economics: " there is a material and social world that exists indepen-
dently of any individual consciousness . .." (Lawson, 1989, p. 61).
J Alexander Rosenberg has emphatically made the point that economics and other social
sciences are heavily dependent on folk psychological notions (e.g., Rosenberg, 1981).
4 This perspective on the matter is not devoid of problems . One of them is the very notion
of folk economics. We do not have a very accurate idea of the contents of economic
folk views ; we are much more enlightened about the contents of folk psychology and
446 USKALI MAKI
folk physics, for example . Some research has been done in the are of a so-called naive
economics, and it should be of some help here. Naive economics, however, is mostly
concerned with children 's views of econom ic mailers. The problem is that a great number
of grown-ups have some education in "scientific" economics and use it in their everyday
conceptualization of economic phenomena . Thus, the dividing line between scientific eco-
nomics and actually held economic folk views would not seem to be entirely sharp. Folk
economics has to be taken as a simplification itself. Note also that I am using 'scien-
tific economics ' as a descriptive, non-evaluative term.
5 This suggestion implies that those economists in the 1960's were wrong who sug-
gested that 'firm ' and other such basic terms be treated as ("partially interpreted")
theoretical terms akin to 'e lectron' (e.g., Ysander, 1961; Puu, 1967; Fabian, 1967).
6 The possible arguments are not exhausted by these. Yet another possibility is to argue
as Rosenberg does, namely that in the background of the predictive and technological
impotency of economics lies its dependence on folk psychological categorial frame -
work: the intentional vocabulary of folk psychology fails to carve the nature into the joints,
i.e., into natural kinds or causally homogenous classes . He suggests this as an explana -
tion for the predictive and technological failures of economics which he takes as an
uncontroversial fact (Rosenberg, 1981).
7 Note that if Hacking is right with his argument from the manipulability of electrons,
then Robbins's statement would have to be qualified .
REFERENCES
Cairnes, John Eliot (1888 [1975]). The Chara cter and Logical Method of Polit ical
Economy, Macmillan, London.
Fabian , Robert (1967) . 'An Empirical Principle for Deductive Theory in Economics ',
Southern Economic Journal 34, pp. 53-66.
Franklin, Allan (1984) . 'The Epistemology of Experiment ' , British Journal for the
Hacking, Ian (1983). Representing and Intervening , Cambridge University Press ,
Cambridge.
Hands, D. Wade (1991). 'Reply to Hamminga and Maki', in Neil de Marchi and Mark
Blaug (eds.), Appraising Economic Theories. Studies in the Methodology of Research
Programs, Edward Elgar, Aldershot.
Hausman , Daniel (1992). The Inexact and Separate Science of Economics, Cambridge
University Press, Cambridge.
Lawson, Tony (1989). ' Abstraction, Tendencies and Stylized Facts: A Realist Approach
to Economic Analysis', Cambridge Journal of Economics 13, pp. 59-78.
Maki, Uskali (1989). ' On the Problem of Realism in Economics ' , Ricerche Economiche
43 , pp. 176-198. [Repr inted in The Philosophy and Methodology of Economic s,
edited by Bruce Caldwell. Edward Elgar 1992.)
Maki, Uskali (I 990a). 'Sc ientific Realism and Austrian Explanation ', Review of Political
Economy 2, pp. 310-344.
Maki, Uskali (I 990b). ' Mengerian Economics in Realist Perspective', History of Political
Economy, Annual Supplement to Vol 22, pp. 289-310.
Maki, Uskali (l992a). 'On the Method of Isolation in Economics ', Poznan Studies in
the Philosophy of the Sciences and the Humanities 25, pp. 319-354.
Maki, Uskali (l992b). 'The Market as an Isolated Causal Process: A Metaphysical Ground
for Realism ', in Bruce Caldwell and Stephan Boehm (eds .), Austrian Economics :
Tensions and New Developments , Kluwer, Dordrecht.
Maki, Uskali (1994). 'Isolation, Idealization and Truth in Economics', Poznan Studies
in the Philosophy of the Sciences and the Humanities 38, pp. 147-168.
Musgrave, Alan (1981) . '''Unreal Assumption" in Economic Theory : The F-twist
Untwisted', Kyklos 34, pp. 377-387.
Putnam, Hilary (1975-76). ' What is "Realism"?' Proceedings of the Aristotelian Society,
pp.I77-194.
Puu, Tonu (1967). ' Some Reflections on the Relation between Economic Theory and
Empirical Reality', Swedish Journal of Economics, pp. 85-114.
Rescher, Nicholas (1987). Scientific Realism, Reidel, Dordrecht.
Robbins , Lionel (1935). An Essay on the Nature and Significance of Economics,
Macmillan, London.
Rosenberg, Alexander (1981). Sociobiology and the Preemption of Social Science,
Blackwell, Oxford.
Ysander, Bengt-Christer (1961). ' Truth and Meaning of Economic Postulates' , in Hugo
Hegeland (ed.), Money, Growth and Methodology and Other Essays in Economics,
Lund.
TIMOTHY SHANAHAN
REALISM AND ANTIREALISM IN

EVOLUTIONARY BIOLOGY
(I) INTRODUCTION
(a) The Realism-Antirealism Debate
Do scientific theories provide us with genuine insight into the causal

structure of the world, or do they merely provide useful models for
organizing observable phenomena into coherent patterns? Scientific
realists mainta in that scientific theories do, at least approximately and
on some occasions, give us insight into the causal structure of the
world, including its unobserved and unobservable parts. Antirealists
claim that scientific theories serve a number of useful functions ,
such as organizing empirical data and facilitating predictions, but
revealing the hidden structure of the world is not one of them. At issue
here is nothing less than the nature of our scientific understanding of
the world.
(b) The Neglect of Biology
The reali sm-antirealism issue has been discussed extensively in the

context of the physical sciences, with quantum mechanics in particular
receiving a lion's share of the attention. One might reasonably expect
that philo sophers, in order to gain a wider perspective, would explore
this issue in the context of other sciences as well. But to a surprising
extent this has not happened. Biology , in particular, has received scant
attention from philosophers interested in the realism-antirealism issue .
This has had two unfortunate effects. First, arguments with regard to
the justification of realist vs. anti realist philosophies of science have
generally ignored the contribution that biology might make to these
debates. Secondly, ongoing controversies in biology have continued
without much attention given to the possibility that they might be
clarified by exploring the realist and antirealist assumptions undergirding
different biological claims and positions.
449
R. S. Cohen, R. Hilp inen and Qiu Renzong (eds.), Realism and Anti -Reali sm in the
450 TIMOTHY SHANAHAN
(c) Aim of this Paper
My aim in this paper is to begin to rectify this situation by exploring

the realism-antirealism debate as it arises in one part of biological science:
evolutionary biology. Even within this restricted domain, my discus-
sion will focus on just one specific problem, that of evaluating competing
claims about the operation of natural selection. Recent work on this issue
in the philosophical literature suggests very strongly that realist and anti-
realist assumptions playa major role in ongoing debates about natural
selection. Further analysis of this work is needed to make clear pre -
cisely what role such assumptions play, and to determine how such
assumptions affect the derivation of substantive conclusions in and about
evolutionary biology. Although the scope of my investigation will be
restricted to a small skirmish in the larger contexts of the philosophy
of science and of evolutionary biology, the conclusions reached bear
on our understanding of both domains.
(2) THE UNITS OF SELECTION CONTROVERSY
(a) The Nature of the Problem
Darwinian evolutionary biology postulates that much of the change in

the living world over the last 600 million years or so has been due to
the differential survival and reproduction of differentially adapted life
forms - that is, to the process of evolution by natural selection. This much
is uncontroversial, at least among professional scientists. Where scien -
tists begin to disagree is on the correct account of the processes that
constitute evolution by natural selection. One of the most persistent
debates in evolutionary biology concerns the question of the level(s) of
nature at which natural selection operates. The living world is orga -
nized hierarchically, with genes, organelles, cells, organisms, kin-groups,
populations, species, and ecosystems (and many other entities as well)
constituting an ascending scale of biological inclusiveness. At which
of these levels does natural selection operate? Which sorts of bio-
logical entities should be thought of as in competition with one another?
Which sorts of biological entities exhibit adaptations? Biologists have
defended a variety of different - and sometimes mutually incompatible
- answers to these questions. To date, it has not been possible to reach
a consensus on this fundamental issue in evolutionary biology .
REALISM AND ANTIREALISM 451
At first glance it might seem a relatively simple matter to resolve

this problem . Write a grant proposal , get some funding, do field research ,
analyze the data, write up the results , publish , and move on to the next
biological problem. Unfortunately, matters are rarely so simple . First ,
empirical research is, by definition, based on observations, and hence
on what is observable. But natural selection is not directly observable;
it is an inference made from observations of nature conjoined with
theoretical principles. What is (at least sometimes) observable is phe-
notypic variation, competition for resources, likeness of offspring to
parent(s) , and differential survival and reproduction. That nature "selects"
some variations over others because of their superior fitness, and that
such selection causes a gradual evolution of traits over time, are claims
based on observations. Selection itself, however, is not directly observ-
able (For more on this point, see Shanahan, 1992). Evolution by natural
selection is a fact, but not all facts are known by direct (non-inferen-
tial) observation of nature .
Secondly, identifying the level(s) at which selection operates in a given
situation is even farther removed from direct observation than is the
process of selection itself. In a given selection process, there may be
any number of kinds of biological entities exhibiting variation. Merely
noting this variation does not reveal which of these kinds of entities is
responsible for evolutionary change. When a kin-group interacts with
some aspect of its environment, for example, so do the organisms that
compose it, as well as the genes that constitute part of the make-up of
the organisms . In addition, all three kinds of entities are causally con-
nected with one another : the properties of the group are at least partially
the result of the properties of the constituent organisms, and the prop-
erties of the organisms are largely the result of the properties of the genes
they contain. Distinguishing the causal role(s) of each of these kinds
of entities does not admit of any straightforward empirical determina-
tion. Evidently, theoretical (and philosophical) considerations must play
an equally important role in resolving this problem.
(b) Hierarchical Monism and the Pluralist Gambit
Recognizing this need, biologists and philosophers have offered a

variety of theoretical analyses of natural selection designed to identify
the essential properties of units and levels of selection (see, e.g., the
papers in Brandon and Burian, 1984 and those listed in Shanahan,
1990a, b). None of these proposals has been entirely successful in

resolving the problem. When an intellectual logjam of this kind occurs,
one begins to suspect that it is due to some mistaken underlying assump-
tion in the way the problem has been formulated. In this case, despite
their differences, all of these proposals share a common underlying pre-
supposition, one that is intrinsic to the formulation of the problem as
stated. They assume that for any given evolutionary phenomenon
requiring a selectionist explanation, it is possible to identify the units
(and levels) of selection that are causally responsible for that phenom-
enon. That is, they assume that units of selection are real, that such
units can be identified with one or another of the entities making up
the biological hierarchy, and that a satisfactory analysis of a given selec-
tion process requires the identification of the unit(s) and level(s) of
selection causally responsible for that process .
In an important contribution to this debate, Kim Sterelny and Philip
Kitcher (1988) challenge this "realist" assumption, and link their chal-
lenge to an explicitly antirealist philosophy of science. They christen
the basic assumption underlying most analy ses of natural selection
Hierarchical Monism, defined as follows:
Hierarchical mon ism claims that. for any selection proce ss, there is a unique level
of the hierarchy such that only repre sentations that depict select ion as acting at
that level are maximally adequate. (Intuitively, representations that see selection as
acting at other levels get the causal structure wrong .) (Sterelny and Kitcher, 1988.
p. 359).
According to Hierarchical Monism, for any selection process there is

at most one analysis that correctly captures the causal structure of that
process, and this analysis depicts selection as operating on one and only
one level of the biological hierarchy.
In place of Hierarchical Monism, Sterelny and Kitcher propose a
form of Pluralism, according to which there are many "maximally
adequate representations" of the causal structure of any given selection
process. They explain the difference between these two basic views as
follows :
Hierarchical monism differs from [pluralism] in an interesting way: whereas the plu-
ralist insists that, for any process, there are many adequate representations ... the
hierarchical monist maintains that for each process there is just one kind of adequate
representation, but that processes are diverse in the kinds of representations they demand
(Sterelny and Kitcher, 1988, p. 359).
Whereas hierarchical monism recommends a "plurality of processes",

they note, pluralism recommends "a plurality of models of the same
process" (Sterelny and Kitcher, 1988, p. 359 note 14).
Central to their argument for pluralism is recognition of the apparent
arbitrariness of segmenting the causal chain constituting selection
processes in one way rather than in another. Consider the following
biological example : Some spiders construct webs for capturing prey. They
do so because they possess certain genes that are causally connected with
web-building behavior. When the spider successfully entraps prey in
its web, both the spider and the genes that caused the web-building
behavior benefit. A consequence of this is that webs of the type that such
spiders construct become more common in the world. In this example
there is a causal chain extending from genes for spiders with web-building
behavior to spiders exhibiting this behavior to spider-webs to the capture
of prey to the biological success of web-building spiders and their web-
building spider genes. By reflecting on such cases, Sterelny and Kitcher
conclude that,
There is no privileged way to segment the causal chain and isolate the (really) real
causal story [about a selection process] . . . . We are left with the general thesis of
pluralism: there are alternative , maximally adequate representations of the causal struc-
ture of the selection process (Sterelny and Kitcher, 1988, p. 358).
On this view, selection can be represented as a function of competition

among spiders who build webs or as a function of competition among
genes for web-building. There is thus no pressing problem in determining
the level of the biological hierarchy at which selection "really" operates.
Selection can be equally well represented as operating at anyone of
several levels of the biological hierarchy. We may pick whichever one
suits our needs - methodological considerations take over. Pluralism thus
(dis)solves the units of selection problem. "[Hjierarchical monism", on
the other hand, will always remain "entangled in spider webs" (Sterelny
and Kitcher, 1988, p. 360).
(c) Pluralistic Antirealism
An especially interesting feature of Sterelny and Kitcher 's proposal

is their explicit linkage of pluralism and various antirealist doctrines
in the philosophy of science. They first align their view with instru -
mentalism:
Pluralism of the kind we espouse has affinities with some traditional views in the phi-
losophy of science . Specifically, our approach is instrumentalist, not of course in denying
the existence of entities like genes. but in opposing the idea that natural selection is a force
that acts on some determinate target, such as the genotype or the phenotype. Monists
err , we believe, in claiming that select ion processes must be described in a particular
way, and their error involves them in positing entities, "targets of selection", that do
not exist (Sterelny and Kitcher, 1988, p. 359).
Immediately following this passage, Sterelny and Kitcher align their view
with conventionalism:
Another way to understand our pluralism is to connect it with conventionalist approaches
to space-time theories. Just as conventionalists have insisted that there are alternative
accounts of the phenomena which meet all our methodological desiderata, so too we
maintain that select ion processes can usually be treated, equally adequately, from more
than one point of view (Sterelny and Kitcher , 1988, p. 359).
Although some previous analyses have hinted at a conventionalist

solution to the units of selection problem (i.e., Cassidy, 1981), Sterelny
and Kitcher provide by far the most explicit argument in support of
such an interpretation. But is it right?
(3) A CRITIQUE OF PLURALISTIC ANTIREALISM
(a) Some Questions
In this section I want to address a number of questions that naturally

arise from the foregoing sketch of Sterelny and Kitcher's "pluralistic
antirealism". First, is pluralism of the sort that they defend cogent?
Is it true that there are (or can be) many maximally adequate repre-
sentations of the causal structure of any given selection process?
Secondly, is there any interesting connection, as they suggest that
there is, between their pluralism and antirealism in the philosophy
of science? Does their pluralism entail instrumentalism, for example,
or vice versa? Finally, have they shown that hierarchical monism is
mistaken? Can this view be reformulated to avoid the criticisms they level
against it?
These questions are important, because if pluralism is, as Sterelny and
Kitcher argue, superior to monism , and if pluralism entails antirealism,
then antirealism gains support as a philosophical interpretation of evo -
lutionary biology - and perhaps gains credibility as a global interpretation
of natural science as well. On the other hand, if pluralism - as Sterelny
and Kitcher argue for it - is not a cogent position, or if it is not linked

to antireali sm in any interesting way, then their arguments for pluralism
are not sufficient to establish the superiority of an antirealist interpre-
tation of evolutionary biology. Hence both a solution to the units of
selection controversy as well as support for specific positions in the
philosophy of science are at stake. I will address the questions concerning
pluralism and antirealism in this section, and reexamine hierarchical
monism in section (4).
(b) On Pluralism
Sterelny and Kitcher offer two different versions of pluralism, which ,

while similar, are not identical. According to the "general thesis of
pluralism: there are alternative, maximally adequate representations of
the causal structure of the selection process" (Sterelny and Kitcher, 1988,
p. 358). But a page later they write that, "the pluralist insists that, for any
[selection] process, there are many adequate representations" (Sterelny
and Kitcher, 1988, p. 359). The first version is considerably stronger than
the second. The first version , in fact , entails the second, but not vice
versa. If there are a number of maximally adequate representations of
a process , then, all of them being maximal , they are necessarily equally
adequate as well. But there might be many equally adequate represen -
tations of a process, all of which fail to be maximally adequate. (Analogy:
If three students do "maximally well" on an exam, by each earning 100%
of the possible points , then all three necessarily did equally well. But
those three students could have done "equally well" among themselves
on another exam, by all earning precisely 80% of the points, without
thereby doing maximally well on that exam.) To capture this important
distinction, I will refer to these two versions of pluralism as Strong
Pluralism and Weak Pluralism. Is either version cogent?
Concerning Strong Pluralism, what does it mean to say that there
are many maximally adequate representations of the causal structure of
any [selection] process? A representation of a process would presumably
be a description which highlights certain features of the process as
being of special importance and significance. If the representation is of
the causal structure of a process , then the description would highlight
the entities and their properties, along with their interactions, that are
constitutive of that process. A maximally adequate representation of
the causal structure of a process would then be the best description
possible of the entities, properties, and interactions that are constitutive

of that process. This much seems fairly clear.
The difficult question, of course, concerns the notion of "adequacy".
Adequate for what? One possibility is: adequate for explaining the
outcome of the given process. A representation would be maximally
adequate on this criterion if it provides the best possible explanation of
the outcome of the given process, by showing how the various parts
of the process together produced the outcome in question. Another
possibility would be: adequacy for making successful predictions and
organizing empirical phenomena. A representation would be maximally
adequate on this second criterion if it provides the best possible instru-
ment for predicting the outcomes of specific (kinds of) processes.
Clearly, a given representation could be maximally adequate when
judged by one of these criteria but not maximally adequate when judged
by the other. The so-called "Symmetry Thesis" is false : Good explana-
tions are not necessarily logically identical to equally good predictions.
(Two models of the solar system may be maximally adequate for pre-
dicting the next solar eclipse, but may differ in their adequacy for
explaining such an event.)
Which kind of adequacy do Sterelny and Kitcher have in mind? From
their various remarks it is clear that their pluralism is committed to the
idea that a number of different representations of a single selection
process can each be maximally adequate in the sense that each repre-
sentation attributes "causal efficacy" to properties of entities at different
levels of the biological hierarchy. This suggests that they are thinking
of explanatory adequacy. So, for example, in the case of the spider-
web example, an explanation of that process that attributes causal efficacy
to organismic properties may be a maximally adequate representation
of the causal structure of the process, while an explanation that attrib-
utes causal efficacy to genic properties might also be a maximally
adequate representation of the causal structure of that process. In short ,
maximally adequate representations and /or explanations of selection
processes may be framed in terms of the entities at anyone of several
levels of the biological hierarchy.
This way of understanding Strong Pluralism, however, gives rise to
a dilemma. If the two (or more) representations are really different in
attributing causal efficacy to different kinds of properties, then it is
not clear how either can really be maximally adequate, because each
representation omits the causal role of the properties emphasized by
the other representation - properties that presumably play some role in

the causal process in question. To the extent that a given representa-
tion omits causally relevant factors , it is only partially, and hence not
maximally, adequate. (In the spider-web example, genic and organismic
properties both contribute to the selection process. Omitting either set
of properties results in a submaximally adequate representation.) A
maximally adequate representation would be one that includes all
the causally relevant properties constitutive of the causal structure of
the process . On the other hand, if we say that there is really only one
maximally adequate representation of the causal structure of the process,
namely, the one that includes all the causally relevant properties (genic,
organismic, etc.), then the view is no longer recognizably pluralist. Hence
the dilemma : Either it is false that the different representations in question
are each maximally adequate, or it is false that there can be more than
one representation that is maximally adequate. Either way, it is not clear
that Strong Pluralism is a cogent view.
Weak Pluralism, on the other hand, rather than being mistaken, seems
to be in danger of being trivial. The claim of Weak Pluralism is that
for any [selection) process, there are many (not necessarily maximally)
adequate representations. Obviously the status of this claim hangs on
what is meant by "adequate representations". If by "adequate" is meant
that a given representation completely captures the causal structure of
a proce ss, then it is identical with Strong Pluralism, and faces the same
objections. On the other hand, if what is meant is that a given repre-
sentation allows one to capture enough of the causal structure of a process
to organize empirical data in a coherent way or to make successful
predictions, or to explain some feature of a process, then the plurality
of equally adequate representations is simply a function of the uniform
incompleteness of each . Like the proverbial blind men examining an
elephant, each description can be "equally adequate" just because each
represents only a partial understanding of the whole. Equivalence can
be achieved in this way, but becomes cheapened as well. Despite their
provocative claims , therefore, Sterelny and Kitcher have not shown that
their brand of pluralism provides a cogent analysis of natural selection.
(c) On Antirealism
Instrumentalism in the philosophy of science is the view that while

the theoretical terms of a scientific theory may be useful (or even
necessary) for structuring empirical data into coherent patterns and/or for
making predictions, we are not on this basis required to believe that
such terms refer to anything real. Indispensability does not entail exis -
tence. Claims about unobservable theoretical entities are merely rules for
making predictions about observable states of affairs. They are to be
judged on their adequacy or inadequacy for this purpose; hence such
claims are not literally true or false. According to this view, scientific
theories are to be judged for their adequacy in organizing empirical obser-
vational data, not for their ability to reveal the unobservable constituents
and structures of the world.
As noted above, Sterelny and Kitcher connect their pluralism with
an instrumentalist philosophy of science "in opposing the idea that natural
selection is a force that acts on some determinate target, such as the
genotype or the phenotype." They add that, "Monists err . . . in claiming
that selection processes must be described in a particular way, and their
error involves them in positing entities, ' targets of selection', that do
not exist" (Sterelny and Kitcher, 1988, p. 359). The inference here
seems to be this: Monists claim that there is an exclusively correct way
to describe any selection process. An exclusively correct description of
any selection process will involve positing "targets of selection". But
it is false that there is an exclusively correct way to descr ibe any selec -
tion process. Therefore "targets of selection" do not exist.
Put this way, the fallacy becomes obvious. Consider an analogy.
"Baseball monists" claim that there is an exclusively correct way to
describe any baseball game . An exclusively correct description of any
baseball game will involve positing entities such as bats, gloves , base-
balls, outfielders, umpires, etc. But it is false that there is an exclusively
correct way to describe any baseball game . Therefore entities such as
bats, gloves, baseballs, outfielders, umpires, etc., do not exist.
Clearly, from the claim that there is no exclusively correct way to
describe a baseball game, it does not follow that bats and so on do not
exist, or that claims about such things are neither true nor false. Such
things may not exist , of course , but their non-existence does not follow
from the claim in question . Likewise , from the premise that there is no
exclusively correct description of any selection process it does not follow
that "targets of selection" do not exist, because the existence of targets
of selection does not depend in any obvious way on the truth of monism .
One could be a monist and deny the existence of "targets of selection".
Likewise, one could be a pluralist and affirm their existence. Despite their
intriguing claim, therefore, Sterelny and Kitcher have not shown that
pluralism (of the sort they espouse) and instrumentalism are connected
in some interesting way.
It seems clear that pluralism does not entail instrumentalism. It perhaps
goes without saying that instrumentalism does not entail pluralism either.
An instrumentalist might hold that there is only one maximally adequate
representation for a given process, and simply maintain that this repre-
sentation organizes the empirical data without revealing anything about
the unobservable structure of this process. "Monistic instrumentalism"
is not self-contradictory.
Sterelny and Kitcher link their pluralism with antirealism in the
philosophy of science . I would suggest, however, that their view actually
presupposes a form of realism. Recall that their pluralism is the view that
there are many maximally adequate representations of any selection
process . That is, there are "a plurality of models of the same process"
(Sterelny and Kitcher, 1988, p. 359). But talking of a plurality of models
of a single causal proce ss, as Sterelny and Kitcher do , implies that
there is a unitary physical phenomenon serving as the benchmark for
scientific representations. If, as argued earlier, the process of natural
selection is, strictly speaking , unobservable, then Sterelny and Kitcher
are committed to holding that we can have genuine insights into the
reality of this process, despite its unobservability. This is, of course, a
central tenet of scientific realism. At the very least, talk of a "maxi-
mally adequate representation of the causal structure of a process" departs
significantly from the spirit of antirealism, because having such a
representation suggests that one does have access to the unobservable
entities and processes responsible for observable phenomena - otherwise
why have confidence that there is just one process being represented
in diverse ways?
Ironically , then, Kitcher and Sterelny 's Pluralistic Antirealism seems,
upon closer inspection, to be based on an underlying assumption of
realism about the process of natural selection. If their pluralism concerned
only the equivalences of different models of natural selection to organize
empirical data and to facilitate predictions, with no attempt to say
anything about the causal structure of selection processes, then their view
would be unexceptional. If one only restricts one's demands on what a
model should do, then it will always be possible to find a plurality of
"maximally adequate" models of any phenomenon. But establishing the
claim that there are a plurality of maximally adequate representations
of a single causal process is a different , and much more difficult, task.

Pluralism and an ecumenical spirit with regard to how we model and
represent natural selection are all very well, especially in cases where
different representations exhibit different methodological virtues. But
it is well to remember that although methodological and ontological
matters are frequently connected in interesting ways, they are just as often
conflated.
(4) HIERARCHICAL MONISM RECONSIDERED
(a) The Problem with Previous Views
I have been arguing that the Pluralistic Antirealism of Sterelny and

Kitcher faces some serious objections. My analysis has been critical
and negative. But by reflecting on the problems with Pluralistic
Antirealism and Hierarchical Monism we can begin to get as glimpse
of the broad outlines of a more adequate analysis.
Sterelny and Kitcher's pluralism is an attempt to provide a princi-
pled alternative to Hierarchical Monism. Recall that Hierarchical Monism
is the claim that, for any selection process, there is a unique level of
the biological hierarchy such that only representations that depict selec-
tion as acting at that level are maximally adequate. Representations
that see selection as acting at other levels get the causal structure wrong .
Sterelny and Kitcher rightly attack this claim, because it ignores the
fact that a given selection process can be understood as the product of
causal interactions at anyone of a plurality of different levels.
Ironically, however, Sterelny and Kitcher 's pluralism embodies a
strong monist assumption. Their view still utilizes the idea that a
representation focusing on anyone level of the biological hierarchy can
be maximally adequate. Their view differs from the one they reject
only in that while Hierarchical Monists hold that a representation of a
selection process at a single level will be exclusively correct, Sterelny
and Kitcher believe that such a representation will be just one of several
maximally adequate possibilities. But both views hold that an analysis
of a selection process at only one level may be maximally adequate.
What undermines both formulations is the fact that each selection
process is a multi-level causal process . That is, both views fail to accom-
modate the fact that any selection process will involve causal chains
connecting entities on a plurality of different levels. Consequently, a
representation (be it "monist" or "pluralist") that focusses on one level

only will fail to be maximally adequate. A maximally adequate repre-
sentation of a selection process should depict selection as the product
of forces operating on and between a plurality of levels, with no single
level standing out as uniquely important. This is precisely the kind of
view that I believe is correct and wish to defend below.
(b) Exclusive and Inclusive Hierarchical Monism
Hierarchical Monism of the sort that Sterelny and Kitcher criticize might
better be called Exclusive Hierarchical Monism. Thi s view claims that
the causal structure of a given selection process can be captured by
focusing on just one level of the biological hierarchy. The view I wish
to propose might be called Inclusive Hierarchical Monism. This view
claims that a maximally adequate representation of the causal structure
of a selection process will have to take into account a number of levels
of the biological hierarchy. Whereas the former view excludes all but one
level, the latter view includes many levels in one analysis.
To repe at this important distinction: According to Exclusive Hier-
archical Monism, for any selection process there is at most one analysis
that correctly captures the causal structure of that process, and this
analysis depicts selection as operating on one and only one level of
the biological hierarchy. According to Inclusive Hierarchical Monism,
for any selection process there is at most one analysis that correctly
captures the causal structure of that process, and this analysis depicts
selection as operating on a plurality of different levels of the bio-
logical hierarchy.
To illustrate Inclusive Hierarchical Monism , consider the spider-web
example again. Spider genes are causally connected with spiders and web-
building behavior, and thus indirectly to webs and to the capture of
prey, resulting in the differential propagation of both the genes and the
spiders. Selection in this case is operating on both the genes and the
spiders, and on the entire gene-spider-web causal process. It is the entire
process linking genes and spiders and webs that results in the differen-
tial biological success of genes, spiders, and even web-phenotypes.
Selection may be best thought of as operating on the entire causal
chain. A maximally adequate representation of the causal structure of
this selection process will include not only a description of how each
of the entities involved in the process fares relative to other entities of
the same kind (i.e., at the same level) , but also how the entire causal
chain contributes to the differential representation of genes, spiders,
and webs in successive generations. Inclusive Hierarchical Monism
includes the various multi-leveled entities constituting a selection process
in a single representation of the causal structure of that process .
Perhaps an analogy will make this view clearer. Consider the auto
industry. The Ford Motor Company builds and sells vehicles such as
the Escort, Taurus, and Probe. The sale of these vehicles is the end
product of a long causal process, involving engineers who design these
vehicles, auto workers who use design instructions to construct these
vehicles, and auto salesmen who display these vehicles in the hope of
persuading consumers to purchase them. Information about customers's
buying habits is then used to improve (or at least change) the design
of future models. Through this process the design features (e.g., anti-lock
brakes) of very successful models are more likely to show up in future
models than are the design features of less successful models.
Within the Ford Motor Company, therefore, there is a kind of selec-
tion operating among the vehicles that they build, resulting in more
Ford vehicles with features attractive within the current consumer envi-
ronment. Designers are competing with one another to come up with more
attractive features, workers are competing with one another to produc e
high-quality vehicles, and salesmen are competing with one another to
outsell one another. There is selection going on at all levels of the
causal process, "driving" it forward.
If we step back and consider the entire auto industry, however, we
realize that the Ford Motor Company is just one part of a much larger
competitive arena . The Toyota Motor Company produces vehicles such
as the Corolla, Camry, and Cressida - models that consumers may choose
to purchase rather than the Ford models mentioned above. Toyota, too,
has its team of engineers, workers, and salesmen. Within the global
auto industry, there is a competition between Ford and Toyota, as well
as a competition between other U.S. and Japanese auto makers . Because
the final product of each manufacturer is the end result of a long causal
chain involving design, production, and marketing, the relative success
of rival auto companies can be seen as dependent on the entire causal
chain resulting in their products. That is, the entire Ford causal chain
is in competition with the entire Toyota causal chain. It is Taurus vs.
Cressida and Ford vs. Toyota.
Since American and Japanese auto makers use somewhat different
processes in the production of their vehicles (e.g., "top-down" manage-

ment vs. "module-team" organization), it is more than just autos that
are in competition - it is also methods of producing autos . American
auto designers are in competition with Japanese auto designers, and
American auto workers are in competition with Japanese auto workers,
and so on, but entire companies (and methods and processes) are also
in competition with one another. Selection operates within all levels of
the "auto hierarchy" as well as between the different causal processes
constituting each hierarchy. Understanding "the auto industry" requires
understanding each of these processes and how they interact to pro-
duce the vehicles television commercials incessantly encourage us to
purchase.
Returning now to biological examples, Inclusive Hierarchical Monism
holds that selection can best be viewed as operating among entire causal
chains. Genes are designers, organisms through development realize
these genetic plans, and phenotypes (including behavior) test these
designs against rivals. Success depends on each stage in the causal
process - the best design in the world will not be successful if it fails
to be translated into reality. Selection operates among different spider
gene-organism-web causal processes, resulting in the differential repre-
sentation of these processes in subsequent generations. A maximally
adequate representation of this process will include the causal role of
all the relevant levels , and show how they are causally interconnected.
(c) Causal Realism
Inclusive Hierarchical Monism, as described above, pre supposes that

for any selection process, there may be several equally adequate repre-
sentations, but there will be at most a single maximally adequate
representation - that representation that correctly captures the multi-level
causal structure of the process. Equally adequate representations are by
their very nature only partial descriptions of the causal structure of
the proce ss. There may, without inconsistency, be any number of such
representations. But a maximally adequate representation, in contrast,
will be singular. One of the aims of inquiry on this view is to assemble
the various partial representations into something more closely approx-
imating a maximally adequate representation. It need not be assumed,
however, that such a maximal representation will ever be achieved.
Such a view is explicitly realist because it assumes that we can have
genuine insights into the causal structure of the world or, at least, that
there is such a causal structure that we can continue to gain informa-
tion about. Hence one can be a pluralist with regard to the currently
available representations of natural selection, but a realist with regard
to the causal process itself. That is, one could be a methodological
pluralist but a metaphysical monist. It is not that there is no causal
structure to be discovered, or that claims about the causal structure of
a process are neither true nor false (as an instrumentalist would hold),
but rather that we often cannot be sure just what the correct causal
structure is. This is not a rejection of realism, but rather just an honest
assessment of our (current) epistemic limitations. (For a view not unlike
the one expressed here, see Cartwright, 1981.)
Note that Sterelny and Kitcher's argument against (Exclusive)
Hierarchical Monism does not affect Inclusive Hierarchical Monism.
It might be (and indeed is) false that there can be a uniquely correct
representation of the causal structure of a selection process that focuses
on just one level of the biological hierarchy, but true that there can be
a uniquely correct representation of the causal structure of a selection
process that includes a number of levels of the biological hierarchy.
Sterelny and Kitcher have not shown that Hierarchical Monism is false,
but only that one, rather implausible version of it, is mistaken. When
Hierarchical Monism is reformulated along the lines suggested above ,
it not only escapes their criticism but also provides a plausible philo-
sophical analysis of the proces s of evolution by natural selection.
(5) CONCLUSIONS
According to scientific realists, scientific theories do, at least approxi-

mately and on some occasions, give us insight into the causal structure
of the world . But even when it is not known which representation of
the causal structure of a part of the world is correct, realists hold that
a correct (i.e., maximally adequate) representation is at least possible,
that there is a truth of the matter and that, with any luck, we will discover
what it is. Such a position does not preclude one from utilizing any
number of alternative models of a process in the pursuit of scientific
goods: empirical adequacy, predictive success, explanatory scope, etc.
But recognizing the (perhaps temporary) empirical equivalence of dif-
ferent models does not entail that there is no underlying reality to learn
more about. As Eman McMullin has noted ,
Scientists have never thought themselves disqualified from pursuing one of a number of
physical models that, for the moment, appear empirically equivalent. As met aphors ,
these models may give rise to quite different lines of inquiry, leading eventually to their
empirical separation. Or it may be that one of the alternative models appears undesir-
able on other grounds than immediate empirical adequacy. If prolonged efforts to separate
the models empirically are unsuccessful, or if it comes to be shown that the models are
in principle empirically equivalent, scienti sts will , of course, turn to other matters. But
this is not a rejection of reali sm. It is, rather , an admission that no decision can be made
in this case as to what the theo ry, on a realist reading, commits us to (McMullin, 1984,
p. II).
We cannot, just by casually observing nature, determine the correct causal

structure of its processes. Empirical adequacy is a necessary but not
sufficient condition for discovering the truth. The difficult task for all
scientists is to state clearly the relationship between their theoretical
models and the causal structure of nature (as currently understood) , and
to avoid the temptation to identify the causal structure of nature with
the structure of their models just because the models "fit" a certain
selected subset of the empirical phenomena and are computationally con-
venient. Doing so is difficult for an additional reason: It requires that
we admit (openly) that our models of natural phenomena are just that
- hopeful human constructs - which do not necessarily capture the causal
structure of nature/reality, no matter how pretty our models may be or
how useful we may find them. Theories are nets that we cast to capture
the world . But as in fishing, so in science, there is no guarantee that
even our best efforts will bring home the catch we are after.
Loyola Marymount University,
Los Angeles, U.S.A.
BIBLIOGRAPHY
Brandon, Robert and Burian , Richard (eds.) (1984). Genes. Organisms. and Populations:
Controversies over the Units of Selection, MIT Press , Cambridge, MA.
Cartwright, Nancy (1981). ' The Reality of Causes in a World of Instrumental Laws' , in
P. Asquith and R. G iere (eds .), PSA /980 , Philosophy of Science Association, East
Lansing, MI., pp. 38-48.
Cassidy, John (1981) . ' Ambiguities and Pragmatic Factors in the Unit s of Selection
Controversy ' , Philosophy of Science 48, pp. 95-111.
McMullin, Ernan (1984). ' A Case for Scientific Realism' , in J . Leplin (ed .), Scientific
Realism , University of California Press , Berkeley , pp. 8-40.
Shanahan, Timothy (l990a). ' Group Selection and the Evolution of Myxomatosis',
Evolutionary Theory 9, pp. 239-254.
Shanahan, Timothy (I 990b). 'Evolution, Phenotypic Selection, and the Units of Selection',
Shanahan, Timothy (1992). 'Selection, Drift, and the Aims of Evolutionary Theory', in
Paul Griffiths (ed.), Trees of Life: Essays in Philosophy of Biology, Kluwer Academic
Publishers, Netherlands, pp. 133-161.
Sterelny, Kim and Philip Kitcher (1988). 'The Return of the Gene', Journal ofPhilosophy
85, pp. 339-361.
PAUL C . L. TANG AND RALPH W . BROWN III
ANTI-REALISM AND THE

COMPLEMENTARITY MODEL OF MIND -BRAIN
We introduce the first steps towards advancing a new, anti-realistic, Com-

plementarity Model of the mind-brain. We begin by discussing a very
current debate within the philosophy of cognitive science, viz., the debate
between Jerry Fodor and Paul and Patricia Churchland concerning the
appropriate model of cognition, i.e., the computational model versus
the neural net model. We suggest that this debate leads us back to the
age-old question of the ontological status of mind and body. At this point
we introduce another contemporary representative theory of mind-body,
viz., that of Karl Popper. We then argue for the underdetermination and
incommensurability of all theories of the mind/body and then present
the first steps toward our instrumentalist, complementarity theory. We
then identify the virtues of our model within philosophical psychology
and suggest that our model may also be capable of helping to resolve
the Fodor-Churchland debate.
II
One of the foremost advocates of the Computational Theory of cogni-

tive phenomena is Jerry Fodor, while two of the strongest advocates of
the Neural Net/Connectionist theory of cognitive phenomena are Paul
and Patricia Churchland. According to Fodor, psychology is concerned
with the formal structure of symbols in the mind and of the manner in
which these symbols are manipulated. These symbols are mental repre-
sentations, a so-called language of thought, and they play the role of
propositions in discourse involving propositional attitudes; moreover,
these symbols are held to represent the phenomena about which a subject
is thinking . In addition, the mind possesses a set of rules - corresponding
to rules of inference - that determine which operations are performed
on these representations. Fodor argues that this classical architecture
can be used to explain rational behavior, concept learning, and percep-
tion. He also argues that this theory can explain , e.g., logical peculiarities
467
R. S. Cohen, R. Hilpinen and Qiu Renzong (eds.), Realism and Anti -Realism in the
468 PAUL C . L. TANG AND RALPH W . BROWN III
of discourse about mental states such as how the mind can understand
such relationships as, e.g., "John loves Mary" and "Mary loves John".
The ease by which human beings can make these systematic transfor-
mations is exactly representable in symbolic terms.
Advocates of connectionism (sometimes referred to as PDP or neural
net theories) claim that cognitive phenomena are not computational in
the manner of a von Neumann computer. Connectionists, such as the
Churchlands, advocate exploring systems known as neural nets . Such
systems consist of nodes, each of which possesses a determinate degree
of activation at any time, and is connected to many other nodes to which
it sends excitatory or inhibitory stimuli. Given an initial pattern of acti-
vation, the excitations and inhibitions passing through the system will
alter the activation states of the nodes until a stable pattern is reached .
Moreover, the strengths of excitatory and inhibitory connections can
be altered as a result of local activity within the system. Such systems
can learn to respond in new ways and settle into different states at later
times . In Matter and Consciousness, Paul Churchland provides a simple
example of a submarine's neural net computer system training up to learn
to recognize the difference between a rock and a mine using sonic
echos, hidden units, automated learning by the back propagation of
error and the generalized delta rule. I Neural nets have been used to model
certain cognitive functions; in certain tasks, such as pattern recogni-
tion , these systems appear to be much more human-like than those of
rule-processing machines . A recent success , NETalk, by Terry Sejnowski,
is a project to get a neural net based machine to learn to speak simple ,
grammatically correct sentences after the researcher has inputted a 1,000-
word transcription of an average child 's conversation. However,
systematic transformations of relationships, guaranteed by classical archi-
tecture , are not a natural property of neural nets.
The debate between Fodor and the Churchlands has been vigorous
indeed. But the underlying problem, we suggest, is not one of which
model - computationist versus neural net - captures the process of cog-
nition in a realist sense, i.e., in a sense in which a scientific theory or
model provides a literally true (or approximately true) account of the
structure of the universe, or of nature, or, as in our present case, of human
cognition. Indeed, the problem is not just that of realism versus anti-
realism, but one that concerns even more deeply the way in which these
two notions of realism and anti-realism pertain to an ontological question,
viz., the classical problem of mind/body dualism, so acutely posed by
THE COMPLEMENTARITY MODEL OF MIND-BRAIN 469
Descartes. This issue must be addressed first, we suggest, before we

can evaluate the Fodorian and Churchlandian cognitive models.
III
On a realist basis, the problem of whether there exists a mind over and
above the brain, or whether there exists only a brain, has given rise
to a great variety of duali st and monist positions, such as substance
dualism, emergentism, epiphenominalism, elemental dualism, interac-
tionist property dualism, popular dualism, functionalism , reductive
materialism, eliminative materialism, and so on . These "isms" repre-
sent broad classes of philosophical theories, and they indicate the
exceedingly complex and subtle nature of the ontological problem. The
arguments for these various positions are well known and we shall not
rehearse them here.
Nevertheless, which general philosophical position is the correct one?
Dualism or monism? This question is important because the former philo-
sophical theory (dualism) grounds - at least from the point of view of
language - the Fodorian position and the latter philosophical theory
(materialism) grounds the Churchlandian position. Granted, Fodor 's
position is somewhat complex, for if he is a functionalist, then he is a
functionalist with very strong Carte sian overtones. This latter charac-
terization stems from his bold talk of the innate language of thought
and in his holding that mental states really exist, that they can interact
with one another, that it is possible to study them, and that they are
nonreductive.
At this point, going beyond just the debate between Fodor and the
Churchlands, and in anticipation of the general, alternative comple-
mentarity model to be advanced later, we can identify unequivocally
contemporary dualists in the persons of Karl Popper and the neuro-
physiologist, John Eccles , although they would prefer to call their position
"interactionism". Popper focuses on aspects of mental activity that
could not easily claim to be accomplished solely by physical bodies, e.g.,
mathematics; in this respect he echoes the philosophy of Descartes. For
Popper, abstract objects, such as mathematical objects, scientific theories
and works of literature assume a type of autonomy, a "life of their
own", and also constitute a distinct realm, which he calls "World 3". This
World is to be distinguished from World I, the realm of physical objects
and also from World 2, the world of mental activity, by the fact that World
470 PAUL C. L. TANG AND RALPH W. BROWN III
3 is governed by normative principles, such as the laws and principles

of logic. Popper's argument that mental activities are distinct from
physical objects follows from the need for an intermediary that can apply
information from World 3 to World I. Popper believes and tries to argue
that no purely physical system can grasp the contents of World 3; thus,
there must be mental activities that can grasp the objects of World 3
and then causally interact with objects and events of World I . For Popper,
only interactionism can properly explain how World 3 principles regulate
World I events and objects. For a contemporary defense of Cartesian sub-
stance dualism, the reader can look to the work of Richard Swinburne,
a distinguished philosopher of religion who has also published much
in the philosophy of science.'
How shall we answer, then , the ontological question? We shall argue
from both an empirically based point of view as well as a conceptually
based one. Let us begin with the empirically based view, and, specifi-
cally, with a clinical example, which is representative of the vast clinical
and experimental literature on mental illness. It is well known that the
illness known as clinical depression is associated with a decrease in
the concentration of neurotransmitters such as norepinepherine and
acetylcholine. It is also well known that certain tricyclic drugs, such as
lithium, imipramine, amitriptiline, chlorpromazine and phenelzine can
restore the level of concentration of these neurotransmitters within 24
to 48 hours, although the antidepressant effects often do not become
apparent for one to three weeks (See Dennis L. Murphy et al., 1975).
Furthermore, according to A. J. Prange (1975) , all antidepressants exhibit
this lag time.
However, drug therapy, or pharmacotherapy is, apparently, not the sole
means of successful treatment, It is also well known that so-call " talk
therapy", or psychotherapy, can be solely and successfully employed.
According to a study by Ernest E. Beckham and William R. Leber
(1985) , psychotherapy, in numerous clinical trials, alone yielded superior
results in three general types of depression; while in three other types ,
medication or pharmacotherapy yielded superior results; while in six other
types, in which either psychotherapy or pharmacotherapy were used , there
was no difference in the success rates . As the investigators write : "The
above results suggest that there is currently no evidence to support strict
adherence to either a psychotherapeutic or pharmacological model of
treating outpatient depression. Each type of treatment has demonstrated
its ability to alleviate symptoms of depression" (Beckham and Leber,

p. 328). From the point of view of a realist ontology, what do these results
indicate?
Before going on to answer this question, we must examine two further
concepts, viz., underdetermination and incommensurability (See Quine ,
1961, 1975). The first thesis claims that, for any given set of empirical
data, or observations, or evidential sentences, an infinite number of
theories or models, some inconsistent with each other, can imply (i.e.,
explain) that set of data, or set of observations, or, better, set of evidential
sentences.' Moreover, we note that these two theories are incommen-
surable in Kuhn's meaning-variance sense. For example, within the
dualist/psychotherapy model, the term 'cause' is used differently from
the way in which this term is used in the materialist/pharmacotherapy
model, where ' cause' refers not to some immaterial event, but to a
material, physically based event. Following Wittgenstein's theory of
meaning as use, we can say that the term 'cause' means differently in
the two different philosophical theories and in the two different, corre -
sponding psychological models. On the basis of this all-important concept
of cause, we conclude that the two models , and the two philosophical
positions of dualism and materialism, are in-commensurable.
We now return to the main issue at hand, and explain how under-
determination and incommensurability apply. Referring back to the
Beckham and Leber study on depression, we see that, in the case of
clinical depression, psychotherapy can effect a successful cure, and this
model implies that certain cognitive states, certain mental belief states
of the patient are causally responsible for the recovery. This technique
of psychotherapy can well be used by, say, a cognitive psychologist.
However, drug therapy can also be used, and, on this model, chemical
action on the brain is ultimately responsible for effecting a cure . This
technique of drug therapy can well be used by, say, a physiological
psychologist. So we see that the two philosophical theorie s, the dualist
theory that underlies the psychotherapy model and the monistic , mate-
rialistic theory that underlies the drug therapy model are underdetermined
by the set of evidential statements, viz. , the patient's recovery from
depression.
472 PAUL C. L. TANG AND RALPH W. BROWN III
IV
We are at an impasse. From a realist point of view, which philosoph-

ical theory is the correct one? The answer is that we may have to accept
the possibility that we may never answer this question, or that we may
never know. Perhaps, again following Wittgenstein, we ought to dissolve
the problem. But how? The clue, we suggest, comes from examining
the nature of quantum theory. In the famous double-slit experiment, a
beam of photons passes through a screen with only one slit open and
produces on a metal plate located behind the screen a pattern that can
only be interpreted as light rays being fundamentally composed of par-
ticles. However, when both slits are open, the photons produce an
interference pattern on the metal plate, and such a pattern could occur
only if each photon had actually gone through both slits, like a wave.
The lesson to be learned from this Complementarity Principle is that light
has a dualism, two fundamental characteristics that are incompatible with
each other. Moreover, from the viewpoint of scientific realism, one model
represents one 'reality' ; however, we now have two models (wave model
and particle model) that represent 'two realities', and this result refutes
again the thesis of scientific realism. The most we can say is that, at
times, light is fundamentally particle-like; at other time s, it is funda-
mentally wave -like.
The Complementarity Principle in Quantum Theory gives rise to a
philosophical problem, for, if we claim, on the assumption of scientific
realism, that our models, or theories, given us a literally true descrip-
tion of the universe, conceived as a "single reality", then we have a
paradox. For now light, in its fundamental character, is both wave and
particle. But this cannot be so.
Within the received Copenhagen Interpretation of quantum theory, this
paradox is obviated by quantum physicists treating the theory as simply
a tool or instrument for explanation and prediction, noting that the theory
does very successfully explain and predict phenomena in the microworld.
Copenhagenists do not claim that the theory provides a literally true
account of these microphenomena. On this anti-realistic, instrumentalist
view, Copenhagenists do not find it problematical to say that light some-
times behaves as if it were composed of particles and at other times as
ifit were composed of waves. There is no claim that the quantum theory
provides a literally true account of the microworld. A key concept is
the notion of the ' as if', which forms the basis of Kant's discussion of
THE COMPLEMENTARITY MODEL OF MIND -BRAIN 473
regulative concepts and regulative principles that help us to order our

experiences in the world . Kant's concept of the 'as if' ultimately gives
rise to the philosophical position known as 'pragmatism' .
v
We now return to the problem of mind/brain. Given the argument for
the underdetermination and incommensurability of dualism and materi-
alism as philosophical theories, and the empirical support provided by
the psychotherapy and drug therapy models that are used to treat mental
illness (e.g., depression), we suggest that the most fruitful approach to
the ontological problem in the context of modern psychology and cog-
nitive science is to follow the lead of the Copenhagen quantum physicists
and adopt an anti-realistic, instrumentalist stance towards the mind/body
problem . When psychotherapy is used, the psychologist is acting as if
there exists a mind over and above the brain, and that this entity, imma-
terial thought it may be, has so-called states, such as belief, that are
causally efficacious in effecting a successful outcome, viz., the absence
of depression. From the point of view of our anti-realistic Complemen-
tarity Model of Mind-Brain that underlies this model of therapy and of
psychological explanation, the therapist is acting as if the theory or
position of dualism is true. However, if psychotherapy is unsuccessful,
the psychologist may prescribe drugs, with or without additional psy-
chotherapy. In extreme cases, the psychologist and physician might
prescribe electroshock therapy . In these cases, from the point of view
of the Complementarity Model of Mind-Brain, the therapist is acting
as if the theory or position of materialism is true. What is important to
each are the pragmatic results, which can also later be discussed in
terms of explanation and prediction.
And now we return to the two competing classes of models of cog-
nition . The Fodorian model accounts well for certain important aspects
of cognition, such as systematicity, while the neural net model accounts
well for perceptual knowledge. But a realist position gives rise to the
dispute as to what 'really' is there, overlooking the important point that
Nancy Cartwright makes, viz ., that all theories abstract from reality
and hence are false. This leads us into our anti-realist, pragmatic, instru-
mentalist Complementarity Model of Mind-Brain. On a realist basis,
Fodor and the Churchlands are basically making assumptions, perhaps
even value judgments, about which ontological position to take, and
474 PAUL C. L. TANG AND RALPH W . BROWN III
then developing their respective, incompatible theories of cognition

from that point, even though each model, on its own, is successful in
its own domain. The Complementarity Model that we begin to advance
here accounts for the success of both the Fodorian and Churchlandian
models of cognition and places a premium on pragmatics rather than
on ontology.
We claim that our use of complementarity in quantum physics is to
be taken as an important springboard, not as a perfect analogy with the
Complementarity Model of mind-brain, which we have begun to develop
in this paper. And we also note that Bohr himself held that some sort
of complementarity is to be found in psychology. In a little-known but
philosophically insightful paper entitled 'The Quantum of Action and the
Description of Nature' (1929), which appears in his collection of essays,
Atomic Theory and the Description of Nature (1961), Bohr writes about
Heisenberg and says that
.. . we must, in general , be prepared to accept the fact that a complete elucidation of
one and the same object may require diverse points of view which defy a unique descrip-
tion . .. The necessity of taking recourse to a complementary, or reciprocal, mode of
description is perhaps most familiar to us from psychological problems (Bohr, p. 96).
Later, in the same article, Bohr writes

It might still be permitted here briefly to refer to the relat ion which exists between the
regularities in the domain of psychology and the problems of causality of physical phe-
nomena . When considering the contrast between the feeling of free will, which governs
the psychic life, and the apparently uninterrupted causal chain of the accompany ing
physiological processes. the thought has, indeed , not eluded philosophers that we
may be concerned here with an unv isualizable relation of complementarity. Thu s. the
opinion has often been expressed that a detailed invest igation of the processes of
the brain, which, although not practicable, is, nevertheless, thinkable, would reveal a
causal chain that formed a unique representation of emotional mental experience (Bohr ,
p. 100).
And, finally, he writes:

According to the above-mentioned view on the relation between the processes in the
brain and the psychical experiences, we must, therefore , be prepared to accept the fact that
an attempt to observe the former will bring about an essential alteration in the aware -
ness of volition. Although , in the present case , we can be concerned only with more or
less fitting analogies, yet we can hardly escape the conv iction that in the facts which
are revealed to us by the quantum theory and lie outside the domain of our ordinary
form s of perception we have acquired a means of elucidating general philosophical
problems (Bohr, 101).
These passages are very revealing, for they show that, at his most philo-
sophical moments, Bohr believed that complementarity or reciprocity
is to be found in psychology and elsewhere."
VI
We began this paper by reviewing the debate between Fodor and the
Churchlands on the correct model of cognition. This review led us to a
more general position, that of the need to examine more closely the onto-
logical foundations of each position - eliminative materialism, in the case
of the Church lands, and functionalism, with strong Cartesian overtones,
in the case of Fodor. We then extended this talk of ontology to include
an exposition of Popper 's position, viz. , a dualist, interactionist view
of mind-brain.
Using these three views as representative of the two general posi -
tions of dualism and monism, we then argued for the underdetermination
and incommensurability of all theories of mind-brain. Our arguments
relied on empirical, clinical support as well as conceptual arguments.
As a way out of the underdetermination impasse, we investigated com -
plementarity in physics and used this principle as a springboard (not as
a strict analogy) to develop our Complementarity Model of mind-brain,
which is instrumentalist and anti-realist. We also cited some passages
from Bohr, one of the founders of quantum theory, to the effect that he
expected to see complementarity in other explanations of the world,
including psychology.
We close by indicating the virtues of our model. The Complementarity
Model of mind-brain may be able to effect a reconciliation between the
Fodorian and the Churchlandian positions. There is no need to make
any claim as to what is ' really there' , but only to develop models that
succeed, on the one hand , in predicting and explaining perceptual knowl-
edge (the Churchlands) or predicting and explaining higher cognition
on the other hand (Fodor). On the Complementarity Model, scientists can
act as if there is a mind, that there is causal interaction between mind
and brain, that mind is irreducible to brain, and so on, for the sake of
explanation and prediction. Also under our model, scientists can act as
if there is only brain, which can function in the manner of neural nets
(the churchlands). In a strategy somewhat reminiscent of Kant, the Com-
plementarity Model may allow for a synthesis of what appears to be
two opposing, irreconcilable points of view.
Indeed, this Kantian synthesis approach reaches even further to the

Kantian 'as if' approach, which grounds other areas, such as psychology,
psychiatry, and law. We have discussed the first two cases in our section
on explaining depression. What is often overlooked is that the law func-
tions as if each individual has mental states such as volitions and beliefs
and is thus responsible for his action s. However, the well known insanity
defense presupposes a materialist ontology, viz., we act as if chemical,
anatomical, and physiological abnormalities causally result in the agent's
not having free will, and thus not having responsibility for the committed
act. On a realist basis , this type of switching poses problems of consis-
tency that are obviated by not claiming either that mind and brain both
exist or that brain only exists. In other words, the Complementarity Model
may not get as 'what's really there', although it does have the virtue
of allowing a consistent model to undergird, and comport with, the way
we do act when we perform our functions as psychologists, therapists,
jurists, lawyers, and so on.
Finally, for those who find that the new model advanced here detracts
from the 'mystery' of the mind-brain problem, we would simply say
that indeed it does. If the Complementarity Model dissolves (a la the later
Wittgenstein) the classical mind-body problem, then so be it. However,
our interpretation of any kind of Wittgensteinian dissolution of meta-
physical problems involves also the notion of reposing of the problem.
And that reposing of the problem allows philosophers to join with neu-
roscientists, psychologists, linguists, biologists, physicists, and so on,
to develop new questions within a newly emerged, interdisciplinary field.
That is precisely what cognitive science is now doing, and the new
problems it raises are probably ones that philosophers, we think , can con-
tribute to solving in a significant way. The main import of the
Complementarity Model of mind-brain may not be so much the dissolving
of a classical question as much as it is the indicator of new ways of
conceiving old problems.
California State University, Long Beach ,
Long Beach , U.S.A.
NOTES
1 A network train s up by successive repetitions of a problem with smaller and smaller
errors produced with each iteration. The mo st commonly used function for the error is:
E = 1/2 :I:(Yi - d;)2, where d, is the desired output of unit i, and Yi is its actual output,
where Yi is the sigmoid funct ion 1/(1 + e'). To minimize the error we take the deriva-
tive of the error with respect to w ij, which is the weight between the units i and j. Thus
we obtain
i)E/aw ij = YiY/I - Yj)~j
where ~j = (Yj - d} for output units and ~j = ~wjkYk(1 - Yk)~k for hidden units, where k
represents the number of units in the next layer that unit j is connected to.
The error can be calculated for the link s going into the output units . For h idden units,
however, the derivative depends on the values calculated at all levels that come after it;
thus ~ must be back-propagated through the network to calculate the derivatives. Based
on these equ ations, the algorithm is (\) choose a step-size I) (used to update the weights;
(2) train the network such that for each sample pattern (a) do a forward pass through
the net, producing an output pattern; (b) for all forward outputs, calculate ~j = (Yj - d.);
(c) for all other units , from last layer to first, calculate ~ using the calculation from the
layer after it: ~j - ~ WjkYk(\ - Yk)~k ; (d) and for all weights in the network , change the
weight by f,w ij = -I)YiY/I - Yj)~j'
2 See, for example , some of Swinburne 's books: The Coherence of Theism (\ 977 ; The
Exi stence of God (\ 979); Personal Identity (with Sidney Shoemaker, 1984); and The
Evolution of the Soul (1986). All these books are publi shed by Oxford University Press,
except Personal Identity, which is published by Blackwell ' s. For another, modern defense
of dualism, see J. B. Pratt, Matter and Spirit (New York : Macmillan, 1922/1957).
3 The underdetermination thesis can be proved simply as follows: If Theory T implies
a set of evidential statements E, then T in conjunction with any sentence not included
in T also implies E; but then, since (T & S) -7 E, and, since T & S -7 T, then (T & S)
-7 E. Now replace S with S' and iterate the argument; then replace S' with S" and iterate
again, and so on to infinity. We conclude that an infinite number of different theories
can imply E.
4 The authors are grateful to Professor Henry Fol se for informing us - in a brief dis-
cussion after the earlier version of thi s paper wa s read at the Beijing International
Conference for Philosophy of Science 1992 - of this article by Bohr.
BIBLIOGRAPHY
Allman, William F. (1989). Apprentices ofWonder: Inside the Neural Network Revolution,
Bantam, New York .
Anderson, James A. (1983). ' Cognitive and Psychological Computation with Neural
Networks', in IEEE Transaction s on Systems, Man , and Cybernetics, Vol. SMC-13,
No.5 (September/October 1983) .
Bechtel, William (1988). Philosophy ofMind: Overview for Cognitive Science, Lawrence
Erlbaum, Hillsdale, N.J .
Bechtel , William (1988). Philosophy of Science : Overv iew for Cognitive Science , Lawrence
Erlbaum, Hillsdale, N.J.
Block, Ned (1980). Readings in Philo sophy and Psychology, Harvard University Press,
Cambridge, Mass .
Bohr, Niels (1961) . 'The Quantum of Action and the Description of Nature', in Atomic
Theory and the Description of Nature, Cambridge University Press, Cambridge.
Cartwright, Nancy (1988). How the Laws of Physics Lie, Clarendon Press , Oxford.
Churchland, Patricia Smith (1988) . Neuroph ilosophy, MIT Press, Cambridge, Mass.
Churchland, Patricia Smith (1980) . 'A Perspective on Mind-Brain Research ' , Journal of
Philosophy 77(4) , pp. 185-207.
Churchland, Patricia Smith (1982) . 'Mind-Brain Reduction: New Perspectives from the
Philosophy of Science', Neuroscience 7(5), pp. 1041-1047.
Churchland, Paul M. (1988) . Matter and Consciousness, MIT Press, Cambridge, Mass .
Churchland, Paul M. (1980). ' Critical Notice : Joseph Margolis: Persons and Minds : The
Prospects of Nonreductive Materialism', Dialogue 19, pp. 461-469.
Churchland, Paul M. (1981) . 'Eliminative Mater ialism and the Propositional Attitudes ' ,
Journal of Philo sophy 78, pp . 67-90.
Churchland, Paul M. (1985). 'Reduction, Qualia and the Direct Introspection of Brain
States' , Journal of Philosophy 82, pp. 8-28.
Churchland, Paul M. and Patricia Smith Churchland (1990) . ' Could a Machine Talk? '
Scientific American 262, pp . 32-37.
Cummins, Robert (1983). The Nature of Psychological Explanation, MIT Press,
Cambridge, Mass.
Dennett, Daniel (1981) . Brainstorms, MIT Press , Cambridge, Mass.
Fodor, Jerry A. et al. (1988). 'Connectionism and Cognitive Architecture ', Cogn ition ,
pp. 236-258.
Fodor, Jerry A. et al. (1975) . The Language of Thought , MIT Pres s, Cambridge, Mass.
Fodor, Jerry A. (1981). Representations : Philosophical Essays on the Foundations of
Cogn itive Science, MIT Press, Cambridge, Mass.
Fodor Jerry A. (1968). Psychological Explanation , Random House, New York .
Fodor, Jerry A. (1988) . 'Connectionism and Cognit ive Arch itecture: A Critical Analysis.
Cognitive Science Memorandum' , COGMEM 29, 1987, Cognition .
Gardener, Howard (1985). The Mind 's New Science: A History ofthe Cognitive Revolution,
Basic Books, New York .
Garson, James W. (1990). 'Cognition Without Classical Architecture ', Unpublished
pape r delivered at the APA Eastern Division meeting, Boston, Mass. December 1990.
Grossberg, S. (1980). ' How Doe s the Bra in Build a Cognitive Code?' Psychological
Review 87, pp. I-5 1.
Haugeland, John. (1980) . Mind Design: Philosophy, Psychology. Artificial Intelligence,
MIT Pres s, Cambridge, Mass.
Hopfield, J. J. (1982) . ' Neural Networks and Physic al Systems With Emergent Collective
Computational Abilities', in Proc. Natl. Acad. Sci. 79, pp. 2554-2558.
Hopfield, J. J. (1986) . 'Computing With Neural Circuits : A Model', Scien ce 23, August,
pp. 625-633.
Jones , W. T. and Hoskins, Josiah (1987). Back-Propagation: A Generalized Delta-Learning
Rule, BYTE, October, pp. 155-162.
Kuhn, Thomas, S. (1962). The Structure of Scientific Revolutions, University of Chicago

Press, Chicago.
Luplin, J. (1984). Synapses. Circuits, and the Beginnings of Memory, MIT Press,
Cambridge, Mass.
McClelland, J. L. and Rumelhart, D. E. (1986) . Parallel Distributed Processing:
Explorations in the Microstructure of Cognition Vol. 2: Psychological and Models,
Bradford Books/MIT Press, Cambridge, Mass.
Murphy, Dennis L. and Costa , Jonathan L. (1975) . ' Utilization of Cellular Studies of
Neurotransmitter-related Enzymes and Transport Processes in Man for the Investigation
of Biological Factors in Behavioral Disorder' , in Joseph Mendels (ed.), The
Psychobiology of Depression, Spectrum Publication, New York, pp. 27-46.
Pinker, Steven et al., ' On Language and Connectionism: Analysis of a Parallel Distributed
Processing Model of Language Acquisition ' , Occasional Paper #33. Request for
reprints : Pinker, Dept. of Brain and Cognitive Sciences, MIT Press, Cambridge, Mass.,
02139 .
Popper, Karl R. and Eccles, J. (1977) . The Self and Its Brain , Springer-Verlag, New
York.
Prange, Arthur , J., Jr., M.D. (1975). 'Pharmacotherapy of Depression', in Frederick F.
Flach, M.D. and Susanne C. Draghi , M.D. (eds.), The Nature and Treatment of
Depression, John Wiley, New York.
Quine , W. V. (1975). ' On Empirically Equivalent Systems of the World ', Erkenntnis 9,
pp.313-328.
INDEX OF NAMES
Abbe of Galiana 17 Boler, J.F. 91, 93

Ackermann, Bob 146 Boring, E. 18
Acquinas, St. Thomas 78, 91, 92 Born, Max 149,285,337
Adams , R.G. 113 Bouguer, Pierre 395
Anderson, J.R. 106 Boyd, R. 19,242,440,443
Araki, H. 311, 312 Boyle, R. 223, 367, 411
Aristarchus 412 Bradley 378
Aristotle 12, 75, 78, 79, 80, 89, 92, 94, Bridgman, P.W. 163
195, 196,205,245,411,415 Briskman 226
Armstrong, D.H. 53, 75, 76, 77, 78, 79, Brown, C. 62
80,81 ,82,83,84,85,86,88,89,90, Brown, James R. 90
91,93,94 Brynjolfsson, Ari 173
Aspect 381 Burian, Dick 19
Avenarius 14
Avicenna 91,93 Cairnes, John Elliot 442
Ayer, Sir Alfred J. xi, xiv, xviii, xxi Canton, J. 274
Camap, Rudolf xiii, xiv, xvii, xix, xx,
Barnes , Winston xii xxi, 109, 110,220
Beccaria, G.B. 274 Carruthers, Peter 71
Beckham, Ernest E. 470, 471 Cartwright, Nancy 20, 131, 132, 139,
Bell, J.S. 281, 293, 299, 313, 317 145,146,147,172,407,473
Berkeley, G. 35, 40, 42, 56, 125, 126, Carus, Paul 2
210,223 Castaneda, Hector-Neri 368
Bernal, J.D. 248, 249 Chalmers, Alan 26, 27, 40
Bjerrum, J. 154 Churchland, Patricia 145, 467, 468, 473
Blackburn, Simon 35, 40,129 Churchland, Paul 467, 468, 473
Blackett, P.M.S. 173, 175, 179 Churchlands 475
Blondlot, Rene 259 Chwolson, a.D. 144
Bogdanov 14 Clark, G.L. 159
Bogen, James 119,120, 171, 184,398 Cline, David 132
Bohm, D. 309, 312, 314, 317, 325, 362, Cohen, J. 71
386,401 Cohen , L.J. 65
Bohr, Niels 11, 12, 15, 17,56, 122, 123, Cohen, Robert S. xi
141, 142, 159, 208, 209, 210, 211, Compton, Arthur 410
212, 216, 217, 224, 279, 280, 281, Confucius 68
282, 283, 284, 285, 286, 289, 290, Copernicus xvii, 289, 296, 412, 414, 415
291, 292, 293, 294, 295, 296, 297, Copleston 91, 92
309, 312, 325, 343, 344, 345, 346, Coulson, Charles Alfred 149, 150
347, 348, 349, 350, 353, 364, 365, Cox, A. 182
366, 369, 381, 388, 389, 408, 414, Crick, F. 410
415,474, 475 Currie , Gregory 242
481
482 INDEX OF NAMES
Ewald, E.E. 155

D' Alembert,1. 338
D'Espagnat, B. 299, 301, 308, 310, 312, Fan Dainian xiii
317,319 Faraday, M. 274
Dalrymple 182 Feigl, Herbert xv, xvii, xx, xxi
Da1ton,1. 325, 415 Feng Youlan xviii, xix
Daneri, A. 305 Feyerabend, P. xxi, 126,411 ,412,420
Darwin, Ch. 221, 222, 327, 410 Field, Hartry 1,25,47,48
Davidson30 Fine, Arthur24, 27, 37, 38, 39, 293, 297,
De Broglie, L. 312, 313, 351, 386 405,415,419
De Sitter, W. 271 Fleischmann, Martin 257
De Vries, Gerard 34 Fletcher, Harvey 260
Dehmelt , H. 146 Fock 282, 369
Democritus325,415 Fodor, Jerry 467, 468, 469, 473, 475
Denitt, M. 62, 71 Folse, H.I. 297
Derrida, 1. 266 Foucault, 1. 410
Descartes. Rene 223, 245, 247,274,367, Frank, Philipp xvii
415,469 Frege, G. 3
Devitt, M. 7, 10, 20, 30, 31 French, A.P. 8
Dickinson, Roscoe 161 Fresnel, A.I. 410
Dieks, D. 386 Friedman, Milton 335, 429, 430
Dirac, P. 152,274,314 Friedmann, A. 271
Dobzhansky, T. 220 Fry, Harry 160
Doell 182 Fung Yu-Ian xv
Duhem, P. 14, 21
Dummett, M. 242, 243 Galen 415
Duns Scotus, J. 78, 79, 80, 82, 83, 84, Galileo 223, 247, 360, 406, 409, 410,
85,86,91 ,92,93 413,414
Durkheim, E. 120 Gell-Mann, Murray 132
Gerlach 390
Eccles, John 469 Gilbert, W. 274
Eddington, A.S. 22, 23, 57, 64 Gilson, E. 93, 94
Ehrenhaft, Felix xvii, 143, 144 Glashow, Sheldon 416
Einstein, Albert xvii 14, 36, 124, 126, Godel, K. xvii, 15
142, 208, 209, 210, 212, 216, 217, Gong-sun Long 286
220, 245, 251, 254, 255, 270, 271, Goodman,Nelson35, 51, 216, 339, 340,
279, 280, 281, 285, 286, 289, 290, 341,375
291, 292, 293, 294, 295, 296, 297, Grim, P. 48
299, 307, 312, 325, 328, 343, 345, Gullstrand, A. 143
353, 361, 364, 369, 385, 406, 410,
414 Haack, Susan 8, 90
Eliot, T.S. 27 Hacking, Ian 20, 45, 52, 53, 131 , 132,
Ellis, Brian 42, 47 136,145,439,446
Engel-Tiercelin, C. 93, 94 Haeckel, Ernst 220, 221
Engels, F. xii Hahn, Hans xvii
Epicurus xv, 415 Hanson 126
Eucken, Rudolf xvii Harkins, D.W. 164
Everett 317,326,392 Harper,W. 129
INDEXOF NAMES 483
Harre, R. 88,93, 367 Jammer, Max 280, 368, 370

Harris, C.R.S. 91, 92 Jennings, R. 40, 41, 42
Hartshorne, C. 8 Jin Yue-lin xix, 317. 325
Harvey, W. 274 Jones, Steven257
Hausman, D. 443 Jordan 149
HeLin xix Joule,JamesPrescott120, 121 , 122,395,
Hegel,G.W.F. xii 402
Heidegger, M. 104 Ju Zhang xi
Heisenberg, W. 15, 142, 149, 156,208,
270, 280, 292, 294, 297, 317, 326, Kaiser, M. 129, 171
365,366,367,388 Kant, Immanuelxv, xviii,xix, 13,35,36,
Heitler, Walter 152, 154, 155, 156, 157, 40,49,51,56,62,104,108,160,197,
158, 159, 161, 162, 164, 165. 166 198,222,245,408,472,473,475
Heller, Mark 6, 7, 8 Kepler,1. 274, 360, 406, 407, 409, 413
Hempel, C.G. xi Keynes, 1.M. 441
Hepp, K. 312 Kitcher, Philip 335, 452, 453, 454, 455,
Heraclitus 59 456,457,458,459,460,461 ,464
Herschel, D. 398 Kohn, Alexander 256
Hertz, H. 269, 274 Konstantinowsky 143
Herzberg, Karl 154 Kraft, Victorxvii, xxi
Hess, Harry 178, 182 Kripke, S. 376
Hilbert, D. 314, 390 Kuhn, ThomasS. xiv,xxi, 35, 45, 47,51,
Hiley, BJ. 314 52,126,325,401 ,411 ,412,420,471
Hintikka, 1. 53 Kuipers, Theo A.F. 242
Hipparchus 405 Kusch, Martin 51
Hobbes, T. 367
Holcomb, H.R. 366 Lakatos, I. 360
Hollitscher, Walterxvii Lamarck, 1. 410
Holmes,Arthur 178, 180 Lambert, Johann Heinrich395
Hong Qian xvii, xviii, xix, xx, xxi Laudan,Larry 21, 265, 297. 325
Hooke,R. 274, 410 Lavoisier, A.L. 269, 271, 410
Hooker, C. 145 Lawson,Hilary 266
Horwich,P. 53 Lawson,Tony445
Howard, Don 293, 297 Layzer, David 319
Howson, C. 146 Leber, WilliamR. 470, 471
Hu Xinhe 367 Lee, Joseph C.T. xi
Hubble, E. 271 Leeds, Stephen 31
Hull, D. 401 Leibniz, G.w. 125,255 ,368
Hume, D. 63 Lemaitre, G. 271
Hund, F. 164 Lenin, Vladimir I. 14, 15
Husser!, E. xviii Leplin 19,20, 145
Huxley, T.H. 221 Leucippus 415
Huygens, Chr. 274, 410 Levin,Michael28
Lewis, David 32
Irvine, AndrewD. 90, 91 Lewis,G.N.153, 154,159,160,161,162
Irving, E. 180 Liang Qi-chao xiv,xvii
Lipton,P. 243
James, William30,221 Locke, J. 223, 247
484 INDEXOF NAMES
Lockhart , C.M. 312 Neurath, Otto xiii, xiv, xv, xvii, xix, xxi,
Loinger, A. 305 216,226
London, Fritz 152, 154, 155, 156, 157, Newton, Isaac, 17, 125, 126, 140, 225,
158,159, 160, 161, 164, 165, 166 245, 255, 270, 271, 272, 274, 330,
Lopatin 14 360, 406, 408, 409, 410, 411, 414,
Lorentz, H.A. 274, 362 415
Lucretius 415 Newton-Smith , W.H. 242
Lycan, Bill 19 Niiniluoto, Hkka 52, 53, 186, 192,242
Nola, Robert 3, 9
Mach, E. xix, xviii, xx, 14, 125, 126,405
Machida, S. 311 Oddie, G. 242
Mackie, John 25, 40 Oersted, H.C. 274
Madden, E.H. 88, 93 Opdyke, N.D. 182
Mao Zedong 64, 270 Oreskes, N. 172
Margenau, H. 366, 367 Osherson, D.N. 113
Markov, M.A. 282 Osiander, Andreas 289,296,297,405
Marx, Karl 14, 438
Matheson, Car129, 42 Parmenides 378
Matthews, D.H. 178, 180 Pauli, W. 283
Maturana, H. 108 Pauling, Linus 153, 155. 161, 162, 163,
Maxwell, Grover 224 164,410
Maxwell , J.C. 141, 147,269,274,361 Pearson , K. 14
Mayo, Deborah 19 Perrce,C.S.8,9,53,82, 87, 88,90,91,
McClelland, J.R. 111 92,93,94
McMullin, Eman 464 Pesch!, M. 112, 116
McTaggart 220 Pfander, Alexander 155
Mendeleev, Dimitri 122, 123, 138,407 Pickering 129
Merrnin, N.D. 299 Pigden, Charles 25
Merton,R.K.252,400 Pitt, J.C. 241
Michael, Fred 93, 94 Planck, Max 254, 333, 388, 407, 415
Michelson, A.A. 269,414 Plato 12,56, 64, 68, 75, 78, 81, 83,90,
Mill, J.S. 433, 442 91,123,368
Millikan, Robert A. 138, 142, 143, 144, Podolsky, B. 280, 299
145,260 Poincare, J.H. 14
Misra, B. 312 Pons, B. Stanley 257
Moore, E. 29, 91, 93 Popper, Karl xxi, 57, 64, 187,207,223,
Morley 414 226,242,467,469,470,475
Morrison, M. 145,146 Posner, MJ. 113
Mosely, Henry 122, 123, 124 Priestley, J. 269
MUller, Johannes 16, 18 Prigogine, 1. 326
Mulliken, R.S. 153, 155, 162, 164, 165, Primas, H. 314
166 Prosperi, G.M. 305
Musgrave, Alan 241 Putnam, Hilary 1,2,3,4,8, 19,30,31,
32,33,34,35,39,45,46,47,48,49,
Namiki, M. 311 52, 53, 62, 69, 227, 236, 237, 241,
Needham xiv 242,270,440
Neurath, Marie xiii
Qiu Renzong xi
INDEXOF NAMES 485
Quine,W.V. xiv, xix, xxi, 265, 267, 408, Stapp 317
411 Sterelny, Kim 452, 453, 454, 455, 456,
457,458,459,460,461 ,464
Rabi,1. 408 Stem 390
Radakovic, Theodor xvii Stillings, N.A. 113
Rayleigh,J.W.S. 259 Stoke, G.G. 269
Reichenbach, Hans xv, xvii Stove, David 40
Rescher, N. 47, 50, 53, 432 Strawson, Peter xxi, 64
Richter,Burton 406 Sun Yat-sen xiii
Rickert xviii Suppes, Patrick 223
Robbins, L. 442, 444, 446 Swinburne, Richard 470
Rodebush, Worth 159 Switinjk, Zeno 146
Rorty, Richard 35, 266, 267 Sykes 183
Rosen, N. 280, 299
Rosenberg, A. 445 Tarski, Alfred xvii, 24, 26, 27, 28, 30,
Rosenfeld, Leon 209, 286 425
Rumelhart, D.E. III Taylor, Barry 243
Rumford, B. 395, 397, 398, 399, 400, Teller, Paul 326, 369
402 Thagard,Paul 412
Runcom, S.K. 175, 179 Thales 59
Russell, B. xv, 63 Thomson, U. 153
Rutherford, E. 224, 415 Ting 406
Tooley, M. 91
Salam 119, 140, 147 Tscha Hung xi. xii, xiii, xiv, xv, xvii
Scheler xviii Tuomela, Raimo 46, 47, 53
Schlick, Moritz xiii, xiv, xv, xvii, xviii Tycho, Brahe 413
xix, xx, xxi
Schopenhauer, A. 51 Ullman-Margalit 437
Schrodinger, E. 149, 154, 155, 158. 159, Unger, Peter 49
163,328,335,390,392
Sejnowski, Terry 468 Van der Waals, J. 155
Sellars, Wilfrid 22, 48, 126, 131, 132, Van Fraassen,Bas 24,34,131,132,134,
138, 140 135, 136, 138, 139, 140, 141, 144,
Senior, Nassau 433, 442 145, 147, 230, 242, 265, 267, 268,
Shapere, Dudley 268, 272 339,341,381 ,405
Shimony 317, 322 Van Vleck, 1. 159
Shu Xingbei 286 Vine, FJ . 178, 180
Sidgwick,Neville 149 von Weizsacker 343
Simon, Herbert 429,430 Von Neumann 389
Sklar, Lawrence 337, 338
Slater 164 Waismann, Friedrich xvii
Sliv 282 Walker, R.C.S. 53
Smith, Adam 437, 438 Wallner, F. 113
Socrates 81, 83 Waterson, U . 259
Sommerfeld, A. 154, 155, 161 Watson, 410
Spencer, H. 97 Wegener, Alfred 172, 177, 180, 190
Stachel, John 145 Weinberg 119, 140, 147
StaW 410 Weiss, P. 8
486 INDEXOF NAMES
Weyl, H.155 Worrall, J. 241

Wheeler, John Archibald 15, 16, 17, 224 Wright, Crispin 243
Wheland, George Willard 162, 163 Wright, Sewall 220
Whitehead, A.N. 12
Whorf411 Xenophanes 12
Wigner, Eugene 13,382, 383, 390
Wilson, E.O. 128 Young 330, 410
Wilson, Tuzo 178, 181, 190
Winde1band xviii Zerner, F. 143
Wittgenstein, Ludwig xx, xxi, 27, 47, 55, Zhang Huaxia 57, 64
103,113,471,472,476 Zhexue Yicong xxi
Woodward, James 119, 120, 129, 171, Zhuang Zhou 71
184,398 Zilsel, Edgar xiv, xvii
Boston Studies in the Philosophy of Science
Editor: Robert S. Cohen, Boston University
1. M.W. Wartofsky (ed.): Proceedings ofthe BostonColloquium for the PhilosophyofScience.

196111962. [Synthese Library 6] 1963 ISBN 90-277-0021-4
2. R.S. Cohen and M.W. Wartofsky (eds.): Proceedings of the Boston Colloquium for the
PhilosophyofScience. 196211964. In Honor of P. Frank. [Synthese Library 10] 1965
ISBN 90-277-9004-0
Philosophy of Science, 1964/1966. In Memory of Norwood Russell Hanson. [Synthese
Library 14] 1967 ISBN 90-277-0013-3
Philosophy ofScience. 1966/1968. [Synthese Library 18] 1969 ISBN 90-277-0014-1
Philosophy ofScience. 196611968. [Synthese Library 19] 1969 ISBN 9()'277-OO15-X
6. R.S. Cohen and RJ. Seeger (eds.): Ernst Mach. Physicist and Philosopher. [Synthese
Library 27] 1970 ISBN 90-277-0016-8
7. M. Capek: Bergson and Modem Physics. A Reinterpretation and Re-evaluation. [Synthese
Library 37] 1971 ISBN 9()'277-o186-5
8. R.C. Buck and R.S. Cohen (eds.): PSA 1970. Proceedings of the 2nd Biennial Meeting of the
Philosophy and Science Association (Boston, Fall 1970). In Memory of Rudolf Camap.
[Synthese Library 39] 1971 ISBN 9().277-0187-3; Pb 9()'277-o309-4
9. A.A. Zinov'ev : Foundations ofthe Logical TheoryofScientific Knowledge (ComplexLogic).
Translated from Russian. Revised and enlarged English Edition. with an Appendix by G.A.
Smimov, E.A. Sidorenko, A.M. Fedina and L.A. Bobrova. [Synthese Library 46] 1973 .
ISBN 90-277-0193-8; Pb 9().277-o324-8
10. L. Tondl: ScientificProcedures. A Contribution Concerning the Methodological Problems of
Scientific Concepts and Scientific Explanation.Translated from Czech. [Synthese Library
47] 1973 ISBN 9()'277-o147-4; Pb 9O-277-0323-X
11. RJ. Seeger and R.S. Cohen (eds .): Philosophical Foundations of Science. Proceedings of
Section L, 1969, American Association for the Advancement of Science. [Synthese Library
58]1974 ISBN 9()'277-039()'6; Pb 9().277-o376-0
12. A. Griinbaum: Philosophical Problems of Space and Times. 2nd enlarged ed. [Synthese
Library 55] 1973 ISBN 9()'277-0357-4; Pb 90-277-0358-2
13. R.S. Cohen and M.W. Wartofsky (eds .): Logical and Epistemological Studies in Contem·
porary Physics. Proceedings of the Boston Colloquium for the Philosophy of Science.
1969172, Part I. [Synthese Library 59] 1974 ISBN 9()'277-0391-4; Pb 90-277-0377-9
14. R.S. Cohen and M.W. Wartofsky (eds.): Methodological and Historical Essays in the
Natural and Social Sciences. Proceedings of the Boston Colloquium for the Philosophy of
Science, 1969n2, Part II. [Synthese Library 60]1974
ISBN 9().277-0392-2; Pb 90-277-0378-7
15. R.S. Cohen, I.I. Stachel and M.W. Wartofsky (eds .): For Dirk Struik: Scientific, Historical
and Political Essays in Honor of Dirk J. Stroik. [Synthese Library 61]1974
ISBN 9()'277-0393-0; Pb 90-277-0379-5
16. N. Geschwind: SelectedPaperson Languageand theBrains. [Synthese Library 68]1974
ISBN 9().277-0262-4; Pb 90-277-0263-2
17. B.G. Kuznetsov: Reason and Being. Translated from Russian. Edited by C.R. Fawcett and
R.S. Cohen. 1987 ISBN 90-277-2181-5
18. P. Mittelstaedt: Philosophical Problems ofModem Physics. Translatedfrom the revised 4th
Germaneditionby W. Riemer and ee1ited by R.S. Cohen. [Synthese Library95] 1976
ISBN 90-277-0285-3; Pb 90-277-0506-2
19. H. Mehlberg: Time. Causality, and the Quantum Theory. Studies in the Philosophy of
Science. Vol. I: Essay on the Causal Theory ofTime. Vol. II: Time in a Quantized Universe.
Translatedfrom French. Edited by R.S. Cohen. 1980
Vol. I: ISBN 90-277-0721-9; Pb 90-277-1074-0
Vol. II: ISBN 90-277-1075-9; Pb 90-277-1076-7
20. K.F. Schaffnerand R.S. Cohen (eds.): PSA 1972.Proceedings of the 3rd BiennialMeetingof
the Philosophy of Science Association (Lansing, Michigan, Fall 1972). [Synthese Library
64] 1974 ISBN 90-277-0408-2; Pb 90-277-0409-0
21. R.S. Cohen and 1.1. Stachel (eds.): Selected Papers of Leon Rosenfeld. [Synthese Library
100] 1979 ISBN 90-277-0651-4; Pb 90-277-0652-2
22. M. Capek (ed .): The Concepts of Space and Time. Their Structure and Their Development.
[SyntheseLibrary74] 1976 ISBN90-277-0355-8; Pb 90-277-0375-2
23. M. Grene: The Understanding of Nature. Essays in the Philosophy of Biology. [Synthese
Library 66] 1974 ISBN 90-277-0462-7; Ph 90-277-0463-5
24. D. Ihde: Technics and Praxis. A Philosophy of Technology. [Synthese Library 130] 1979
ISBN9o-277-0953-X; Pb 90-277-0954-8
25. J. Hintikka and U. Remes: The Method of Analysis. Its Geometrical Origin and Its General
Significance. [Synthese Library75) 1974 ISBN 90-277-0532-1 ; Pb 90-277-0543·7
26. J.E. Murdochand E.n. Sylla (eels.): The Cultural Context ofMedieval Learning. Proceedings
of the First International Colloquium on Philosophy, Science, and Theology in the Middle
Ages, 1973. [Synthese Library76] 1975 ISBN 90-277-0560-7; Pb 90-277-0587-9
27. M. Grene and E. Mendelsohn (eds.): Topics in the Philosophy ofBiology. [SyntheseLibrary
84] 1976 ISBN90-277-o595-X; Pb 90-277-0596-8
28. J. Agassi: Science in Flux. [Synthese Library80] 1975
ISBN90-277-0584-4; Pb 90-277-0612-3
29. J.J. Wiatr (ed.): Polish Essays in the Methodology ofthe Social Sciences. [Synthese Library
131] 1979 ISBN90-277-0723-5; Pb 90-277-0956-4
30. P. Janich: Protophysics of Time. Constructive Foundation and History of Time Measure-
ment.Translatedfrom German, 1985 ISBN90-277-0724-3
31. R.S. Cohenand M.W. Wartofsky(eels.): Language, Logic, and Method. 1983
ISBN 90-277-0725-1
32. R.S. Cohen, C.A. Hooker, A.C. Michalos and J.W. van Evra (eds.): PSA 1974. Proceedings
of the 4th Biennial Meeting of the Philosophy of Science Association. [Synthese Library
101] 1976 ISBN90-277-0647-6; Ph 90-277-0648-4
33. G. Holton and W.A. Blanpied (eds .): Science and Its Public. The Changing Relationship.
[Synthese Library96] 1976 ISBN90-277-0657-3; Ph 90-277-0658-1
34. M.D. Grmek, R.S. Cohen and G. Cimino (eds.): On Scientific Discovery. The 1977 Erice
Lectures. 1981 ISBN 90-277-1122-4; Pb 90-277-1123-2
35. S. Amsterdamski: Between Experience and MeUJphysics. Philosophical Problems of the
Evolutionof Science. Translatedfrom Polish. [Synthese Library77] 1975
ISBN90-277·0568-2; Pb 9o-277'{)580-1
36. M. Markovic and G. Petrovic (eels.): Praxis. Yugoslav Essays in the Philosophy and
Methodology of the SocialSciences.[Synthese Library 134] 1979
ISBN90-277.{)727·8; Pb 90-277-0968-8
37. H. von Helmholtz: Epistemological Writings. The Paul Hertz I Moritz Schlick Centenary
Edition of 1921. Translated from German by M.F. Lowe. Edited with an Introduction and
Bibliography by R.S. Cohen and Y. Elkana. [Synthese Library 79]1977
ISBN 9O-277-029O-X; Pb 90-277-0582-8
38. R.M. Martin: Pragmatics, TruthandLanguage. 1979
ISBN 90-277-0992-0; Pb 90-277-0993-9
39. R.S. Cohen, P.K. Feyerabend and M.W. Wartofsky (eds.) : Essays in Memory of Imre
Lakatos. [Synthese Library 99]1976 ISBN 90-277-0654-9; Pb 90-277-0655-7
40. Not published.
41. Not published.
42. H.R. Maturana and FJ. Varela: Autopoiesis and Cognition. The Realization of the Living.
With a Preface to •Autopoiesis' by S. Beer. 1980
ISBN 90-277-1015-5; Pb 90-277-1016-3
43. A. Kasher (ed.): Language in Focus: Foundations. Methods and Systems. Essays in Memory
ofYehoshua Bar-Hillel. [Synthese Library 89]1976
ISBN 90-277-0644-1; Pb 9Q-277-0645-X
44. T.D. Thao: Investigations into the Origin ofLanguage and Consciousness. 1984
ISBN 90-277-0827-4
45. Not published.
46. P.L. Kapitza: Experiment. Theory. Practice. Articles and Addresses. Edited by R.S. Cohen.
1980 ISBN 90-277-1061-9; Pb 90-277-1062-7
47. M.L. Dalla Chiara (ed.): ItalianStudies in the Philosophy ofScience. 1981
ISBN 90-277-0735-9; Pb 90-277-1073-2
48. M.W. Wartofsky: Models. Representation and the Scientific Understanding. [Synthese
Library 129]1979 ISBN 90-277-0736-7; Pb 90-277-0947-5
49. T.D. Thao: Phenomenology and Dialectical Materialism. Edited by R.S. Cohen. 1986
ISBN 90-277-0737-5
50. Y. Fried and J. Agassi: Paranoia. A Study in Diagnosis. [Synthese Library 102]1976
ISBN 90-277-0704-9; Pb 90·277-0705-7
51. K.H. Wolff: Surrender and Cath. Experience and Inquiry Today. [Synthese Library 105]
1976 ISBN 90-277-0758·8; Pb 90-277-0765-0
52. K. Kosik: Dialectics of the Concrete. A Study on Problems of Man and World. 1976
ISBN 90-277-0761-8; Pb 90-277-0764-2
53. N. Goodman: The StructureofAppearance. [Synthese Library 107]1977
ISBN 90-277-0773-1; Pb 90-277-0774-X
54. H.A. Simon: Models of Discovery and Other Topics in the Methods of Science. [Synthese
Library 114] 1977 ISBN 90-277-0812-6; Pb 90-277-0858-4
55. M. Lazerowitz: The Language of Philosophy. Freud and Wittgenstein. [Synthese Library
117]1977 ISBN 90-277-0826-6; Pb 90-277-0862-2
56. T. Nickles (ed.): Scientific Discovery. Logic. and Rationality. 1980
ISBN 90-277-1069-4; Pb 90-277-1070-8
57. 1. Margolis: Persons and Mind. The Prospects of Nonreductive Materialism. [Synthese
Library 121]1978 ISBN 90-277-0854-1; Pb 90-277-0863-0
58. G. RadnitzkY and G. Andersson (eds.): Progress and Rationality in Science. [Synthese
Library 125]1978 ISBN 90-277-0921-1; Pb 9Q-277-0922-X
59. G. Radnitzky and G. Andersson (eds.): TheStructure and Development ofScience. [Synthese
Library 136]1979 ISBN 90-277-0994-7; Pb 90-277-0995-5
60. T. Nickles (ed.): Scientific Discovery. Case Studies. 1980
ISBN 90-277-1092-9; Pb 90-277-1093-7
61. M.A. Finocchiaro: Galileo and the Art of Reasoning. Rhetorical Foundation of Logic and
Scientific Method. 1980 ISBN 90-277-1094-5; Pb 90-277-1095-3
62. W.A. Wallace : Prelude to Galileo. Essays on Medieval and 16th-Century Sources of
Galileo's Thought. 1981 ISBN 90-277-1215-8; Pb 90-277-1216-6
63. F. Rapp: Analytical Philosophy ofTechnalogy. Translated from German . 1981
ISBN 90-277-1221-2; Pb 9O-277-1222..<J
64. R.S. Cohen and M W. Wartofsky (eds.): Hegel and the Sciences. 1984
ISBN 9O-277..<J726-X
65. J. Agassi: Science and Society. Studies in the Sociology of Science. 1981
ISBN 90-277-1244-1 ; Pb 90-277-1245-X
66. L. Tondl : Problems of Semantics. A Connibution to the Analysis of the Language of
Science. Translated from Czech. 1981 ISBN 9O-277..<J148-2; Pb 9O-277..<J316-7
67. J. Agassi and R.S. Cohen (eds.): Scientific Philosophy Today. Essays in Honor of Mario
Bunge. 1982 ISBN 9O-277-1262-X; Pb 90-277-1263-8
68. W. Krajewski (ed.): Polish Essays in the Philosophy of the Natural Sciences. Translated
from Polish and edited by R.S. Cohen and C.R. Fawcett. 1982
ISBN 90-277-1286-7; Pb 90-277-1287-5
69. J.H. Fetzer. Scientific Knowledge . Causation, Explanation and Corroboration. 1981
ISBN 90-277-1335-9; Pb 90-277-1336-7
70. S. Grossberg : Studies of Mind and Brain. Neural Principles of Learning, Perception,
Development, Cognition, and Motor Control. 1982
ISBN 90-277-1359-6; Pb 90-277-1360-X
71. R.S. Cohen and M.W. Wartofsky (eds.): Epistemology, Methodology, and the Social
Sciences . 1983. ISBN 90-277-1454-1
72. K. Berka: Measurement. Its Concepts, Theories and Problems. Translated from Czech. 1983
ISBN 90-277-1416-9
73. G.L. Pandit The Structure and Growth of Scientific Knowledge . A Study in the Methodol-
ogy of Epistemic Appraisal. 1983 ISBN 90-277-1434-7
74. A.A. Zinov 'ev: Logical Physics. Translated from Russian. Edited by R.S. Cohen. 1983
[see also Volume 9] ISBN 9O-277..<J734-0
75. G-G. Granger. Formal Thought and the Sciences ofMan. Translated from French. With and
Introduction by A. Rosenberg. 1983 ISBN 90-277-1524-6
76. R.S. Cohen and L. Laudan (eds.): Physics , Philosophy and Psychoanalysis. Essays in Honor
of Adolf Grunbaum. 1983 ISBN 90-277-1533-5
77. G. Bohme , W. van den Daele, R. Hohlfeld, W. Krohn and W. Schafer: Finalization in
Science. The Social Orientation of Scientific Progress. Translated from German. Edited by
W. Schafer, 1983 ISBN 90-277-1549-1
78. D. Shapere: Reason and the Search for Knowledge. Investigations in the Philosophy of
Science . 1984 ISBN 90-277-1551-3; Pb 90-277-1641-2
79. G. Andersson (ed.): Rationality in Science and Politics. Translated from German. 1984
ISBN 90-277-1575-0; Pb 90-277-1953-5
80. P.T. Durbin and F. Rapp (eds.): Philosophy and Technology. [Also Philosophy and
Technology Series, Vol. 1]1983 ISBN 90-277·1576-9
81. M. Markovic : Dialectical Theory ofMeaning . Translated from Serbo-Croat. 1984
ISBN 90-277-1596-3
82. R.S. Cohen and M.W. Wartofsky (eds.) : PhysicalSciences and HistoryofPhysics. 1984.
ISBN 90-277-1615-3
83. E. Meyerson: The Relativistic Deduction. Epistemological Implications of the Theory of
Relativity. Translated from French. With a Review by Albert Einstein and an Introduction by
Mili6 Capek. 1985 ISBN 90-277-1699-4
84. R.S. Cohen and M.W. Wartofsky (eds.): Methodology, Metaphysics and the History of
Science. In Memory of Benjamin Nelson. 1984 ISBN 90-277-1711-7
85. G. Tamas: The LogicofCategories. Translated from Hungarian. Edited by R.S. Cohen. 1986
ISBN 90-277-1742-7
86. S.L. de C. Fernandes: Foundations of Objective Knowledge. The Relations of Popper's
Theory of Knowledge to That of Kant. 1985 ISBN 90-277-1809-1
87. R.S. Cohen and T. Schnelle (eds.): Cognition and Fact. Materials on Ludwik Fleck. 1986
ISBN 90-277-1902-0
88. G. Freudenthal: Atom and Individual in the Age ofNewton. On the Genesis of theMechanis-
tic World View. Translated from German. 1986 ISBN 90-277-1905-5
89. A. Donagan, A.N. Perovich Jr and M.V. Wedin (eds.): Human Nature and Natural
Knowledge. Essays presented to Marjorie Grene on the Occasion of Her 75th Birthday. 1986
ISBN 90-277-1974-8
90. C. Mitcham and A. Hunning (eds.) : Philosophy and Technology 11. Information Technology
and Computers in Theory and Practice. [Also Philosophy and Technology Series, Vol. 2]
1986 ISBN 90-277-1975-6
91. M. Grene and D. Nails (eds.): Spinoza and the Sciences. 1986 ISBN 90-277-1976-4
92. S.P. Turner. The Searchfor a Methodology of Social Science. Durkheim, Weber, and the
19th-CenturyProblem of Cause, Probability, and Action. 1986. ISBN 90-277-2067-3
93. I.C. Jarvie: Thinking aboutSociety. Theory and Practice. 1986 ISBN 90-277-2068-1
94. E. Ullmann-Margalit (ed.): The Kaleidoscope of Science. The Israel Colloquium: Studies in
History, Philosophy, and Sociology of Science, Vol. 1. 1986
ISBN 90-277-2158-0;Pb 90-277-2159-9
95. E. Ullmann-Margalit (ed.): The PrismofScience. The Israel Colloquium: Studies in History,
Philosophy, and Sociology of Science, Vol. 2. 1986
ISBN 90-277-2160-2; Pb 90-277-2161-0
96. G. MArkus: Languageand Production. A Critique of the Paradigms. Translated from French.
1986 ISBN 90-277-2169-6
97. F. Amrine, FJ. Zucker and H. Wheeler (eds .): Goethe and the Sciences: A Reappraisal.
1987 ISBN 90-277-2265-X; Pb 90-277-2400-8
98. J.C. Pitt and M. Pera (eds.): Rational Changes in Science. Essays on Scientific Reasoning.
Translated from Italian. 1987 ISBN 90-277-2417-2
99. O. Costa de Beauregard: Time, the Physical Magnitude. 1987 ISBN 90-277-2444-X
100. A. Shimony and D. Nails (eds.): Naturalistic Epistemology. A Symposium of Two Decades.
1987 ISBN 90-277-2337-0
101. N. Rotenstreich; Timeand Meaning in History. 1987 ISBN 90-277-2467-9
102. D.B. Zilberman: The Birth ofMeaning in HinduThought. Edited by R.S. Cohen. 1988
ISBN 90-277-2497-0
103. T.F. Glick (ed.): The Comparative Reception of Relativity. 1987 ISBN 90-277-2498-9
104. Z. Harris, M. Gottfried, T. Ryckman. P. Mattick Jr, A. Daladier, T.N. Harris and S. Harris:
The Form of Information in Science. Analysis of an Immunology Sublanguage. With a
Preface by Hilary Putnam. 1989 ISBN 90-277-2516-0
105. F. Burwick(ed.): Approachesto OrganicForm. Permutations in Scienceand Culture. 1987
ISBN 90-277-2541-1
106. M. Almasi: The Philosophy ofAppearances. Translatedfrom Hungarian. 1989
ISBN90-277-2150-5
107. S. Hook, W.L. O'Neill and R. O'Toole (eds.): Philosophy, Historyand SocialAction. Essays
in Honorof LewisFeuer.With an Autobiographical Essay by L. Feuer. 1988
ISBN90-277-2644-2
108. 1. Hronszky, M. Feher and B. Dajka: ScientificKnowledge Socialized. SelectedProceedings
of the Sth Joint International Conference on the Historyand Philosophy of Scienceorganized
by the IUHPS(Veszprem, Hungary, 1984). 1988 ISBN 90-277-2284-6
109. P. Tillers and E.D. Green (eds.): Probability and Inferencein the Law ofEvidence. The Uses
and Limits of Bayesianism. 1988 ISBN90-277-2689-2
110. E. Ullmann-Margalit (ed.): Sciencein Reflection. The Israel Colloquium: Studies in History,
Philosophy, and Sociologyof Science,Vol. 3. 1988
ISBN 90-277-2712-0; Pb 90-277-2713-9
111. K. Gavroglu, Y. Goudaroulis and P. Nicolacopoulos (eds.): Imre Lakatos and Theories of
ScientificChange. 1989 ISBN 90-277-2766-X
112. B. Glassner and 1.0. Moreno (eds.): The Qualitative-Quantitative Distinction in the Social
Sciences. 1989 ISBN 90-277-2829-1
113. K. Arens:Structures ofKnowing. Psychologies of the 19thCentury. 1989
ISBN 0-7923-0009-2
114. A.Ianik; Style, Politicsand the Future of Philosophy. 1989 ISBN 0-7923-0056-4
115. F. Amrine (ed.): Literatureand Scienceas Modes ofExpression. With an Introduction by S.
Weininger. 1989 ISBN 0-7923-0133-1
116. I.R. Brown and J. Mittelstrass (eds.): An Intimate Relation. Studies in the History and
Philosophy of Science. Presentedto RobertE. Buttson His 60th Birthday. 1989
ISBN 0-7923-0169-2
117. F. D'Agostino and I.C. Jarvie (eds.) : Freedom and Rationality. Essays in Honor of John
Watkins. 1989 ISBN 0-7923-0264-8
118. O. Zolo: ReflexiveEpistemology. The Philosophical Legacyof Otto Neurath. 1989
ISBN 0-7923-0320-2
119. M. Kearn, B.S. Philips and R.S. Cohen (eds.): Georg Simmel and Contemporary Sociology.
1989 ISBN 0-7923-0407-1
120. T.H. Levere and W.R. Shea (eds.): Nature, Experiment and the Science. Essays on Galileo
and the Nature of Science. In Honourof StillmanDrake. 1989 ISBN 0-7923-0420-9
121. P. Nicolacopoulos (ed.): GreekStudiesin the Philosophy and HistoryofScience. 1990
ISBN0-7923-0717-8
122. R. Cooke and O. Costantini (eds.): Statistics in Science. The Foundations of Statistical
Methods in Biology,Physics and Economics. 1990 ISBN0-7923-0797-6
123. P. Duhem: The Originsof Statics. Translatedfrom French by G.F. Leneaux, V.N. Vagliente
and G.H. Wagner.With an Introduction by S.L. Jaki. 1991 ISBN0-7923-0898-0
124, H. Karnerlingh Onnes: Through Measurement to Knowledge. The Selected Papers, 1853-
1926. Editedand with an Introduction by K. Gavroglu and Y. Goudaroulis. 1991
ISBN0-7923-0825-5
125. M. capek: The New Aspects of Time: Its Conti1Ulity and Novelties. Selected Papers in the
Philosophyof Science. 1991 ISBNQ.7923-0911-1
126. S. Unguru (ed .): Physics, Cosmology and Astronomy, 1300·1700. Tension and Accommoda-
tion. 1991 ISBN 0-7923-1022-5
127. Z. Bechler: Newton's Physicson the Conceptual Structure ofthe Scientific Revolution. 1991
ISBN 0-7923-1054-3
128. E. Meyerson: Explanation in the Sciences. Translated from French by M-A. Siple and D.A.
Siple. 1991 ISBN 0-7923-1129-9
129. A.I. Tauber (ed.) : Organism and the Origins ofSelf. 1991 ISBN 0-7923-1185-X
130. FJ. Varela and J-P. Dupuy (eds .): Understanding Origins. Contemporary Views on the
Origin of Life, Mind and Society. 1992 ISBN 0-7923-1251-1
131. G.L. Pandit: Methodological Variance. Essays in Epistemological Ontology and the
Methodology of Science. 1991 ISBN 0-7923-1263-5
132. G. Munevar (ed.): BeyondReason. Essays on the Philosophy of Paul Feyerabend. 1991
ISBN 0-7923-1272-4
133. T.E. Uebel (ed.): Rediscovering the Forgotten Vienna Circle. Austrian Studies on Otto
Neurath and the Vienna Circle. Partly translated from German. 1991 ISBN 0-7923-1276-7
134. W.R. Woodward and R.S. Cohen (eds.): World Views and Scientific Discipline Formation.
Science Studies in the [former] German Democratic Republic. Partly translated from
German by W.R. Woodward. 1991 ISBN 0-7923-1286-4
135. P. Zambelli: The Speculum Astronomiae and Its Enigma. Astrology, Theology and Science
in Albertus Magnus and His Contemporaries. 1992 ISBN 0-7923-1380-1
136. P. Petitjean, C. Jami and A.M. Moulin (eds.): Science and Empires. Historical Studies about
Scientific Development and European Expansion. ISBN 0-7923-1518-9
137. W.A. Wallace: Galileo's Logic ofDiscovery and Proof The Background, Content, and Use
of His Appropriated Treatises on Aristotle's PosteriorAnalytics. 1992 ISBN 0-7923-1577-4
138. W.A. Wallace: Galileo's Logical Treatises. A Translation, with Notes and Commentary, of
His Appropriated Latin Questions on Aristotle's Posterior Analytics. 1992
ISBN 0-7923-1578-2
Set (137 + 138) ISBN 0-7923-1579-0
139. M.J. Nye, J.L. Richards and R.H. Stuewer (eds.): The Invention of Physical Science.
Intersections of Mathematics, Theology and Natural Philosophy since the Seventeenth
Century. Essays in Honor of Erwin N. Hiebert. 1992 ISBN 0-7923-1753-X
140. G. Corsi, M.L. dalla Chiara and G.c. Ghirardi (eds.): Bridging the Gap: Philosophy,
Mathematics and Physics. Lectures on the Foundations of Science. 1992
ISBN 0-7923-1761-0
141. C.-H. Lin and D. Fu (eds.): Philosophy and Conceptual HistoryofSciencein Taiwan. 1992
ISBN 0-7923-1766-1
142. S. Sarkar (ed.): The Founders ofEvolutionary Genetics. A Centenary Reappraisal. 1992
ISBN 0-7923-1777-7
143. J. Blackmore (ed.) : ErnstMach -A DeeperLook. Documentsand New Perspectives. 1992
ISBN 0-7923-1853-6
144. P. Kroes and M. Bakker (eds.): Technological Development and Science in the Industrial
Age. New Perspectives on the Science-Technology Relationship. 1992 ISBN 0-7923·1898-6
145. S. Amsterdamski: BetweenHistory and Method. Disputes about the Rationality of Science.
1992 ISBN 0-7923-1941-9
146. E. Ullmann-Margalit (ed.): The Scientific Enterprise. The Bar-Hillel Colloquium: Studies in
History, Philosophy, and Sociology of Science, Volume 4. 1992 ISBN 0-7923-1992-3
147. L. Embree (ed.) : MetaaTchaeology. Reflections by Archaeologists and Philosophers. 1992
ISBN0-7923-2023-9
148. S. French and H. Kamminga (eds.): Correspondence. lnvariance and Heuristics. Essays in
Honourof HeinzPost 1993 ISBN0-7923-2085-9
149. M. Bunzl: The ContextofExplanation. 1993 ISBN0-7923-2153-7
150. I.B. Cohen(ed.): The NaturalSciences and the SocialSciences. SomeCriticaland Historical
Perspectives. 1994 ISBN0-7923-2223-1
151. K. Gavroglu, Y. Christianidis and E. Nicolaidis (eds.): Trends in the Historiography of
Science. 1994 ISBN0-7923-2255-X
152. S. Poggi andM. Bossi(eds.): Romanticism in Science. Sciencein Europe, 1790--1840. 1994
ISBN0-7923-2336-X
153. J. Fayeand HJ . Folse(eds.): NielsBohr and Contemporary Philosophy. 1994
ISBN0-7923-2378-5
154. C.C. Gould and R.S. Cohen (eds.): Artifacts, Representations, and Social Practice. Essays
for MarxW. Wartofsky. 1994 ISBN0-7923-2481-1
ISS. R.E. Butts: Historical Pragmatics. Philosophical Essays. 1993 ISBN0-7923-2498-6
156. R. Rashed: The Development of Arabic Mathematics: Between Arithmetic and Algebra.
Translated fromFrenchby A.F.W. Armstrong. 1994 ISBN0-7923-2565-6
157. I. Szumilewicz-Lachman (ed.): Zygmunt Zawirski: His Life and Work. With Selected
Writings on Time, Logicand the Methodology of Science. Translations by FeliksLachman.
Ed. by R.S. Cohen, with the assistance of B. Bergo. 1994 ISBN0-7923-2566-4
158. S.N. Haq: Names, Natures and Things. The Alchemist Jabir ibn I;IayyAn and His Kitab ale
A1)jar (Bookof Stones). 1994 ISBN0-7923-2587-7
159. P. Plaass: Kant's Theory of Natural Science. Translation, Analytic Introduction and
Commentary by Alfred E. andMariaG. Miller. 1994 ISBN0-7923-2750-0
160. J. Misiek (ed.): The Problem of Rationality in Science and its Philosophy. On Popper vs.
Polanyi. The PolishConferences 1988-89. 1995 ISBN0-7923-2925-2
161. I.C. Jarvie and N. Laor (eds.): Critical Rationalism, Metaphysics and Science. Essays for
JosephAgassi, Volume I. 1995 ISBN0-7923-2960-0
162. I.C. Jarvie and N. Laor (eds .): CriticalRationalism, the SocialSciences and the Humanities.
Essaysfor JosephAgassi, Volume 11.1995 ISBN0-7923-2961-9
Set (161-162)ISBN0-7923-2962-7
163. K. Gavroglu, J. Stachel and M.W. Wartofsky (eds.) : Physics, Philosophy, and the Scientific
Community. Essaysin the Philosophy and History of the Natural Sciences and Mathematics.
In Honorof RobertS. Cohen. 1995 ISBN0-7923-2988-0
164. K. Gavroglu, J. Staeheland M.W. Wartofsky (eds.): Science, Politics and Social Practice.
Essayson Marxism and Science, Philosophy of Culture and the SocialSciences. In Honorof
RobertS. Cohen. 1995 ISBN0-7923-2989-9
165. K. Gavroglu, J. Stachel and M.W. Wartofsky (eds.): Science, Mind and Art. Essays on
Science and the Humanistic Understanding in Art, Epistemology, Religion and Ethics.
Essaysin Honorof RobertS. Cohen. 1995 ISBN0-7923-2990-2
Set (163-165)ISBN0-7923-2991-0
166. K.H.Wolff: Transformation in the Writing. A Caseof Surrender-and-Catch. 1995
ISBN0-7923-3178-8
167. AJ. Koxand D.M. Siegel (eds.): No TruthExcept in the Details. Essays in Honorof Martin
J. Klein. 1995 ISBN0-7923-3195-8
168. J. Blackmore: LudwigBoltzmann. His Later Life and Philosophy, 1900-1906. Book One: A
Documentary History. 1995 ISBN 0-7923-3231-8
169. R.S. Cohen, R. Hilpinen and R. Qiu (eds.): Realism and Anti-Realism in the Philosophy of
Science. Beijing International Conference. 1992. 1996 ISBN 0-7923-3233-4
170. I. Kucuradi and R.S. Cohen (OOs.): The Concept ofKnowledge. TheAnkara Seminar. 1995
ISBN 0-7923-3241-5
171. M.A. Grodin (00.): Meta Medical Ethics: The PhilosophicalFoundations of Bioethics. 1995
ISBN 0-7923-3344-6
172. S. Ramirez and R.S. Cohen (eds.): Mexican Studies in the History and Philosophy of
Science. 1995 ISBN 0-7923-3462-0
173. C. Dilworth; The Metaphysics of Science. An Account of Modem Science in Terms of
Principles, Laws and Theories. 1995 ISBN 0-7923-3693-3
174. J. Blackmore: Ludwig Boltzmann, His Later Life and Philosophy, 1900-1906 Book Two:
The Philosopher. 1995 ISBN 0-7923-3464-7
175. P. Damerow: Abstraction and Representation. Essays on the Cultural Evolution of Thinking.
1996 ISBN 0-7923-3816-2
176. G. Tarozzi (ed.): Karl Popper, Philosopher ofScience. (in prep.)
177. M. Marion and R.S. Cohen (eds.): Quebec Studies in the Philosophy of Science. Part I:
Logic, Mathematics, Physics and History of Science. Essays in Honor of Hugues Leblanc.
1995 ISBN 0-7923-3559-7
178. M. Marion and R.S. Cohen (eds.): Quebec Studies in the Philosophy of Science. Part II:
Biology, Psychology, Cognitive Science and Economics. Essays in Honor of Hugues
Leblanc. 1996 ISBN 0-7923-3560-0
Set (177-178) ISBN 0-7923-3561-9
179. Fan Dainian and R.S. Cohen (eds.): Chinese Studies in the History and Philosophy of
Science and Technology. 1996 ISBN 0-7923-3463-9
180. P. Forman and J.M. SAnchez-Ron (eds.): NationalMilitaryEstablishments and the Advance-
ment of Scienceand Technology. Studies in 20th Century History. 1996ISBN 0-7923-3541-4
181. EJ. Post: Quantum Reprogramming. Ensembles and Single Systems: A Two-Tier Approach
to Quantum Mechanics. 1995 ISBN 0-7923-3565-1
182. A.I. Tauber (ed.): The ElusiveSynthesis: Aesthetics and Science. 1996 ISBN 0-7923-3904-5
183. S. Sarkar (ed.): The Philosophy andHistoryofMolecular Biology: New Perspectives. 1996
ISBN 0-7923-3947-9
184. J.T. Cushing, A. Fine and S. Goldstein (eds .): Bohmian Mechanics and Quantum Theory: An
Appraisal. 1996 ISBN 0-7923-4028-0
185. K. Michalski: Logicand Time. An Essay on Husserl's Theory of Meaning. 1996
ISBN 0·7923-4082-5
186. G. Munevar (00.): Spanish Studies in the Philosophy of Science. 1996 ISBN 0-7923-4147-3
187. G. Schubring (00.): Hermann GuntherGraj3mann (1809-1877): Visionary Mathematician.
Scientistand Neohumanist Scholar. Papers from a SesquicentennialConference. 1996
ISBN 0-7923-4261-5
Also ofinterest:
R.S. Cohen and M.W. Wartofsky (eds.): A Portraitof Twenty-Five Years Boston Colloquia for the
Philosophy ofScience. 1960-1985. 1985 ISBN Pb 90-277-1971-3
Previousvolumes are still available.
KLUWER ACADEMIC PUBUSHERS - DORDRECHTI BOSTON I LONDON

Boston Studies in The Philosophy of Science

Uploaded by

Copyright:

Available Formats

Boston Studies in The Philosophy of Science

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Boston Studies in The Philosophy of Science

Uploaded by

Copyright:

Available Formats

BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE

ROBERTS. COHEN, Boston University

mOMAS F. GLICK, Boston University

Realism and anti-realism in the philosophy of seienee Be t j Inq

ISBN 978-90-481-4493-8 ISBN 978-94-015-8638-2 (eBook)

Printed on acid-free paper

All Rights Reserved

ROBERT S. COHEN / Recollections of Tscha Hung xiii

RISTO HILPINEN / On Some Formulations of Realism, or How

REALISM AND GENERAL METHODOLOGY

JAMES ROBERT BROWN / Phenomena 117

CHIN-TAl KIM / Science as Ideology 195

FAN DAINIAN / Niels Bohr and Realism 279

SOCIAL AND BIOLOGICAL SCIENCES

Name Index 481

engineering, like Western medicine and surgery, so philosophy of the

RECOLLECTIONS OF TSCHA HUNG

I first learned of Tscha Hung in 1955 during conversation with Marie

apt to try 'rational reconstruction' of traditional, or 'received', Chinese

HONG QIAN AND THE VIENNA CIRCLE

Hong Qian (Tscha Hung , 1909-1992), a member of the Program

many different places became the professor of philosophy at National

Hong agreed with Schlick :

Metaphysics is meaningful to ethics and the philosophy of life , so it

Then Hong was made editor-in-chief of a series of western philosophy

Neurath (1982) in Australia, and lectured in Tokyo (1986) and Hong

Hong hoped that analytical philosophers including logical empiricists will

ON SOME FORMULATIONS OF REALISM, OR

In his Paul Carns Lectures Putnam says:"

Sentence (P I) can be paraphrased as follows :

The concept of an object, unlike the concept of a chair or the concept

The word 'object' is not an individuating expression, but if F is any indi-

the expression 'object' in (P3) is supplemented by some individuating

(P7) The number of F-objects in the world depends on the choice

(P9) If our conceptual scheme were different from what it actually

thesis is not meaningless (or semantically confused) in the same way

(M3) The world contains a fixed totality of mind-independent F-

where F is some individuating description. By generalizing with respect

By a similar generalization we get from (P7) the following thesis of

It is not quite clear what 'mind-independent' means here. If the unavoid-

When philosophers have discussed real ism relativized to individuat-

(PI2) The existence of F-objects in the world depends on the choice

20 I am indebted to Professor Alan Musgrave for this observation.

This paper investigates the way to dissolve the adversary either/or

A great insight is a statement whose denial is another

Philosophical and scientific problems manifest their intrinsic intercon-

hypothetical, open-ended. It found its corresponding world-view in the

When the province of physical theory was extended

successful way of thinking in physical sciences inspired a speedy appli-

As we can see it in retrospect, Lenin's crusade against the spirit of

Physics does not deal with physics . Physics deals with

cultural tradition of the West which is entirely foreign to the traditional

This paper attempted to expose the cultural, philosophical, and scien -

scientific advances at the beginning of the 20th Century which forced

4. Lenin, Vladimir 1. (1952). Materialism and Empiriocriticism, Critical comments on

REALISM, TRUTH AND OBJECTIVITY*

Bill Lycan once called me a mad-dog realist. Whereupon Deborah Mayo

I . REALISM AND TRUTH