Kusch, Martin - Foucault's Strata and Fields

FOUCAULT'S STRATA AND FIELDS
SYNTHESE LffiRARY
STUDIES IN EPISTEMOLOGY,
LOGIC, METHODOLOGY, AND PHILOSOPHY OF SCIENCE
Managing Editor:
JAAKKO HINTIKKA, Boston University
Editors:
DONALD DAVIDSON, University ofCalifornia, Berkeley

GABRIEL NUCHELMANS, University ofLeyden
WESLEY C. SALMON, University ofPittsburgh
VOLUME 218
Library of Congress Cataloging-in-Publication Data
Kusch, Martin.
Foucault's strata and fields ап lnvestigatjOn into
archaeological and genealogical science studies / Ьу Martin Kusch.
р. ст. -- (Synthese 1 ibrary : v. 218)
Includes bibliographical references and indexes.
ISBN 978-94-010-5567-3 ISBN 978-94-011-3540-5 (eBook)
DOI 10.1007/978-94-011-3540-5
1. Foucault. Mjchel. 2. Science--Philosophy--Hjstory--20th
century. 3. Social sciences--Philosophy--History--20th century.
1. Title. П. Series.
B2430.F724K87 1991
194--dc20 91-33950
ISBN 978-94-010-5567-3
Printed оп acid-free paper
АН Rights Reserved
© 1991 Springer Science+Business Media Dordrecht
Originally published Ьу Кluwer Academic Publishers in 1991
Softcover reprint of the hardcover 1st edition 1991
No part of the maierial protected Ьу this copyright notice тау ье reproduced or
utilized in any form or Ьу апу means, electronic or mechanical,
including photocopying, recording or Ьу any information storage and
retrieval system, without written permission from the copyright owner.
CONTENTS
ACKNOWLEDGEMENTS. . . . . . . . . . . . . . . . . . . . . . .. ix
INTRODUCTION xi
PART I: FOUCAULDIAN ARCHAEOWGY . . . . . . . . . . . 1
1. INTRODUCTION.......................... 1
2. ON THE VERY NOTION OF 'ARCHAEOLOGY' . . . . . 5
3. THE NEW HISTORIES IN FRANCE 12

THE ANNALES EXEMPLARS . . • . • • . • . . . . . . . . • . .. 12
Against traditional history . . . . . . . . . . . . . . . . . .. 13
The history ofmentalities 15
Serial history and the construction of the historical fact. 19
THE EXEMPLARS OF HISTORICAL EPISTEMOLOGY . • • . .. 24
Duhem vs. Koyre 24
Bachelard . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 27
Canguilhem . . . . . . . . . . . . . . . . . . . . . . . . . . .. 32
Althusser 35
4. ARCHAEOLOGY, THE NEW HISTORIES, AND THE

HISTORY OF IDEAS 40
CORRELATING THE NEW HISTORIES . • . . . . . . • . . . . . 41
A CRITICAL RE-EvALUATION OF THE HISTORY OF IDEAS. 43
Foucault and Lovejoy . . . . . . . . . . . . . . . . . . . . .. 44
The mythology of ideas . . . . . . . . . . . . . . . . . . . .. 47
The mythology at work: the examples of Holton and
Cohen 50
The mythology of coherence and contradiction 54
v
vi CONTENTS
5. THE ARCHAEOLOGICAL MODEL I: IDENTIFYING

DISCURSIVE FORMATIONS , 58
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . .. 58
WHAT Is A FOUCAULDIAN STATEMENT? 60
THE FOUR SERIES 64
Objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 65
Enunciative modalities 67
Organizations ofthe field of statements . . . . . . . . . .. 68
Strategies ..... . . . . . . . . . . . . . . . . . . . . . . .. 70
INTERRELATIONS, DERIVATION TREES AND CONTRADIC-
TIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 71
COUNTERPARTS AND ACCESSIBILITY RELATIONS - COM-
PARING DISCURSIVE FORMATIONS 75
Discourses and discursive formations . . . . . . . . . . .. 75
Counterparts . . . . . . . . . . . . . . . . . . . . . . . . . .. 78
Accessibility relations . . . . . . . . . . . . . . . . . . . . .. 80
6. THE ARCHAEOLOGICAL MODEL II: BEYOND CON-

TINUITY AND DISCONTINUITY . . . . . . . . . . . . . .. 83
PHILOSOPHER'S vs. HISTORIANS' HISTORY OF SCIENCE .. 83
THE DEBATE OVER CONTINUITY AND DISCONTINUITY
RECONSTRUCTED . . . . . . . . . . . . . . . . . . . . . . . . .. 86
The positivist-empirist model 87
The antipositivist model 88
Adding strata: the antipositivist revisionism 90
Laudan and Foucault as postmodernists 95
THE ARCHAEOLOGICAL DIMENSION 102
7. ARCHAEOLOGY OF KNOWLEDGE AND OTHER HIS-

TORIES OF SCIENCE ". . . . . . . . . . . . . . . . . . . . . .. 109
NOTES TO PART I 112

CONTENTS vii
PART ll: FOUCAULDIAN GENEAWGY . . . . . . . . . . . .. 115
8. INTRODUCTION.......................... 115
9. THE CONCEPT OF POWER 117

SOME STANDARD DISTINCTIONS 118
A FOUCAULDIAN DEFINITION OF POWER . . . . . . . . . .. 122
10. THE GENEALOGICAL CONCEPTION OF POWER I:

FIELDS AND NETWORKS 127
MODELS OF POWER 127
INTERNAL RELATIONS. . . . . . . . . . . . . . . . . . . . . .. 130
POWER AND PERSONAL IDENTITY 134
NETWORKS OF POWER 138
NETWORKS OF POWER AND INVISIBLE HANDS 143
11. THE GENEALOGICAL CONCEPTION OF POWER II:

SOCIAL POWER AND SCIENTIFIC KNOWLEDGE ... 149
THE PRODUCTIVITY OF POWER . . . . . . . . . . . . . . . .. 149
BARNES' PROPOSAL 152
DEFENDING "POWER/KNOWLEDGE" 155
A disquieting suggestion . . . . . . . . . . . . . . . . . . .. 157
Science and exclusion . . . . . . . . . . . . . . . . . . . . .. 157
An argument from underdetermination 161
REALISM AND MODALITY 163
12. GENEALOGICAL RESEARCH STRATEGIES 165

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . .. 165
RECOVERING AGONISTIC EVENTS 169
The return ofthe event . . . . . . . . . . . . . . . . . . . .. 169
Recovering contingency 172
Causal multiplication . . . . . . . . . . . . . . . . . . . . .. 174
The agonistic dimension 175
PHYSICALISM 178
Histories ofthe body 178
Body, power, knowledge . . . . . . . . . . . . . . . . . . .. 181
viii CONTENTS
DISTRUST AND IRONY . . . . . . . . . . . . . . . . . . . . . .. 186

Interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 187
Ironic hypotheses 191
13. GENEALOGICAL PERSPECTIVISM 193

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . .. 193
SKEPTICISM 194
CONSTRUCTIVISM 196
EPISTEMOLOGICAL RELATIVISM 200
"Relativism Refuted" - refuted . . . . . . . . . . . . . . .. 200
The "not-to-be-neglected truth" 207
Foucault's relativism 210
14. GENEALOGICAL CRITICISM OF POWER AND RA-

TIONALITIES 213
GENEALOGY AS CRITICISM AND ANALYSIS OF RATION-
ALITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 213
Is RESISTANCE POSSIBLE? 217
CRITICISM WITHOUT NORMATIVITY? 221
NOTES TO PART II . . . . . . . . . . . . . . . . . . . . . . . . . . .. 227
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 235
INDEX OF NAMES . . . . . . . . . . . . . . . . . . . . . . . . . . .. 250
INDEX OF SUBJECTS 257

ACKNOWLEDGEMENTS
I am indebted, most of all, to my colleagues and students in the Depart-

ment of History of the University of Oulu. I have especially profited
from discussions with Jouko Aho, Heini Hakosalo, Hippo Johansson,
Jouko Jokisalo, Vaino Louekari, Matti Lukkarila, Juha Manninen, Aarne
Nikkonen, Erkki Routti, Arja-Liisa Raisanen, Minna Uimonen, and
Erkki Urpilainen. I am also grateful to "my" librarians, both Paula
Viander of the Department of History and the people of the interlibrary
loan office of the University Library. Outside of my alma mater, special
thanks are due to Timo Airaksinen, Harri Kantele, Esa Itkonen, Karin
Knorr-Cetina, Lothar Lasker, Tuomas Nevanlinna, Ilkka Niiniluoto,
Calvin Normore, Kari Palonen, Outi Pasanen, Tuija Pulkkinen, Yves
Roussel, Kari Saastamoinen, Hartmut SchrOder, and Georg Henrik von
Wright for their criticism and/or encouragement. My wife, Riitta Korho-
nen-Kusch, deserves special thanks not only for her commenting, but
also for her crucial support at a point when I was ready to give up on
the whole project. I am also grateful to an anonymous reviewer for
Kluwer Academic Publishers.
The University of Oulu has supported my project with a travel-grant
to Paris. The last stages of the work I carried out while holding the
position of a younger fellow at the Academy of Finland, and while
enjoying the hospitality of the University of Leuven (Belgium).
Finally, I wish to thank Ian Morris-Wilson and Jan Jagiellowicz for
help with my English, Auli Kaipainen for excellent secretarial assistance,
Jaakko Hintikka for accepting the book into Synthese Library, and Annie
Kuipers of Kluwer Academic Publishers for her friendly cooperation.
I gratefully acknowledge the permissions to quote from the follow-
ing books: Michel Foucault, The Archaeology of Knowledge, translated
by A.M. Sheridan, Tavistock, London, 1974, and Michel Foucault,
Power/Knowledge: Selected Interviews & Other Writings 1972-1977,
copyright (c) 1972, 1975, 1976, 1977 by Michel Foucault, edited by
C. Gordon, Pantheon Books, Random House, Inc., New York, 1980.
I dedicate this book to two friends who have had a large share in its
creation and completion: Heini Hakosalo and Marja-Liisa Kakkuri-
Knuuttila. Marja-Liisa Kakkuri-Knuuttila has been a close friend and
supporter for many years now, and during the writing of this book we
IX
x ACKNOWLEDGEMENTS
have spent many hours on the phone and in Chinese restaurants in

Helsinki discussing Aristotle, rhetoric, feminism, vegetarianism, music
and what not. As concerns the topic of this specific study, even more
important has been Heini Hakosalo's cooperation. Since we set out to
work on Foucault roughly around the same time, we had planned to
publish our results as one book. As both of our parts kept growing in
size, however, this plan had to be abandoned. Heini Hakosalo's part
(Hakosalo 1991) therefore had to be published separately. Being my
nextdoor neighbour in the history department, Heini Hakosalo has been
not only a central source of inspiration and good humour, but also the
first hurdle and final stumbling block for most of my ideas on Foucault.
Qulu, June 1991

Martin Kusch
INTRODUCTION
In recent years, a large number of books and articles on Foucault has

been published. Almost all of the book-size studies are expository and
introductory. Indeed, there seems to be no other modern philosopher
with reference to whom a comparable number of introductions have been
produced in such a short period. Most of the articles too provide over-
views, rather than critical assessments or rational reconstructions, even
though there exists by now a small number of fine papers also in the two
latter genres. Moreover, more often than not, writers on Foucault
approach his work as part and parcel of so-called "postmodern" philo-
sophy. They concentrate on topics like the "death of the subject", the
relation of Foucault's work to .Derrida or Habermas, or its significance
for postmodern art and culture.
Without wanting to deny the merits, either of introductory exposi-
tions, or of studies that read Foucault as a postmodern thinker, it seems
to me that these received perspectives have tended to leave central areas
and aspects of Foucault's work somewhat underexposed. As I see it, the
most important of these areas are such as would suggest reading Fou-
cault from the vantage point of recent developments in the philosophy,
sociology and history of science.
During the last two decades or so, a number of new approaches
have emerged and taken shape within philosophical, sociological and
historical science studies. These new endeavors run under such catch-
word-labels as "anthropological" science studies, the "strong pro-
gramme" of the sociology of science, "constructivism", "new experi-
mentalism", "the reticulated model of scientific rationality", and "the
feminist critique of science". While these approaches differ in many
respects from one another, it still seems possible to point out two main
features which many of them share.
First, a number of philosophers and historians have - at last -
moved beyond the fruitless dispute over continuity and discontinuity in
the history of science. The predominant question is no longer whether a
scientific field develops either cumulatively or through total breaks.
Instead, the crucial problem now is how to distinguish, within a scien-
tific field of study, different forms of changes within different strata or
layers of theoretical assumptions, problems, values, methods and tech-
xi
xii INTRODUCTION
niques within a scientific culture.

Second, sociologists and feminists have begun to study scientific
communities as "tribes", "cultures" and "power networks". For many
writers in the sociology of science as well as for many feminist critics,
science is no longer the purely cognitive, socially disinterested and
gender-neutral, rational enterprise as which it was often presented by
traditional philosophies of science. Rather, in a number of investigations
published during the seventies and eighties, science is looked upon as
one social culture (or bundle of cultures) among others, a culture whose
beliefs, relations and actions can be studied with methods known from
sociology and anthropology.
Now, to claim that Foucault can be fruitfully studied as contributing
to these recent developments, is no radically new and contested thesis.
The present study merely takes this thesis and tries to develop it further.
That is to say, in this book, I present a reconstruction of those parts of
Michel Foucault's writings that I feel are of considerable interest for
both of these two main concerns in contemporary science studies. Going
beyond a mere mechanical comparison, I shall propose various ways in
which these more recent lines of thought in science studies, and a
reconstructed, conceptually sharpened and, with respect to several key
theses, newly defended Foucauldian project, can be used to illuminate
and strengthen one another.
As concerns the move beyond continuity and discontinuity, Foucault
is, according to the received view, a historian or philosopher who works
on the assumption of radical ruptures in the history of science. However,
while this conception might perhaps be partially justified as far as some
of Foucault's less successful historical investigations are concerned, it
certainly does not hold for his discours de la methode, The Archaeology
of Knowledge. In the first part of this study, I shall show that The
Archaeology of Knowledge contains arguments against both one-sided
continuist and excessive discontinuist models of scientific change. I shall
also point out that The Archaeology includes a detailed discussion of
various approaches to the history of science, ranging from the French
Annales School and the historical epistemology of Bachelard and Can-
guilhem to the History of Ideas. Most importantly, the work outlines a
theory of historical constructivism, a theory that is informed by the
methodological writings of Annales historians.
INTRODUCTION xiii
In the second part of this study, I shall turn to the central concepts
and ideas of Foucault's "genealogical" works. In these Nietzsche-
inspired studies, written roughly between 1970 and his death in 1984,
Foucault focusses more directly than in his earlier, archaeological phase
on phenomena of power, especially the relations between scientific
knowledge and social power. As I shall show, Foucauldian genealogy
includes several of those theses that can roughly be grouped around the
second main ingredient of the new science studies, to wit, the study of
science as a social field. Foucault proposes a theory of social power and
its relations to scientific knowledge as well as detailed methodological
suggestions on how these interrelations might be studied from a histori-
cal point of view.
As should be clear from the above, my central interest in this study
is topical, rather than historical or biographical. In other words, my
overriding concern is not to determine what Foucault really meant or
said - even though I obviously cannot ignore such questions completely
- but rather to use his ideas for the purpose of contributing to the
debates in the philosophy and sociology of science. In adopting this
attitude towards Foucault's texts, I actually abide by his own view on
how best to approach his writings. Commenting on his relation to
Nietzsche, Foucault once wrote that "the only valid tribute to thought
such as Nietzsche's is precisely to use it, to deform it, ... And if com-
mentators then say that I am being faithful or unfaithful to Nietzsche,
that is of absolutely no interest" (p/K 54).
Finally, a word is due on the relation between the two main parts of
this study. I follow the received wisdom of Foucault research and divide
his oeuvre into two phases. The first comprises Foucault's "archaeologi-
cal" work, issuing in the afore-mentioned discours de la methode, The
Archaeology ofKnowledge, published in 1969. The second phase is that
of "genealogy", i.e. Foucault's writings from 1970 onwards. Of course,
within both phases further distinctions could easily be made. However,
since I am not primarily interested in Foucault's intellectual biography,
questions of development will be discussed only occasionally, and only
as far as the major changes are concerned.
I have not tried to make my reconstructions of Foucauldian archae-
ologyand genealogy part of one overall, unified argument. Consequent-
ly, both parts can also be read separately. This is not to say, that the
xiv INTRODUCTION
archaeological and the genealogical projects are without links and

interrelations. To the contrary, already the key notion of Foucauldian
archaeology, the notion of "discursive formation" , refers to social condi-
tions of science, and in the genealogical writings, Foucault spells out
more clearly many of the philosophical tenets at work in his archaeolo-
gy. Moreover, the critical thrust of The Archaeology of Knowledge, its
criticism of anachronisms and taken for granted units like "discipline",
"author" , or "oeuvre", carries over into the genealogical edifice. Indeed,
Foucault could have left behind the general historiographical proposals
of The Archaeology only by paying the high price of falling back to
traditional naive history. However, all of these interrelations notwith-
standing, and despite my reluctance to follow those interpreters who
construct something of a "break" between archaeology and genealogy,
I have thought it best to discuss the two projects separately. Two reasons
have been pivotal. First, the terminology within the two phases is rather
different; and even though I did not intend to put forward merely an-
other introductory book on Foucault, I still want to provide the reader
with some help in finding her way around the Foucauldian conceptual
jungle. Second, and more importantly, it seems to me that a separation
of the two projects is called for by the compartmentalization of present
day science studies themselves. Not everyone who is interested in the
broad questions of continuity and discontinuity in the history of science,
is likewise convinced, that the relation between scientific knowledge and
social power is a question worth pondering upon, and not everyone who
shares the genealogical interest in identifying the effects of exercises of
power upon the human body, is likewise interested in detailed considera-
tions of the relations betweens scientific statements of different dis-
ciplines. To bind together all of Foucault's suggestions into one single
bundle, I fear, would have lessened the accessibility and the convincing-
ness of what I take to be the most important tenets of his work; it would
have made it difficult for me to argue that Foucault's theoretical work is
not a diffuse tangle of ideas, but rather a clearly delimited set of pro-
posals that are separately open to critical assessment.
PART I:
FOUCAULDIAN ARCHAEOWGY
1. INTRODUCTION
Foucault's archaeological discours de la methode, The Archaeology of

Knowledge (AK) is much less widely read than all of his other books. In
part this is due to the fact that this methodological treatise does not
contain the kind of grand historical - speculative or sweeping - theo-
rizing that readers have come to expect from this writer. The compara-
tively 'sober' nature of the book has not helped it much, however, in
reaching the audience of specialists for whom it was written, historians
in general and historians and philosophers of science in particular. One
reason is that the majority of historians and philosophers - in France as
well as elsewhere - have formed their predominantly negative perception
of Foucauldian historiography on the basis of The Order of Things (OT).
This work, published three years before The Archaeology, hardly lives
up to strict standards of historical scholarship, and thus it is small
wonder that practicing historians and methodologists do not feel any urge
to study Foucault's methodological ideas.
Another reason for the small amount of attention that the book has
received, is undoubtedly its difficulty. 1 Its intricacies are partly due to
the fact that Foucault outlines his methodological suggestions by and
large via negationis - a fact that has led Maurice Blanchot to speak of a
"negative theology" (1987: 74), and Jean-Fran~ois Revel to propose
"Neither ... nor" as a more appropriate title (O'Farrell 1989: 15). The
book is also hard to follow because it is written for an audience familiar
not only with twentieth century French historiography (Le. with various
approaches within the Annales school, with the history of science of
Duhem and Koyr~, with the epistemology of Cavaill~, Bachelard and
Canguilhem, the structuralist marxism of Althusser), but also with the
history of ideas as represented by e.g., Lovejoy, Cassirer and Gold-
mann. Since general history and the history of science traditionally live
separate lives, presupposing acquaintance with both cannot but further
restrict the circle of potential readers.
1
2 PART I
Unfortunately, existing book-size presentations of Foucault reflect

rather than remove these difficulties. While one author, Clare O'Farrell,
simply bypasses the book with the disarming confession that The Ar-
chaeology "is a compelling book" only for those "who appreciate
intricate formal geometrical structures in thought and method" (1989:
14), other interpreters see The Archaeology as something of an appendix
to The Order of Things, as a series of corrections with respect to Fou-
cault's earlier historical studies, or as an anticipation of his genealogical
writings.
The first interpretative strategy is neatly exemplified not only by
Gary Gutting, who writes that the "leitmotif of AK is ... the death of
man" (1989: 227), but also by the best known commentary on Fou-
cault's 'oeuvre', Hubert Dreyfus's and Paul Rabinow's Michel Foucault.
Beyond Structuralism and Hermeneutics (1983). 'Bracketing' the metho-
dological-historiographical dimension of The Archaeology ofKnowledge,
Dreyfus and Rabinow read the book as a systematic theory of language,
compare it with Hussed, Heidegger, Kuhn and Searle, and ultimately
fault Foucault for not having provided the "phenomenology to end all
phenomenologies" which, as the authors see it, the book sets out to (or
should?) develop (1983: 44-103).
The second tendency, i.e. reading The Archaeology as a mere series
of afterthoughts on Foucault's earlier work, surfaces for instance in
Gilles Deleuze or Charles Lemert and Garth Gillan. For Deleuze "[...]
in this archaeology, Foucault [perhaps] offers us less a discourse on his
method than the poem of his previous works, and reaches the point
where philosophy is necessarily poetry, the severe poetry of what is said,
which subsumes both nonsense and the greatest profundities" (1988: 18).
Lemert and Gillan draw the conclusion that since The Archaeology is but
a bunch of after-thoughts, they need to "disabuse readers of the idea that
his Archaeology of Knowledge is of special importance. Archaeology of
Knowledge is not a methodological statement ..." (1982: ix-x).
Finally, and to come to the third strategy, most interpreters struggle
with the problem of how to situate The Archaeology within Foucault's
intellectual biography. In the seventies Foucault adopts a different label,
"genealogy", for his historical studies, and drops the term "archaeolo-
gy". Commentators take this as a challenge to address the question
whether Foucault is an "archaeologist of ideas or [a] genealogist of
FOUCAULDIAN ARCHAEOLOGY 3
power?" (Smart 1985: 47). The usual answer is that - in part because of
the events of May 1968 - Foucault turns Nietzschean, overcomes "the
illusion of autonomous discourse" and his earlier role of "the detached
spectator", and starts doing what he should have been doing, or intended
to do, all along: study the relation between power and discourse (Drey-
fus & Rabinow 1983: 103).
Here I do not wish to criticize these various interpretations in any
detail, even though it is hard to suppress the remark that these studies
could easily be used as illustrations for all of the categories of interpreta-
tory excesses which The Archaeology ofKnowledge enumerates. Instead,
suffice it here to outline briefly my own concerns and preoccupations.
My reconstruction of The Archaeology is motivated by the view that
this work contains a number of suggestions and criticisms concerning the
writing of the history of science that are - or should be - of interest to
present-day philosophers, sociologists and historians of science alike. To
put it in a nutshell, The Archaeology ofKnowledge merits such attention
because its proposals and discussions link up with many of the topics
that researchers in science studies have come to address over the last
decade or so. To mention only a few catchwords, such topics and new
interests are discourse analysis, research mentalities, styles of reasoning,
the anthropology of science, different strata of assumptions involved in
theorizing and experimenting, the relation between the history of ideas
and the history of science, the historical links and structural similarities
between disciplines, conceptualizations of science as a field of struggle
over the modality of statements, and a new attitude with respect to the
notorious problem of revolutions in science: the central question is no
longer whether science is continuous or not, but rather which elements
do and which do not accumulate.
Many of these new concerns have arisen from a heightened aware-
ness of the anachronistic dangers in writing the history of science. This
awareness, in turn, has been triggered, or has at least been reinforced,
by a loosening of the borders between the history of science and other
domains of historical research. Given this development, Foucault's signi-
ficance need no longer come as a surprise. After all, Foucault was
intimately familiar not only with the achievements of historians and
philosophers of science like Bachelard and Canguilhem, but he was
equally at home in the substantial historical as well as the methodological
4 PART I
labors of the most important school in historical scholarship of this

century, the Annales historians. Once we add the further observation
that, already in the thirties, Bachelard formulated views on the develop-
ment of science and the role of experimentation, more recent counter-
parts of which have sometimes been hailed as novelties in the sixties and
the eighties respectively, it should be even less unexpected that a French
text of 1969 can be of interest to scholars in the nineties.
It follows from the above-said that my reconstruction of The Archae-
ology needs to be "historical" as well as "rational". The reconstruction
will be historical not in that it seeks to place this book within Foucault's
intellectual biography or that it wants to present the book's ideas as
successors to anticipations in the writings of others. Rather, the focus of
the following reconstruction must be historical in so far as it attends not
only to those views that, according to a somewhat unsophisticated
categorization, are "Foucault's own", but also to ideas and theories "by
others", i.e. ideas and theories which the book endorses. On the other
hand, the reconstruction will be and must be "rational" in order for me
to succeed in my attempt to convince the skeptical reader of the book's
interest, that is, in order to make the book - or rather my reconstruction
of it - a contribution to present concerns in the philosophy, history and
sociology of science. Thus I do not abide by the view that "Foucault is
not as opaque as Joyce, but like Joyce his style and his thought are
inseparable" (Lemert & Gillan 1982: xii). Instead, I seek to formulate
what I take to be the central theses of the book in a language that is as
clear and precise as the inherent difficulty of the very topic allows for.
Rational reconstruction also implies a criterion of selection. I do not
attempt to provide a full commentary on all aspects of the book, nor do
I engage the book in a debate over those concepts and ideas that are less
important for my present concerns or that I regard as obscure; to add yet
another deconstruction of the notoriously opaque notions like episteme or
archive to the criticisms already found in the literature seems super-
fluous.
To conclude this introduction with the setting of some signposts, I
shall start with a brief comment on the notion of archaeology, its use as
a metaphor by other writers in France, and its employment by Foucault.
Here I shall also make a brief suggestion on how to understand this
metaphor. Subsequently, I shall review the exemplars of Annalist and
historical epistemology, focussing especially on notions like mentality,

discontinuity, series and time. This chapter will be followed by an
analysis of Foucault's criticism of various anachronistic tendencies in the
history of ideas and the history of science. As will be shown by way of
examples, this criticism has not become obsolete. Finally, in the last two
chapters, I shall reconstruct the "archaeological model" for studying the
history of science. Here I shall compare the archaeology of knowledge
with theories proposed more recently in the philosophy of science. My
main point will be to argue that Foucauldian archaeology includes a
model of scientific change that is not only a far cry from the excesses of
The Order oj Things, but also of some interest for present-day concerns.
2. ON THE VERY NOTION OF 'ARCHAEOLOGY'
In choosing the term 'archaeology' as a label for his approach to the

history of science, Foucault took advantage of the fact that this term had
already gained currency in France. Since the fifties, the concept of
'archaeology' (as well as the concept of 'geology') had been used to
characterize the various fashionable intellectual movements of the day,
movements like phenomenology, structuralist anthropology, psycho-
analysis, Bachelardian epistemology, and Annalist history.
For instance, in his widely-read Tristes Tropiques, first published in
1955, Claude Levi-Strauss had written of his "intense interest [in]
geology". Geology allegedly allows us to pass beyond "the vast chaos"
and "the various accidents of history and prehistory" in that it ultimately
provides us with "the most majestic meaning of all ... [a meaning] that
precedes, commands, and, to a large extent, explains the others". Levi-
Strauss had also explained his interest in anthropology, marxism and
psychoanalysis by referring to their common geological dimension. In all
four cases, what initially appear to be "impenetrable phenomena", or "a
seemingly incoherent mass", tum out to have an "order" which is
"neither contingent nor arbitrary"; in all these instances, "understanding
consists in reducing one type of reality to another", where this reduction
is based upon "a kind of superrationalism" that "will integrate the first
[meaning] with the second, without sacrificing any of its properties"
(1976: 69-71).
6 PART I
Psychoanalysis was also regarded as an essential archaeological

enterprise by such renowned figures as Maurice Merleau-Ponty and Paul
Ricoeur, both of whom couched the parallel between Husserlian phe-
nomenology and psychoanalysis precisely in terms of the joint "archae-
ological" features of these two theories. 2 To speak: of phenomenology
as an archaeology had been common in France at least since Jean
Cavaill~'s Sur la logique et la theorie de la science, a work published
in 1947.3 Cavaill~, perhaps best known as a historian of mathematics,
was one of the central figures in the early stages of French historical
epistemology, a school of thought whose best known representatives are
Gaston Bachelard and Georges Canguilhem. In 1959 Canguilhem charac-
terized Bachelard's writings as amounting to an "archaeology of sci-
ence" , implying that this type of historiography is more interested in
fields - or strata - of problems than in individual achievements (1959:
5).
Finally, the "new history" of the Annales school, a school of
historiography that became dominant in French historical scholarship
after World War II, also made use of archaeology as a methodological
metaphor. In his classic Apologie pour I'histoire ou Ie Metier d'historien
(1953), Marc Bloch proposed that historians cure themselves of their
excessive preoccupation with narratives by attending to archaeologists'
methods, and the model of postwar Annales scholarship, Fernand Brau-
del's La Mediterranee al'epoque de Philippe II, was of course based on
the notion of different time strata (Braudel 1972-74).
Foucault himself seems to have been drawn towards the archaeo-
logical metaphor right from the start. Thus already in his early Maladie
mentale et personnalite of 1954, he speaks of "neurosis [as] ... a spon-
taneous archaeology of the libido" (1954: 26), and in his these comple-
mentaire pour Ie doctorat des lettres, "Introduction ~ L'Anthropologie de
Kant", submitted in 1960, he writes of the "sedimentations" of Kant's
work that call for an "archaeology of the text" (1960: 3-4). In fact,
Foucault later even claimed to have adopted the notion of archaeology
from Kant's "Fortschritte der Metaphysik". As a reading of this text
shows, however, Foucault's claim that there one finds "the word, the
text and the meaning to which I refer" (l971b), is not believable; neither
the term nor any meaning even vaguely resembling Foucauldian ar-
chaeology appears in the Fortschritte.
In Folie et deraison: Histoire de la folie a l'age classique (1954)

(MC), Foucault speaks only en passant of the "archeology of that
silence" to which madmen were reduced in and after the Classical age.
It is only in the subsequent books, Naissance de la clinique: une archeo-
logie du regard medical (1963) (BC), Les mots et les choses: une ar-
cheologie des sciences humaines (1966) (OT), and L'archeologie du
savoir (1969) (AK), that Foucault begins to use 'archaeology' as the
central label for his own approach. It thus seems appropriate to say that
whereas the employment of the archaeological metaphor was by and
large coincidental and arbitrary in Foucault's earliest works, it was only
in the sixties that he came to regard it as sufficient to characterize his
overall project. This assessment is confirmed by Foucault himself where
he admits to having used this term "initially ... a bit blindly" or that "it
seems to me that chance has not guided me badly" (1969: 23).
Interestingly enough, Foucault has never attempted to spell out the
archaeological metaphor in any detail, confining himself rather to the
'translation' of 'archaeology' as "description of the archive" (ibid.). In
order to give something of a first general overview over what Foucauld-
ian archaeology of knowledge amounts to, it seems a useful exercise to
point out some parallels between some of its central claims, on the one
hand, and key aspects of archaeology (proper), especially more recent
archaeological stratigraphy, on the other hand.
As one is informed by any textbook, archaeology typically is the
study of prehistory, Le. of periods from which no texts have remained.
Archaeology is the study of the "archaeological record", Le. of the
"fossilized results" of human behavior. Furthermore, archaeology is
neither the history of individual men, nor the history of individual
fossils. First, archaeology "cannot aspire to be biographical, and ar-
chaeologists are excluded from the school of 'great man' history"
(Childe 1962: 14). Second, rather than in the unique implement, the
archaeologist is interested in "types" of objects that can be found in
significant archaeological contexts. In other words, archaeology seeks to
order arbitrary aggregates of objects into assemblages of fossils that
belong temporally together. Assemblages of the same type encountered
at different sites are called "cultures", and thus the archaeologist deter-
mines "culture sequences" by investigating successive levels of different
stratified sites (ibid., 19).
8 PART I
It is also characteristic of the archaeologist's work, that she must

meticulously define and delimit series and classifications. This is because
absolute chronology was - until the invention of radiocarbon dating -
impossible, and is still only one of several forms of dating. In principle,
the archaeologist "need keep only archaeological time. For archaeologi-
cal time exhibits succession but not duration" (ibid., 30). Classificatory
principles include the function of objects, the ordering of objects made
for similar purposes at different times according to their degree of
efficiency and sophistication ("typological series"), as well as geogra-
phical, chorological distribution. The construction of typological series
needs to take into account not only the possibility of differential rates of
change, but also the correlations between several such series (Childe
1956: 70-74). This constructive procedure thus presupposes the distinc-
tion between the relative age of a given artefact in its typological se-
quence, its relative age as an element of a stratum, and its absolute age
as its date of origin.
Foucauldian archaeology also deals primarily with prehistory, at
least if we understand prehistory in the sense which is frequent in
writings on the historiography of science where the prehistory of a
science is often taken to be a discourse (or a bundle of discourses) that
are eventually transformed into the science in question (e.g. Hacking
1975: 9; Kuhn 1970: 21). Or, put differently, the prehistory of a current
science is a discourse that is not immediately recognizable as part and
parcel of the modem discipline (Lepenies 1978: 443). Foucauldian ar-
chaeology is preoccupied foremost with prehistory in this sense, even
though it relates somewhat critically to this very notion, and even though
it does not confine itself to it. Furthermore, like archaeology proper,
Foucauldian archaeology is no biographical history, and rejects the focus
upon"great men". It is interested in defining series and types of objects,
statements, discourses and concepts rather than concerned with individu-
als or unique achievements. It deplores the practice of many historians
of science who take certain units like discipline, period, or oeuvre, for
granted, and it calls for an explicit methodology for the construction of
temporal series, discourses, their interrelations and correlations, as well
as their rates of change. In so doing, it puts little emphasis upon dating
scientific events by means of the calendar; rather, it seeks to define
different time-scales in the history of science, different rhythms of
change, and different relative ages of contemporaneous sciences. In

distinguishing different levels, Foucauldian archaeology wants to show,
for instance, that the social construction of objects like madness need not
be contemporaneous with the genesis of a scientific discourse about these
objects.
Even though these parallels between archaeology proper and Fou-
cauldian archaeology already suffice to explain the naturalness of the
metaphor in the latter's case, we can still go one step further by compar-
ing Foucauldian archaeology with a more recent 'revolutionary' advance
in archaeology proper, Le. Edward Harris's "archaeological strati-
graphy" (Harris 1979). This is captivating especially since Harris's
criticism of traditional archaeology and its leaning towards geological
stratigraphy has a counterpart in Foucault's criticism of the history of
ideas, the philosophy of history and - at least indirectly - structuralism.
First, whereas Harris argues that archaeology needs a systematic
theory of archaeological stratification, and whereas Harris accuses his
colleagues of too often relying on a problematic conception of geolo-
gical, Le. natural, strata (1979: xiii), Foucault calls for a genuine
historical theory of historical strata, and renounces the explicit or impli-
cit employment of periodizations taken over from the philosophy of
history. In like manner he distances himself from the structuralist-
geological search for "the most majestic meaning of all" (Lt!vi-Strauss)
by claiming that the archaeology of knowledge "does not imply the
search for a beginning; it does not relate analysis to geological excava-
tion" (AK 131).
The attack on the assimilation of the subject matter of archaeology
to natural objects is also common to both Harris and Foucault in so far
as both wish to emphasize the inanimate character of their basic 'fossils':
Harris faults archaeologists for tacitly assimilating archaeological arti-
facts to geological, animate, fossils (1979: 5); Foucault in turn deplores
the organistic metaphors that historians of ideas use when talking of the
alleged "birth", "life" and "death" of ideas or themata.
Furthermore, whereas Harris sees the stratum as the basic unit to be
reconstructed by archaeology, earlier archaeologists were often exclu-
sively searching for the chief relics, artifacts or monuments (ibid., 15).
Similarly, Foucault is as outspoken a proponent as any of what has
become standardly referred to as "vertical history" (e.g. Kragh 1987:
10 PART I
81). To refer to the vertical history of science as the history of strata is,
of course, not unique to Foucault; e.g. David Knight writes that histo-
rians of science "have tended to drill a small hole down from the present
through the strata of history; they should be well advised instead to look
much more closely at the contents of one particular stratum" (Knight
1975: 25).
To drill "small holes down" is of course also a standard method in
archaeology and geology where such holes are produced in order to
obtain "sections" of given excavation sites. Harris is highly skeptical of
this method, since it constitutes again a transferring of geological meth-
ods to archaeology. Since "most sites produce multilinear stratigraphic
sequences of a complexity which would baffle many geologists ... it
would be difficult to obtain a section which would be representative of
little more than the section itself" (1979: 51). This shortcoming, Harris
claims, cannot even be overcome by following the advice of one well-
known textbook: " ... look[..] at the section upside-down (standing that
is, with the back to the section and bending to look through the legs);
from this unaccustomed posture it is frequently possible to notice details
not apparent to the normal view" (ibid., 54). While I prefer to leave it
to the reader to find a counterpart to this advice in the writings on the
history or philosophy of science, it should be mentioned in any case that
Foucault mutatis mutandis not only criticizes the historiographical
practice of following just one idea or discipline through its history, but
also denounces the tendency (for instance of Bachelard and Althusser) to
draw conclusions about the development of all sciences on the basis of
one or two, namely mathematics or physics. To put it differently,
according to Foucault, the history of mathematics and physics does not
provide us with a representative section for the histories of other sci-
ences.
Of course, neither the archaeologist proper nor the archaeologist of
knowledge can avoid the drawing of sections, even though, in both
Harris's and Foucault's case, section drawing is only one method and
needs to be supplemented by investigations into "boundary contours",
i.e. vertical boundary lines of archaeological units. When sections are
drawn, however, Harris calls for the drawing of "stylized sections" with
"interfacial lines and its layers ... numbered" (1979: 58). (Realistic
sections record only the artifacts found.) Likewise, The Archaeology of
Knowledge demands that the historian attend to discontinuities in the

history of science, even if such discontinuities are never total and instan-
taneous.
This last parallel between Harrisian and Foucauldian stratigraphy
can be strengthened by attending to three further points of contact. First,
Harris calls upon archaeologists to take account of the fact that the usual
order of geological stratification, according to which the strata lay upon
one another in a natural order of age, does not hold true for archaeolo-
gical stratification. Because of human intervention, or the intervention of
natural processes, the order of strata need not correspond to age (1979:
xiii). Second, and in part as a corollary, the archaeologist must pay
special attention to the shapes 'of strata. She should try to excavate one
stratum at a time rather than excavate arbitrarily or metrically, say, ten
centimeters throughout the site (ibid., 16). Third, archaeology needs a
theory to explain the interfaces between strata, and this theory cannot be
taken over unmodified from geology. Interfaces can just be boundaries
of a given stratum ("period interfaces"), but they can also constitute
"feature interfaces", Le. strata in their own right (that perhaps have been
squeezed into a thin layer by later depositions) (ibid., 43).
Harris' attention to the order of strata and his stress upon the non-
horizontal shape of strata interfaces coincides well with Foucault's
emphasis upon the non-synchronic character of changes and transforma-
tions in or between discourses. For instance, two sciences can develop
along similar patterns, yet this development can take much more time in
the case of one as compared with the other. Furthermore, the different
elements needed for the formation of a scientific discourse of some given
type may well have different and distinct histories and they may be
correlated and interrelated to form this discourse only in some specific
span of time. The shape of the stratum to which this discourse belongs
will thus possibly include these earlier elements. This implies that
Foucauldian archaeology (at least in 1969) has to reject the arbitrary
usage of period labels like "renaissance" or "Classical age", and that it
refuses to have any of the metrical stratigraphy of George Sarton who
once cut up the history of science in Europe and Asia in strata of fifty
years (Sarton 1975).
The Archaeology of Knowledge also proposes an apparatus for the
conceptualization of what appear to be breaks or discontinuities. Other
12 PART I
than, say, Althusser or Kuhn, Foucault is not satisfied with the alterna-
tive 'total break versus continuity'. The Archaeology introduces a list of
new terms and categories by means of which different forms of continu-
ity and discontinuity can be distinguished.
Finally, and to conclude this comparison between Harris's stratig-
raphy and Foucault's archaeology, both reproach their respective col-
leagues for putting an exaggerated emphasis upon the dates of origin of
their fossils (Harris 1979: 97). As important as is the origin of some
artefact or statement, as important is, for both authors, its subsequent
use, transformation and deposition.
3. THE NEW HISTORIES IN FRANCE
Few commentators have failed to notice the fact that Annales historio-
graphy and the historical epistemology of Bachelard and Canguilhem
figure centrally in The Archaeology of Knowledge. Indeed, this fact is
somewhat hard to overlook since the book mentions both schools re-
peatedly. What I regard as the main shortcoming of existing commenta-
ries is rather that they focus on one of the two schools at the expense of
the other, and that they do not pay sufficient attention to parallels in the
concerns, theories and problems of the two. To set the record straight
seems to be of some interest, not only as far as Foucault is concerned.
Now that historians and philosophers of science have begun to study
science from a more anthropological-cultural perspective, the work of
Annales historians and of the epistemological school has to loom large;
after all, in these fields of scholarship these allegedly new topics have
been among the central preoccupations for the last fifty years.
THE ANNALES EXEMPLARS
In order to see the importance of the Annales School for Foucauldian

archaeology in particular, and for the historiography of science in
general, two central topics of this school of historians need to be at-
tended to in more detail: the notion of mentality and the methodology of
serial history.4
Against traditional history
The Annales School dates back to the turn of the century, roughly to the
founding of the Revue de synthese historique by the philosopher Henri
Berr. Berr together with a group of like-minded historians and social
scientists reacted to the challenge posed by Durkheimian sociologists
who had denied history the status of a science and who had claimed the
scientific study of society in its historical dimension to be solely the
domain of sociology. In answering this challenge, Berr's group differed
from more traditional historians. The more conservative historians
belonging to I'ecole methodique, like Charles-Victor Langlois and
Charles Seignobos, and centered on La Revue historique, were ready to
capitulate in the face of the sociological challenge, Le. willing to con-
cede that history as a discipline is nothing more than the study of the
particular and that the scientific dimension of historical study consists
exclusively in the application of a series of analytical methods of text-
criticism (internal and external critique of documents). Unsatisfied with
this deferential attitude, a band of historians grouping around Berr's new
journal aspired to the status of a science by adopting the methods of
those very sciences (sociology, geography, psychology, economics) that
threatened history's traditional role.
It was in Berr's journal and in his monograph series L'evolution de
I'Humanite that Lucien Febvre and Marc Bloch, the leading first-genera-
tion figures of what later became the Annales School, published their
first studies, before founding, in 1929, the revue from which the name
of the school derives, Les Annales d 'histoire economique et sociale.
Whereas before World War II the school was long denied access to the
leading institutions in Paris (Febvre and Bloch worked in Strasbourg),
after 1945 it became the leading school in French historical scholarship.
In 1946 the journal was renamed as Annales: Economies - Societes -
Civilisations, and in 1947 it acquired its own institutional bastion, the
Sixieme Section at the Ecole Pratique des Hautes Etudes. The influence
of the group soon became so pervasive that critics have come to speak
of "the Annalist holding" (Bourd~ & Martin 1983: 249).
What the Annales historians never tired of attacking was the pre-
occupation of the ecole methodique and other forms of "traditional"
history with producing narratives that concentrate on (mainly political)
14 PART I
events and that are based, almost exclusively, on written documents. For
Febvre, Bloch and their pupils, such narrative history neglects long-term
or large-scale phenomena like material culture, trading routes, economic
or demographic patterns of development, and mentalities. For instance,
Fernand Braudel who inherited the leadership of the school from Febvre
in the late fifties, renounces traditional history in the following way:
Traditional history, with its concern for the short time span, for the
individual and the event, has long accustomed us to the headlong,
dramatic, breathless rush of its narrative. ... a short time span,
proportionate to individuals, to daily life, to our illusions, to our
hasty awareness - above all the time of the chronicle and the jour-
nalist (Braudel 1980: 27-28).
Marc Bloch, in his Apologie pour I'histoire, identifies another theme

of traditional history: "the obsession with origins" (1953: 29). In search-
ing for the ultimate origins of events and ideas, traditional history typi-
cally conflates beginnings and causes. It looks for the ultimate starting-
points of events and ideas and neglects the historical strata in which
these events and ideas are situated. Thus it "confuses ancestry with
explanation": "As if the main problem were not to understand how and
why the transition had taken place" (ibid., 33). To provide an alterna-
tive, Bloch suggests that "a historical phenomenon can never be under-
stood apart from its moment in time. ... As the old Arab proverb has it:
'Men resemble their times more than they do their father'" (ibid., 35).
Bloch is also unconvinced by the key assumption of the ecole
methodique, according to which the historian obtains knowledge only
indirectly, i.e. via reports of contemporaneous witnesses to historical
events. It is here that Bloch proposes that historians would do better to
take their lead from archaeology:
If the best-known theorists of our methods had not shown such an

astonishing and arrogant indifference toward the techniques of
archaeology, if they had not been as obsessed with narrative in the
category of documents as they were with incident in the category of
actions, they would doubtless have been less ready to throw us back
upon an eternally dependent method of observation (ibid., 53).
Lucien Febvre criticizes not only traditional history in general, but

also intellectual history, or the history of ideas in particular. He rejects
the use of catch-all categories like feudalism or renaissance, and de-
mands that historians reconstruct different systems of thought before
advancing to generalizations (Chartier 1982: 16). Moreover, Febvre
attacks, in harsh terms, historians of ideas who either follow singular,
isolated ideas through the past, or attempt to analyze systems of thought
in isolation from their larger cultural context:
Of all the workers who cling to the generic title of historian, there
is only one group that cannot in some way justify it in our eyes.
They are those who, applying themselves to rethink for their own
purposes systems that are sometimes several centuries old, and
without the slightest care to show the links with other manifestations
of the epoch which saw them come into being, end up doing pre-
cisely the opposite of what a historical method demands. They
engender concepts from disincarnated minds which live their lives
beyond time and space. Historians who deal with them forge strange
chains whose links are at the same time unreal and closed. (Quoted
from Chartier 1982: 17.)
The history of mentalities
However, Febvre distances himself from the Burckhardtian way of

explaining similarities of contemporaneous historical phenomena in terms
of a common Zeitgeist. Or rather, Febvre holds that the Zeitgeist itself
stands in need of explanation (Chartier 1982: 18). In order to improve
upon the traditional methodologies, Febvre as well as Bloch relied on a
notion that had become central in the ethnology of Lucien L~vy-Bruhl,
in the psychology of Henri WaHon (Le Goff 1985: 171), and that also
appeared en passant in J. Huizinga's The Waning of the Middle Ages
(Huizinga 1970). This notion is the notion of "mentality" (mentalite).
With what may be regarded as their respective major works, Bloch's
Feudal Society (1936) and Febvre's The Problem of Unbelief in the
Sixteenth Century. The Religion of Rabelais (1942), Bloch and Febvre
provided the two classic loci of histoire des mentalites in Annales
16 PART I
scholarship. Under the title "modes of feeling and thought" (Bloch 1962:
72-87), Bloch describes how medieval men and women experienced
nature as threatening and unfamiliar, human life as endangered by
disease, violence and the lack of hygiene, how they accepted the super-
natural, why they could not doubt its existence, and how they related to
myths, rites, theology, religion. Moreover, Bloch studies the passion
with which emotions were expressed, how irrationalism surfaced both in
the limited role left for reasoned arguments and in the importance given
to dreams, how time was taken to be both circular and linear, and why
human action was regarded as predetermined. Bloch also attends to the
lack of precision in thought, a feature due in part to the absence of a
developed system of numerals, the lack of honesty, the low degree of
literacy, the strong role of dialects, and the ambiguity of words. Bloch
concludes by commenting on the low esteem for learning and the typi-
cally low self-consciousness. In dealing with these elements of the
collective mentality, Bloch remains close to anthropological methods,
studies how these attitudes were expressed in everyday practices, and
makes no concessions to individual differences (Burgui~re 1982: 433).
Febvre's The Problem of Unbelief runs a somewhat different course
in that it takes its starting point from an individual, Rabelais, whose
"limits of what is thinkable" Febvre seeks to determine by placing him
within his time. Even though Febvre is keenly interested in feelings,
sensibility and emotions, his approach to the mentality of sixteenth
century France is by and large slightly more intellectualist than Bloch's.
This difference of emphasis is clear from the fact that for Febvre the
core of a mentality is its "mental tools". These tools are first and
foremost "the ensemble of categories of perception, conceptualization, of
expression and action which structure individual as well as collective
experience" (Revel 1986: 451). With respect to the mentality of the
sixteenth century, for example, Febvre argues that the outillage mentale
was marked by a lack of crucial words, like "absolute", "relative",
"abstract", "concrete", "complex", "confused", "adequate", "virtual",
"intentional" or "transcendental"; by the absence of a nomenclature; by
a syntax which made the expression of perspective rather difficult and
which did not clearly distinguish between co-ordination and subordina-
tion; by a low degree of formalization in the sciences and mathematics;
and by the inaccessibility of important books. Febvre explains the
prevailing attitudes and beliefs of the sixteenth century in terms of these

mental tools. It is in this manner that he investigates religion (assump-
tions about God, soul and immortality, the possibility of doubt and
choice in religious matters), logicality (the treatment of conflicts in
intellectual questions, the awareness of contradictions, modes of rea-
soning, requirements of proofs), time (types of chronometers, the
historical sense), images of sciences and the scientist (compilation vs.
explanation, dogmatism vs. skepticism, veracity, professional images,
the relation between theoretical knowledge and practical skill, the role of
observation), cosmology (ways of reasoning in questions of cosmology),
emotions (ways of expressing one's feelings), senses (the role of diffe-
rent senses), art (which genres were central), and the category of the
possible (to what extent are physical and logical impossibility held
distinct).
In some places, Febvre seems to suggest that two historical periods
with diverging mental tools are incommensurable: " ... there is really no
common measure between sixteenth-century man's way of feeling,
thinking and speaking and our own" (Febvre 1982: 1(0). This viewpoint
motivates Febvre's accusation of massive anachronism directed against
those historians of ideas who have tried to present Rabelais as an atheist,
and who have made no distinction between the atheism of the sixteenth
and later centuries. Against these earlier interpretations, Febvre seeks to
show that there is, as one recent commentator puts it, "an epistemologi-
cal break, a conceptual caesura between the unbelief of the sixteenth and
that of the eighteenth century" (Wootton 1988: 703). In other words,
Febvre replaced "an idea of incremental progress with one of radical
discontinuity" (ibid., 729). Thus in his Rabelais, Febvre tries to show
that for the sixteenth century thinker atheism meant merely deviation
from some specific doctrinal elements of christian religion rather than
disbelief in the supernatural.
While strongly advocating radical breaks in the history of thought,
Febvre does not hold, however, that mental tools change overnight:
Every civilization has its own mental tools. Even more, every era of
the same civilization, every advance in technology or science that
gives it its character, has a revised set of tools, a little more refined
for certain purposes, a little less so for others. A civilization of an
18 PART I
era has no assurance that it will be able to transmit these mental

tools in their entirety to succeeding civilisations and eras. The tools
may undergo significant deterioration; or, on the contrary, more
improvement, enrichment, and complexity. They are valuable for
the civilization that succeeds in forging them, and they are valuable
for the era that uses them; they are not valuable for all eternity, or
for all humanity, not even for the whole course of development
within one civilization (Febvre 1982: 150).
Even though Febvre thus allows not only for the development of
mental tools but also for their being (at least in part) inherited by sub-
sequent eras, ever since Febvre's and Bloch's classics, the histoire des
mentalites has continuously been haunted by the problem of accounting
for change and development. Emphasizing strongly the pervasive in-
fluence of prevailing mentalities, the histoire des mentalites has been
unable to distinguish between the change within one given mentality and
the replacement of one mentality by another (see e.g. Burke 1986: 443;
Chartier 1982: 31; Wootton 1988: 703). Since they tend to conceive of
mentalities as "prisons" (Braudel 1982: 31), Annales historians make it
difficult to understand how any culture could ever move on to 'inhabit'
a new one.
Nevertheless, some suggestions concerning the problem of change
have been made. For instance, as the above quotation shows, Febvre
leaves room for the possibility that mental tools change piecemeal. Some
passages of The Problem of Unbelief also suggest that Febvre reckons
with variation within a given mentality; some groups, like scientists or
artists, make full use of the available "idea materials", while others are
able to employ only a small part of the mental tools of their time (Char-
tier 1982: 21). Furthermore, and interestingly enough, in order to cope
with change, historians of mentalities permit the history of ideas re-
entrance into the Annales paradigms. For example, Jacques Le Goff,
who together with Emmanuel Le Roy Ladurie inherited the leadership of
the school from Fernand Braudel in the early seventies, writes that "the
mentality of anyone individual ... is precisely what that individual
shares with other men of his time" (Le Goff 1985: 167). This characteri-
zation leaves room for individual originality and difference as an area of
thought not shared with others. Whereas Le Goff proposes historical
psychology ("archaeopsychology") and ethnology as the proper ap-

proaches for the study of mentalities, he provides space for the history
of ideas as the study of great thinkers. The study of mentalities and the
history of ideas have to be related, however, for instance in order to
investigate the origins of elements of a given mentality:
The history of mentalities must also be distinguished from the

history of ideas, out of opposition to which it also grew. It was not
the ideas of Thomas Aquinas or Saint Bonaventura which, from the
thirteenth century on, governed people's minds, but mental nebulae
in which such ideas played only a part as deformed echoes, de-
valued fragments and words taken out of context (1985: 176).
Finally, in a book that otherwise closely follows the outline of

Febvre's opus magnum, R. Mandrou's Introduction to modern France
1500-1640. An essay in historical psychology (Mandrou 1975), mental-
ities are made more dynamic by introducing different time scales. Thus
Mandrou distinguishes between long-term "common features ... as the
basic features common to every mentality at the time" (ibid., 236), i.e.
features like hypersensitivity of temperament, social aggressiveness, and
man's feeling of impotence in the face of the natural world; "views of
the world: mental structures" as the ensemble of "all the notions -
intellectual as well as ethical - within which individuals and groups daily
developed their thoughts and actions" (ibid., 243) (at this level there are
numerous differences between various groups, classes and individuals);
and "climates of sensitivity: mental conjunctures": these are periods of
shorter or longer duration, from a few years up to a generation, periods
of optimism or pessimism, of crisis or confidence. According to Man-
drou, it is at this level that"gradual changes in outlook" originate (ibid. ,
249).
Serial history and the construction of the historical fact
Mandrou's distinction between different time scales within a mentality is

of course but an application of quantitative or serial history to the
histoire des mentalites. The differentiation between different time-scales
20 PART I
does not come from Febvre or Bloch but from the work of F. Simiand
and C.-E. Labrousse, two economic historians who exerted a strong
influence on the Annales School. Nevertheless, the locus classicus within
Annalism for the distinction between different times is the monumental
La Mediterranee of Braudel. Since we have already encountered his
characterization of the time of events as the time of the short time span,
it remains to mention his two other time strata, the time of cyclical
movements or conjunctures, and the time of structures, Le. the duree
longue. The time of cyclical movements is the time of"a price curve, a
demographic progression, the movement of wages, the variations in
interest rates, the study (...) of productivity, ... [or of] money supply"
(Braudel 1982: 29). Structures qua phenomena that persist over centuries
function as "hindrances ... [or] limits (...) beyond which man and his
experiences cannot go. Just think of the difficulties of breaking out of
certain geographical frameworks, certain biological realities, certain
limits of productivity, even particular spiritual constraints ... " (ibid.,
31). Braudel divides La Mediterranee in accordance with this three-
partite distinction. However, even though he argues that "these different
time spans ... are all interdependent" and that" ... to be able to achieve
an imaginative understanding of one of these time spans is to be able to
understand them all" (ibid., 48), he in fact does not succeed in explain-
ing the interrelations between these time spans.
While the works on structures and conjunctures within Annales
scholarship make for fascinating reading, I shall not try to summarize
the numerous studies that the Sixieme Section has carried out in the field
of the quantitative history of conjunctures of different durations, studies,
for instance on prices, salaries, trade, demography, epidemics or the
climate. For our purposes it is more important to take note of the new
historiographical philosophy that takes shape in these studies. This new
view of the nature of historical research comes out especially clearly in
a remark by Emmanuel Le Roy Ladurie in his TImes ofFeast, TImes of
Famine: A History of the Climate since the Year 1000 (Le Roy Ladurie
1972) and a programmatic article by Fran<;ois Furet (Furet 1985).
Le Roy Ladurie proposes that the new serial history is no longer
bound to "the anthropocentric approach" (1972: 16). Serial history can
be applied to any series of data, regardless of whether these data involve
- directly or indirectly - human beings. As a consequence, Le Roy
Ladurie no longer underwrites Marc Bloch's famous claim, according to

which "behind the tangible features of landscape ... it is human beings
that the historian is trying to discern" (ibid., 18). As the historian of the
climate sees it, it "is mutilating the historian to make him into no more
than a specialist of humanity ... he may also in certain cases be interest-
ed in nature for its own sake, and make known by his own irreplaceable
methods nature's own special Time" (ibid., 20). What makes this thesis
remarkable is that with it Annalist criticism of the traditional political
event reaches its peak. Whereas the first and the second generation of
the school 'deconstructed' the event by enlarging the time scale, Le Roy
Ladurie goes one step further and challenges the connection between
history and the human being. Needless to say, such a viewpoint was well
received at a time when antihumanism loomed large, and Foucault
announced the "end of man" .
The programmatic article of Fran~ois Furet that I referred to above
is his "L'histoire quantitative et la construction du fait historique". This
article was published in the Annales in 1971, and later translated into
English as part of a collection of articles by Annales historians, fittingly
entitled Constructing the Past. s And indeed, it is this aspect of "con-
struction" that succinctly sums up the new historiographic attitude.
Furet's starting point is the distinction between quantitative and
serial history, a distinction first used by Pierre Chaunu (e.g. Chaunu
1970). Whereas quantitative history refers to any kind of history that
employs mathematical tools, serial history (histoire serielle) is interested
in the quantification of repeated facts; it constructs "temporal series of
homogeneous and comparable units" in order to measure "the evolution
of that reality at given intervals, usually annually". In this way, serial
history, right from the start, rejects the traditional historian's preoccupa-
tion with the individual fact. Its subject matter is not found but con-
structed; it is not the unique, but the repeatable. Historical reality, or the
historical record, can be broken up or constructed in several ways and at
different levels, and it is up to the historian to establish correlations
between these different levels (Furet 1985: 14).
The serial historian must also distinguish between change and
mutation, must define thresholds and breaks: as long as cyclical patterns
of development prevail, the serial historian is led to speak of "change
within stability". Yet when the serial historian encounters, for instance,
22 PART I
tendencies towards indefinite growth, she must assume that the latter
may alter "the very structure of time and rhythms of change". In order
to be able to distinguish the later sort of mutations from changes within
stability, i.e. in order not to assimilate mutations to cyclical changes, the
serial historian must work with temporal units of various length (ibid.,
15).
Another problem that the serial historian must cope with is the fact
that European archives have been set up mainly for the study of political
events rather than for the study of long-term developments ("an archive
is the memory of nations"). The big European archives contain the
documents that the traditional historian treats as evidence for, or testimo-
ny of, historical facts: "In this sense, a historical document is a unique,
discrete, particular moment within a global evolution which either
remains temporally indeterminate or is divided into centuries, reigns,
ministries." For the serial historian, on the other hand, the elements of
the historical record are no longer documents about the 'significant turns
in history'. External criticism, like the comparison of a document with
other contemporary texts, is no longer the most crucial historiographic
operation. Rather, what becomes central is the comparability and homo-
geneity of a given text with other texts in the same temporal series
(ibid., 16).
Furthermore, the document or text is no longer treated as a unit of
meaning in and by itself. The serial historian rejects the idea according
to which "the sources ... define the questions asked by a discipline";
instead, it is the historian's query that "determiners] the sources" (ibid.,
18). For instance, rather than asking for the meaning of texts, the serial
historian can investigate handwriting and thus arrive at statistically
obtained conclusions about the development of literacy (ibid., 17).
These new procedures and methodological problems of serial
history, Furet holds, amount to "a revolution in historiographical con-
sciousness" (ibid., 21). The core of this revolution is the idea of con-
struction. The essence of the matter is not that serial historians construct
their subject matter, whereas more traditional, document or event-
oriented history found it. Rather, the emergence of serial history, and the
discussion of its methodology, has brought to the fore that all historical
research is inevitably constructive:
... today's historian must give up methodological naivety and think

about the conditions under which historical knowledge is estab-
lished.... the mask of some kind of historical objectivity hidden in
the 'facts' and discovered at the same time as them, has been re-
moved for ever; the historian can no longer avoid being aware that
he has constructed his 'facts' ... (ibid., 20).
More specifically, Furet claims, the coming of age of serial history

implies a transformation of the conception of time in history. For
traditional history with its prejudice in favor of the unique and irreplace-
able, an event cannot be placed in a series of comparable, measurable
data. It follows that an event is only studied in the short term perspective
or that it is interpreted in terms of a teleological meaning: "it has no past
but it can have a future" (ibid., 21). The teleological meaning in ques-
tion has been the progress towards modern forms of the state or modern
forms of rationality and reason. Moreover, even when events were
related to one another in traditional history, this relating typically
happened in terms of a unifying meaning, like vision du monde or
Zeitgeist, that was postulated independently of the events. Looked at
from the angle of temporality, these commitments of traditional history
imply a curious way of understanding historical time: a series of un-
repeatable and unique events is narrated in the form of a story whose
plot has been pre-established on philosophical and ideological grounds;
in Furet's terms: "a series of discontinuities described in the continuous
mode - a story" (ibid., 21).
The conception of historical time involved in serial history, how-
ever, is diametrically opposed to that of traditional history: "Serial
history ... describes continuities in the discontinuous mode: it is history
as problem rather than history as story" (ibid., 21). Even though serial
history starts from constructing homogeneous series (qua continuities),
it distinguishes between different levels, thresholds and cycles (qua
discontinuities). It works with series that have different speeds and
temporal rhythms, it identifies series not all of which can be correlated.
Serial history also compares cycles taking place at different times or in
different regions: in the first case, similarities may point to common
geographical factors, in the latter, nonsimilarity may suggest difference
in economic structures (ibid., 24).
24 PART I
Taking the advances of serial history seriously demands then,

according to Furet, that history "must now start from the hypothesis that
the nature of historical time may be different according to the level of
reality or the partial historical system which is being analyzed" (ibid.,
25). From this it follows, first, that the tacit assumption of traditional
history, according to which different elements of a historical period
develop approximately in parallel, has to be kept in suspense. Rather
than start from a given periodization, like the Renaissance, and study
different elements within this period, the historian is better advised to
investigate elements over a much longer period. And second, only when
using such a procedure, Furet claims, can the historian answer the
question as to which elements and patterns of a given period undergo
rapid development and which are marked by stagnation (ibid., 26).
THE EXEMPLARS OF HISTORICAL EPISTEMOLOGY
Duhem vs. Koyre
The study of the history of science in France has been decisively shaped
by the famous disagreement between Pierre Duhem and Alexandre
Koyre over whether the advances in the natural sciences during the
sixteenth and seventeenth centuries qualify for the title of a scientific
revolution. Duhem repeatedly argues that the answer should be negative
by pointing out that central concepts of Galilei can already be found in
the writings of the Parisian masters of the fourteenth century:
The mechanical and physical science, of which modern times is

rightfully so proud, derives, in a hardly perceptible manner, from
an uninterrupted series of improvements of doctrines professed in
the womb of medieval schools; the alleged intellectual revolutions
were most often nothing but slow and long-prepared evolutions; the
self-proclaimed renaissances frequently nothing but unjust and sterile
reactions. Respect for tradition is an essential condition of scientific
progress (Duhem 1905: iv).
The development of mechanics must be declared "an evolution; each of

the stages of this evolution is a natural corollary of the stages which
precede it, and a part of the stages that follow it" (quoted from Fichant
1969: 73). This observation leads Duhem to oppose talk of scientific
revolutions as "brisk jumps" (Duhem 1906b: 156). While there can be
acceleration of the speed of development, this occasional acceleration is
simply a "condensation" of the evolution (Fichant 1969: 74). Indeed,
according to Duhem, the study of the history of science is usually dis-
torted by "two prejudices", to wit, that scientific progress is but "a
sequence of sudden and unforeseen discoveries", and that it is "the work
of geniuses who have no precursors at all" (Duhem 1906a: 1). For
Duhem, a new physical theory is "not the sudden product of a creation,
it is the slow and progressive result of an evolution" (1906b: 365).
Moreover, scientific theory does not originate from a sudden break with
prescientific thought: "in the genesis of a scientific doctrine there is no
absolute beginning"; it will always be possible to identify earlier opin-
ions which have "prepared, suggested and announced" the theory in
question (1913: 5). Duhem also believes in the existence of eternal
themes and ideas, for instance, the questions concerning the value of
physical theory and its relation to metaphysics "belong to all times"
(1969: 3).
Duhem's downgrading of Galileo's achievements is informed by his
strong catholicism and his positivism. Or rather, his positivism - the
sharp distinction between physical-mathematical theory and metaphysics
- is his device for "saving" the achievements of medieval scholasticism.
This religious prejudice comes to the fore not only in peculiar asides like
"Kepler was a Protestant, but [I] deeply religious" (1969: 100), but also
in Duhem's overall strategy of replacing the scientific by a "theological
revolution". As he alleges, it was the Church Fathers' attack on the
pagan philosophers, their rejection of the notion of eternal prime matter
and the belief in the domination of planets over sublunary events, that
cleared the ground for modern science; "it was necessary that a theologi-
cal revolution take place", and this revolution was "the work of Chris-
tian theology" (1914: 453). Thus, what Duhem denies the scientists, he
readily concedes to the theologians: the making of a revolution in
thought.
Koyre, on the other hand, holds that "the apparent continuity in the
26 PART I
development of physics, from the Middle Ages to modern times ... is

illusionary" (1973: 172). He speaks of the advances of the sixteenth and
seventeenth century as "a deep revolution which changed the framework
and patterns of our thinking" (1957: vii), and allows in general for
"veritable 'mutation[s]''' in science (1978: 1). Koyr6 studied the history
of science in its interrelations with philosophical thought. For him the
scientific revolution meant the dissolution of the Greek and medieval
conception of the cosmos:
... the founders of modern science ... had to destroy a world and
replace it with another. They had to reform the structure of our very
intelligence, reformulate and revise its concepts, envisage Being in
a new manner, elaborate a new concept of knowledge, a new con-
cept of science - and they also had to substitute a quite natural point
of view, Le. that of common sense, with another [point of view]
that is not at all natural (1973: 172).
Despite this direct disagreement with Duhem, and despite his

criticism of the pervasive use of the categories of precursor and influ-
ence in Duhem (Koyr~ 1957: 5-6), Koyr~ does not depict scientific
revolutions as abrupt ruptures. For example, Koyr~ is ready to write that
Kepler anticipated modern empirism and positivism, and he defends this
claim by arguing for continuity with respect to basic positions in episte-
mology (ibid., 58, 286). He also presents Archimedes as "the veritable
precursor of modern physics" (1973: 172), and interprets the Newtonian
synthesis of mathematical physics as the result of a "resurrection" of
Democritos' atomism and Plato's mathematics (ibid., 333). For Koyr6,
revolutions too need preparation, they "too, have a history" . In fact, the
destruction of the cosmos and the geometrization of space, the two
central elements of the scientific revolution, needed about 150 years to
be accomplished (1957: ix). Finally, Koyr6 also allows for the non-
uniformity of progress and the non-homogeneity of contemporaneous
works:
Palingenius and Copernicus are practically contemporaries. Indeed,

the Zodiacus vitae and the De revolutionibus orbium coelestium must
have been written about the same time. Yet they have nothing, or
nearly nothing, in common. They are as far away from each other
as if they were separated by centuries (ibid., 28).
Bachelard
Gaston Bachelard's epistemology emphasizes revolutions, breaks or

discontinuities in the history of the natural sciences even more than
Koyr~. However, Bachelard is less concerned with the sixteenth and
seventeenth centuries than with the physical "revolutions" in our cen-
tury. It was the period in physics from Einstein to Chadwick, Le. the
period of the special and general theory of relativity, wave mechanics,
quantum mechanics, the discovery of cosmic rays, the probabilistic
theory of the electron, the uncertainty principle, the relativist interpreta-
tion of wave mechanics, the discovery of neutrinos, positrons, neutrons
and mesons, which convinced Bachelard that "one decade in our epoch
is equal to centuries in earlier epochs" (Lecourt 1975: 32, 95), and that
this new physical science is "without antecedents" (ibid., 36).
Bachelard is especially fascinated by the fact that microphysics
breaks radically with common sense. Whereas in earlier times the natural
scientist was "one of us", and whereas in earlier centuries "natural
science and philosophy spoke the same language" (Bachelard 1971: 16),
modern physics forces us to adopt a completely new ontology in which
objects are no longer basic. This new ontology also implies "an absolute
break" with the conceptions and the "reism" of classical physics (ibid.,
59).
Modern physics is characterized by Bachelard with notions like
"noumenology", "active empirism", "phenomenotechnique", "rational
materialism", and "applied" or "dialectical rationalism". The common
denominator of these various labels is the idea that twentieth century
physics does not start from observation but from "an organization of
objects of thought" that only subsequently "become objects oftechnical-
experimental experience, in a pure artificiality of experience" (ibid., 62).
In other words, the objects of physical science are constructed by
experimental means, not found; modern physics is "a science of effects
rather than facts", "experience is no longer a starting point, it is not a
simple guide any more, experience is a goal" (ibid., 75).
28 PART I
The experimentalist is not, however, the mere handyman of the

theoretician. Rather modern physicists have an "abstract-concrete men-
tality" (ibid., 122) that allows challenges to go both ways: the experi-
mentalist will search for the vindication of a phenomenon which the
theorist announces as a possibility, whereas the theorist will, in the light
of new experimental phenomena, modify her theory in order to show
that her theory should have predicted the phenomenon in question (ibid. ,
123). Bachelard justifies the assumption of a "dialectics" between theory
and experiment by pointing out that the instruments of modern physical
experimentation are but "reified theorem[s]" (ibid., 149).
What gives Bachelard's reflections upon the history of modern
physics in particular, and the development of science in general, their
specific "epistemological" dimension, is first of all his thesis that diffe-
rent forms of scientific practice involve different epistemological views,
and that the order in which the discourse of physics has gone through
these positions provides us with "a principle of classification for diffe-
rent philosophies and for the study of the progress of reason" (ibid., 34).
In La philosophie du Non (Bachelard 1978), Bachelard explains this
classification in terms of the notion of mass. For the child mass is
something which is judged in terms of size, "mass is only a quantity if
it is fairly big" (ibid., 36). This is the level of "animism". A higher
level, that of "realism" , is reached when the notion of mass is linked to
the use of a scale. In this stage, thought is "empirical, solid, clear,
positive, [yet] immobile" (ibid., 39). The Newtonian notion of mass
represents rationalism, since here the absolute use of the notion is
replaced by a correlative use of several concepts. Mass is now defined
in terms of other concepts and no longer regarded as an immediate
element of direct experience. This marks the transition from "a realism
of things to a realism of laws" (ibid., 41). With the theory of relativity,
we attain the level of "complex realism"; now the earlier belief that the
mass of an object can be determined independent of its speed has to be
given up; the notion of absolute mass turns out to be meaningless.
Furthermore, mass is no longer independent of energy (ibid., 44).
Finally, with Dirac's mechanics we arrive at the stage of "dialectical
surrealism", the stage where we are dealing with two masses for one
object: "One of these masses comprises in a perfect way everything
which the four earlier philosophies knew about mass ... But the second
mass, the dialectical counterpart of the first, is a negative mass" (ibid.,

48).
In Le rationalisme applique (1949), Bachelard's "philosophical
spectrum" differs somewhat; here he presents (naive) realism, empirism
and positivism, on the one hand, and idealism, conventionalism and
formalism, on the other hand, as two diametrically opposed series of
philosophies, both of which are unable to grasp the nature of modem
physics as "applied rationalism": formalism, conventionalism and
idealism represent a series which increasingly loses the empirical element
of science, while positivism, empirism and (naive) realism are successive
stages of losing the constructive-experimental and mathematical-theoreti-
cal side (Bachelard 1971: 127).
While castigating these various epistemologies as inadequate philo-
sophical stands, Bachelard holds that in everyday life as well as in
science, we all are philosophical pluralists. For instance, in different
situations we employ different notions of mass, Le. we rely tacitly on
different philosophies. Bachelard suggests that we can draw "episte-
mological profiles" of our understanding of scientific terms, profiles
which depict the extent to which we abide by different philosophies
(1978: 57).
In order to clarify transitions from one epistemological stage to the
next, Bachelard introduces the dichotomy "epistemological obstacle"
versus "epistemological act". Epistemological obstacles are what hinders
the scientist in advancing to the next stair of Bachelard's epistemological
staircase, while epistemological acts are successful steps made on the
way up the staircase. Bachelard never provides any detailed typology of
different epistemological obstacles, even though in La formation de
1'esprit scientifique he distinguishes between "first experience", the
"realistic obstacle", the "animistic obstacle" and "the libido". The first
experience constitutes an obstacle in so far as its vividness and manifold
character might hinder us to pose scientific questions; realism fixes our
attention on things and substances; animism tempts us to treat active
substances as animate; and the libido is responsible for our inclination to
depict natural processes in terms of sexual metaphors (Bachelard 1971:
178-86).
By attending to different uses of the notion of "epistemological
obstacle" in Bachelard's many books, one can distinguish at least the
30 PART I
following cases: (1) "the primary experience" (the object of knowledge

is taken to be immediately given in itself); (2) "the substantialist obsta-
cle" (the assumption that the order of the real mirrors the order of
thought; thus, e.g. the subject/predicate scheme is imposed upon nature
as the substance/accidens scheme), (3) "general knowledge" (a system of
concepts is composed based on vague analogies); (4) "pragmatic knowl-
edge" (non-scientific interests and inclinations); (5) "the animist obsta-
cle" (confusion of regions of objects, e.g. between physics and biology);
and (6) "the quantitative obstacle" (the overestimation of quantification
and measurement) (Briihmann 1980: 143-44).
These various forms of epistemological obstacles also provide via
negationis some content for the opposite concept, the epistemological
act, of which Bachelard only tells us that they are "those blows of
scientific genius that bring unexpected impulses in the course of scienti-
fic development" (1971: 213). By means of epistemological acts the
scientist cuts herself free from various inclinations and prejudices. In this
sense, the epistemological act constitutes a leap forward to a level of
thought that comes closer, or ultimately reaches, the stage of applied
rationalism. More precisely, in escaping the epistemological obstacles,
the scientist develops a system of concepts and experimental techniques,
in other words, a "problematique", which by an internal dynamic -
which Bachelard leaves unanalysed -leads to ever new questions, radical
reorganizations of theories, and new instruments (Briihmann 1980: 144;
Bachelard 1971: 32, 131, 135).
The dichotomy "epistemological obstacle vs. epistemological act"
also shapes Bachelard's views on the issue of continuity vs. discontinuity
in history, and thereby his conception of how the history of science
should be written.
To begin with, Bachelard's above-mentioned stress on discontinuity
is hardly surprising given his belief that decisive scientific advances
involve epistemological acts, on the one hand, and a re-organization of
systems of concepts and experimental techniques, on the other hand.
However, for the epistemologist these discontinuities do not make talk of
progress impossible. Of course, using the philosophical spectrum as an
indicator of progress would be circular, since this spectrum has after all
been derived from the factual development of physics. What allows
Bachelard to stick to the perspective of progress despite the existence of
discontinuities is rather his belief that the rejection of a past science does
not rule out the preservation of its results in the form of special, limited
cases of later theories:
Non-Euclidean geometry was not invented in order to contradict

Euclidean geometry.... First constructed in the margins of Euclid-
ean geometry, non-Euclidean geometry sheds a revealing light on
the limitations of its predecessor. The same may be said of all the
new varieties of scientific thought, ... (Bachelard 1984: 8).
Put differently, Bachelard denies that the new knowledge can be ex-
pressed in the old, rejected framework (" ... one must not hope to find a
simple formula for converting the new doctrines into terms comprehensi-
ble within the framework of the old", (ibid., 8)), but he allows for the
old knowledge to be expressed in the new. Perhaps it is this asymmetry
which also informs his opposition to the categories of influence and the
predecessor. Bachelard rejects as fruitless - at least for modern science
- to speak of the influence of major ideas and discoveries across the
borders of problematics (1971: 203), and he castigates attempts to make
Hegel the precursor of Maxwell, and Raspail the forerunner of Bohr (see
Lecourt 1975: 84-85).
Rather than search for influences and anticipations, the epistemologi-
cal study of the history of science must construct two series of events,
the series of epistemological acts and the series of epistemological
obstacles. The former is the constructed object of "recurrent" or "ac-
cepted history" (histoire sanctionee), the latter the constructed subject
matter of "obsolete history" (histoire perimee) (Bachelard (1951: 24-27).
Neither of these two forms of writing history confines itself to a neutral
narration of the history of science dated by the calendar; rather, both
amount to an evaluative reconstruction of that history. Accepted history
reconstructs history in terms of how it should have happened, obsolete
history in terms of how it should not have developed. Accepted history
identifies the steps that approach the truths of present-day science,
obsolete history studies those elements that have no place in the teleo-
logical scheme of accepted history.
Finally, Bachelard's profile of philosophical attitudes involved in
science enables him to speak of different rhythms and times of sciences,
32 PART I
depending upon the rates at which these sciences cross the various
thresholds. By the same token, the historical epistemologist can also
propose that two sciences that are contemporaries when judged by the
calendar, are not contemporaneous when judged by their epistemological
profile (Briihmann 1980: 157): "Science is like a half-renovated city,
wherein the new (the non-Euclidean, say) stands side by side with the
old (the Euclidean)" (Bachelard 1984: 7).
Canguilhem
Bachelard's student and successor as the director of the Institut d'His-

toire des Sciences et des Techniques at the University of Paris, Georges
Canguilhem, continues on several of Bachelard's central lines while
taking exception to some others. To begin with those aspects of the
epistemological program which Canguilhem underwrites or pushes
further, Canguilhem shares his teacher's conviction that "the history of
science is the history of the defeats of irrationalism" (Canguilhem 1968:
183). He stresses the importance of the epistemological obstacle notion
as a means of avoiding the illusion that progress is linear and that the
history of science can confine itself to the recording of results (ibid.,
177). Canguilhem too holds that different sciences have different
rhythms of development and different speeds at different times, and that
the "epistemology of continuity" is insufficient and needs to be comple-
mented by an "epistemology of discontinuity" (1988: 16). For instance,
Canguilhem suggests that biophysics and biochemistry constitute a
radical break with earlier biology. A crystal of DNA is "a 'superreal',
nonnatural object, the product of considerable technical and theoretical
labor ... of microextraction and microdissection, of combinatorial
algebra, of statistics, of electron microscopy, of enzyme chemistry". In
order to arrive at this new object, central elements of what formerly
were taken to be crucial features of living things, like sexual reproduc-
tion and enzymatic reactions, had to be left aside. The new object of
research was inconceivable for earlier biology, and its constitution in and
through these new various techniques amounts to the "conversion to a
new view of the world" (ibid., 117-18).
In agreement with Bachelard, Canguilhem also emphasizes the
importance of problematics. Thus he argues, for example, that Mendel's

discovery of the laws of heredity could not have hastened progress in
biological research even if they had been recognized by scientists of the
day. At the time, Mendel's work could not have been linked with
Darwin's theory of variations: interest in evolution was then running
high, and Mendel's laws would have appeared to his contemporaries as
an outdated form of fix ism (ibid., 116). Furthermore, Darwin's work
could not have been combined with a chemical study of the cell nucleus
in a single problematic, since unicellular species were not considered as
organisms, and since no-one could yet conceive of the possibility that the
laws of life might be discovered by studying bacteria and microbes, Le.
discovered in the very threats to life itself (ibid., 115).
Yet another aspect of Bachelard's epistemology that Canguilhem
develops at some length is the criticism of the search for predecessors.
For Canguilhem, the "virus of the predecessor" destroys the history of
science, since "if there really were precursors, then the historical dimen-
sion of science would be an illusion" (1968: 21). It is the lack of episte-
mological critique, of analyses of problematics and obstacles which
result in the anachronistic construction of a few central lines of influ-
ence. Such a relation, i.e. the relation of influence, between two authors
may be assumed only once epistemological critique has shown that these
authors have identical questions as well as research strategies and that
the leading concepts of their respective systems of concepts are identical
(ibid., 22).
Other than Bachelard, however, Canguilhem does not employ the
normative scheme of the "philosophical spectrum" and "epistemological
acts" of genius, and he focuses more on social factors, ideology and the
interrelations between disciplines. Canguilhem will have none of Bache-
lardian "strokes of genius" not only because the work of every scientist
is always a mixture of the old and the new (1988: 15), but also because
the individual scientist is usually dispensable. Had Darwin died during
the voyage of the Beagle, his discoveries would have been made by
Alfred Russel Wallace (ibid., 110-11).
An extensive use of the notion of the flash of insight easily leads to
a conceptualization of the history of science as a series of accidents. This
is a viewpoint, however, which Canguilhem seeks to exorcise from
history. He argues, for instance, that the discovery of iodine by Clemont
34 PART I
and Desormes in 1812 was not - as is often thought - the result of an

accident. Chemists at the time were searching for active substances
present in organic compounds; thus even if a Parisian manufacturer of
saltpetre had not brought the unknown substance which caused extensive
corrosion to Clemont an':! Desormes, iodine would have been discovered
sooner or later anyway (ibid., 274-94; see also Lecourt 1975: 169-70).
Canguilhem's attention to the interplay among disciplines and the
interrelations between sciences and society comes out clearly in his
remarks on the history of statistics as well as in his two best-known
studies, Le Normal et Ie pathologique (1943) and La formation du
concept de reflexe aux XVlr et XVIIr siecles (1955).
The emergence of statistics, we are told, cannot be derived from the
mathematics of Laplace, the biology of Darwin, the psychophysics of
Fechner or the ethnology of Tylor. Biometrics and psychometrics
become possible once new objects of observation became accessible
outside of science. Quetelet's statistical study of human extremities
became possible through the creation of national, standing armies, and
Binet's study of intelligence was made possible by the introduction of
elementary schools. Put differently, "the history of science is not dealing
with merely a group of sciences without inner relations among them, but
it is also dealing with non-science, with ideology and political and social
practice" (Canguilhem 1968: 18-19).
In his study of different notions of disease in nineteenth century
medicine and physiology, Canguilhem demonstrates that the shift from
a quantitative notion of disease (disease as merely gradually different
from the normal, healthy state) to a qualitative notion of disease (disease
as principally different from the healthy state) must take into account not
only medicine but also sociology. This is because the quantitative notion
of disease and normality is first taken over by Comtean sociology,
transformed for sociological purposes, only to re-enter medicine later in
the century, where it is then soon replaced by the qualitative conception
(Lepenies 1978: 446).
Finally, in his investigations into the origins of the notion of reflex,
Canguilhem provides his reader with an excellent example of his theses
that epistemological critique must carefully construct its historical
trajectories, pay special attention to concepts, and not be misled by
superficial similarities (cf. Lecourt 1975: 171-77; Gutting 1989: 35-37).
Canguilhem seeks to reject the assumption that a description of reflex

movements within a mechanistic biology suffices to attribute the concept
of reflex to Descartes. This is because, for Canguilhem, a concept is
more than a description. It provides an interpretation of data that makes
fruitful scientific work possible. In the case of reflex motion, the concept
must include the ideas that the two movements from the sense-organs to
the response-center and back are alike, and that the brain is not directly
involved. Neither of these two decisive elements are part of Descartes's
account, however. As Canguilhem shows, the concept of reflex motion
originates instead within a framework which is foreign to mechanical
biology. The concept of reflex has its origin in the work of Thomas
Willis and Georg Prochaska, two vitalists of the seventeenth and eigh-
teenth century. Since they likened life to light, they could model reflex
motion on the reflection of light and conceive of a symmetry of the two
movements involved. Furthermore, since they distinguished between
cerebellum as the center of involuntary movement and cerebrum as the
brain, they could conceive of movements originating from "eccentric"
centers (Canguilhem 1955: 68, 127). Only in the late nineteenth century
did physiologists place this concept of reflex within a mechanistic
physiology. Thus Canguilhem can argue that concepts need not be homo-
geneous with the theories in which they ultimately will figure; concepts
are "theoretically polyvalent" (ibid., 6).
Althusser
Bachelard's epistemological oeuvre was not only read and used by

historians of science like Canguilhem. Bachelardian notions and theses
can also be found in the writings of the leading figure of French struc-
turalist marxism, Louis Althusser. Perhaps it is no exaggeration to claim
that at least part of Althusser's work can be read - and certainly was
read so in the France of the late sixties - as an attempt to formulate,
within a marxist framework, conclusions about the nature of science
deriving in part from premisses that Althusser found in Bachelard and
Canguilhem.
To begin with, Althusser rejects the notion, also repeatedly criticized
by Bachelard and Canguilhem, that science finds its objects 'ready-made
36 PART I
out there'. For Althusser, we have to free ourselves of such "mirror

myths" by conceiving of knowledge as a production (Althusser 1970a:
24). In conceptualizing this production process, Althusser uses Marxian
as well as Bachelardian terms. He suggests that this "theoretical produc-
tion" always happens according to a specific, historically changing
"mode o/production of knowledges", Le. "an apparatus o/thought", or
"a structure which combines (...) the type of object (raw material) on
which it labors, the theoretical means of production available (its theory,
its method and its technique, experimental or otherwise) and the histori-
cal relations (both theoretical, ideological and social) in which it pro-
duces" (ibid., 41). Here "the raw material" is not a bunch of sense data
or "the real object" out there, but rather the subject matter of science as
it is constituted by the prevailing state of scientific knowledge, its
technical means, and by relations between science and society (ibid., 43;
cf. Benton 1984: 36-39). Althusser also writes that this raw material, as
object and problem, is what is "visible" for a given scientific field at a
given moment in time. More precisely, this object is visible "within the
horizon of a definite theoretical structure, its problematic, which con-
stitutes its absolute and definite condition of possibility, and hence the
absolute determination of the forms in which all problems must be posed,
at any given moment in the science ... " (Althusser 1970a: 25). This field
of the visible, however, is not opposed merely to what remains as yet
unknown, but also - with reference to Foucault's works on madness and
the clinic - also and primarily set apart from "the inner darkness of
exclusion", Le. from forms of knowledge that are ruled out and sup-
pressed as unscientific within a given science (ibid., 26-30).
In full accordance with Bachelard, Althusser proposes that there is
no absolute standpoint from which the internal object of a science can be
compared to the real object. Instead, theoretical practice itself defines its
own criteria of knowledge and truth since it "contains in itself definite
protocols with which to validate the quality of its product". The model
here is the practice of mathematics:
No mathematician in the world waits until physics has verified a

theorem ... although whole areas of mathematics are applied in
physics: the truth of his theorem is a hundred per cent provided by
criteria purely internal to the practice of mathematical proof ...

(ibid., 59).
Althusser also makes use of the Bachelardian notions of rupture and

discontinuity. In contrast with Bachelard, he applies these concepts
foremost to mark the difference between ideology and science. Accord-
ing to Althusser a new science can emerge only by rejecting "the pre-
scientific theoretical ideologies that occupy the 'terrain' in which it is
establishing itself" (1977: 13). Put differently, the prehistory of a
science is always an "ideological theoretical practice" which is "qualita-
tively" distinct from, and discontinuous with, its history (ibid., 168).
Althusser is not very outspoken on what precisely defines an ideological
theoretical practice but he does tell us that in this practice the validation
of claims and theories does not yet happen by criteria internal to the
theoretical practice, or that in this practice only such problems can be
formulated whose "solution" has already been provided by "extra-
theoretical instances and exigencies (by religious, ethical, political or
other 'interests')" (1970a: 52). A further hint is provided where Althus-
ser contrasts the scientific concept of socialism with the ideological
concept of humanism: whereas the latter can only "designate" certain
existing relations, the former "does ... provide us with a means of
knowing them" (1977: 223). Perhaps this distinction between the ideol-
ogical concept and the scientific concept is modelled upon the Canguil-
hemian opposition, encountered above with respect to reflex motion,
between the description and the concept qua adequate interpretation of a
phenomenon.
In Reading Capital, Althusser also suggests an "outline of a concept
of historical time" (Althusser 1970b). Althusser seeks to show that
Marx's Capital contains a theory of historical time which is able to
improve upon the work of Braudel and other Annalists, and which is also
of immediate relevance to the history of science. The ideological con-
cept, from which this new theory has to radically separate itself, is the
Hegelian concept of history and historical time. This concept is charac-
terized by two main features, the "homogeneous continuity" and the
"contemporaneity" of time. The first characteristic means that time is
understood as a continuum "in which" the dialectical development of the
Hegelian Idea through its various stages takes place. To conceive oftime
38 PART I
in this way, Althusser claims, makes it natural to see the defining and
delimiting of historical periods as the central task of the historian. The
second feature of Hegelian time, the contemporaneity of time, reinforces
this tendency. The contemporaneity thesis is the assumption that all parts
of a given whole, i.e. all elements of a given level of the development
of the Hegelian Idea, coexist. None of them can lag behind the others,
and none can run ahead of its time. In other words, this postulate assures
the historian that wherever she makes a vertical cut through history, she
will inevitably find "all the elements of the whole revealed by this cut ...
in an immediate relationship with one another" (ibid., 94).
Althusser claims that this Hegelian-ideological concept of time is
still with us, and that not even Braudel, Febvre and serial history have
fully broken with it. This is because even though they "observe" that
there are different time spans and different temporal shapes, these
historians have no theory on how these different times are interrelated.
In the absence of such a theory, however, they have to fall back on the
Hegelian notion of time and define their different times in terms of their
duration, i.e. in terms of the "ideological time continuum" (ibid., 96).
Opposing such a drawback, Althusser demands that the various
temporal rhythms and shapes of different social practices and processes
be interrelated not by the common measure of duration, but rather by a
marxist theory of society that explains the specific nature or "relative
autonomy" of each of these times, while ultimately, "in the last in-
stance", explaining them in terms of the economic structure. Thus, the
common measure of the various times that historians want to speak about
is provided by one - or a bundle - of those times - namely the time(s)
of the economy - rather than by an 'absolute time' like the Hegelian
continuum (ibid., 97).
Althusser speaks of economical time as the rhythms in which
different modes of production develop their productive forces (ibid., 99).
This time is a specific time, yet "complex and non-linear":
... a time of times, a complex time that cannot be read in the con-
tinuity of the time of life or clocks, but has to be constructed out of
the peculiar structures of production. ... The concept of this time
must be constructed out of the reality of the different rhythms ... :
out of the concepts of these different operations, e.g. the difference
between production time and labor time, the difference between the
different cycles of production (the turnover of fixed capital, of
circulating capital, of variable capital, monetary turnover, turnover
of commercial capital and of financial capital, etc.) (ibid., 101).
To construct the concept of a history of any domain, Althusser

proposes, is to define the "historical facts", Le. those events "which
cause[..] a mutation in the existing structural relations" (ibid., 102). For
instance, in order to construct "the concept of the time of the history of
philosophy" (ibid., 101), we have to reject historical chronology, clarify
the relation of philosophical practice to the sciences and ideology, and to
structure the time of philosophy in terms of "philosophical facts" or
"philosophical events" that "cause real mutations ... in ... the existing
theoretical problematic" (ibid., 102). This will lead us, Althusser
believes, for instance, to regard Locke and Spinoza as more important
than German Idealism, and to draw a direct line from Spinoza to Marx;
after all, for Althusser, Spinoza is the only "direct ancestor" of Marx
(ibid., 102).
Althusser holds, furthermore, that his - or the Marxian - concept of
time also passes beyond the continuity vs. discontinuity opposition as far
as culture as a whole is concerned. Since his concept of time renounces
the idea of the contemporaneity of time, Le. since he rejects the idea that
the elements of any time slice are all expressions of some core meaning
like the Hegelian or Burckhardtian Zeitgeist, in any time slice there will
always be both continuities and discontinuities: "The break 'valid' for a
determinate level ... does not correspond to anything of the kind in the
other levels ... which live in different times and know other breaks,
other rhythms and other punctuations" (ibid., 104). Likewise, Althusser
is even skeptical about irregular or "stepped or multiply toothed cuts" or
strata, since the construction of such strata is easily informed by the
assumption that the elements standing out from the otherwise vertical
surface are but further expressions of the one whole or core meaning
(ibid., 105).
Moreover, the notions of synchrony and diachrony can only be
retained when understood along the lines of the new conception of
historical time as construction. Synchrony is not the co-presence of two
processes, events or things, but the existence of a construction of rela-
40 PART I
tions that connect these processes, events or things. Mutatis mutandis,

the existence of diachrony between facts is nothing but the existence of
a construction that defines transformations between them (ibid., 108).
Finally, Althusser submits that the inability of historians to conceive
of historical time as an interrelation of different times in the way sug-
gested by his outline, is due to their lack of theories: "To put it crudely,
history lives in the illusion that it can do without theory in the strong
sense, without a theory of its object ... What acts as its theory is its
methodology ..." (ibid., 109). A methodology, however, that is not
grounded in a theory of historical time - which in turn is built upon a
theory of society - can only be ideological. In other words, as long as
the discipline of history has no theory of its object, as long as it has no
problematic, the discipline remains at the level of a prehistory of a
scientific history still to come. It is bound to having its problems and
objects determined by religious, ethical and political interests, and it is
unable to define its own criteria of validation.
4. ARCHAEOLOGY, THE NEW HISTORIES,

AND THE HISTORY OF IDEAS
It is one of the peculiar facts of twentieth century French intellectual

history that the two new strands in historical scholarship, the Annales
School and the epistemologists, have by and large ignored each other's
work. Lucien Febvre's 1939 review of Bachelard's Psychanalyse dujeu
(Bachelard 1964) remained the only review of Bachelard's many books
in the Annales, and none of Koyre's or Canguilhem's books was ever
reviewed in this journal (Chartier 1982: 31). Furthermore, in a program-
matic article on how the history of science should be written, "Sur
Einstein et sur l'histoire" (Febvre 1955), published in the Annales,
Febvre makes no mention of Koyre's, Bachelard's and Canguilhem's
work. Febvre suggests that the study of the history of science be inte-
grated into general history (ibid., 306). Likewise, references to the
works of the Annales historians are missing from the texts of the episte-
mologists.
The two schools of thought meet each other, it seems, only in the
writings of authors that stand somewhat outside of both, Le. in the
writings of philosophers like Althusser and Foucault. Foucault's relations

with both groupings are visible in that he published Folie et deraison in
a series edited by Philippe Ari~, one of the leading figures of the
Annales school, and La naissance de la ciinique in a series edited by
Canguilhem. As evidenced by favorable comments coming from Ari~s,
Braudel, Revel, Canguilhem and Althusser, Foucault also managed to
gain recognition in both camps.'
Foucault's close awareness of developments in both schools is
especially obvious in The Archaeology of Knowledge; in fact, large
portions of this book can be read as a philosophical reflection of, and a
critical comment upon, the state of the art in both areas of study.
CORRELATING THE NEW HISTORIES
In the introduction of The Archaeology of Knowledge (AK 3-17) Fou-

cault is mainly concerned with stressing the points of convergence of
Annalist and epistemologist paradigms. Indeed, he states as his aim, "to
uncover the principles and consequences of an autochthonous transforma-
tion that is taking place in the field of historical knowledge" (AK 15).
As Foucault sees it, these points of convergence are the concern with the
definition of series, a new attitude towards the document, an attention to
discontinuities, and a rejection of totalizations; in short, what unites both
schools is the awareness that historical scholarship cannot but construct
its subject matter.
While Foucault starts by depicting Annalist and epistemologist
scholarship as somewhat diametrically opposed tendencies - concern
with long periods in the one camp, eagerness to break up long periods
into discontinuous strata in the other - he soon makes clear that both
features are the natural result of the same basic constructivist trend. In
both cases historians no longer stick to traditional, received periodiza-
tions and levels of events, but seek to define new trajectories in terms of
new criteria. In the case of serial history, this leads to the isolation of
ever new strata and series, to new criteria of periodizations and the
search for novel interrelations between types of series (AK 4); in the
case of the epistemologists it results in an attention to ruptures and trans-
formations, in the breaking up of allegedly self-evident lines of tradition,
42 PART I
and in the questioning of unities like the scientific discipline, the oeuvre
of an author, the theory or the concept (AK 6).
Foucault characterizes the transformation, which issues in serial
history and epistemological historical study in terms of the changed
attitude towards the document. The document is no longer the primary
unit of interpretation and of internal as well as external text-criticism.
Instead, documents are organized into series, and broken up into parts
and elements. Alluding to the archaeological practice of describing and
analyzing monuments (Le. archeological relics that, unlike artifacts,
cannot be removed from the site of excavation, (Childe 1962: 11»,
Foucault sets up an opposition between the new and the old histories in
the following way:
... let us say that history, in its traditional form, undertook to

'memorize' the monuments of the past, transform them into docu-
ments, and lend speech to those traces which, in themselves, are
often not verbal, ... in our time, history is that which transforms
documents into monuments. ... in our time history aspires to the
condition of archaeology, to the intrinsic description of the monu-
ment (AK 7).
Whereas the traditional historian took for granted the existence of

certain 'natural' series and saw as her main task the placing of event-
units within them, the modern historian first defines different types of
events (e.g. according to their duration) and only then organizes the
tokens of the respective types into series of different length and struc-
ture. Furthermore, whereas for the traditional historian "the discon-
tinuous was both the given and the unthinkable", for the modern histo-
rian it is "one of the basic elements of historical analysis" (AK 8). For
the traditional historian the discontinuous was given in the form of
individual and unique events and documents. Subsequently, this discon-
tinuity was transformed into continuity by drawing on totalizing concepts
like historical evolution, Zeitgeist or We/tbi/d. In the hands of the
modernes, however, discontinuity figures as operation, result and con-
cept to be specified: it figures as "deliberate operation" in so far as the
historian must start from the supposition that certain events can be
organized into series in partial isolation from others; it is "the result of
his description" in so far as the historian searches for ruptures in knowl-

edge or points where cyclical changes start and end; and it is "the
concept that the historian's work never ceases to specify" in so far as
discontinuity is an umbrella concept: "one does not speak of the same
discontinuity when describing an epistemological threshold, the point of
reflexion in a population curve, or the replacement of one technique by
another" (AK 9).
Taken together, Foucault concludes, the various ingredients of the
conscious constructivism of the new histories amount to a new methodo-
logy of historical inquiry - and this despite the fact that many of the
now-prevailing methodological problems have existed before. It also
implies a deadly blow to the earlier project of "total history" and the
philosophy of history by which it was informed. The various attempts of
total history to secularize, by means of notions like Zeitgeist or Weltan-
schauung, the teleological Hegelian scheme turn out to be impossible
projects once confronted with the serial historians' descriptions of ever
new discontinuities, diversities, and asymmetries. The projects of total
history must thus be replaced by a "general history" that confines itself
to the description and explanation of ever more series in their partial
interrelation (AK 9).
A CRITICAL RE-EVALUATION OF THE HISTORY OF IDEAS
Outlining what he takes to be the decisive turn of more recent French

historical scholarship is only one of several strategies which Foucault
employs in order to define the character of his archaeology of knowl-
edge. Besides placing himself in the vicinity of serial history and episte-
mology, he also seeks to draw a line of separation between archaeology
and a variety of other historical inquiries into knowledge and science.
This variety of approaches is labelled, in The Archaeology of Knowl-
edge, "history of ideas", even though the criticism that is levelled
against the "history of ideas" is also, and by the same token, directed
against certain viewpoints of the Annalist and epistemological schools,
against assumptions prevailing in some quarters of the history of science,
and against a number of philosophical approaches to history.
Rather than merely criticizing certain particular shortcomings of the
44 PART I
history of ideas, Foucault wants to go further and establish that these

shortcomings are grounded in the very character of this approach to
history itself. Since this brand of historical research lacks a clearly
defined object, its relation to other forms of historical studies remains
obscure, its methods borrowed, and its criteria and categorizations
unreflected. Moreover, it is based upon a pervasive evaluative scheme
that it has inherited from the nineteenth century. To be sure, this is not
to say that all studies carried out under the title "history of ideas" have
produced nothing of value whatsoever. Rather it is to say, first, that the
history of ideas lags behind many other forms of historical research in
rigor and scholarship and, second, that its occasional misconstructions
are not as coincidental as they might appear upon first sight.
In order to see that Foucault's analysis of the history of ideas is not
unfair, we need to compare his claims with work actually done under
this heading. As far as Foucault's more general claims are concerned,
the natural reference point - at least for an English speaking audience -
is of course Arthur Lovejoy. Lest it be suspected that other historians of
thought have avoided the pitfalls of Lovejoy's work and program, I shall
relate Foucault's central points to some other critics of the history of
ideas, on the one hand, and to more recent studies in the history of
ideas, on the other hand.
Foucault and Lovejoy
To begin with Foucault's claims that the history of ideas has "difficulties
in demarcating [its] ... domain", "in defining the nature of its object",
and in determining its relation to other disciplines (Foucault 1978: 18),
it is not difficult to validate his assessment. For instance, in his program-
matic paper "The Historiography of Ideas" (Lovejoy 1960: 1-13),
Lovejoy starts from the observation that "historical study having to do,
more or less, with ideas" can be found "under at least twelve different
labels": the history of philosophy, the history of science, the history of
language, the history of religious beliefs, literary history, comparative
literature, the history of arts, economic history and the history of eco-
nomic theory, the history of education, political and social history, as
well as the historical part of sociology (ibid., 1-2). Lovejoy wavers
somewhat on whether the history of ideas sui generis is an interdisci-

plinary project involving all of these disciplines or whether it possesses
a domain of its own. In the end, he comes down on the side of the latter
option, arguing that it is the study of "unit-ideas" which distinguishes
the history of ideas from the other disciplines mentioned. These unit-
ideas, we are told, must be studied in relative separation from the twelve
other disciplines because they can manifest themselves in all of the
different domains of thought that those disciplines study:
There is here another distinct realm of historiography, which needs

to be added to the dozen mentioned at the outset, ... Until these
units [Le. unit-ideas] are first discriminated, until each of them
which has played any large role in history is separately pursued
through all the regions into which it has entered and in which it has
exercised influence, any manifestation of it in a single region of
intellectual history, or in an individual writer or writing, will, as a
rule, be imperfectly understood ... (ibid., 9).
Unit-ideas, as Lovejoy explains in the introduction to his famous

The Great Chain of Being (Lovejoy 1973: 3-23), are a "decidedly
limited" group of ideas that live through the centuries, combine in
different ways in the works of philosophers, scientists, theologians and
artists, are only rarely invented and die slowly (ibid., 3-4). Unit-ideas
can be "implicit or incompletely explicit assumptions or more or less
unconscious mental habits" (ibid., 7), "dialectical motives ... [e.g. a]
turn of reasoning, [a] trick of logic, [a] methodological assumption"
(ibid., 10), "kinds of metaphysical pathos" (ibid., 11), "sacred words
and phrases", or "a single specific proposition or 'principle' expressly
enunciated by the most influential of early European philosophers"
(ibid., 14).
Given this diversity one will certainly need to add even more
disciplines to Lovejoy's bundle, for instance, disciplines like historical
psychology, anthropology and the history of mentalities. Furthermore,
and more importantly, note that Lovejoy manages to save the history of
ideas from resolving into other disciplines only by means of two rather
awkward assumptions: that a historical study of the surfacing of an idea
in different domains can precede the study of the role of that idea within
46 PART I
the respective fields, and that ideas can somehow be decontextualized. It

is also worth recording that Lovejoy remains silent on the question as to
the methodology by which such units are to be identified, or as to the
criteria by which we decide when two texts express "the same" idea. To
say that the history of ideas is "somewhat analogous to ... analytical
chemistry" (ibid., 3) is merely a metaphor, and Lovejoy's call for
"philosophical analysis" remains somewhat empty as long as we are not
told just what kind of philosophical analysis is involved (Lovejoy 1960:
8).
Foucault also claims that it is possible to identify "three great
themes" within the history of ideas, to wit, "genesis, continuity, totaliza-
tion" (AK 138): genesis in so far as the origin of ideas or themes is
regarded as a crucial question; continuity and totalization in so far as
different eras or different contemporaneous fields of thought are made
homogeneous by the assumption of omnipresent elements.
Foucault regards these themes as guiding presuppositions in the
choice of topics and the style of analysis of historians of ideas. With
respect to the former, Foucault holds that the history of ideas favors the
study of knowledge "which could never in the whole of its long, persis-
tent life attain the form of scientificity" (AK 136). The history of ideas
tries to find its niche between the histories of the sciences and literature
by studying Diderot and Fontenelle rather than Descartes and Leibniz
(cf. Foucault 1971a: 144), by attending less to "the great discursive
monuments" than to widespread opinions, and by studying errors rather
than truths, or mentalities rather than "forms of thought" (AK 137).
As concerns the typical style of investigation of the history of ideas,
Foucault sees it as characterized by the dissolving of the works of
historical figures into elements which are common to many fields or men
in the respective period, on the one hand, and by a reconstruction of the
life-histories (birth, gradual growth, maturity, diffusion, death) of
themes, on the other hand (AK 137). The pervasive goals of the enter-
prise are to distinguish the original from the regular, and to reduce the
multitude of differences or contradictions to an underlying harmony.
First, the historian of ideas wants to determine who originally came up
with a certain theme or idea or who combined prevailing ideas in the
most unusual way (AK 141). Second, she is also eager to derr.onstrate
that what at the surface seem to be crucial differences, can be traced
back as different surface effects to an underlying, perhaps unconscious

unit, or that diversity can be grouped around a few, or a single, opposi-
tions (AK 149).
Again it seems hard not to concede to Foucault that his analysis is
on the mark, at least as far as Lovejoy is concerned.' First, Lovejoy
wants his followers to be "especially concerned with the manifestations
of specific unit-ideas in the collective thought of large groups of per-
sons" and to pay attention to common "beliefs, prejudices, pieties,
tastes, aspirations, current among the educated classes through, it may
be, a whole generation, or many generations" (Lovejoy 1973: 19). Even
though he assumes that most unit-ideas can be traced back eventually to
philosophers, he suggests that historians of ideas "seek for [their] most
significant manifestations in art, and above all in literature" (ibid., 17).
Lovejoy also wants the history of ideas to lay bare the fact that the
history of thought is "in great part, a history of confusions of ideas"
(ibid., 22). Second, Lovejoy reduces, in both synchronic and diachronic
perspective, the diversity of philosophical systems, works of arts and
theories of science to an antithetical conflict of a handful of unit-ideas.
Moreover, he also calls for a sound knowledge of philosophy in order to
be able to correctly find the philosopher-originators (Lovejoy 1960: 8)
and he hopes to be able to "penetrate" behind "surface-dissimilarities"
to "the common logical and pseudo-logical or affective ingredients"
(Lovejoy 1973: 4). The urge to reduce plurality to harmonic simplicity
is also evident where Lovejoy writes that the history of ideas "aims at
interpretation and unification and seeks to correlate things which often
are not on the surface connected" (ibid., 21). Genesis, continuity and
totalization thus figure indeed centrally in Lovejoy's program.
The mythology of ideas
Needless to say, Foucault does not stand alone in his criticism of the
history of ideas. 8 Unfortunately, the amount of criticism levelled against
the history of ideas, on the one hand, and certain of its practices with
counterparts in other domains of historical scholarship, on the other
hand, has not led to a situation in which this criticism has become
obsolete.
48 PART I
Given his general view of the history of ideas, it is not surprising

that Foucault directs his polemics against certain concepts and strategies
which the historian of ideas (or of science, or of mentalities) uses in
attempts to reduce differences, to operationalize her opposition between
the original and the regular, and to vindicate some teleological scheme
of progress.
To begin with the notion of tradition, Foucault suggests that the
application of this concept is often marked by the preoccupation to
reduce, in an arbitrary fashion, a variety of phenomena to the "form of
the same" (AK 21). An unqualified use of this notion, Le. a use that is
not checked by the laying down of criteria for its use beforehand, makes
it possible to relate almost any contemporaneous or noncontemporaneous
elements. The Archaeology of Knowledge also proposes that the notion
of tradition is linked to the search for origins and the category of the
"genius": By means of an assumed tradition the historian of ideas can
link ("at a distance and through time - as if through the mediation of a
medium of propagation") a series of later individuals, notions or theories
back to the founder as the ingenious originator of the chain to follow
(AK 21).
Typically, the diachronic coherence of the sequence is established by
reference to the slow and gradual evolution of one or several - typically
ancient - ideas or themes, or by employing the concepts of influence,
anticipation and precession. Like other critics of the history of ideas,
Foucault is unsatisfied with all of these categories. First, the assumption
of "unit-ideas", or "themata" leads away from the study of a specific
discourse at a given time: this discourse is not "treated as and when it
occurs" but referred back to a "semi-silence that precedes it" (AK 25).
The employment of these notions and the interpretative strategies con-
nected to them tacitly presupposes that all major ideas have always
somehow been present below discourse until, finally, they "gradually
come together and suddenly condense into the fine point of the work"
(AK 138). Or, to put it in less metaphorical terms, for every important
idea, it is assumed that it was "always in some sense immanent in
history, even if various thinkers failed to 'hit upon' it, even if it 'drop-
ped from sight' at various times, even if an entire era failed (note the
implication that they tried) to 'rise to a consciousness' of it" (Skinner
1969: 10).
Second, the belief that every major idea has always been already
immanently present - gradually evolving and maturing - also invites the
notorious search for predecessors and anticipations. Every time some
historian of ideas presents a claim of the form "X was the first to
present idea Y", a chase starts to find anticipations, resulting in papers
proposing an earlier instance of the same idea (Condren 1985: 101). The
interpretative technique behind this procedure has been fittingly labelled
"the emergence technique" and characterized in terms of two impera-
tives:
Every idea has a predecessor which resembles it more closely than

any other predecessor; find it! (Agassi 1963: 32).
If a thinker's idea can be only partly traced to his predecessor, try
to track the rest of his idea to another predecessor (ibid., 34).
Thus historians have chased Bohr's planet system model back to the
atom of Newton's alchemist writings, the Einsteinian concept of space to
Aristotle, the Newtonian ether to the Stoics' pneuma, and the field
theories of modern physics to Spinoza (Kragh 1987: 84-85, 99). As
Foucault and others point out, however, to speak straightforwardly of the
resemblance or identity between ideas or formulations, is to overlook
that such resemblance and identity claims presuppose criteria which
define "what is identity, partial or total, in the order of discourse" (AK
143).
Like Quentin Skinner, Foucault is also annoyed by the undifferen-
tiated use of the category of influence. For both critics, this notion
provides "a support - of too magical a kind to be very amenable to
analysis - for the facts of transmission and communication" (AK 21).
Both writers suggest that this notion should either be suspended alto-
gether, or be applied only "in very precisely defined series" (AK 26,
143). Skinner's conditions are that the influenced writer could have got
his doctrine only from the alleged source of influence, and that the
probability of a coincidental similarity is very low (1969: 26).
Obviously, the search for predecessors, anticipations and influences
is equivalent to the determination of degrees of originality: the "stock of
originality" is raised or lowered by "degrees of nobility that are mea-
sured here by the absence of ancestors" (AK 143), and the significance
50 PART I
of an author is evaluated in terms of the difference between the number

of writers she has been influenced by and the number of writers influ-
enced by her (Condren 1985: 131). For Foucault, these are "amusements
for historians who refuse to grow up" (AK 144).
However, whereas Skinner sees the main genuine objective of the
historian as being the identification of the "intention in writing" of
authors (1969: 49), Foucault regards such a kind of goal as being part
and parcel of the same bundle of ideas which characterizes the history of
ideas. For him it is one of the central ingredients of this tradition to try
"to rediscover beyond the statements themselves the intention of the
speaking subject, his conscious activity, what he meant" (AK 27). In
part this suspicion might be due to the observation that the endless
debate over what this or that thinker "really" meant is an essential
element of the evaluative game of lowering and raising the stock of
originality, in part the suspicion is also motivated by Foucault's deeming
it more important to study what in discourse becomes accessible to all or
most writers of a given field and time, rather than to inquire what the
specific intentions of a single writer in dealing with some topic are. 9
The mythology at work: the examples of Holton and Cohen
Before going further and reviewing Foucault's criticism directed against

concepts which figure as much in Annales and epistemological histories
as in more traditional approaches, it might be useful to see that the
arguments of Foucault and other critics of the history of ideas do still
apply to the work done by leading historians of ideas and science of the
eighties. For instance in the 1988 revised edition of his Thematic Origins
of Scientific Thought. Kepler to Einstein, Gerald Holton restates his
viewpoint according to which the study of "themata" provides a crucial
key to the history of science. Like Lovejoy's unit-ideas, these themata
can be catch-words, "quasi-aesthetic judgments", feelings, and prin-
ciples. Furthermore, Holton's themata are long-living, appear jointly
with one or two "antithemata", and "a total of fewer than fifty couples
or triads seem historically to have sufficed for negotiating the great
variety of discoveries" (1988: 17). Applying his theory, Holton suggests,
among other things, that from the times of Milesian cosmology right
down to the contemporary debates over evolutionary and steady-state

cosmology, the same themata of evolution vs. constancy have been at
work (ibid., 45) and that the theme of discontinuity has been influential
from antiquity to modern physics (ibid., 17). In the light of such as-
sumptions, one will no longer be surprised to find Holton express views
like the following:
... at bottom, the work of major scientific 'revolutionaries', like

Einstein, ... can usually be seen to be projections back to an ideal-
ized, purified state of the past - and, in particular, a return to a
state of imagined classical purity, ... (ibid., 93).
Not surprisingly, Holton makes no attempt to define criteria for the

identity of themata in different authors, and he in no way answers the
criticisms that have been levelled against the Lovejoyan program.
Since Holton's views have been criticized often enough, and since
they might therefore be easily regarded as an exception, it seems more
interesting to turn briefly to another leading figure in the historiography
of science. The study I have in mind is I. Bernard Cohen's The New-
tonian Revolution (1980). As the subtitle of the book, With illustrations
ofthe transformation ofscientific ideas, already suggests, Cohen argues
that all creative scientific work in general, but also all revolutionary
advances in particular, consist of"a series oftransformations" of ideas.
For instance, "Darwin transformed Lyell's concept of an interspecific
competition or 'struggle for existence' to an intraspecific one, ... " (ibid.,
168).
Cohen proposes that an idea can either be transformed into a diffe-
rent idea, or then, like the notion of atom, "undergo successive transfor-
mations and continue to live on for a long time in science" (ibid., 197).
Furthermore, a series of transformations, starting from an idea, say p,
can eventually lead to another idea, say q, which logically implies ""p.
Cohen speaks of the negative and positive "potentialities" of ideas,
depending on whether a given idea is able "to serve a purpose quite
opposite to any beliefs of an author" (negative potentiality) or not. For
instance, the central ideas of Plato's TImaeus dialogue had negative
potentiality since they led to the formation of a mathematical physical
science based on experimentation and observation, an idea quite foreign
52 PART I
to Plato himself. Along the same line, Cohen claims that different ideas
have different potentialities of useful transformation, and that this
usefulness is to be measured by the ideas' ability "to serve in the further
advancement of science" (ibid., 162). Moreover, the transformation of
an idea can be conscious or unconscious, fortunate or unfortunate,
correct or incorrect, and genius-dependent or genius-independent (ibid.,
168, 194, 195, 203, 221). The transformation is fortunate or unfortunate
depending on whether or not the transformation leads to "good or useful
or fruitful" ideas; the transformation is correct or incorrect depending on
whether or not the transforming individual understands the original idea
according to the intentions of its author; and the transformation is
genius-dependent or independent depending on whether the transforma-
tion in question is "the last great step" of a revolutionary advance.
Cohen writes that this last step "requires a mind of truly heroic creative
proportions", a "supreme human genius" (ibid., 162, 221). Finally, it
deserves to be mentioned that Cohen regards Ernst Mach as his most
important predecessor in stressing the importance of the notion of
transformation. Mach held that ideas are gradually transformed "as in all
likelihood one animal species is gradually transformed into a new
species". Mach combined this gradualism with the conception of survival
of the fittest: "Many ideas arise simultaneously. They fight the battle for
existence not otherwise than do the Ichthysaurus, the Brahman, and the
horse" (ibid., 283-84).
Applying the criticism of the last section to Cohen, we can start
from the observation that the biological-evolutionist metaphor is central
in his approach. Cohen writes that historians and philosophers often treat
"ideas as if ideas were persons" (ibid., 327), but makes no critical
comment on this style of writing history. He approvingly quotes Mach,
who explicitly likens ideas to species. Cohen also writes that ideas "live
on", have "existence", "survive", and "have potentialities" (ibid., 197).
More importantly, in Cohen's talk of the potentialities of ideas it is
not difficult to identify the tendency to read history backwards, Le. to
trace back a current "idea". Tacitly this procedure assumes that the
current doctrine has always been in some sense immanent in history.
That this is not an unfair accusation can be seen from the facts that
Cohen wants to assess the fruitfulness of ideas and transformations in
terms of the way they have advanced science, and that he allows for
positive as well as negative potentiality: after all, the introduction of the

negative potentiality of ideas makes it possible to say that earlier ideas
are the seeds even of those later ideas which contradict them. Cohen's
talk of the potentiality of ideas strongly suggests a vitalist's view of the
entelechy, Le. a power in the organism - here: in the idea - that directs
the latter's development and completion. The Aristotelian notion of
actuality, as resulting from a coincidence of potentia activa and potentia
passiva also comes to mind because of Cohen's remark that a revolu-
tionary advance presupposes the meeting of an idea "with potentialities
for useful transformation ... with a mind capable of making the transfor-
mation" (ibid., 168).
As concerns the "virus of the predecessor" - as Canguilhem liked
to put it - Cohen seems to think that he can endorse this chase without
falling prey to its implausible implications. Since the step from the
predecessor to the successor involves a transformation of ideas, the
historical dimension of science appears to be preserved (ibid., 162-66).
However, the credibility of this assurance is seriously undermined by
Cohen's failure to provide us with a criterion for deciding whether a
given transformation results in a new variant of an old idea, or whether
the transformation results in a totally new idea. In one place, Cohen
seems to suggest that the decisive criterion is the intention of the original
creator of that idea which enters into the transformation (ibid., 159). But
this hint - which neatly supports the Foucauldian skepticism concerning
intentions - clearly flies in the face of Cohen's example of the atom as
one and the same idea despite successive transformations. It certainly
seems absurd to say that Bohr remained truthful to the intentions of
Democritos, and that thus Democritos can be regarded as a predecessor
of Bohr's theory.
Finally, as concerns Cohen's interest in genius, it must be men-
tioned that he conceives of scientific revolutions as starting from the
accumulation ofvarious smaller-scale transformations, the resulting ideas
of which are finally synthesized into a radical new set of ideas by a
"supreme human genius" (ibid., 221). As needs to be pointed out, this
scheme is strongly reminiscent of the Lovejoy school's eagerness to
distinguish between "radical originality" as the creation of new ideas,
and "synthetic originality" as the creation of a new synthesis (see
Condren 1985: 106). Since the number of ideas are regarded as decisive-
54 PART I
ly limited and long-living, the only way to increase their stock is by the
implausible introduction of geniuses. Be this as it may, Cohen's preoccu-
pation with genius and originality also shows, in any case, how central
the evaluative game around the original vs. regular axis is to his frame-
work.
The mythology of coherence and contradiction
The strategy of many historians of ideas and science to create coherence

and downgrade differences comes to the fore not only in the evolutionary
metaphor and the concepts linked to it. Foucault and Skinner also see it
at work in certain interpretative strategies employed vis-A-vis texts, and
Foucault identifies it moreover in concepts like Weltanschauung, menta-
lite, Zeitgeist, and in the science vs. non-science opposition proposed by
Bachelard and Althusser.
Foucault's rejection of books and oeuvres as natural units is well in
line with his general criticism of traditional document-based histories.
Since the serial historian can accept units only in so far as she has first
established the criteria for series and units, she cannot leave unques-
tioned the criteria by means of which texts or statements are grouped
around the figure of the author. In the case of books, Foucault claims,
it is not sufficient to base their coherence on "material unity", since this
unity cannot account for the differences of, say, "an anthology of
poems, a collection of posthumous fragments, Desargues' Traite des
Coniques, or a volume of Michelet's Histoire de France" (AK 23).
Moreover, a book is always caught up on a "network of references" to
other texts, and in this network of statements the division in terms of
book-units is only one of many.
The category of the oeuvre also needs to be critically examined.
Criteria are needed for deciding why the historian includes or excludes,
say, texts written under a pseudonym, various drafts of latter works,
unpublished writings in general, letters, testaments, and what not. For
instance, "the name 'Mallarm~' does not refer in the same way to his
thtmes (translation exercises from French to English), his translations of
Edgar Poe, his poems, and his replies to questionnaires" (AK 24). Put
differently, the oeuvre of any author can be defined in several different
ways, none of which is self-evident in itself. What is more, the focus on

oeuvres in the history of philosophy, science and literature, is often
informed by the tacit or open assumption that the whole production of a
writer must be the expression of cp single - or small set of - idea(s)
(ibid.). Used in this way, an oeuvre, like already a book, can be the
chief tool for reducing the number of differences and contradictions, that
is, the main device for establishing coherence (AK 150). Thus the
historian seeks to show that the different points of view of an author can
be made to cohere by explaining them in terms of an underlying -
explicit or implicit - theme or system; that the author's early, 'imma-
ture' works are simply so many attempts to formulate the views found in
later ones; or that, in those texts which do not harmonize with his
central doctrine, he was perhaps even 'hiding' his real views (Skinner
1969: 16-22; AK 150).
However, the quest for coherence is also operative on a more
general level. It surfaces in the project of much of the traditional history
of ideas to find below the different and contradictory phenomena of a
given civilization, a common and unifying principle (AK 9, 22, 150).
This is the project which we have already seen Foucault renounce as
total history. The classic of this kind of writing history is of course
Burckhardt's study of the Italian Renaissance, in which Burckhardt
interpreted the Italian Renaissance as the age of the awakening individ-
ual. Foucault does not mention Burckhardt, but refers to the works of
Cassirer, Goldmann, Lukc1cs and Dilthey (Foucault 1969: 24; Foucault
1966). At various places, he also suggests that the notion of mentality
can playa role similar to that of Weltanschauung, vision du monde, or
Geist (AK 137, 209).10
In a review of the French translation of Ernst Cassirer's Die Philo-
sophie der Aujklllrung (Cassirer 1932), a book that promised to interpret
the Enlightenment "in the unity of its intellectual origin and its deter-
mining principle ... out of One Central Point" (ibid., vii-viii), Foucault
argues that Cassirer's work exemplifies well the general tendency of
"traditional histories", namely the tendency to reduce the multiplicity of
forms of a given period first to the texts in which they were reflected at
the time ("a theory of the world rather than positive science, ... an
esthetics rather than the work of art, ... a philosophy rather than an
institution"), and second to a philosophical principle common to these
56 PART I
reflections (Foucault 1966: 4).

The same leaning, combined with a teleological perspective of
progress and the strong preference given to the original over the regular,
is of course pervasive in Goldmann. For instance, in his The Human
Sciences and Philosophy (1969), Goldmann justifies the study of history
as necessary for understanding the path from the past through the present
towards "an authentic and universal community in the future" (ibid.,
29). The history of ideas is, for Goldmann, the study of the "conscious-
ness of a class" (ibid., 59), and he defines "the maximum of potential
consciousness of a social class" as "a coherent world-view which may be
expressed on the plane of religion, philosophy, literature or art" (ibid.,
103). Here potential consciousness is distinguished from "real conscious-
ness" as "the totality of states of individual consciousness" and their
interrelations (ibid., 127). Potential consciousness is the collective
consciousness that would prevail were it not for "multiple obstacles and
deviations, ... the infinite variety and multiplicity of actions of other
social groupings, and also of cosmic factors" (ibid., 118). For Gold-
mann, it is the outstanding work of art, science and philosophy which is
able to reach beyond the real to the possible consciousness. From this he
infers that world-views enlarge due to the efforts of outstanding indivi-
duals, that we can best study the world-view of a period by attending to
its outstanding works, and that these works provide us with the under-
lying coherence of real world-views (ibid., 129-30).
Foucault claims that such schemes constitute a "histoire pour phi-
losophes" rather than historical research proper. With respect to these
schemes he also accepts the accusation made against him by Sartre, to
wit, that he "is killing History" (Foucault 1968: 22).
The tendency to reduce contradictions and increase coherence can
also lead to misconstructions in the (archaeologically and epistemologi-
cally orientated) study of science. In The Archaeology of Knowledge
Foucault turns critically against Folie et deraison and Naissance de la
clinique by observing that in these books he tried excessively to reduce
the variety of objects in psychiatric discourse and the multiplicity of
forms of statements in 19th century medicine to the singularity of one
object, namely madness, in the first case, and one modus or style,
namely description, in the latter case (AK 32-33). He now also feels
more need to emphasize that the notorious epistem~ of Les Mots et les
choses, is not to be misunderstood as a unifying principle (AK 191).

As concerns epistemological science studies from Bachelard to
Althusser, an exaggerated downplaying of differences can be seen at
work. After all, the total break between common sense and science in
Bachelard's case, or then the contradictory opposition between science
and ideology in Althusser, are but versions of the traditional historian's
project of either resolving various contradictions altogether, or reducing
them until "the fundamental contradiction emerges ... as the founding,
secret law that accounts for all minor contradictions" (AK 151).
Foucault argues, primarily against Bachelard, that in studying the
emergence of a scientific discipline more than just one threshold must be
distinguished. When in a given society a discourse with specific objects,
styles and concepts has emerged, this discourse can be regarded as
having passed the "threshold of positive discourse". This discourse need
not yet be confined to anyone institution (or clearly delimited group of
institutions), for indeed at this level the discourse is typically spread out
over many different domains and institutions. The discourse passes the
"threshold of epistemology" once norms and models exist which allow
the initially vague knowledge of the discourse to be checked. When the
discourse has established rules as to how scientific knowledge is to be
obtained, it has crossed the "threshold of science". And finally, once its
theories are axiomatized, the scientific discipline has passed the "thre-
shold of formalization" (AK 186-87).
Foucault suggests that different fields traverse these thresholds at
different speeds, and that sometimes two or more thresholds can be
passed at once. Disregarding these thresholds leads one to adopt "a
model to be applied at all times and for all forms of knowledge: a
division between what is definitively or what is not yet scientific" (AK
188). As The Archaeology of Knowledge proposes, this tendency to
homogenize the development of sciences to one threshold is due to a
mistaken assimilation of all sciences to mathematics; only in its case the
four thresholds were indeed crossed at once. Yet precisely this fact
makes mathematics a "bad example, an example at least from which one
cannot generalize" (AK 189).
More specifically against Althusser, Foucault points out that ideolo-
gy does not wear away once a problematic is established. First, as the
examples of clinical discourse and political economy show, disciplines
58 PART I
may well be deeply involved in the production or strengthening of

ideologies and yet produce knowledge which is not throughout "under-
mined by error, contradiction, and a lack of objectivity". And second,
by establishing more rigorous methods of proof and research, a disci-
pline does not necessarily break with ideology: "The role of ideology
does not diminish as rigor increases and error is dissipated" (AK 186).
5. THE ARCHAEOLOGICAL MODEL I:

IDENTIFYING DISCURSIVE FORMATIONS
INTRODUCTION
Having reviewed where The Archaeology ofKnowledge identifies friends

and foes, we can turn to Foucault's own proposals on how the history of
science might be written. Foucault's own, archaeological, model is most
naturally reconstructed in two steps. I shall begin by introducing the
main conceptual tools that Foucault employs for analyzing what he terms
"discursive formations" and "statements". These notions will be related
to the methodology of serial history, and sharpened by drawing on some
basic notions from analytical philosophy. It is only subsequently that we
can take up the Gretchenjrage for any philosophy or history of science,
to wit, the question how continuities and discontinuities are conceptual-
ized and accounted for. I hope to show that the archaeological model
contains suggestions for describing change in science that are still of
systematic interest today.
The negative, critical part of the archaeological enterprise is, as seen
above, something like a 'deconstruction' of some of the allegedly 'natu-
ral' categories and units which historians of science and culture have
traditionally employed. The positive, or constructive, side of The Ar-
chaeology of Knowledge consists of the introduction of a new set of
concepts and questions. To put it in a nutshell, Foucault seeks to define,
in a by and large systematic fashion, new kinds of events, series of these
events, their parts as well as their conditions, and he wants to determine
different series of series, Le. different interrelations between series.
Now what remains, after we have gotten rid of the traditional units
like tradition, influence, Zeitgeist and vision du monde, are simply
aggregates of events. As Foucault sees it, these events, or fossils of the

archaeological record, are "statements" (enonces):
... the totality of all effective statements ... the material with which
one is dealing is, in its raw, neutral state, a population of events in
the space of discourse in general (AK 27).
To be exact, even statements are not something immediately given,

of course, since looked upon without prior categories like book, oeuvre
or tradition, the archaeological record, or "the discursive field" (AK
28), consists of just a mass of inscriptions. To speak of statements as
events thus calls for a theory of what statements are, and how they are
to be identified.
Relying for the moment on an intuitive understanding of 'state-
ments', we might note, in any case, that to treat statements as events is
not implausible; after all, acts of writing down or uttering statements are
events, their printing or recording are events, and their repetition,
translation, transformation, reactivation, or forgetting are also events.
Furthermore, Foucault's speaking of the statements themselves as events
too, can be justified by the parallel between statements and speech acts,
a parallel that Foucault, after some initial hesitation, was ready to accept
(Dreyfus & Rabinow 1983: 46). Since speech acts are - according to
Austin's famous formula - acts of "doing things with words", they
obviously are events in time.
While statements, their conditions, their parts and the objects to
which they refer, form the basic, initial units of archaeological study, its
series, or series of series, are "discursive formations" and their interrela-
tions. As a first approximation, we can say that a discursive formation
is simply a fairly large population of statements. In order to delimit this
series, however, we need to identify several other series that ultimately
mark out the conditions of existence of those statements. Put differently,
to explain how a given series of statement-events is possible, demands
that statement-events be analyzed in various directions, such as with
respect to the objects to which they refer, the people who formulate
them, the concepts they contain, and the different ways they can be
combined. This investigation then leads to a more refined concept of
discursive formations. From this more analytic perspective, discursive
60 PART I
formations are practices displaying specific regularities, or, more pre-

cisely, discursive formations are practices constituted by the interrelation
and correspondence between several series of factors and conditions.
In The Archaeology of Knowledge, Foucault initially leaves the
statement undefined, and introduces discursive formations by way of
examples from his earlier books. From the vantage point of my recon-
struction, however, it seems more natural to start from the notion of the
statement.
WHAT IS A FOUCAULDIAN STATEMENT?
A reconstruction of Foucault's notion of the statement is by no means an

easy task. The 6O-odd pages of The Archeology of Knowledge (AK
79-134) that elaborate on this concept are not only the most difficult part
of the whole book, but they also contain some of the least transparent
passages of Foucault's whole oeuvre. The reader's patience is challenged
by definitions that are proposed only to be subsequently withdrawn, by
a long list of what a statement is not supposed to be, and by conceptual
distinctions that remain unclear to the end. Here I shall not try to discuss
the twists and turns, nor try to criticize the obscurities of Foucault's text.
Rather, I shall present my reading of what I take to be the hard -
intelligible - core of Foucault's treatment of statements without going
into every detail of his elaboration.
The fairly simple step of my interpretation is to take seriously
Foucault's suggestion that dealing with statements means "dealing with
an enunciative function" that relates signs "to a field of objects ... a
number of possible subjective positions ... a domain of coordination and
coexistence [and] ... a space in which they are used and repeated" (AK
106). The natural way of rendering this passage intelligible is, obvious-
ly, to take the notion of a ~nction at its mathematical face value.
Whether a string of signs is a statement or not, Foucault tells us, de-
pends on its relation to other factors. What then is more inviting than to
express this dependence with a function in its most general, set-theoreti-
cal sense? After all, functions in this sense are precisely the kind of
things by means of which dependencies can be formally expressed. My
proposal then is that a Foucauldian statement is an ordered quintuple
and that any set of statements (E) can be represented by a function ifJt}
that maps various combinations of elements of sets R, P, A, L into
elements of set S:
It: R x P x A x L - S
Here R is the class of "referentials", P is the class of "subjective posi-
tions" , A is the class of "associated domains", L is the class of limits of
repeatability, and S is the class of strings or systems of signs. All of
these five coordinates, or ingredients, of a statement call, of course, for
some explanation.
(1) To begin with the notion of referentials, Foucault informs us that in

order to qualify as a statement, a string of signs must "refer to some-
thing", Le. must refer to objects and states of affairs. Each statement has
as its "correlate ... a group of domains [e.g. of objects and states of
affairs] a domain of material objects, ... a domain of fictitious objects ...
a domain of spatial and geographical locations" (AK 91). This condition
for statementhood is of course tantamount to the central idea of inten-
sional semantics. In the latter, a proposition specifies a set of possible
worlds in which it is true, viz. in intensional semantics a proposition
specifies a set of possible worlds that contain states of affairs and objects
to which the proposition refers. Similarly, a Foucauldian statement is
related to a set of domains of states of affairs and objects: statements are
correlated with domains in which they have reference.
There is, however, one additional complication in Foucault's notion
of referentials that is not immediately captured by this parallel. Foucault
draws a distinction between "correlate" and "referential" to the effect
that whereas the correlate consists of the domains of objects, the referen-
tial consists of "laws of possibility" or "rules of existence" of these
various domains (AK 91). Thus different domains are not simply pre-
given. Instead, they themselves depend upon various factors and opera-
tions, they themselves are the result of a construction: what kinds of
domains of objects are possible or accessible at a given time depends on
a whole variety of relations between institutions, instruments, and
62 PART I
systems of concepts. In order then to account for the differentiation

between correlate and referential we might say that whereas the correlate
of a statement consists of the domains in which it has reference, the
referential comprises the conditions of the possibility of these domains.
(2) To turn to the class of subjective positions, P, Foucault writes that it

is constitutive of a statement to prescribe "a particular, vacant place that
in fact may be filled by different individuals ... To describe a formula-
tion qua statement does not consist in analyzing the relations between the
author and what he says ... but in determining what position can and
must be occupied by any individual if he is to be the subject of it" (AK
96). In other words, each statement prescribes certain conditions that a
speaker/writer must fulfill in order to qualify as the "author" of that
statement.
In order to flesh out the P coordinate somewhat, it might be fruitful
to take the P index as an expression of modality in the sense in which
Latour and Woolgar use this term. According to these authors, state-
ments in science can be classified by means of a fivefold distinction
between modalities, these modalities being: (1) wild speculation, (2)
plausible suggestion, (3) reporting the finding of others, (4) fact-stating,
and (5) taking-for-granted. (Latour & Woolgar 1986: 76-82). Since in
each of these cases the statement obviously prescribes different roles and
positions for the speaker or writer, the parallel between these modalities
and Foucault's subjective positions does not seem artificial.
(3) Each statement is always related to a domain of other statements that

figure as its "associated domain", "collateral space", "background of a
whole verbal network", "associated field", "enunciated field", or
"enunciative network" (AK 96-99). No statement can ever appear in
isolation, i.e., it is constitutive of a statement to be situated within, and
to be determined by, a class of other statements. Although Foucault does
not specify the upper limits of the associated field, he nevertheless makes
it clear that the associated domain of a statement is much more than just
an immediate context or co-text (AK 97). In fact, what counts as a
possible context of some statement is determined by the rules and the
structure of its associated domain. Thus, for instance, restrictions on
context work differently in poetry and mathematics: whereas the two
statements "1 love Ronny" and "2+2=4" can immediately follow each
other in a poetic text, they can hardly do so in a mathematical treatise.
(4) In order to approach the class of limits of repeatability, L, I have to

introduce some more of Foucault's terminological machinery. Especially
his distinction between "statement", "enunciation", and "formulation"
is important here. Both formulation and enunciation are bound to a
single spatiotemporal location~ but whereas a formulation is the act of
stating a statement, an enunciation is merely the act of uttering a string
of signs (AK 98, 101). Now, because an enunciation is merely an
utterance of signs - and not yet an utterance of a statement - it allows
for an abstraction over its spatiotemporal constants: " ... this uniqueness
[of the enunciation] allows for a number of constants - grammatical,
semantical, logical - by which one can, by neutralizing the moment of
enunciation and the coordinates that individualize it, recognize the
general form of a sentence, a meaning, a proposition" (AK 101).
In the case of the statement, however, the type-token distinction
works somewhat differently. Each formulation of a statement is a token
of a type, but whether different formulations are of the same type or not,
Le. whether they are formulations of the same statement, is not deter-
mined by grammatical structure or logical form. Instead it is to be
decided by studying the statement's "status as a thing or object", on the
one hand, and by attending to its "field of stabilization", on the other
hand (AK 103). To investigate the first is to study the limits and restric-
tions that a given culture, community or institution imposes upon the
"possibilities of reinscription and transcription" of statements (AK 103).
For instance, a psychoanalytically oriented literary critic might treat two
formulations, one of which a novelist utters in an interview, and the
other of which she places into the mouth of one of her characters, as
constituting the same statement. A legal expert, however, who has to
determine whether the formulations in question qualify as an insult
against someone, might, in the same case, speak of two different state-
ments of which only one can lead to legal action.
To investigate a statement's "field of stabilization", on the other
hand, is to attend to the series of statements to which a given statement
belongs. Here one seeks to determine whether two sentences which,
when looked upon in isolation, seem to express the same proposition, do
64 PART I
in fact refer to the same (kind ot) objects, have the same (kind ot)
subject position, and belong to the same discursive series. Attending to
such series, Foucault holds, forces us to say, for instance, that "the
affirmation that the earth is round or that species evolve does not consti-
tute the same statement before and after Copernicus, before and after
Darwin ... [and that] the sentence 'dreams fulfil desires' ... is not the
same statement in Plato and Freud" (AK 103).
(5) To tum, finally, to the sign coordinate of the statement, it suffices to

mention that Foucauldian statements are not restricted to formulations in
natural language; formalizations and graphical representations too can be
statements (AK 86).
THE FOUR SERIES
To delimit a population of statements as belonging to one and the same

discursive formation is to identify a structure of interrelations holding
between the conditions of existence of the statements' coordinates.
Varying his terminology slightly, Foucault suggests that to study a
discursive formation as to the conditions of its possibility is to identify
the historical conditions of the possibility of the formation of its "ob-
jects", "enunciative modalities", "concepts" and "strategies". He
emphasizes that discursive formations cannot be identified, and thus
delimited from one another, by identifying for each of them a single
central object, e.g., madness, by identifying for each a single common
mode of language or "enunciative modality", e.g., descriptive or pre-
scriptive language, by identifying for each a clearly confined small set of
concepts, or by identifying for each a single theory or theme. Instead, a
discursive formation is identified once a law-like regularity of depen-
dence or interrelation can be detected, a regularity that expresses the
dependencies and interrelations between conditions of the possibility for
a variety of objects, a variety of enunciative modalities, a variety of con-
cepts, and a variety of theories or themes (AK 32-35).
These four coordinates of the discursive formation do not corre-
spond exactly to the five coordinates of the statement, but as we shall
see this asymmetry has little bearing upon substantial issues. It is more
important to turn to Foucault's proposals as to how each of these coor-

dinates are to be studied.
Objects
As concerns the constitution of objects, The Archaeology of Knowledge

submits that we attend to three factors: "surfaces of emergence", "au-
thorities of delimitation", and "grids of specification" (AK 41-44).
Surfaces of emergence are the places and spaces, Le. institutions or
cultural fields, in which certain elements emerge as objects for dis-
course. In the case of the rise of psychopathological discourse in the
nineteenth century, Foucault claims, such surfaces were, among others,
the family, the immediate social groups, the working place, the religious
community and art (AK 41). It was in these social and cultural contexts
that madness was first systematically looked for and - alas - found.
Authorities of delimitation were those individuals, groups and disciplines
that came to be regarded as qualified for the identification of these new
objects, foremost doctors, judges, art critics and priests. Finally, grids
of specification were the systems of concepts "according to which the
different 'kinds of madness' are divided, contrasted, related, classified,
[and] derived from one another as objects of psychiatric discourse" (AK
42).
Or, to take as another example Steven Shapin's and Simon Schaf-
fer's study on the emergence of experimentalism in natural philosophy
of mid-seventeenth century England, we might say that the surfaces of
emergence of a new scientific object like Boyle's "spring of the air"
were the laboratory and the assembly room of the Royal Society with its
crucial instrument, the air-pump. Authorities of delimitation were Boyle
and his followers, the "physiologists", who severely restricted access to
this new space, who evaluated and judged the credibility of observations
reported, and who determined who was to publish in their Philosophical
Transactions. Their grid of specification consisted not only of the
traditional corpuscular and mechanical notions, but also of new concepts
relating to different aspects of the spring of air. Their grid was also
marked by a tight distinction between the language of theory and the
language of facts (Shapin & Schaffer 1985: 38-39, 45, 57-58, 71,
66 PART I
135-136, 180-181).
What Foucault wishes to claim is that the objects of a (scientific)
discourse are not waiting 'out there', but are rather the result of a group
of relations that exist within and between surfaces, authorities and grids.
In other words, "it is not enough for us to open our eyes, to pay atten-
tion" in order to encounter new scientific objects, rather these new
objects exist only "under the positive conditions of a complex group of
relations". Furthermore, these relations hold "between institutions,
economic and social processes, behavioral patterns, systems of norms,
techniques, types of classification, modes of characterization; and these
relations are not present in the object ..." (AK 45). Foucault calls these
object-constituting relations"discursive relations" and distinguishes them
from "primary" and "reflexive relations". Primary relations are the
relations between, say, the doctor and the family, relations that can be
described without attending to the doctors' medical discourse. However,
both reflexive and discursive relations are relations pertaining to the
doctors' purported knowledge. Reflexive relations are relations that
figure explicitly in the doctors' discourse, are those relations that he
describes and analyzes. (For instance, the relations between family life
and criminality.) Discursive relations differ from reflexive relations in
not being formulated in the doctors' discourse: discursive relations make
certain objects of the doctors' discourse possible, constitute objects (like
madness) that the doctors come to identify, amongst other places, in the
families (AK 45).
The isolating of discursive relations is crucial to the archaeological
enterprise. To see why this is so we need to recall the central problem
of serial history, to wit, the delimitation of series. Prima facie, the
wealth of different objects of different discourses, and the continuous
emergence of new ones, might make it seem impossible to pick out a
group of objects as forming one distinct series. Foucault proposes,
however, that this pessimism turns out to be unfounded once we attend
to discursive relations. According to his suggestion, we identify a series
of objects once we have shown for a group of non-contemporaneous
objects that they emerge on the basis of the same set of discursive
relations; in other words, that they have the same (social, institutional
and conceptual) conditions of possibility:
... a discursive formation is defined (as far as its objects are con-
cerned, at least) ... if one can show how any particular object of
discourse fmds in it its place and law of emergence; if one can show
that it may give birth simultaneously or successively to mutually
exclusive objects, without having to modify itself (AK 44).
Enunciative modalities
Turning to the second series that archaeology seeks to define, the "sub-
jective positions", or "enunciative modalities", we encounter the same
principle of identification albeit now defined in terms of three different
factors. The first of these factors might be called 'speaker's position' and
includes criteria of competence, the relations of the competent speakers
to other social groups and the characteristics that define their status in
society (AK 50-51). In the case of mid-seventeenth century 'physiology'
an analysis of this element would have to include a description of the
following norms: in order to qualify as competent members, researchers
have to communicate a candidate matter of fact, have to accept experi-
ments as the touchstone of natural philosophy, have to reject metaphysics
and have to abide by specific rules of criticizing and writing. One would
also have to mention the role of experimentalists in Restoration England,
as well as their self-conscious posing as social arbitrators and as "priests
of nature". These latter roles are important for understanding the support
Boyleans received from the Church and the State (Shapin & Schaffer
1985: 70, 174, 184, 216, 319).
To comprehend the conditions of a given group of enunciative
modalities also calls for a study of "institutional sites" and "positions of
the subject". The former refer to places from which the competent
subject makes his discourse (e.g. the assembly hall, the laboratory, the
book written in English or Latin), the latter to his or her position in the
information network and to the different roles the subject takes up in
relation to his or her objects (AK 52-53). In the case of the "physiolo-
gists", such typical postures were that of the eye-witness, of the faithful
reporter, of the modest narrator of successes and failures, and of the
receiver and writer of letters circulating within a small group of natural
philosophers, theologians and philosophers (Shapin & Schaffer 1985: 56,
68 PART I
59, 60, 68, 319).
Organizations ofthe field ofstatements
Under the title of a study of "the formation of concepts", Foucault

provides a longish list of categories for the investigation into various
relations between statements. This is because The Archaeology ofKnowl-
edge deems a study of statements an indispensable groundwork for the
study of concepts. Since a det~iled scrutiny of this list lies somewhat to
the side of our main interest, suffice it here to summarize these distinc-
tions by way of a table (AK 57-68). (Figure 1)
For instance, the discourse of the English physiologists showed the
following "organization of the field of statements": its statements were
ordered according to the logic of description, Le. the narrative and the
letter. Demonstrative reasoning and the form of the dialogue were
rejected as orderings of the dogmatic philosopher. The central type of
dependence was the hypothesis I verification dyad, and the typical
rhetorical scheme was the essay, as "the piecemeal reporting of the
experimental trials". The architecture of Boyle's text was also marked by
the "conspicuous interval" that Boyle wished to leave between his
narratives of experimental findings and his "discourses" on their inter-
pretation (Shapin & Schaffer 1985: 39,40,44,66,67, 143). Concerning
forms of coexistence of statements, the field of presence of Boyleans
contained the philosophical discourse on vacuism and plenism only as
something to be rejected and transformed into the language of experi-
ments. An intriguing example of a Boyleanfield ofconcomitance was the
legal discourse on eye witnessing, a discourse which was applied per
analogiam to the witnessing of experiments. The field ofmemory of the
new experimental science contained philosophical-metaphysical demon-
strations, on the one hand, and· some references to the alchemists, on the
other hand. The statements of the former were those in relation to which
Boyle established a relation of discontinuity, whereas the latter figured
occasionally as predecessors of his own enterprise (ibid., 45, 56, 71).
Finally, with respect to characteristic procedures of intervention, one
should at least mention the use of tables, and the central role of geome-
try rather than arithmetic as an important measure for transcribing
FORMS OF SUCCESSION
ordering ofstatement series
e.g. demonstrative reasoning
descriptions
-
types ofdependence
e.g. hypothesis I verification
assertion I critique
law I application
rhetorical schemata
the architecture of the text
FORMS OF COEXISTENCE
field ofpresence
statements taken up in a given discourse positively or
negatively; e.g. as correct description, good reasoning,
necessary presupposition, as authoritative statement ...
field of concomitance
statements of other discourses taken UK as analogical
confirmation, general principle, model, igher authority
...
field ofmemory
statements no lonter accepted as valid, but in relation to
which relations 0 filiation, continuity and discontinuity
can be defined
PROCEDURES OF REWRITING
techniques of rewritins
e.g. linear descnptions in tables
method of transcribing statements from natural into fonnal
languages
modes of translating quantitative into qualitative statements
and vice versa
means ofincreasing the approximation ofstatements. to extend
and restrict their validity
ways of transferring a statement to a new field of application
methods of systematizing statements
Figure I
70 PART I
statements from natural language into another sign-system (Shapin 1988:

23-58).
Strategies
The fourth kind of series that Foucauldian archaeology seeks to define is

the series of "strategies" . Under this heading The Archaeology o/Knowl-
edge groups theories and themes and proposes that in their case too we
can define certain regularities of conditions of possibility. Defining these
regularities involves the identification of "possible points of diffraction
of discourse", of "the economy of the discursive constellation" and of
the discourse's "function in a field of non-discursive practices". Points
of diffractions are points where a given discourse branches off into sub-
discourses: these junctures are "points of equivalence" in the sense that
the two or more alternatives are determined by the same regularities
concerning the conditions of objects, concepts and modalities; they are
"points of incompatibility" in the sense that at the juncture a choice has
to be made; and, they are "link points of systematization" in that both
alternatives can give rise to a new series of statements (AK 65-66). For
instance, Hobbes's interpretation of some of the results obtained in
Boyle's air-pump experiments meet these conditions. Phenomena which
Boyle regarded as clear proofs of the existence of the spring of air, like
the dying of animals in the allegedly 'evacuated' receiver, Hobbes
interpreted as effects of the uneven mixtures of extremely small, earthy
particles moving rapidly in a circular motion (Shapin & Schaffer 1985:
122).
The study of the conditions of theories is also the study of the
conditions for, and the regularities of, the mechanisms which close
disagreement and choice. These mechanisms of closure are the "econ-
omy of the discursive constellation" and the "function in a field of non-
discursive practices". The first factor is meant to capture the relation of
a given (scientific) discourse to other, contemporaneous discourses which
function as model, analogue, opposite or complement. For example, in
the case of the seventeenth century physiologists, theory-choice was
influenced by the partial rejection of the alchemists, and the total rejec-
tion of metaphysics. Theology, however, was seen as an ally (ibid.,
139). The second factor comprises elements like the role of the respec-
tive discipline in non-scientific practices, e.g. in education or manufac-
turing, the intervention of non-scientific institutions into science, and
"the possible positions of desire in relation to discourse" (AK 68). For
example, Boyle's victory over Hobbes was due in part to the fact that
priests as well as lawyers were opposed to Hobbes. In order to translate
the theologians' concerns into support for his theories, Boyle justified his
nescience as to real causes by the theological argument that God can
produce the same effect by a number of different causes. As far as desire
was concerned, be it mentioned that the air-pump was repeatedly used
for the entertainment of royalty and politicians, and that - at least in
iconography - the air-pump promised an approximation of God's know-
ledge and the reaching for the invisible (Shapin & Schaffer 1985: 30,
32, 36, 37, 153, 313).
INTERRELATIONS, DERIVATION TREES AND CONTRADICTIONS
Defining series of objects in terms of their common conditions of

possibility, and doing likewise for enunciative modalities, concepts, and
strategies, does not yet suffice for identifying a discursive formation. In
order to isolate a discursive formation, The Archaeology of Knowledge
claims, we have to go further and study correlations and interrelations
between the four kinds of series. Foucault also writes that a discursive
formation is "a scheme of the correspondence between several temporal
series" (AK 74). I take this to mean that of the four temporal series
defined each has its own rhythm of change and its own specific limits,
its own duration. In other words, for us to speak of a discursive forma-
tion, it is not necessary that the limits of all of the four series coincide
in time. What is called for, as a justification for calling a given group of
statements a discursive formation, is rather that the four series of ob-
jects, modalities, modes of organization of statements and strategies
coincide or overlap for some time, and that we can establish interrela-
tions between these series during this time (Figure 2). (Writing "0" for
a series of objects, "EM" for a series of enunciative modalities, "C" for
a series of concepts and organizations of statements, "r for a series of
theories and strategies, "*" for interdependence between given series).
72 PART I
EM ,--I_ _ O_-----:C~I-----:-T~I-----:--I--..L...-----.
discursive
time ...... formation
Figure 2
To quote Foucault more fully:
A discursive formation ... presents the principle of articulation

between a series of discursive events and other series of events,
transformations, mutations, and processes. It is not an atemporal
form, but a scheme of correspondence between several temporal
series (AK 74).
Foucault does not provide us with a list of possibilities as to how

these strata can influence or determine each other, but he stresses, in any
case, that the direction of conditioning, restricting and delimiting does
not necessarily take a uniform direction from upper strata to lower ones,
or from lower to upper ones. These interrelations are rather of various
kinds, and can change during the existence-span of the discursive forma-
tion (AK 73-75).
The last mentioned point is worth underlining, since it shows that a
discursive formation is not a static set of fixed regularities. This point
should already be clear in so far as the four series defined are series of
simultaneous as well as successive, even in part mutually exclusive,
objects, modalities, concepts and theories. To this we must add that the
interplay of the four series can even involve the altering of some local
regularity, (e.g. of a regularity holding within the conditions of the
emergence of some objects), without, however, the discursive formation
being itself changed fundamentally. In fact, the existence of such kinds

of influences is as good evidence as any that one is in fact dealing with
a discursive formation (qua interrelation between various series). To cite
Foucault's own example:
The hospital field ... did not remain unaffected when clinical dis-
course was put into relation with the laboratory: the body of rules
that governed its working, the status accorded to the hospital doctor,
the function of his observation, the level of analysis that can be
carried out in it, were necessarily modified (AK 74).
Allowing for changes on the level of local regularities within one

and the same discursive formation obviously presupposes a distinction
between regularities of different generality. And indeed, in several places
Foucault distinguishes between more fundamental or "general rules" and
"more specific" or "derived" rules (AK 147, 168, 177). This distinction
can be criticized as far as this talk of rules (here as elsewhere in the
archaeological edifice) is somewhat misplaced; after all, if rules are to be
more than regularities they must somehow be internalized by someone.
Nevertheless, the distinction between different levels of regularities is
certainly called for and helpful, especially when we tum to discontinui-
ties and revolutions in science. Since this is a topic to be taken up
further below, suffice it here to mention two other uses to which Fou-
cault puts this distinction.
The first case concerns what Foucault calls the construction of "the
tree of derivation" of a discursive formation. The Archaeology ofKnowl-
edge is not especially clear on how this tree is to be drawn, but it is
plain that at the roots of this tree one places such statements that display
the most general regularities of the discursive formation in question.
These "guiding statements" (AK 147) do not formulate the discursive
relations, since these are, as earlier heard, inaccessible for the discourse
they regulate. Instead, guiding statements might be looked upon as
counterparts of discursive relations within discourse itself. Guiding state-
ments do not formulate but mirror discursive relations; the way they
refer to objects, use concepts and form theories displays, on a general
level, the outer limits of what it is possible to speak about in a given
discursive formation. Or, put differently, guiding statements formulate
74 PART I
in the forms of general norms, presuppositions, axioms or exemplars the

limits of discourse in so far as these limits are accessible to discourse
itself. (perhaps we can say that Boyle's statements on how experimen-
talists are to go about doing science and how they are to achieve agree-
ment are in this sense guiding statements of seventeenth century experi-
mentalist science. The social and institutional conditions of the possibili-
ty of this experimental form of life are not directly, or only distortedly,
present in his discourse, but his programmatic statements make them
easy to identify.)
Derived statements, on the other hand, are further up on the
branched tree structure and display more local regularities. They are less
fundamental than guiding statements in the sense that both their accept-
ance and rejection is consistent with the guiding statements, while they
themselves are consistent only with the acceptance of the guiding state-
ments. l l
The relation between guiding and less fundamental statements need
not be one of temporal succession, however. Foucault also allows for the
possibility that the guiding statements of a given discursive formation are
formulated only late within the existence-span of a discursive formation:
"For certain discursive formations, the archaeological order [of deriva-
tion] is perhaps not very different from the systematic order, as in other
cases it may follow the thread of chronological successions" (AK 148).
Only once we understand the structure of a given discursive forma-
tion in this way does it also become possible to distinguish between
different types of contradictions or disagreements. "Derived contradic-
tions" concern branchings fairly high up in the tree-structure, "intrinsic
contradictions" are located in mid-air, above the roots and below the
finer branches, and "extrinsic contradictions" are contradictions between
statements that belong to two different discursive formations (AK 154).
To take a more contemporary example, we might say that we encounter
a derived contradiction where two linguists, both working within the
framework of Chomsky's Theory of Government and Binding, disagree
over the question whether or not to allow for rightward movements of
elements in the process of transformation; we are dealing with an
intrinsic contradiction in the case of two linguists that take different sides
on whether or not to accept transformations within a generative frame-
work (e.g. Gazdar's Generalized Phrase Structure Grammar versus
Chomsky's GB Theory); and we have an instance of a still more funda-

mental difference, perhaps an extrinsic one, where two linguists disagree
over whether the study of language should take its starting point from
Chomsky-style syntactical theories or the ethnomethodological descrip-
tion of conversations. 12
COUNTERPARTS AND ACCESSIBILITY RELATIONS

COMPARING DISCURSIVE FORMATIONS
Having reviewed the central conditions or "rules" by means of which

archaeology delimits discursive formations, we can turn to what Foucault
calls "the comparative facts" (AK 157-65), Le. relations of similarity
and difference between discursive formations. Foucault's suggestions
concerning these topics deserve attention, since in recent years several
historians and philosophers of science have called for theories and
historical studies on relations between disciplines.tJ
Discourses and discursive formations
Above I pointed out that a discursive formation can be characterized

either by listing the statements of which it is composed, or by describing
the system of interrelations between the four series, the system which
constitutes the condition of the possibility of these statements. In other
words, even though discursive formations are made up of statements, not
every statement can be part of just any discursive formation. Foucault
expresses this conception by writing that each discursive formation is
characterized by a specific "law of coexistence", "principle of rarifica-
tion", or "law of enunciative poverty" that restricts the discursive
formation to precisely the statements it in fact is constructed to have (AK
116-20).
When couched in this way, it is inviting to suggest that what appears
as "enunciative poverty" from one perspective, might just as well be
regarded as "enunciative plenitude" from a different angle. While we
perhaps cannot say that every statement which is possible with respect to
the regularities of a given discursive formation must be actual sooner or
76 PART I
later, we can say, in any case, that every statement that is possible,
according to the laws of the discursive formation in question, can only
become actual within this same discursive formation.
This interpretation also invites the assumption that there can be no
identity of statements across the borders of a discursive formation. First
of all, Foucault proposes that the statement is determined in its identity
by "conditions and limits ... that are imposed by all other statements
among which it figures" and that identity of wording is not enough to
safeguard identity on the statement level. Thus, for example, the state-
ment that the earth is round, is a different statement in the discursive
formations before and after Copernicus (AK 103). In other words, a
statement is repeatable only within one and the same discursive forma-
tion. 'Transdiscursive identity' is ruled out.
However, lest discursive formations begin to look like closed
Leibnizian possible worlds, it cannot be stressed too much that this
denial of transdiscursive identity holds only for statements, not for
objects, modalities, concepts or even theories. This point does not
always come out very clearly in The Archaeology of Knowledge, and
deserves some closer scrutiny.
As Foucault has it, statements are grouped into discursive forma-
tions on the basis of whether or not their coordinates are restricted, or
made possible, by one and the same system of interrelations between the
four series. In other words, it is not enough for two statements to refer
to the same objects in order for them to belong to the same discursive
formation. It is also necessary, first, that their other coordinates are
elements of the same series as well, and, second, that the four series
from which both statements draw their coordinates, are interrelated
according to one and the same" system of interrelations.
Now while there is nothing implausible about this suggestion as
such, it should nevertheless be made explicit that it must also leave room
for statements that do not belong to any discursive formation. If we
accept the tight Foucauldian conditions that a set of statements has to
satisfy in order to qualify as a discursive formation, then we obviously
must leave room for many discourses which are unstructured by archae-
ological standards (which of course does not mean that in such dis-
courses "anything goes"). This observation suggests that discourse
should be defined as a broader concept than discursive formations, and
that discursive formations are best regarded as discourses of a special

kind (as defined by Foucault and reconstructed above). One is then led
to say that there can be statements outside of discursive formations, but
not outside of discourses. 14
Since for Foucault a discursive formation only exists once all of its
characteristic series are present in a characteristic interplay, it also seems
natural to assume that two different successive formations will often or
typically be separated by a discursive phase in which such characteristic
interplay, such a system of interrelations, does not exist. Of course to
assume such a less structured phase is not the same as claiming that both
discursive formations form one historical line of tradition or discipline
and that the intermediate phase, or "feature interface", is one of a
Kuhnian crisis.
Finally, as the example of Foucault's own historical studies on the
emergence of different discursive formations suggests, it is often useful
to describe the statements of an intermediate discourse as "fuzzy"
elements either of prior or of subsequent discursive formations. Accord-
ing to this suggestion, a fuzzy element of a discursive formation is an
element of an intermediate discourse which displays the regularities of
the respective discursive formations only in part. Elements of the inter-
mediate discourse are then not totally set apart from full members of a
prior or later discursive formation. (Using, for the purpose of illustra-
tion, numerical values, we might say that full members of a discursive
formation are assigned the value 1, full members of a successor forma-
tion assigned the value 0, and statements of the intermediate discourse a
value ex, such that 1 < ex ~ O. (Figure 3»
78 PART I
o
EM
C
T
J J < a So 0 o
value assignments to statements for
time ~~~ the predicate 'being an element of df1 '
Figure 3
Counterparts
Although Foucault rejects the notion of transdiscursive identity of

statements, he does, nevertheless, allow for transdiscursive similarity.
My employment of the notion "statement-counterpart" should thus be
clear enough. After all, it is due to the writings of David Lewis that the
idea of transworld similarity has gained currency under the label of "the
counterpart theory" (e.g. Lewis 1983a). Simply put, a counterpart of
some individual, say i, of a given world, v, is the counterpart in v of
another individual, j, in another world, W, if and only if, i resembles j
more than any other individual of v.
The naturalness of transferring Lewis's notion to historical inquiry
- even quite apart from Foucault - can also be defended by briefly
indicating that Lewis's characterization of this relation as not transitive,
not symmetric, and not a one-to-one relation is also intuitively convinc-
ing for the relation between statements. Let 1 to 9 be statements of three
discursive formations, A, B, and C, as indicated in Figure 4.
A B c
1
i I
6
1
i
1 8
3 - ---------- ---- - 6 9
I
Figure 4
The counterpart relation is not transitive for even though 5 might be

more like 2 than any other statement of B, and even though 8 is more
like 5 than any other element of C, it still does not follow that 8 is more
like 2 than any other element of C. In fact, it might be 9 that is more
like 2 than any other element of C. Nor is the counterpart relation
symmetric: 4 might resemble both 1 and 2 very strongly, more than any
other statement of B. Thus 4 is a counterpart of 1. Yet it might be that
2 resembles 4 more than 1. Thus the counterpart of 4 in A is 2 and not
1. Furthermore, the counterpart relation is not a one-to-one relation: 4
and 5 might resemble 7 more than anything else in B and they might
resemble it to equal degrees. Thus they are both counterparts of 7.
Obviously, it is then also natural to allow that one statement is the
counterpart of two statements. Finally, 3 might resemble 6 more than
anything else in A, but still not be very like 6 at all. Then 3 is no
counterpart of 6.
In Foucault, the main instances of what I have been calling discur-
sive counterparts are "archeological isotopia":
80 PART I
[They are] entirely different concepts Qike those of value and spe-
cific character [in the discursive formations of 17th century econom-
ics and of natural history respectively] ...) [that] occupy a similar
position in the ramification of their system of positivity [L e., discur-
sive formation] - although their domain of application, their degree
of formalization, and above all, their historical genesis makes them
quite alien to one another (AK 161).
It is true that Foucault is talking here about concept counterpart, not

statement counterparts. However, the transition from the first to the
second can be justified by Foucault's pronouncements that concepts
function only within statements. Furthermore, in another passage,
Foucault argues that even though a statement cannot appear in two
different discursive formations, it can still be related to statements within
another discursive formation that are either "linguistically analogous",
Le., that have the same syntactic-lexical surface structure, or that are
"logically identical", Le., that are equivalent (AK 145). Thus altogether,
we are provided with three criteria of statement similarity across discur-
sive borders.
Accessibility relations
It also seems useful to maintain some kind of parallel between Foucault

and Lewis as far as the relations between discursive formations are
concerned. In Lewis's possible worlds conception every possible world
is surrounded by a sphere of overall similar worlds. This notion of
similarity of worlds is needed to account for the semantics of counterfac-
tual sentences. For instance, to say that 'If Saddam Hussein were an
American, Saddam Hussein and George Bush would be compatriots' is
to say that there is a set (sphere of accessibility) of possible worlds, such
that they are all similar to our world and such that Saddam Hussein - or
his counterparts - figure as American citizens (Lewis 1973: 8-9).
Foucault, in turn, is interested in the question as to what conditions
are required for two or more discursive formations to form a "region of
interpositivity" or an "interdiscursive group" (AK 159). This question
subsequently leads to the problem as to what conditions are needed for
two or more discursive formations to be accessible to one another. In

other words, Foucault holds that communication between discursive
formations is possible only when the discursive formations in question
are similar in some respects. Contrary to Lewis, in whose Counterjac-
tuals the notion of similarity between possible worlds remains a primi-
tive, unspecified concept, Foucault provides us with three criteria for
similarity between discursive formations. We have already referred to
the first one, to wit, the existence of "archeological isotopia", Le.,
concepts or statements with a similar function. The second and third
criteria are "archeological isomorphisms" and "archeological models".
Foucault introduces these notions somewhat telegraphically, but it would
seem that what he is trying to capture by means of these notions are
certain parallels on the level of discursive relations. Even though two
discourses might deal with different objects, have different enunciative
modalities, different concepts and different theories, it is still possible
that both owe their respective existences to similar conditions and
regularities. Foucault speaks of "archaeological isomorphisms" when we
find such common conditions in two discursive formations, and of the
same "archaeological model" when even the rhythms of internal devel-
opment and change are similar. (Foucault holds that Natural History, the
Analysis of Wealth and General Grammar of the latter half of the
seventeenth and of the eighteenth centuries show archaeological iso-
morphisms but have dissimilar. archaeological models (AK 160-61).)
Now, since similarity between discursive formations is the condition
for the possibility of their "law of communication", which in turn is the
condition for the possibility of "archaeological correlations" as "relations
of subordination or complementarity" between discursive formations
(AK 161), it is inviting to ask what kind of accessibility relations Fou-
cault's similarity criteria specify .15
First, the accessibility relation between discursive formations must
certainly be reflexive since every discursive formation must be similar to
itself, and thus accessible to itself. Yet the accessibility relation is neither
symmetric nor transitive. 16 To see why it cannot be transitive, note that
a discursive formation, dh, can be similar to another discursive forma-
tion, dh, with respect to one set of regularities, R], and similar to
another discursive formation, dh, with respect to another set of regular-
ities, R 2 , without dh and dfj sharing any common, or similar, regular-
82 PART I
ities. Foucault himself mentions that a discursive formation can "enter

simultaneously into several fields of relations", viz., several "regions of
interpositivity" (AK 159). Clearly, for these "several fields" to remain
distinct, for them not to collapse into one larger region of interpositivity,
it must be possible for a discursive formation to be accessible from, and
have access to, at least two other discursive formations that are not
accessible to one another. Finally, the question whether the accessibility
relation is symmetric must also be answered negatively. This is impor-
tant since it suggests that similarity between discursive formations is only
a necessary but not a sufficient condition for accessibility. Certainly,
similarity is a symmetric relation: if A is similar to B, then B is also
similar to A. Yet when one discursive formation is subordinated to
another - a possible situation mentioned by Foucault himself (AK 161)
- it seems natural to allow for the possibility that the communication
channel is used only in one direction. In such a case the accessibility
relation is not symmetric, however similar the two discursive formations
happen to be.
Before leaving the problem of communication between different
discursive formations, one obvious critical question has to be addressed,
at least briefly. This question turns on the apparent contradiction be-
tween Foucault's allowing for interdiscursive communication, on the one
hand, and his rejection of transdiscursive statement identity, on the other
hand. How can one discursive formation influence another, if no state-
ments can cross their border? Two answers can be suggested to this
question. First, one might suggest that influence in this case does not
have to involve any exchange of statements. The influence of one
discursive formation, say dh, on another, say d,h, might simply involve
d,h adopting similar structures of institutions and theories. Second, recall
that Foucault allows for "linguistic analogy" between statements of
different discursive formations, Le., identity "from the point of view of
grammar (vocabulary, syntax, and the language (langue) in general)".
He also speaks of this linguistic analogy as "translatability" (AK 145).
Perhaps we should say that even though a statement (e j E dfJ cannot be
transferred to another discursive formation, it can at least be translated
into a statement of another discursive formation (ej E dfj), where the
second statement (e) may then be regarded as a counterpart of the first
(e;).
6. THE ARCHAEOLOGICAL MODEL II:

BEYOND CONTINUITY AND DISCONTINUITY
PHILOSOPHERS' VS. HISTORIANS' HISTORY OF SCIENCE
Needless to say, the touchstone of respectability for any conceptualiza-

tion of the history of science is its ability to capture phenomena of
change. That is to say, the relative success of any proposal for how to
write the history of science cannot but be measured in terms of its skill
in distinguishing between phases of rapid and slow change, its ingenuity
in dealing with continuity and discontinuity, and its ability to separate
apparent or superficial changes and tendencies from deep-going and far-
reaching ones. Therefore, only by reconstructing and discussing Fou-
cault's suggestions with respect to these topics can we put ourselves in
a position where we can evaluate the strengths and weaknesses of the
whole archaeological edifice.
Reviewing the literature on changes and revolutions in the history of
science, published outside of France, one inevitably notices the peculiar
fact that philosophical and historical scholarship have lived, by and
large, separate lives. On the one hand, philosophers of science have
developed a number of theories on the logic of scientific change, focus-
sing centrally on the question of how scientific revolutions are to be
accounted for, or whether or not they exist at all. Historians of science,
on the other hand, have actually done, that is, written the history of
science, without, however, showing much awareness of, or at least
without using systematicalIy, and testing critically, the philosophers'
ideas. 17
The negligence of historians in this area has often been deplored by
philosophers, who have argued that the dearth of theory of historians
predisposes them to write the history of science in terms of unreflected
and philosophically long-outdated and obsolete terms. IS Even when
historians have themselves come up with some theorizing, their sugges-
tions tend to be somewhat unsophisticated. To pick just one instance, let
us take a brief, critical look at I. Bernard Cohen's monumental work
Revolution in Science (Cohen 1985). Cohen's study is an exceptional
achievement for a historian of science not only for the vastness of
material covered: the study analyzes more than twenty smaller and larger
84 PART I
scientific revolutions since the 17th century and reports an endless

number of views on what constitutes a scientific revolution. But Cohen's
book is also remarkable in that it amounts to one of the rare cases where
a practicing historian of science seeks to spell out explicitly his views on
what constitutes a scientific revolution and how it is to be identified.
The four criteria that Cohen comes up with are quickly stated. First,
a revolutionary advance in science is always claimed to be such by the
scientist who causes it. Second, the new idea immediately floods into the
literature of the field in question. Third, "competent" historians today
confirm that the new idea - or the bundle of new ideas - indeed amount-
ed to a revolution. And fourth, scientists practicing in the same field
today agree that the idea in question really meant a revolution (ibid.,
40-47).
As is easy to show, these criteria are highly unsatisfactory even if -
in fairness to Cohen - we take note of his disarming admissions that
these criteria are "somewhat subjective" and "do not cover every possi-
ble contingency" (ibid., 47). In fact, it is even somewhat amusing to
note that Cohen's own case-studies of various scientific revolutions in
the very same book provide ample material to challenge his methodologi-
cal suggestions. As concerns the first condition, one can object that, as
Cohen himself impressively demonstrates, the whole conception of
science as advancing in terms of revolutions and ruptures is not older
than the 17th century. Obviously, the criterion is also based on the
somewhat questionable views that a scientist has to have a full under-
standing of her work's importance, and that modesty is a negligible
factor among scientists. With respect to the second criterion, the claim
that the new idea must immediately flood into the literature, it has to be
said that Cohen fails to tell us just how many years may come to pass
for the flooding still to qualify as an immediate one. Mendel's discovery
of the founding laws of genetics had to wait for almost fifty years to
flood into the literature, and even Freud's ideas - and Cohen speaks
explicitly of "the Freudian revolution" - were widely discussed only
after a delay of some fifteen years. In the first six years after the publi-
cation of Die Traumdeutung, only 351 copies of the book were sold
(Freud 1982: 19). Hardly surprising, Cohen's third condition, the
judgment of "competent historians", fares even worse. Clearly, if there
ever were a need for a good example of a circular criterion, Cohen's
suggestion will do as good as any. Since the competence of a historian

depends in great part on his ingenuity in identifying breaks and changes,
we cannot identify revolutions by relying on the historian's alleged
competence. Finally, as concerns the fourth criterion, the view of
scientists today, we need to do little more than recall the well-known fact
that natural scientists usually have a rather rudimentary understanding of
the history of their discipline, an understanding largely based on superfi-
cial, and not markedly "competent" textbooks. 19
Now whereas the unhealthy split between the philosophy and the
history of science is fairly obvious in the academic world outside of
France, such an abyss does not seem to exist in the French context.
After all, in France, from Koyr~ to Canguilhem and Foucault, the
history of science has been written by men trained in philosophy. This
fact alone might be regarded as a prima facie reason for philosophers to
acquaint themselves with French thought. Unfortunately, however,
philosophers of science who have dared to delve into French thought
have barely scratched the surface. This is especially noticeable in Fou-
cault's case, who has typically been read - at least by philosophers of
science - as something of an exaggerated, if not to say perverted,
version of Kuhn. For instance, Larry Laudan, in his Science and Values
(1984) refers to Foucault as a historian of science who, like Kuhn,
conceives of scientific change as "simultaneous rather than sequential"
(ibid., 69). Since Kuhn is wrong, Foucault's excessive Kuhnianism can
then only be "unintelligible ... nonsense", as Laudan informs his reader
in his earlier Progress and Its Problems (Laudan 1977). Laudan tells us
that for Foucault, "the archaeology of ideas" should be written in terms
of "ruptures of human consciousness" and by "invocation of the Zeit-
geist". Laudan concludes, unsurprisingly, that "with its twin emphases
on the mystery and the opacity of human thought, with its stress on
'history as poetry', Foucaultian structuralism must rank as one of the
most obscurantist historiographical fashions of the twentieth century"
(1977: 241). Given such an assessment, it is hardly surprising that philo-
sophers of science, with the notable exception of Ian Hacking, have
displayed little interest in Foucault's work.
Back in France, the situation has been different. There philosopher-
historians like Canguilhem and Dominique Lecourt have commented
favorably on The Archaeology ofKnowledge. For instance, Lecourt hails
86 PART I
Foucault's discours de la methode as a major advance over Bachelard

and Althusser. Lecourt commends Foucault especially for having out-
lined a way of writing the history of science "which refuses both the
continuity of the subject and the structural discontinuity of 'ruptures'"
(1975: 189), and which looks upon science and ideology "as a material,
historically determinate instance in a complex social whole, itself histori-
cally determinate" (ibid., 208). Canguilhem, on the other hand, stresses
the importance of Foucault's conception of ideology for his own work
and professes to have perhaps overlooked earlier the Foucauldian distinc-
tion between different thresholds (1988: x).
In my own reconstruction of the Foucauldian conceptualization of
change in The Archaeology of Knowledge I shall take my lead from
Canguilhem and Lecourt rather than from Laudan. In brief, I shall try to
show that assimilating Foucault to Kuhn is mistaken, first, because
Foucault rejects the continuity vs. discontinuity opposition, second
because Foucault's discursive formations are quite different from Kuhn's
paradigms, and third because The Archaeology of Knowledge has some-
thing of substance to say on the social conditions of science.
THE DEBATE OVER CONTINUITY AND DISCONTINUITY RECONSTRUCTED
In order to situate Foucault adequately with respect to the debate on

scientific revolutions or discontinuities in general, and Kuhn's The
Structure of Scientific Revolutions in particular, it seems imperative to
provide a brief reconstruction of the main stages of this dispute both
inside and outside of France. Developing further a recent suggestion by
Peter Galison (1988), I submit that this debate can be structured around
four models for the philosophy and periodization of science: the posi-
tivist-empiristic model, the antipositivist model, the revised antipositivist
model, and the critical postmodern model. By and large, these four
conceptions can be put in an order of succession at least as far as their
becoming an object of scholariy debate is concerned.
Galison speaks of these models as constituting different types of
philosophies of science. Each type is characterized by one " central
metaphor" which, by the same token, "has deep consequences for the
structure of narrative offered by the history of science" (ibid., 198).
While the suggestion of an interrelation between different philosophies

of science and different ways of writing the history of science is not
new, of course, Galison manages to capture this interrelation in terms of
somewhat oversimplified yet highly suggestive archaeological section
representations.
The positivist-empirist model
The positivist-empirist model operates mainly with two strata, a lower

and independent-determining stratum of observations and experiments
and a higher and dependent-determined stratum of theories (Figure 5).
Positivist periodizatiOD
-,C
_th_eo_ry_I_----'I'-_th_eo_ry_2_ _-..L.I_th_eo_ry_3_ _I_th_eo_ry_4_ _
observation, experiment
time ~~~
Figure 5
The empirist or the positivist conceive of science as being founded upon

accumulating observations that, as time goes by, get more precise and
enlarge in scope. These observations and the observational terms in
which they are formulated are taken to be independent from theory, or
theoretical terms. In other words, observations provide the foundations
that have to be "saved" by different, successive theories.
As Galison is right to propose, this picture of the nature of science
leads to a definite view on how the history of science should be written.
Since observations and experiments are the crucial stratum, and since on
this level no radical breaks or reorganizations occur, the history of
science is a history of accumulation and continuity. To be sure, discon-
tinuity is met with on the level of theories, but since theories are just
codifications of observations, these discontinuities are but phenomena of
88 PART I
the surface or superstructure with no bearing upon the underlying deep

continuity (ibid., 202-3).
As other writers have pointed out, this picture of science also has
the further historiographical consequences of encouraging a belief in
progress, a search for predecessors and anticipations, a belief in the
unity and progressiveness of science, and the writing of the history of
science as a history of methods and results (Agassi 1963: 33-40; Fichant
1969: 96-97). First and trivially, a conception of science working with
the postulate of accumulation will have no problems defending the
progressiveness of science. Second, because of its continuity assumption,
this model suggests looking for predecessors. Furthermore, since the
continuity assumption derives from the belief in an independent stratum
of observations, it is also tempting to think that - at least in principle -
many later-day observations could actually have been made, or were
actually made, much earlier. Third, given these premisses it is also
natural to speak of the unity of science as grounded on the common
observational, theory-independent and discipline-neutral stock of obser-
vations. And fourth, the positivist section metaphor encourages the
historian to concentrate on methods for, and the results of, doing sci-
ence: science is a continuous piecemeal revision of theories, a piecemeal
revision into which major political and social events can intervene only
as temporary obstacles.
Whereas the French locus classicus of this kind of historiography
and philosophy is the work of Duhem, in the English speaking world,
the underlying philosophy is of course foremost represented by logical
empiricism which dominated philosophy in the United States until the
196Os. While little historical work has actually been done by members
of this line of persuasion, it has been suggested that George Sarton's
work on the history of science displays in practice a naive form of what
logical empirism or earlier and later positivism formulate as a much
more sophisticated theory (Fichant 1969: 67-70).
The antipositivist model
The interesting difference between the English speaking world and the
French scene is that the positivist metaphor or section drew fire much
earlier on the old continent. Koynfs study of Galileo came from the
press in 1939, and between 1928 and 1940, Bachelard published alto-
gether ten books outlining and defending the antipositivist philosophy of
science as well as its historiographical counterpart.
Central in the antipositivist model is the reversal of the order
between theory and observation; now theories form the fundamental and
determining stratum, and observations lose their theory-independence
(Figure 6).
Antipositivist periodization
observation) observation2 observation) ...
theory) theory2 theory) ...
time ~~~
Figure 6
To be sure, to make Bachelard a genuine proponent of this picture

is something of an oversimplification; his theory of the relations between
theoretical and experimental physics allows for more than unidirectional
influence between the two strata. Nevertheless, placing him under the
antipositivist Idealtyp can be justified by the fact that the way he was
read in the France of the fifties and sixties fits well with this figure.
As few will need to be reminded, the antipositivist "revolution" in
the English speaking world occurred in the early sixties when Kuhn's
Structures appeared. Kuhn, Feyerabend and their fellow travellers argued
that observation is inevitably "theory-Iaden", and that a theory-neutral
observational "protocol language" ~ la Carnap is an impossibility. It
follows of course that the history of science is the history of (bundles of)
theories or more general theoretical and methodological assumptions.
Furthermore, now that observations are refused the role of the under-
lying stratum common to successive paradigms or grand theories, the
antipositivists tend to go further - or at least are interpreted by their
critics as tending to go further - and deny that there is, or even could
90 PART I
be, any other substratum that connects successive sets of general assump-
tions or paradigms. The endpoint of this argument thus is that adherers
to different paradigms live in "different worlds", and that the question
of whether science is progressive or not is a futile one (Galison 1988:
205).
As to the historiography written according to this model, I have
already noted that for instance the Kuhnian model has inspired few if
any historians in their own research. Or, if it has inspired them, at least
few if any actually employ the Kuhnian scheme of 'paradigm - crisis -
new paradigm'. In fact, in his own substantial historical work on black-
body radiation, not even Kuhn himself draws on this model. As far as
other antipositivist studies of the history of science are concerned, the
major contributions come again from France, to wit, the works of Koyr~
(who inspired Kuhn), Bachelard and Canguilhem. 20 These authors,
however, are not bound to anyone scheme for writing the history of
science, and their antipositivism surfaces more in the forms of a general
attention to problem situations of a science of a given period, a rejection
of the search for predecessors and anticipations, a heightened awareness
of the influence of metaphysical views upon science, and a general
willingness to speak of discontinuities and breaks.
Adding strata: the antipositivist revisionism
To proceed to the next stage of the debate (a stage not distinguished by

Galison), a first wave of nonpositivist critics of the antipositivist model
used mainly two strategies, often in tandem; they introduced new strata
into the antipositivist section drawing and they argued for more fuzzy
borders between successive paradigms (Figure 7). In combination, these
strategies were meant to preserve the rationality and progressiveness of
science. As critics from Popper, Lakatos and Toulmin, to Laudan,
Canguilhem and Hacking saw it, the Kuhn-Feyerabend model of revolu-
tionary scientific change reduces the latter to being "a mystical conver-
sion which is not and cannot be governed by rules of reason ... a kind
of religious change", and "a matter for mob psychology" (Lakatos 1974:
93, 178).21
Revised antipositivist periodization

observation I observation2 observation3
theory I
disciplinary principles I empirical problems
styles of reasoning I research mentalities
time ~ ... ~
Figure 7
Since some of these revisions of the antipositivist section drawing

will turn out to be helpful in my reconstruction of Foucault, let us take
a closer look at Stephen Toulmin's, Larry Laudan's and Ian Hacking's
new proposals.
Toulmin's Human Understanding employs both of the two main
revisionist strategies I have distinguished. As concerns the argument
from fuzziness, he argues against Kuhn that the changeover from New-
tonian to Einsteinian physics cannot be described as "a complete rational
discontinuity" (foulmin 1972: 103). Einstein's work did not constitute
"a fullscale scientific revolution" because every single one of the modifi-
cations that Einstein suggested was discussed in detail, and accepted only
after an intensive debate. Indeed, the more radical the proposed revisions
are that a scientist puts forward, the more prolonged will the ensuing
debate be (ibid., 104). Furthermore, Toulmin plays out "the writings of
T.S. Kuhn the historian" against "the theories of T.S. Kuhn the philo-
sophical sociologist". As Kuhn's own work on Copernicus makes
abundantly clear, Toulmin alleges, "... the so-called "Copernican
Revolution took a century and a half to complete, and was argued out
every step of the way". Even though it resulted in something of a radical
change in "physical and astronomical ideas and theories", the debate
over these ideas "implied no comparable break in the intellectual meth-
ods of physics and astronomy" (ibid., 105). Thus we do best to follow
political theorists, and abolish the use of the notion of revolution, at least
for explanatory purposes, altogether. The distinction between normal and
revolutionary change turns out "to have little real theoretical signifi-
92 PART I
cance": "At most, the two sorts of conceptual changes differ only in
degree; and certainly they must, in the last resort, be accounted for in
terms of the same set of factors and considerations" (ibid., 118).
The new substratum that Toulmin wishes to introduce, and the
substratum that is meant to negotiate between different theories and
ideas, are "disciplinary principles". Other than "theoretical principles"
like Newton's Principle of Universal Gravitation or Mendel's Principles
of Segregation and Recombination, disciplinary principles define "the
basic intellectual goals of a science", e.g. like the explanation of physi-
ological processes in terms of chemical ones. Toulmin suggests that we
confine Kuhnian paradigms to the level of theoretical principles, and
concludes on this basis that the incompatibility between Newtonian and
Einsteinian physics did not extend to the deeper level of disciplinary
principles: "... supporters of the two positions shared enough disci-
plinary aims for them to be able to discuss, in a vocabulary intelligible
to both sides, which of the two theories 'did the better explanatory job'
for theoretical physics" (ibid., 124). Radical incompatibility, like a
disagreement over disciplinary principles, does occur, e.g. Goethe's
disagreement with Newton over colors, but such cases are not really
diachronic breaks within a discipline. Precisely because Goethe does not
subscribe to Newton's disciplinary principles, he does not belong to the
same discipline of theoretical physics (ibid., 124-25).
In Laudan's Progress and Its Problems (1977), the new substratum
consists of "empirical problems" rather than disciplinary principles.
Laudan believes that the historian of science can identify a slowly
changing set of empirical problems that form "a permanent fixture of the
science changes" (ibid., 140). While "research traditions" like Aristo-
telianism, Darwinism or Freudian psychology change or are displaced
and replaced, Laudan submits, this permanent fixture remains and
provides the touchstone with respect to which the relative success of
different research traditions can be evaluated. Different research tradi-
tions can be compared, since "with respect to any two research traditions
(or theories) in any field of science, there are some joint problems which
can be formulated so as to presuppose nothing which is syntactically
dependent upon the specific research traditions being compared" (ibid.,
144). Laudan believes that the introduction of this new substratum
enables him to provide a partial justification to both "revolutionary" and
"gradualist" conceptions of the development of science: discontinuities

occur on the level of "explanation or problem solution", whereas on the
level of empirical problems the assumption of continuity or slow accu-
mulation is justified (ibid., 140).
However, this partial rehabilitation of the discontinuist viewpoint is
to be taken with a grain of salt, since Laudan also employs the second
strategy of the revisionist antipositivist. Laudan rejects the Kuhnian
Gestalt switch not only because of his assumption of paradigm neutral
empirical problems; another line of argument of his weakens the assump-
tion of discontinuity on the theoretical level itself. This line has two
ingredients: on the one hand, Laudan allows for extensive change
already within a research tradition; on the other hand, he suggests the
co-existence of competing research traditions. First, Progress and Its
Problems modifies Lakatosian research programmes by "relativizing the
'essence' of a research tradition with respect to time" (ibid., 1(0). In
other words, the hard core of the most central, general assumptions of
a research tradition is not sacrosanct, but can change over time:
... when it can be shown that certain elements, previously regarded

as essential to the whole enterprise, can be jettisoned without com-
promising the problem-solving success of the tradition itself, these
elements cease to be a part of the "unrejectable core" of the re-
search tradition (ibid., 1(0).
Second, Laudan rejects the Kuhnian notion that a new paradigm emerges
only after a prolonged crisis which lead to the abandoning of its prede-
cessor. Instead, Laudan believes in a "perennial co-existence of conflict-
ing traditions", claiming that it is just this coexistence which makes "the
focus on revolutionary epochs so misleading" (ibid., 136). While not
denying the applicability of the concept of revolution to certain changes
in the history of science, Laudan sees revolutions as phases where a
large number of scientists of a given field start to take seriously a rival
research tradition because of its success in problem-solving (ibid., 138).
The third revisionist notion, or new substratum, is "style". To be
sure, only some uses of this concept can be regarded as revisionist, since
the same term has also played the role of an important metaphor in the
writings of some antipositivists themselves. 22 More clearly as a critical
94 PART I
weapon against the Kuhnian conception, the notion of style is employed

by Alistair Crombie and Ian Hacking. 23
Crombie wants to apply to the study of the history of science a
distinction familiar from Annalist historiography, to wit, the distinction
between the history of ideas or theories and the history of (research)
mentalities (Crombie 1981, 1986, 1988). The later takes as its subject
matter very general commitments concerning conceptions of nature,
science and morals, and relates these to one another via a "taxonomy of
scientific styles":
... [1] the simple postulation established in the mathematical sci-

ences, [2] the experimental exploration and measurement of more
complex observational relations, [3] the hypothetical construction of
analogical models, [4] the ordering of variety by comparison and
taxonomy, [5] the statistical analysis of the regularities of popula-
tions and the calculus of probabilities, and [6) the historical deriva-
tion of genetic development. The first three of these methods con-
cern essentially the science of individual regularities, and the second
three the science of the regularities of populations ordered in space
and time (Crombie 1981: 284).
In a number of papers, Hacking (1982, 1983a, 1985; cf. 1983b: 55,

71) has spelled out the philosophical implications of Crombie's notion of
"styles of reasoning". He uses this conception for two purposes: in order
to defend some sort of relativism and in order to preserve more continu-
ity than the original antipositivist picture of science seems to allow for.
Hacking proposes that the assumption of styles of reasoning forces us to
allow for relativism, simply because "the very candidates for truth and
falsehood have no existence independent of the styles of reasoning that
settle what it is to be true or false in their domain" (1985: 146). How-
ever, the forms of knowledge which different styles of reasoning define
are not to be equated with Kuhnian paradigms. Whereas Kuhnian para-
digms can be shared by only a small group of researchers, styles of
reasoning are "vastly more widespread" (ibid., 149). They evolve much
more slowly, can combine and overlap, separate, and - most important
of all - accumulate:
I can agree with [Kuhn:] ... Knowledge is less cumulative than we

thought. But other things accumulate. In addition to technology and
experimental technique, we can add one style of reasoning to an-
other ... and if a style dies out, the death is slow and often passes
by unnoticed (ibid., 148).
For Hacking as for Crombie, styles of reasoning thus provide an addi-

tional substratum, below theories, or paradigms, a substratum whose
breaks and reshuffles need not, and typically do not, coincide with
changes on these upper levels.
lAudan and Foucault as postmodernists
Hacking's as well as Peter Galison's work (Hacking 1983; Galison 1987)

is also the natural starting point for approaching, finally, the critical
postmodern model, as Galison calls it. Even though this model too
emerged from a critical reflection upon antipositivism, it deviates more
radically from its basic assumptions than the revisionist viewpoint. Eo
ipso, the criticism of the postmodernists thus also applies to the re-
visionists.
By and large, the postmodern approach is characterized by two
steps: it denies that (revolutionary) changes in science affect all strata
simultaneously, and that certain elements, like observations or theories,
are always situated on the determining, deeper level. In fact, Galison
argues that the antipositivist model is but a "flip-side version" of the
positivist section, in that both models "have a well-established hierarchy
that lends unity to the process of scientific work ... in their mirror
reflections there is a good deal of similarity" (1988: 207). For Galison
there is no "reductive hierarchy" to be established between different
levels of physics, Le. "theoretical traditions", and "experimental tradi-
tions", and therefore "it is quite possible for continuity to persist at one
level while being broken at another" (ibid., 209). (Figure 8)
96 PART I
Critical postmodern periodization

theory\
instrument\
experiment\
Figure 8
Contrary to what Galison implicitly alleges, this critical postmodern

model does not only emerge with "the new experimentalism", Le. with
philosophers of science like Hacking and Galison himself who redirect
their attention from the way theories represent reality towards the
problems connected with how scientists intervene in it by way of experi-
ments. 24 This model also surfaces in Laudan's Science and Values
(1984) and, much earlier - alas - in Foucault's Archaeology.
Laudan is not so much concerned with the interplay of theory,
experiment and instrument, as with the interrelations of the three main
components of a Kuhnian paradigm, Le. theories, methods, and aims. As
Laudan sees it, philosophers of science tend to make two mistaken
assumptions concerning these components. First, and here Laudan
includes pre-Kuhnian as well as Kuhnian conceptions, philosophers
suppose that the interrelation between the three components is a hierar-
chical one, such that scientific debates over theories are resolved at the
level of methodology, and methodological debates at the level of values.
Naturally, debates over values must thus inevitably remain undecidable
on rational grounds (1984: 23-43). Second, Laudan claims that Kuhn
and like-minded thinkers, including Foucault, fall prey to "the covari-
ance fallacy", that is, the fallacy of assuming that all three components
can only be replaced simultaneously, not successively (ibid., 43-50). In
other words, the assumption that theories, methods and aims come in
one "inseparable package", making paradigm changes turn out to be
"abrupt and global ruptures", is wrong (ibid., 70-71).
Laudan's arguments against Kuhn are directed against precisely this
inseparable package view. Drawing on examples from the history of

physics and chemistry - examples earlier interpreted by Kuhn as consti-
tuting paradigm shifts - Laudan (ibid., 76-77) shows that the changes
involved are not instantaneous ones from
to
(tim~:) theori~ - method~ - aim~
but rather step-by-step revisions of the three constituents. Laudan sug-

gests that only a superficial historian - Laudan baptizes him "Tom" -
can see abrupt change where in fact the change is almost always gradual.
Tom will be so struck by the shift from timet to tim~, perhaps hap-
pening in just one decade, that he will be blind to the intermediate
stages:
If Tom decides to call the view that scientists eventually come to

hold 'Paradigm 2', and the view from which they began 'Paradigm
1', then he will be able to document the existence of a massive
paradigm shift between what (at our remoteness in time) appear to
be conceptually distant and virtually incommensurable paradigms
(ibid., 78).
To correct these mistakes, Laudan proposes his own "reticulated model

of scientific rationality", which allows not only for different ways in
which theories, methods and aims constrain each other (aims must
harmonize with theories, and they justify methods, which in turn exhibit
the realizability of aims, and justify and are constrained by theories), but
also for slow erosions and sudden overturnings. (Figure 9)
With this rough outline of the postmodern models of Galison and
Laudan, we can turn to the question whether Tom would feel comforta-
ble in the company of Michel. Are we to conceive of a Foucauldian
"rupture" or "discontinuity" as a Kuhnian "irreversible Gestalt-shift", or
"conversion experience", as an instantaneous break that renders discur-
sive formations incommensurable, as a barrier that makes researchers
working in historically subsequent discursive formations live in different
worlds? Some of Foucault's statements indeed suggest this kind of
98 PART I
Laudan's Kuho
theorie8z
methodSz
Laudan's reticulated model

I methodSz I aimS) I
I aimSz I theorieS] I
I theorieSz I methodS) I
time ~~~
Figure 9
interpretation. For instance, The Order of Things speaks of "the sudden-

ness and thoroughness with which sciences were sometimes reor-
ganized", and the same book tells us that after the "great discontinuity"
at the beginning of the nineteenth century, "Classical thought ceased ...
to be directly accessible to us" (OT xii, xxii, 304). As I have already
indicated, my answers to the questions posed are nevertheless negative.
They must be negative not only because of Foucault's explicitly stated
theory of rupture in The Archaeology of Knowledge, but they must also
be negative in the light of his historical-archeological investigations, Le.,
Madness and Civilization, The Binh of the Clinic, and The Order of
Things.
To review briefly the latter works, what Madness and Civilization
studies is not the immediate, break-like emergence of our modern
conception of madness, but rather the gradual, step-by-step emergence of
this conception from the Renaissance to the 19th century. In The Binh of
the Clinic, Foucault studies the rupture in medical thought roughly
between 1780 and 1825. Again the change under scrutiny is gradual, not
immediate: the period under investigation is divided into four phases and
the transformations taking place from one phase to the next are every-
thing but revolutionary. It is only when taken together that these phases
illuminate the radical, rupture-like change in medical thinking between
the outer limits of the whole period. Finally, The Order of1hings, even
though it speaks of "two great discontinuities in the episteme of Western
culture: the first inaugurates the Classical age (roughly half-way through
the seventeenth century) and the second, at the beginning of the nine-
teenth century" (OT xxii), suggests that the emergence of modern
thought - the second rupture - is "possible to follow step by step" (OT
217). Unfortunately, however, in his best known work, Foucault does
not go as far in this analysis as he does in his two earlier books. Yet
even here it must be obvious, even to a superficial reader, that the
change from the classical to the modern period is not understood as
instantaneous and package-like. After all, the transformation is dated
roughly between 1775 and 1825 and a phase of gradual transition is
explicitly allowed for. In this phase, the pillars of Classical thought are
replaced gradually (OT 221).
The distance between Tom and Foucault increases still more radi-
cally, if we turn to Foucault's more theoretical and programmatic texts.
First of all, it deserves to be mentioned that the transition from one
discursive formation to another does not involve the Kuhnian scheme of
'normal science - crisis - normal science'. This is because for Foucault
"the opposition between periods of stability or of universal convergence
and moments of effervescence when minds enter into crisis, ... when all
notions are revised, overturned, revivified, or for an indefinite time, fall
into disuse" is the most important, and worst of the "ill-considered
oppositions" that haunts traditional history (1978: 17). Kuhn's cyclical
picture of the development does not fit the archaeological framework,
since for the latter it is axiomatic that the "appearance and disappearance
of positivities, the play of substitutions to which they give rise, do not
constitute a homogeneous process that takes place everywhere in the
same way" (AK 175). It must also be mentioned that Foucault never
denies that researchers of successive discursive formations are - in
principle - able to understand one another. What Foucault does claim is
that those discursive formations that are separated from us by a rupture
are not "directly [I] accessible to us" (OT 304). This statement must, in
100 PART I
the light of Foucault's Archeology as well as in the light of his historical

studies, be read as a warning against anachronistic assimilations of
earlier scientific concerns to our own, rather than as a plea for incom-
mensurability.
Furthermore, in the Archeology, we read that a rupture "is not an
undifferentiated interval - even a momentary one - between two mani-
fest phases; it is not a kind of lapsus without duration that separates two
periods" . Rather, a rupture between two successive discursive formations
is "always a discontinuity specified by a number of distinct transforma-
tions" (AK 175). Foucault also writes that "rupture is the name given to
transformations that bear on the general rules of one or several discur-
sive formations" (AK 177), or that a "discontinuity ... is a play of
specific transformations different from one another (each one having its
conditions, its rules, its level) and linked among themselves according to
schemes of dependence" (1978: 13).
To go into greater detail, we can start from the observation that
archaeology distinguishes between different forms of change according
to the levels or layers on which they are situated. Change can be studied
on the level of statements, on the level of the four series (objects,
modalities, concepts, strategies), on the level of their regularities, and on
the level of interrelations between different contemporaneous discursive
formations (AK 171).
For instance, on the level of strategies, or "theoretical options",
Foucault distinguishes between seven different types of transformations
within a discursive formation: A theory T2 can derive from, or be
motivated by, a theory T] by "deduction or implication", "generaliza-
tion", "limitation" of T], "passing to the complementary" (e.g. theories
of formal languages motivating the study of natural languages), "passing
to the other term of an alternative", (e.g. from action by contact to
action at a distance), "permutation of dependencies" (e.g. grounding
mathematics in logic leads over into the attempt to ground logic in
mathematics), and "exclusion or inclusion" (1978: 11).
Furthermore, Foucault tells us that to describe the change from one
discursive formation to another must include the identification of trans-
formations on several levels: "[I] how the different elements of a system
of formation were transformed (...); [2] how the characteristic relations
of a system of formation were transformed (...); [3] how the relations
between different rules of formation were transformed (...); [4] lastly,

how the relations between various positivities were transformed (...)"
(AK 172). Using the example of object formation, and employing self-
explanatory abbreviations (with 'conc' for concepts and 'form' for
formation), these different levels can, more succinctly, be rewritten as
follows:
[1] surface of emergence ~ surface of emergencenew

[2] Rei(surface, authority) ~ Relnew(surface, authority)
[3] Rei(formobj' formCQflc) ~ Relnew(formobj , form conc)
[4] Rel(dfm' din) ~ Relnew(dfm' din)
[3] here is of course well in line with my earlier claim that for ar-
chaeology the relation between the different series is not fixed, Le. that
archaeology does not assume, pace Laudan, any fixed hierarchy between
different strata.
Moreover, let us also recall that a discursive formation is "a scheme
of the correspondence between several temporal series", a view which I
interpreted to mean that for us to speak of a discursive formation it is
not necessary that the time limits of all of the four series coincide. In the
central chapter on "Change and Transformations" of The Archaeology of
Knowledge, further evidence to the effect that the four central ingredients
of a discursive formation do not change wholesale, that is to say, that
they do not form one "inseparable package", is easily forthcoming. It is
true that Foucault claims that "the appearance of a discursive formation
is often correlative with a vast renewal of objects, forms of enunciation,
concepts, and strategies" (AK 171), but not only does he hasten to add
a counterexample ("General Grammar was established in the seventeenth
century without much apparent alteration in grammatical tradition"), but
he also writes that the replacement of one discursive formation by
another does not entail "that all objects or concepts, all enunciations or
all theoretical choices [of the earlier discursive formation] disappear"
(AK 173).
It is also important to note here that archaeology regards the dis-
placement of discursive formations as "rare ... events", and that al-
though they are its central topic they are by no means the only one (AK
171). Indeed, archaeology concerns itself also with "continuity, return,
102 PART I
and repetition", too:
... elements - or several of them - may remain identical (...), yet

belong to different systems of dispersion, and be governed by
distinct laws of formation. One can find in such phenomena there-
fore: elements that remain throughout several distinct positivities,
... ; elements that are constituted, modified, organized in one discur-
sive formation, and which, stabilized at last, figure in another;
elements that appear later, as an ultimate derivation in a discursive
formation, and which occupy an important place in a later formation
(...); elements that reappear after a period of desuetude, oblivion, or
even invalidation (...) (AK 173-75).
As all of this evidence suffices to show, Foucauldian archaeology

knows of no fixed hierarchical relations between different layers, does
not treat different layers as forming an "inseparable package", does not
assume "incommensurability" and rejects explicitly the Kuhnian cyclical
model of the history of science. Thus Foucault is as clear an instance of
the critical postmodern model of science as any.
THE ARCHAEOLOGICAL DIMENSION
It hardly needs to be argued at length that to work with fairly abstract

models like Galison's types of periodizations easily leads one to overlook
important differences between those philosophers of science that are
grouped together as adherers' to one and the same 'model'. Indeed,
Galison's, Laudan's and Foucault's central interests with respect to
science differ considerably, despite their common "postmodernism".
Galison is mainly concerned with showing that experimentalists in the
physical sciences are not mere underworkers to theoretical physicists,
Le. that experimentation has a "life of its own", and that a historical
study of the instrument-building tradition of science should be related to
the wider context of material culture (Galison 1988: 208). Laudan's
main interest, on the other hand, is the defence of science as a rational
enterprise that only in exceptional cases stands in need of sociological
reconstruction. Laudan tries to achieve this goal by developing a concep-
tion of rationality according to which rationality itself can change in a

rational way. Here it remains to show how Foucault's main problem
differs from both Galison's and Laudan's.
That Foucault occupies himself with topics somewhat remote from
the questions that philosophers of science debate can be most easily
demonstrated with reference to what I propose to call 'The philosophy of
science questionnaire', or PSQ for short. The PSQ demands of the
proband to define her position with respect to the role, function and
working of "guiding assumptions", rationality, and methodology in
science, where by guiding assumptions I understand things like Thomas
Kuhn's paradigms, Imre Lakatos's research programmes, Larry Lau-
dan's research traditions, and Paul Feyerabend's global theories (Laudan
et al. 1986: 161-62). Using the initials of these four leading figures, one
can summarize their answers to the PSQ in the following table. (Figure
10)25
It is the impossibility of filling out the PSQ for Foucault which
should have made commentators more careful in their eagerness to
assimilate Foucault's work to Kuhn's historical and philosophical studies.
Perhaps one can say that Foucault's texts suggest answering negatively
to (1), (2) and (7) and positively to (12), but even here one feels unsure
whether the guiding assumptions of Kuhn & co. can really be paralleled
to the four series of a discursive formation.
Of course this negative result can lead to two different attitudes with
respect to Foucault's work. Some might take the absence of clear an-
swers to the PSQ simply as the final confirmation of their already
entrenched view that we are indeed dealing here with "one of the most
obscurantist historiographical fashions of the twentieth century" (Laudan
1977: 241). More charitable philosophers will argue that the difficulty in
squeezing Foucault into the PSQ is simply as clear a sign as any that the
PSQ problems are not his. These philosophers will hold that without a
proper grasp of Foucault's - or anyone's - central interests, criticizing
and rejecting him "is not unlike playing one of those parlour games in
which one is given an answer (often a bizarre one), without knowing the
question to which it is an answer!" (Laudan 1977: 176).
To turn then to what I take to be the main problem for archaeology,
we must come back to the notion of threshold. I already mentioned the
idea of various thresholds in an earlier chapter on Foucault's criticism of
104 PART I
Tit PF IL LL
(I) guiding assumptions (GAs) are explicit at the outset n n y y
(2) the core of a set of GAs can change piecemeal ... n n y
(3) comparison between different sets of GAs is the rule
rather than the exception . . . . . . . . . . . . . . . . . . . n y y y
(4) new sets of GAs are seriously explored even before
they succeed . . . . . . . . . . . . . . . . . . . . . . . . . . n y y
(5) given a successor set of GAs (S*) and its predecessor
(S):
- S* accommodates 8's explanatory successes ...... n n y n
- S* is more general than S · ................ n y
- S* is as precise as S .................... nly
...
- S* accommodates S's solved empirical problems n/y n
- S* solves S's anomalies .................. n y n
- S* accommodates S's observational consequences .. n y
(6) disputes over GAs occur constantly .......... n y
(7) during a change in GAs (a scientific revolution) GAs
change abruptly and totally · ................ y n
(8) opponents of different Ss cannot easily communicate:
the meanin~s of their observational terms are different,
th'J have dIfferent problems and employ similar language
an experiments in different ways .. . . . . . . . . . . . y y n n
(9) scientists switch to a different set of GAs because of
propaganda ........................... n y n n
(10) methodological rules are formulated as a matter of
routine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . n y
(11) they always change when GAs change ....... yIn n n
(12) they change over time · ................ y n y
Figure 10
over-simplified oppositions in general, and Bachelard's and Althusser's

strict dichotomy between science and non-science or ideology, in particu-
lar. Now we have to see in greater detail how crucial this distinction is
to the whole archaeological enterprise.
The first threshold (the threshold of positive discourse), Le. the
emergence of a discursive formation out of the discursive field in gene-
ral, was defined above as the arising of a more stable interplay between
the four series. More can and needs to be said about the other three
thresholds, and the types of discourse they characterize.
The second threshold, the threshold of epistemology, issues in what
we might call an 'epistemology-based discourse'. The Archaeology of
Knowledge tells us that at this stage within a discourse, "a group of
statements is articulated, claims to validate (even unsuccessfully) norms
of verification and coherence, [a group which] ... exercises a dominant
function (as a model, a critique, or a verification) over knowledge" (AK
186-87). To flesh out this sketch it is inviting to relate it to Kuhn's
preparadigmatic phase as well as to some other notions in the philosophy
and anthropology of science.
As concerns Kuhn, it can be pointed out, first of all, that his notions
of "immature" or "preparadigmatic" science do not differentiate between
the different types of discourse that Foucauldian archaeology distin-
guishes in the "prehistory" of a science. In the first edition of the
Structure, Kuhn, like Bachelard and Althusser, works only with a simple
dichotomy, Le. a dichotomy between fields having one (or very few)
paradigms, and fields that are, in the absence of a paradigm, "relatively
free" in that there is "no standard set of methods or of phenomena that
every ... writer [feels] forced to employ and explain" (1970: 13).
However, in the 1970 "Postscript" to the second edition of The Struc-
ture, Kuhn introduces a modification: the difference between maturity
and immaturity is no longer linked to the presence of a paradigm:
'" it may help to point out that the transition [from immaturity to
maturity] need not (I think now should not) be associated with the
first acquisition of a paradigm. ... What changes with the transition
to maturity is not the presence of a paradigm but rather its nature.
Only after the change is normal puzzle-solving research possible
(ibid., 179).
106 PART I
This quotation suggests that we can distinguish between three phases:

first, a phase where models of research are so vague, general and
conflicting that we cannot yet speak even of a paradigm; second, a phase
where paradigm exemplars exist, but not yet as well-defined puzzles; and
third, a phase where the number of paradigms has been drastically
reduced, and the "normal puzzle-solving research" has become possible.
The second stage of this conception might then be looked upon as
corresponding roughly to Foucault's epistemology-based discourse.
Other notions naturally situated on this level are perhaps Toulmin's
disciplinary principles, Crombje's and Hacking's styles of reasoning or
research mentalities. 26
Moving on to the "threshold of science", Foucault's initial charac-
terization of the resulting 'scientific discourse' as one which "obeys a
number of formal criteria ... [and] certain laws for the construction of
propositions" (AK 187), is again rather uninformative. However, to give
this form of discourse more content, it helps to note that Foucault later
situates Bachelard's and Canguilhem's epistemological history of science
on this level (AK 190), a fact that suggests that "scientific discourse" is
characterized by the presence of one (or very few) problematics. Equat-
ing problematics with Kuhnian paradigms as "disciplinary matrixes", it
thus seems natural to say that on this level there exists a "community of
specialists" who are committed to certain "symbolic generalizations ...
which can readily be cast in a logical form like (x)(y)(z)<p(x,y,z)",
"beliefs in particular models" which among other things "supply the
group with preferred or permissible analogies and metaphors", cognitive
values like commitment to accurate predictions, simplicity, consistency,
and a fixed set of exemplars (Kuhn 1970: 182-91). It also seems natural
to situate Toulmin's aforementioned theoretical principles on this level.
Finally, suffice it to mention that Foucault wants to make a further
distinction by differentiating between the threshold of science and the
"threshold of formalization". As the term already indicates, the formal-
ized discourse is one in which theories are formulated in a logical or
mathematical language (AK 187).
Now as it stands up to this point, the Foucauldian distinction be-
tween these thresholds and types of discourse is an interesting suggestion
for a more fine-grained grid for the study of the emergence and develop-
ment of mature, scientific disciplines. It cannot be stressed too much,
however, that this way of reading the threshold theory is still one-sided.
What this reading overlooks is that the Foucauldian distinction between
the various forms of discourse is not only a distinction between succes-
sive stages of the prehistory and history of a scientific discipline, but
also - and first and foremost - a distinction between different perspec-
tives with respect to, or different strata of, one (or several) scientific
discipline(s).
To begin with, the Archaeology claims that the chronology of the
different thresholds and types of discourse "is neither regular nor homo-
geneous" (AK 187). This is to say, first, that not all fields of learning
pass through these different stages simultaneously, "thus dividing up the
history of knowledges into different stages". Second, the order in which
different discourses pass through these stages does not follow a fixed
scheme: different discourses cross different thresholds at different inter-
vals, and for some discourses the passing of two or more of these
thresholds can coincide. For example, mathematics has crossed all of the
four thresholds at once (AK 188).
Third, and most importantly, even though for each discourse one
can identify the moments when it crosses one of these thresholds for the
first time, the structure it obtains at these thresholds is not necessarily
abolished when another, later or higher, of the thresholds is passed.
Furthermore, we must reckon not only with thresholds where a discourse
receives orders (of the four kinds) for the first time, but also with
thresholds where for one and the same level, the orders are changed or
replaced. Put differently, the specific structure that a discourse receives
at a given level is preserved even when the following 'higher' level
emerges. Subsequently, what originally came about - at least typically -
successively, now interacts, and constitutes different interrelated strata
that can, but need not necessarily, develop simultaneously. (Figure 11)
In other words, changes on any of the four levels can, but need not
necessarily, go hand in hand. Thus the introduction of, or the change in,
formal mathematical theories need not be accompanied by a change of
paradigms, as evidenced by Maxwell's mathematization of Faraday's
field theory. The introduction of, or the change in, paradigms need not
coincide with changes in disciplinary principles, as Toulmin shows
concerning the debate between Newtonians and Einsteinians. However,
the introduction, or emergence, of a new style of reasoning can be
108 PART I
Dimensions of the history of science

formalized
discourse ••••••••••.••.••••••••• 1 fd l fd 2
scientific ,...----'-----r------+----
discourse •••••••••••••••••• 1 sd l I sd2 sd)
,..--_--'L...-_-..,,..--_.1.-
I
~ist.-based ...L- _
discourse •••• • • • • • • • ed I ed2

discursive
I
formation •••• dfl df2 I df)
time •••
Figure 11
accompanied by the first introduction of a paradigm and a mathematical

tool, as Ian Hacking's study of the emergence of probability in the 1660s
demonstrates (Hacking 1975). Finally, the structure of the epistemology-
based discourse can change without paradigms or formalisms being
replaced (immediately); for instance, the image of (physical) knowledge
of some physicists has changed as a result of the atom bomb, or of the
Star Wars program, without this change having resulted in any new
paradigm or formalism.
Having already seen how archaeology defines the relation of the
level of discursive formations to weakly structured discourses, the
question that remains is how the introduction and change of a discursive
formation relates to other changes. To answer this question, we have to
recall that a discursive formation is not just a series of series of objects,
enunciative modalities, concepts and themes or strategies, but also the
interrelation of regularities of social conditions that constitute these
series. (This is obvious in Madness and Civilization, The Birth of the
Clinic, and The Archaeology ofKnowledge, but obscure in The Order of
Things.) Thus we saw that central in the constitution of objects like
madness were, e.g., the role and function of institutions like the family
and the working place, as well as the authority of the doctor, the judge
and the priest; crucial for the emergence of a specific series of enunci-
ative modalities in medical discourse were the social structure within
science and the role of the scientist in society; important for the role of
certain concepts like the 'spring of air' are specific orderings of state-
ments which in turn reflect the agonistic struggle between different
schools and different disciplines (Boyle vs. Hobbes); and, finally, all
important for the ending of debate over theory choice are competition for
important alliances and the intervention of non-scientific interests. What
all this suggests, therefore, is that the level of discursive formations is
the stratum of the social conditions for the emergence of a structured
discourse within which - perhaps only at a much later date - emerge one
or several disciplines with varying disciplinary principles, paradigms and
theories.
Obviously, given this role of discursive formations, it is then only
natural to assume that with the emergence of further structures like
paradigms, the earlier level of discursive formations has by no means
become superfluous. As these new structures evolve, there are new
regularities and series to be accounted for, series that do not neatly
coincide with changes in the 'social conditions of series of objects or
theories. Indeed, Foucault claims that these new series do not necessarily
have the same limits as successive systems of social conditions, Le.
discursive formations (AK 188). It is clear, e.g., that the existence span
of a discursive formation is considerably longer than the span of theories
of even paradigms; after all, a discursive formation is defined as a series
of even conflicting theories, objects, modalities and concepts. On the
other hand, it is also compelling to allow that there are cases where
paradigms or theories remain, regardless of a change in the social
conditions; mathematics or theoretical physics are as good examples as
any to establish this point.
7. ARCHAEOLOGY OF KNOWLEDGE AND

OTHER HISTORIES OF SCIENCE
Discursive formations as series of series of objects, enunciative modal-

ities, concepts, as well as their conditions, are thus themselves just one
series, one level to be described by histories of science. Not only is this
claim strongly suggested by my interpretation above, but it is also stated
straightforwardly by Foucault ~imself. In fact, he distinguishes between
three types of histories of science, according to the level or series of
110 PART I
events they are primarily concerned with. Thus the history of mathe-
matics writes the history of the· series of formal theories, where typically
later theories are generalizations of earlier ones (AK 189). Foucault does
not mention examples of such histories, but one might mention the work
of Jean Cavail1~ as an outstanding example of this type of historio-
graphy. A second way of writing the history of science is the study of
the emergence of problematics or paradigms. As Foucault sees it, "G.
Bachelard and G. Canguilhem have provided models of this kind of
history" (AK 190). Interestingly enough, The Archaeology ofKnowledge
claims that this kind of history can only be written by taking a normative
perspective, that is by judging the past of science in terms of its present
("this description takes as its norm the fully constituted science" (AK
190)). Finally, the description of the emergence of a discursive forma-
tion as well as its acquiring of epistemological principles, Foucault
reserves for the archaeology of knowledge (AK 191).
Here is seems inviting to go further and suggest that the distinction
between the four levels or strata in the history of science (both as object
and discipline) can also be related to the better-known fields of historical
research in the English speaking world. Thus Imre Lakatos's work on
the history of mathematics would be an obvious counterpart of Cavail1~,
and his consciously anachronistic "rational reconstructions" of "research
programmes" parallels Bachelard's "histoire sanction~"; like Bachelard
Lakatos wants to write the history of science "in the light of his ratio-
nality theory", Le. Lakatos's objective is an internal, "radically im-
proved version" of the history of science, a history that mentions "'false
beliefs'" at most "in a footnote" (Lakatos (1978: 119). Laudan's evalu-
ative scheme for research traditions, a scheme based on problem-solving
power, is also naturally situated on this level. Finally, Crombie's work
moving towards an anthropology of science seems to be concerned
mostly with the level of epistemologized discourse; after all, he describes
and explains how and when (new) disciplinary principles, styles of
reasoning, images of knowledge and research mentalities emerge.
Be it noted also that Foucault's distinction between these various
ways in which the history of science can - and should - be written sets
him favorably apart from the somewhat imperialist tone of voice with
which other methodologists present their own model to the exclusion of
others. 27
However, it would be a mistake to think that because for Foucault

the archaeology of knowledge is concerned with just one possible
stratum of historiography, its criticisms and methodological suggestions
are of no avail to other approaches. Certainly the criticisms and sugges-
tions of The Archaeology do carry over into other domains of the his-
toriography of science: its criticism of the history of ideas stands and
must be answered whether one adopts the archaeological edifice or not;
its identification of anachronistic tendencies is irrelevant only for re-
search programs that consciously adopt the viewpoint of the histoire
sanetionee; its Annalist call for greater attention on the identification and
delimitation of series holds for every historical project; its proposals for
studying statements and textual strategies has not become obsolete due to
recent work in discourse analysis; its suggestions on the interrelations
between disciplines should be of interest even when studying discourses
on the level of the epistemological or the scientific; and its refinement of
the simple scheme 'prehistory vs. history' (of a science) might be
relevant to historians and philosophers of science alike.
It is for this reason, Le. the importance of the archaeological sug-
gestions for any historiography of science, that there also is no need to
follow those interpreters who construct something of a "break" between
Foucault's archaeological and genealogical works. As I see it, Foucault
could have left behind the general historiographical proposals of The
Archaeology only by paying the high price of falling back to traditional,
naive history. What genealogy amounts to then cannot be a rejection of
the basic premisses of historical constructivism marked by the identifica-
tion of series and the investigation into their correlations. Instead,
genealogy implies the introduction of further strata. To put it in a
nutshell, while The Archaeology is mainly concerned with the series of
objects, modalities, concepts and strategies, while The Archaeology treats
social factors more as conditions of these series rather than as series in
their own right, genealogy analyzes more explicitly the series of social
conditions themselves. To do this calls for definitions of new kinds of
events, new kinds of relations, and new kinds of interactions between
series and series of series. In short, it calls for a political theory of
science, for a theory on the relations between power and knowledge.
NOTES TO PART I
1 Sheridan (1980: 90): "The austerities - the aridities, some said - of

L 'archeologie du savoir could not fail to disappoint."
2 For Merleau-Ponty, see Waldenfels (1983: 191); cf. Gutting (1989:
5). For Ricoeur, see Waldenfels (1983: 308).
3 Cf. Gutting (1989: 5). Gutting does not give the reference to Ca-
vailles' text. The passage is in Cavailles (1987: 76) .
.. For general background of the school, I rely on chapters nine and ten
of Bourde and Martin (1983). Cf. Stoianovich (1976).
5 Furet (1985). A parallel reading of this text and Foucault's Archae-
ology is suggested in Bourde and Martin (1983: 325).
fi On Braudel, Aries and Revel, see Revel (19800); Althusser (1970a:
45); Canguilhem (1988: ix-x). For a more critical voice of an Annalist, see
Vilar (1976: 13). See also Vilae (1985: 62-63).
7 Even though the biological metaphor does not loom large in Love-
joy's writings.
8 Above we already encountered the critique of members of the Annales
school, of Bachelard, Canguilhem and Althusser. To this group of critics we
can further add, among others; Michel Fichant (1969), Michel P6cheux
(1969), Bruno Latour (1987, 1988), Fran~ois Jacob (1974), Joseph Agassi
(1963), Larry Laudan (1977), Conal Condren (1985), John Dunn (1968),
W. F. Greenleaf (1966), J.G.A. Pocock (1972), and Quentin Skinner
(1969).
9 Interestingly enough, in writing about the Foundations of Modern
Political Theory, Skinner himself deals more with ideologies and styles of
legitimation than with the intentions in writing. (See Boucher 1985: 223.)
10 Of course, this parallel should not come as a surprise; after all, the
attempts of the first generation of Annales historians (Febvre, Bloch) to
model the various mentalities of medieval and early modem times on the
illogical and hypersensitive mentalite primitive of Uvy-Bruhl has been
criticized even by members of the Annales School itself (Chartier 1982).
11 This reformulation is influenced by Komer (1984: 118-19).
12 lowe this example to Riitta Korhonen-Kusch.
13 However, over and above an illuminating explication of the notion of
"antidiscipline" little theoretical work has so far been done in this field
(Lepenies 1978; Wilson 1977).
14 In fairness to Foucault, I should mention that this observation is at
least hinted at in the Archaeology in his distinction between different
112
NOTES TO PART I 113
thresholds.
15 It is also inviting to ask to what system of modal logic these accessi-
bility relations correspond. After all, different systems of modal logic, like
S4, SS, T, and B, are semantically distinguished precisely by whether the
accessibility relation between possible worlds is reflexive, and/or symmet-
ric, and/or transitive (see, e.g. Hughes and Cresswell 1968: 62-81).
16 Thus it turns out to correspond to the accessibility relation that
characterizes system T of modal logic.
17 Even a work like Thomas Kuhn's The Structure ofScientific Revolu-
tions (Kuhn 1970) has had little impact on the historian's practice, even
though some occasional critical remarks on Kuhn's scheme can be found
here and there (see e.g. Reingold 1980).
18 This is one of the central themes in Agassi (1963). See also Fichant
and Pecheux (1969).
19 For more criticism of Cohen, see Agassi (1988a).
20 Ludwik Fleck's Entstehung und Entwicklung einer wissenschaftlichen
Tatsache (1935/1980) is perhaps the most important exception.
21 For Popper's criticism, see Popper (1974). For Canguilhem's criti-
cism, see Canguilhem (1988: 23). For Laudan, Toulmin and Hacking see
below.
21 Thus Paul Feyerabend has recently couched his radical antipositivism
in terms of a comparison between the history of style periods and the
history of science. (Feyerabend 1984). Already in the thirties, Ludwik
Fleck, a later rediscovered early proponent of a similar model, spoke of
"Denkstil" as a conglomerate of taken-for-evident judgments, methods and
technical as well as literary devices. For Fleck, the style of thought prevail-
ing within a period or a science puts constraints upon the ideas of its
members, it "determines 'what cannot be thought otherwise'" (Fleck 1980:
130). Fleck's favorite example of a different style of thought that has
become inaccessible to us were Paracelsus's writings (ibid., 45).
23 w. L. Wisan's work also deserves to be mentioned. He tries to
preserve continuity in the history of science by shifting attention away from
explicitly formulated theories towards "the emergence of new scientific
styles" (Wisan 1981). Here a style is characterized by a "structure", which
can be, e.g., classificatory or algebraic, by "content" as the subject matter,
by "techniques" like philosophical analysis or geometrical methods, and by
"expressive quality", e.g. "a feeling for the world of substances and
essences" (ibid., 325-28). Wisan suggests that whereas theories might
emerge suddenly, the scientific styles upon which they are based evolve in
114 NOTES TO PART I
a slow and piecemeal fashion (ibid., 312-14).

24 The term "new experimentalism" was coined by Robert Ackermann.
See Ackermann (1989).
2S The table is constructed on the basis of Laudan et al. (1986).
26 Yehuda Elkana's "images of knowledge" might also be situated on
this level: "beliefs held about the task of science (understanding, prediction
etc.), about the nature of truth (certain, probable, attainable, etc.), about
sources of knowledge (by revelation, by ratiocination, by experiments
through the sense)" (1981: 14).
27 For instance, despite the parallels that exist between Foucault's and
Quentin Skinner's criticisms of certain anachronistic tendencies in the
history of ideas, Foucault's allowing for several registers in writing the
history of thought enables him to avoid Skinner's absurd Baconist conse-
quence according to which, as one critic puts it, "we can either become
historians, and renounce the attempt to make connections with the present;
or else, in Mr. Skinner's (strikingly Baconian) words, 'We must learn to do
our own thinking for ourselves'" (Leslie 1970: 435).
PART II:
FOUCAULDIAN GENEALOGY
8. INTRODUCTION
As I have already pointed out in the introduction of this study, my

reconstruction of Foucauldian genealogy is meant to be, by and large,
independent of my reconstruction of the archaeology. This stated inten-
tion notwithstanding, and in order to provide the reader with some hints
of what is to come, a few additional comments on the relations between
the two projects are called for, nevertheless.
In his genealogical writings, Foucault's central interest does no
longer lie with developing an apparatus for conceptualizing relations
between statements and their indices. This type of series is not rejected
as insufficient or unsatisfactory, but the focus is now upon defining a
new set of series of entities and processes. These entities and processes
are power mechanisms, power networks, interests, systems of exclusions
and prohibitions, and coercive institutions. To be sure, the archaeologi-
cal edifice does not rule out attention for such entities. Indeed, for a
precise delimitation of series of objects, already The Archaeology calls
for an investigation into institutions and social relations. Yet The Ar-
chaeology tends to treat such institutions and relations more as social
conditions of discourses, rather than as series of entities and events in
their own right. While in The Archaeology social conditions are invoked
mainly as elements of explanations for the existence of certain forms of
scientific discourses, genealogy raises the domain of the social to a level
that itself stands in need of explanation and analysis. Put differently,
genealogy seeks to provide a new conceptualization for social relations
in general. Scientific discourse as a subject matter does not go by the
board, but the set of principles of selection, the set of principles for
constructing series of texts and discourses, is being enlarged. In Fou-
cauldian genealogy, the existence of a discourse, at a given point in
time, is only one of two possible phenomena to be explained. The new
phenomenon attended to are relations and organizations in the social
domain, and here the existence of discourses can itself be part of the
115
116 PART II
explanation.
It is well in line with the archaeological-methodological demands
that this new or additional principle of selection calls for some kind of
theory as to what kind of entities or processes power mechanisms, power
networks, etc. are, and how they are to be identified and studied.
Unfortunately though, Foucault's genealogy is a highly complex
bundle of ideas, difficult to present coherently, and hard to assess. This
is due partly to the fact that Foucault has not given us for genealogy,
what he has provided us with for archaeology, Le. a booksize elabora-
tion of its main concepts, theses and methodological assumptions.
Rather, his views have to be gathered from several historical mono-
graphs, his editions of historical source materials, numerous articles and
- perhaps not surprisingly for a "French Mandarin" - also from dozens
of interviews. Thus it is understandable that many writers in the field of
science studies ignore or quickly bypass his contribution, and that studies
of Foucault's genealogy either reflect the partial impenetrability of his
texts or confine themselves to rather limited aspects dealt within them.
My discussion too will confine itself to only some facets of Fou-
cault's later work. Ignoring Foucault's dislike for systematicity and
stringency, I shall try to show that his various pronouncements on power
and science can be reconstructed in a way that makes the different
assumptions and parts of genealogy not only distinguishable but also
separately open to assessment. In other words, Foucault's genealogy
need not be the diffuse tangle of ideas which it easily seems to be upon
first reading. I will also seek to show that the different theses of genealo-
gy are often defensible either as they stand, or when supported by
further arguments. To be sure, engaging in this reconstructive work fre-
quently implies going beyond Foucault's own wording: occasionally I
impose a different vocabulary upon his ideas, at several points I establish
links between ideas of his and others that he himself did not attend to,
and sometimes I search for new arguments that he might not have
accepted himself. Obviously, this procedure can only lead to the con-
struction of a 'Foucauldian genealogy', rather than to the faithful unfold-
ing of 'Foucault's' genealogy.
I shall start from a reconstruction and defense of Foucault's theory
of power. For want of a better term, I propose grouping under this
heading of 'theory' those of Foucault's pronouncements that suggest
FOUCAULDIAN GENEALOGY 117
what constitutes a power relation, how power interacts with knowledge,

what types of power can be distinguished, and how power works. For
analytical purposes, this kind of theory of power might be distinguished
from three other ventures vis-tl-vis power, to wit, a methodology of how
power and its effects are to be investigated, a political criticism of
vicious forms of power, and a historical analysis of specific forms of
power in different historical places and periods.
In my discussion of the first two ventures, I shall occasionally draw
attention to parallels and differences between Foucauldian genealogy and
recent work in the sociology of science (Barnes, Bloor, Collins, Latour,
Schaffer, Shapin). On the one hand, the parallels between Foucault and
the sociologists of science are noteworthy because even where Foucault
does reach conclusions strongly reminiscient of the sociology of science,
he arrives at them via a different route. Thus his independent arrival at
similar views might be looked upon as providing independent support for
certain theses of the sociologists. On the other hand, the differences
between Foucault and the sociologists are interesting since they might
lead to advancing our understanding of what a social reconstruction of
the history of science is or should be. (E.g. what role a theory of power
should play in social studies of science.)
Finally, in a brief, concluding, chapter, I shall defend a Foucauldian
strategy for political criticism of science, a strategy that stresses the
importance of questioning rationalities and that abstains from basing
itself upon an elaborate normative, or utopian theory of science or
society.
9. THE CONCEPT OF POWER
My treatment of what I propose to call Foucault's theory of power is

structured around a distinction first advanced with respect to justice
(Rawls 1972: 5-6) and later applied to power (Lukes 1974: 26-27). This
is the distinction between the concept and conceptions of power, where
the concept of power refers to the basic core or primitive notion of
power lying behind all (or most) talk of power in the social sciences and
philosophy, a basic core that is developed or fleshed out by different
authors into distinct theoretical conceptions.
118 PART II
Even though I do not believe that there is one single concept of

power underlying all theories of power, I shall try to show that at least
the best-known sociological work on the notion of power (Lukes 1974),
on the one hand, and Foucault, on the other hand, roughly share the
same concept of power. In establishing this point, I hope to convince the
reader that, when speaking of power, Foucault and Lukes, are not
referring to totally different phenomena.
It is only subsequently that we shall go into the details of Foucault's
conception of power. As we shall see, this conception is marked not only
by its employment of - partly incompatible - models of power, but also,
for instance, by its stress on mechanisms of power, its emphasis upon
the relations between social power and scientific knowledge, and its idea
of power as an omnipresent element of social life.
SOME STANDARD DISTINCTIONS
Let us start by recalling, somewhat telegraphically, some of the main

notions in the standard literature on power, namely, the counterfactual
analysis of power, the different constituents and aspects of power rela-
tions, and the various forms or 'species' of power. (The reader familiar
with these notions can proceed directly to the next section.)
We attribute power not only to people - individuals or collectives -
but also to animals, machines, God(s), and abstract things. We say, for
instance, that boxers have more power in their fists than most other
men, and we speak of the power of teachers over their students, of the
power of a crane, or the power of love. Even though it is hardly infor-
mative, we might say that all these attributions are attributions of capac-
ity to bring about or hinder the coming about of states of affairs or
processes (Barnes 1988: 2).
Contrasting the power of a crane and the power of a boxer's fists
with the power of a teacher suggests a first distinction, Le. the distinc-
tion between social and physical power. Of course, a boxer's physical
power is connected in various ways to his social power, and even the
physical power of a crane can be involved in all kinds of social relations.
Still, the ability, power, or capacity to do physical work can be set apart
from the ability to get someone to do, or abstain from doing, something.
However, 'getting someone to do, or abstain from doing, something'

is still an insufficient characterization of what we mean by social power.
This observation has led writers on power to conceptualize power
relations in terms of counterfactuals and interests. The underlying
intuition seems clear enough. Were it not for the actions of the one who
"has" or "exercises" power, Le. the actions of the "power holder", the
victim, or "power subject", would have acted otherwise, that is, in a
way conceived by her to be more advantageous, more in accord with her
own interests.1
However, linking the attribution of power to the possibility of the
power subject to have acted otherwise and linking it to her interests is
not without problems. In practice it might often be rather difficult to
determine just how the power subject would have acted, or what her
interests are. While admitting these problems, Stephen Lukes, the writer
whose views have perhaps been most widely discussed in recent years,
has suggested that these problems can nevertheless be overcome. Histori-
cal knowledge on how people act under various circumstances and a
notion of real interests, Le. interests that agents do not have to be aware
of themselves, allow us, or so Lukes alleges, to formulate the relevant
counterfactuals even in cases where prima facie we have no direct
knowledge of the deliberations of the power subjects' themselves (Lukes
1974: 46-50).
However, to analyze relations of power does not mean to study only
the interests of the power subjects. The study of power is also concerned
with distinguishing between different constituents of power relations, and
with driving conceptual wedges between different forms of power.
Different authors have suggested and developed various distinctions
between different constituents of power relations, but it seems fair to say
that the following list has emerged as something like the standard view
on this topic2:
(a) the base of a power relation refers to the resources (e.g. economic
wealth or prestige) that the power holder can draw on in order to
influence the actions of power subjects';
(b) the means of the power relation are the specific actions (e.g. com-
mands or threats) the power holder employs in the attempt to subdue
the power subject;
120 PART II
(c) the scope of power is the set of actions that the power holder, by
employing her means and resources, can get the power subject to
perform or abstain from performing;
(d) the amount of power denotes the net increase in probability of the
power subject's performing (or abstaining from performing) a
specific action due to the actions of the power holder;
(e) the extension of power is the set of individuals over which a given
power holder has power;
(t) the costs of a power relation are the various costs that arise for the
power holder from establishing and maintaining the power relation;
(g) the strength of a power relation refers to the various costs that arise
for the power subject in case she refuses to enter and remain within
the relation of power; and
(h) the intensity or zone of acceptance of a power relation is the set of
actions which the power holder can get the power subject to perform
without loss of the latter's compliance.
Turning from constituents to forms of power relations, philosophers

and sociologists have much debated the relations between the notions of
authority, coercion, deterrence, force and manipulation. While different
writers disagree as to how precisely we are to draw the lines between
these concepts, they generally agree roughly as to the completeness of
this list. There also is no general agreement as to whether force or
violence should fall under the concept of power. Authors who answer
negatively draw on the intuition that to be a victim of a power exercise
seems to leave us with some - however small - possibility of choice
between compliance and non-compliance, a possibility absent in the case
of violence. 3 This is not to deny, of course, that violence and power can
be, and often are, tightly coupled; for instance there are cases where
violence figures as the ultimate threat by means of which the power
holder upholds his power. It is in part because of this link:, in part
because of a reliance on a broader definition of power as "the capacity
of some persons to produce intended and foreseen effects on others" that
some authors see force as one species of (social) power rather than as a
distinct category. 4
Disregarding the fact that different writers disagree over details, let
us briefly summarize, for later reference, one way of distinguishing
between forms of power. According to Lukes (1974: 32), coercion and

force involve open and observable conflicts of interest between power
holder and power subject. Manipulation also presupposes a conflict of
interests, but in its case the conflict of interests is not perceived by the
power subject. Authority is many-facetted in that we tend to attribute it
not only in cases where we observe or assume a conflict of interests but
also where we do not even assume any conflict of interests at all. In
general, we speak of authority where the power subject conceives - at
least in the long-run - of the power holder's demands as rationally
grounded and/or as beneficial to herself as the power subject. The
following picture summarizes these distinctions (Figure 12).
CONFLICT OF INTERESTS NO CONFLICT OF INTERESTS
observable latent
Coercion
Force
Manipulation
A u h 0
Figure 12
122 PART II
A FOUCAULDIAN DEFINmON OF POWER
With some of the more central notions and distinctions of more tradi-
tional study of power recalled, we can return to Foucault. Relying
mostly on his 1982 article "The Subject and Power", the following
'Foucauldian definition of power' can be formulated. This thirteen-part
definition can serve us as a first brief overview of our author's concep-
tion of power. More specifically, this definition allows me to substantiate
my claim concerning a concept of power found both in Foucault and
Lukes, and to highlight, in an anticipatory fashion, some central features
of Foucault's theory.
Between two individuals (or group of individuals), a and b, there

exists a relation ofpower (such that a exercises power over b) iff
between a and b there exists a relation R such that
(1) due to R, the actions of a reorder the set of possible actions of
b;
(2) a recognizes b as a person who makes decisions and acts;
(3) b consents to the existence of R, believing that a has a winning
strategy if she (b) does not comply to a's demands or expecta-
tions;
(4) b is free to contest the existence of R;
(5) R holds in place of an open conflict;
(6) R is interrelated with other relations between a and b, such
other relations being relations of communication, kinship,
exchange and production;
(7) the existence of R is dependent upon an objective of a, such as
the maintenance of privileges, the accumulation of profits, etc.;
(8) R is made possible by, and reinforces (possibly increases or
modifies), a system of differentiation (determined by law,
tradition, status, privilege, economic differences, linguistic or
cultural differences, competence) operating between a and b;
(9) the existence of R is dependent upon specific means at a's dis-
posal, means that bring the power relation into being, and
uphold it (threat of arms, effects of words, economic disparity,
means of control, surveillance, rules);
(10) R has a specific form or degree of institutionalization;
(11) R is rationalized to some degree (effective, certain in its re-

sults, related to, and productive of, knowledge);
(12) the existence of R allows for changes with respect to conditions
7-11, and it is related to further relations of power (R', R",
...) that hold between a and b (or between a and other indi-
viduals, or between b and other individuals);
(13) The identity of a and b is (at least in part) determined by R.
Especially theses (1) to (4), (8), (9), and (11) of the Foucauldian
definition parallel some of those ideas referred to in the beginning of the
present chapter.
As concerns (1), Foucault says that power "is a way in which
certain actions modify others ... " (S&P 219). To exercise power is "to
structure the possible field of actions of others" (S&P 221). Two points
are worth noting here. On the one hand, the stress on actions is meant to
distinguish power from violence, while, on the other hand, the notion of
structuring - or rather re-structuring - of actions points to a counterfac-
tual analysis of power.
Foucault draws the distinction between power and violence in a
pretty standard fashion. Violence or force leave the victim no choice
between compliance or non-compliance; as Foucault puts it, in cases of
violence the victim is treated as a body and not as a decision-maker, Le.
as a physical object rather than as a person whose actions are to be
influenced:
[power] ... is a mode of action which does not act directly and im-
mediately on others. Instead it acts upon their actions: an action
upon an action ... A relationship of violence acts upon a body or
upon things (S&P 220).
That Foucault accepts a counterfactual analysis of power is not only

hinted at in his notion of a "structuring of actions"; he also expresses
this idea more straightforwardly in the following passage: "[if e.g. a
man] can be induced to speak, when his ultimate recourse could have
been to hold his tongue ... [then his] freedom has been subjected to
power" (pPHC 83). In other words, speaking of power presupposes that,
were it not for the action of the power holder, the power subject would
124 PART II
or could have acted otherwise.

Condition (2), Le. the recognition of b by a as someone who acts
and makes decisions, is naturally seen as a corollary of this distinction
between violence and power. In the case of a power relation one of the
actors tries to get another to do something, tries to get another to act
(or, to abstain from acting). In Foucault's words, it is an "indispen-
sable" element of the power relation that "'the other' (...) be thoroughly
recognized and maintained to the very end as a person who acts" (S&P
220).
Ingredients (3) to (5) can be taken up jointly. Again, Foucault's
suggestions are not only well in line with mainstream sociology, but they
are also further elaborations of (1) and (2). Since in the case of power -
in counterdistinction to the case of violence - the victim is involved as
an actor and not just as a body, he or she remains able to choose be-
tween either compliance or non-compliance: "Power is exercised only
over free subjects, and only insofar as they are free ... slavery is not a
power relationship when a man is in chains" (S&P 221). The last point
is interesting, because in claiming that the slave is no longer free,
Foucault distances himself from Sartre who once suggested that the
radical freedom of the human being extends even to the slave in chains,
in so far as the latter still has to decide what his chains mean to him
(Sartre 1969: 483-84). Instead, Foucault proposes that a power relation-
ship demands the possibility of "refusal or revolt", even when the latter
might consist only in preferring death over submitting (pPHC 83). In
thus stressing that "resistance is never in a position of exteriority in
relation to power" (ibid.), Foucault comes of course close to restating
Max Weber's well-known suggestion according to which power is the
probability of individuals realizing their wills despite the resistance of
others. 5
Furthermore, Foucault puts much emphasis on the observation that
a power relation exists always in place of an open conflict (5) and that
the submission of the power victim is due to the latter's belief in her
lack of a winning strategy for such a conflict (3).
More precisely, Foucault sees the existence of a power relation as
an unstable equilibrium of conflicting interests in both the power holder
and the power subject. It is the inevitable instability of power relations
which justifies the idea that "at every moment the relationship of power
may become a confrontation between two adversaries" (S&P 226).

Looked at from the side of the power holder, the instability can be
expressed by saying that the very existence of the power relation is an
indication of limitations on her part. While, ideally, the power holder
would like to do away with the power subject's freedom altogether -
thus replacing the relation of power by one of violence - the very
existence of the power relation is the indication of her inability to obtain
this most favorable result. As long as the power relations holds, it is
always possible for the power subject to revolt and resist, i.e. to engage
in a struggle costly for both sides (even where the power holder is likely
to win). In other words, as long as the power relation obtains, the power
holder is forced to accept her victim as a decision-maker and actor. The
power subject, on the other hand, is faced with the alternatives either of
entering into a confrontation she feels will be costly - at least given the
price of the demanded action or non-action - or of yielding to an uncom-
fortable or unpleasant, perhaps humiliating, demand.
In "The Subject and Power", Foucault stresses with respect to
condition (6) that relations of power are interwoven with, but distinct
from, relations of communication between the power holder and the
power subject. He also suggests that relations of power between human
beings are to be distinguished from the 'power' of natural objects or
machines, or 'power to', in general. Like other authors before him,
Foucault refers to the latter relations as relations of "capacity":
Power relations, relationships of communication, objective capacities

should not ... be confused.... This is a question of three types of
relationships which in fact always overlap one another, support one
another reciprocally, and use each other mutually as means to an
end (S&P 217).
References to the fact that relations of power are interwoven with eco-
nomic, kinship and sexual relations can be found in other places of
Foucault's oeuvre (e.g. P/K 142). As we shall see below in greater
detail, it is Foucault's aim to defend the genuineness of power relations,
that is, he seeks to argue foremost that relations of power cannot be
reduced to economic relations.
As concerns (7), i.e. "the types of objectives pursued by those who
126 PART II
act upon the actions of others" (S&P 223), suffice it to say here, for the
purpose of a first overview, that even though Foucault thinks of power
relations as being due to the objectives of power holders, he does not
wish to maintain that all relations of power and all effects of power are
intended or foreseen. Indeed, the allowing for some kinds of invisible
hand or counterfinality phenomena is central to a Foucauldian theory of
power. Later we shall return to this point in greater detail.
Foucault's "system of differentiation" (8) and his "means of bring-
ing power relations into being" (9) (S&P 223) need little explanation at
this point. Obviously, these systems and means correspond fairly
straightforwardly to what the received view refers to as "bases" or
"means" of power.
Conditions (10) to (13) are of the utmost importance to Foucault's
conception of power. (10) captures Foucault's idea that relations of
power never exist in a social vacuum. Power relations not only presup-
pose institutions like e.g. the family, the school, the university, the
police, but they also uphold these institutions and make them possible in
the first place:
... the great strategies of power [Le. the systems of institutions]

encrust themselves and depend for their conditions of exercise on
the level of the micro-relations of power. But there are always also
movements in the opposite direction ... (p1K 199-2(0).
(11) is concerned with the essential relation between knowledge,

power and costs. The means by which a relation of power is upheld can
be more or less effective, Le. more or less costly to the power holder.
The latter will seek to keep the costs of control as low as possible by
constantly increasing the effectiveness of these means (S&P 223-24).
This is one of the several ways in which power and knowledge are
interconnected, ways that will later occupy our attention at some length.
(12) captures Foucault's constant reminder that power relations form
"networks" , that they "are superimposed, ... cross, ... sometimes cancel
one another out, sometimes reinforce one another" (p/K 99). Put diffe-
rently, "relations of power-knowledge ... are matrices of transforma-
tions" (HS 199).
Finally, (13) refers to the idea that our identity as individuals cannot
be separated from the power relations in which we live and act; it is in

and through power that "human beings are made subjects" (S&P 208).
10. THE GENEALOGICAL CONCEPTION OF POWER I:

FIELDS AND NETWORKS
MODELS OF POWER
In the last chapter I attended to parallels between Foucault and the

received analysis of power only insofar as the basic skeleton or core
assumptions were concerned. Further parallels can be drawn and will be
taken up below at least at those points where such allusions can serve to
illuminate or strengthen the Foucauldian position. In any case, even the
parallels referred to up to this point suffice to show that Stephen Lukes's
assessment of Foucault's theory of power as radically different from
Lukes's own tradition, more precisely, as "diffuse" and as unable to
allow for any links between the attribution of power and the attribution
of "causal or moral ... responsibility", is exaggerated at best (Lukes
1986a: 15).'
Turning from parallels to differences, be it noted first that Foucault
shows little interest in distinguishing between different "species" of
power. This is due to the fact that he is mainly concerned with mecha-
nisms of power exercises, a concern that, as we shall see, cuts across
various forms of power. This is not to say, however, that the reconstruc-
ted Foucauldian definition of a power relation cannot be related to the
distinction between various "species" of power, even though this attempt
calls for some refinements. While certainly most of the ingredients of the
Foucauldian definition hold for coercion and authority alike, conditions
(3) to (5) obviously cannot apply in the case of manipulation. This is
because manipulation, like force, does not allow for a decision over
compliance or non-compliance by the power subject. Nevertheless, there
can be no doubt that Foucault takes manipulation to be a form of power.
Perhaps the most natural way to resolve this problem is to think of the
above definition as defining the paradigmatic case of power, Le. coer-
cion, and as being extendable to cases like manipulation and force by
relaxing or dropping some of the thirteen conditions.
128 PART II
A second feature that sets Foucault apart from the better known
analyses of power is his 'models of power', Le. his use of various,
divergent, and in some respects even contradictory conceptualizations of
power. It is hard to decide to what degree the employment of such
models of power is the result of a systematic and deliberate choice or the
result of conceptual confusion and vagueness. Thus one cannot refute
Nancy Fraser, perhaps Foucault's most perceptive critic, when she
charges him with ending up "with a curious amalgam of amoral milita-
ristic description, Marxian jargon, and Kantian morality" (Fraser 1981:
284; cf. Fraser 1985). Actually, one can even add to Fraser's list insofar
as Foucault occasionally also uses the terminology of physics to charac-
terize aspects of power: he does not only call his analyses a "micro-
physics" of power, but his references to "forces" also often have over-
tones more of physical field theories of force rather than of military
action.
Yet even if one is dissatisfied with the specific manner and degree
of sophistication with which Foucault employs his models, and even if
one disagrees with his choices for models from the start, it still seems
possible and reasonable to defend the use of a plurality of metaphors,
analogies and vocabularies (and even conflicting ones) in the study of
power.
A parallel with developments in the philosophy of physics provides
support for such a position - and the fact that it does so already in part
justifies calling Foucault's conception of power a "microphysics". Nancy
Cartwright (1983) and Ian Hacking (1983) have stressed the importance
of the use of a variety of - often mutually inconsistent - models in
modern physical theories, e.g. in quantum mechanics. Physicists usually
work with several models for different purposes, models that often are
not deducible from their common general theory. Taking these and
related observations on the character of physical theory as their starting
point, Cartwright and Hacking deny that the most general laws of
physics can intelligibly be regarded as true. What can meaningfully be
regarded as true are rather the very local, simple laws that are bound to
a given model. Cartwright, with her tongue in her cheek, concludes that
God cannot possibly have the elegant, unifying mind of a French mathe-
matician, but that He necessarily must have the untidy mind of an
Englishman (1983: 19). Hacking suggests that this God is not very far
from the Leibnizian God who creates a maximum of phenomena govern-

ed by the greatest possible simplicity of laws: "the best way to maximize
phenomena and have the simplest laws is to have the laws inconsistent
with each other, each applying to this or that but none applying to all"
(1983: 219).
The way one can transfer these ideas to social science and thus to a
defense of Foucault is obvious enough. If social reality was not designed
by a French mathematician, then it just will not do to claim that either
game theory or Marxist notions or theories of communication provide
the one and only key to the study of power. To assume that social reality
- and the realm of power relations that is coextensive with it - is at least
as messy as the subject matter of physics, leaves us no choice but to
come to grips with its various, often conflicting features in terms of
various, often incompatible, models. This is admitted even by Stephen
Lukes, a writer on power who strives for simplicity and stringency in his
own theory. Lukes writes that "what unites the various views of power
is too thin and formal to provide a generally satisfying definition applica-
ble to all cases" (1986: 4-5).
To repeat, this defense of a plurality of models in a theory of power
does not imply a defense of the Foucauldian models. Let it be said,
though, that there certainly are no apriori arguments against his models
either. There might well be insights to be gained from looking upon the
social body as a Faradayan field of forces, thereby assuming forces -
alias powers - to be primary with respect to physical bodies - alias
individuals. And the model of war - e.g. in a Clausewitzian concep-
tualization - might be fruitful in coming to grips with certain features of
the political or scientific struggle.
It is true that these models are at least partially incompatible with
the standard conceptualization of power in terms of agency and responsi-
bility, a conceptualization that as we saw above also appears in Fou-
cault's own definition of power. Yet Foucauldians have at least two
possible ways of answering this observation. On the one hand, they
might hold that even though all power is analytically linked to agency,
we might still gain new insights into power by bracketing this link at
least occasionally. On the other hand, they might argue that although
there are forms of power where the link between power and agency
holds, there are still other forms, effects and structures of power that
130 PART II
have little or nothing to do with either agency or responsibility. On this

latter view one thus has a double reason for employing agency-excluding
models: for illuminating agency-involving power structures from a new
angle, and for' conceptualizing phenomena of power where agency has no
role to play.
Foucault's use of these various tools becomes more transparent as
we tum to key theses of his theory of power, theses that constitute the
basis of his conception of power. These theses concern the nature of a
power relation, the omnipresence of power, and the distinction between
micro- and macropower. Discussing these theses will pave the way
towards a later defense of Foucault's view on the relation between social
power and scientific knowledge.
INTERNAL RELATIONS
To start with the nature of the relation of power, note that Foucault
stresses time and again that we should not look upon power as being
possessed by some individuals or groups but as being exercised by some
individuals over other individuals:
Power is not something that is acquired, seized, or shared, some-

thing that one holds on to or allows to slip away; power is exercised
(HS 194).
[...] power is neither given, nor exchanged, nor recovered, but
rather exercised, and [...] it only exists in action (p/K 89).
We only need to add to these quotations Foucault's further remark

to the effect that a relation of power must be taken as "strictly rela-
tional" (HS I 95) to see how naturally Foucault's claim is rephrased in
terms of the classical Leibnizian distinction between "external relations"
(of comparison) and "internal relations" (of interaction). 7 A power
relation, for Foucault, is an internal relation of interaction, not an
external relation of comparison.8
Recall that for Leibniz the distinction in question is one that is based
on the criterion of the reducibility of relations to one-place predicates of
the relata. For example, "Titius is wiser than Caius" can be reduced to
the two sentences "Caius is somewhat wise" and "Titius is very wise"
by relying on the general truths according to which "wiser" equals
"superior in point of wisdom"; and "'very' represents a degree superior
to 'somewhat'''. Because of the possibility of this reduction, "Titius is
wiser than Caius" qualifies as an external relation of comparison (relatio
comparationis). However, "Paris loves Helen" cannot be reduced in this
way. This is because in this sentence Paris does not only have the
feature of "Being a lover", but also the feature of "Being a lover in
virtue of - or insofar as, or since (propter or eo ipso) - Helen being the
beloved". In other words, a reduction to one-place predicates is not
possible in this case, since "Being the lover of someone" is still a
relational property. Thus "Paris loves Helen" is a relation of connection
(relatio connexionis), Le. an internal relation of interaction.
Attributing to Foucault the conception of power relations as internal
enables us to explain his emphasis on "mechanisms" of power, his pro-
nounced "anti-economism" and his frequent reference to metaphors of
war.'
Because Foucault is above all interested in power relations as forms
of interaction, Le. because he centers attention on the exercise of power,
he focusses less on the relata of that relation than on the relano itself.
That is to say, he concerns himself with the mechanisms (of control,
coercion, manipulation, etc.) that figure in power relations:
History has studied those who held power [...]; contrasted with this
there has been the history of economic processes and infrastructures.
[••.J But power in [...] its mechanisms has never been studied (p1K
51).
Thus Foucault does not confine himself to conceptualizing the power

relation (R) as an act issuing from the power exerciser (a) to the power
subject (b):
R
a-b
but he rather proposes that the power relation be conceptualized
'inside-out'. In other words, the mechanisms necessarily involved in
132 PART II
relations of power are taken as determining a and b, i.e. as shaping a

and b in different ways. This implies that we must attend not only to a
and b, but also to their respective relations (R', R") to the mechanism
(m) involved in R:
R' R"
a-m-b
R(a,b)
At this point I shall not go into the details as to what kinds of

mechanisms Foucault has in mind here. io Be it mentioned here only
that Discipline and Punish deals, for example, with what one might call
mechanisms of "ecological control"11, that is, mechanisms for limiting
and pre-structuring the spatial environment, temporal limits, and social
entourage of the actions of individuals. Thus in the prison or the work-
place, individuals are confined - inter alia by architectural measures -
to limited spaces, trained to carry out actions in specific and meticulous-
ly planned time intervals, and placed in ways so as to be under constant
surveillance. Going beyond Foucault's own interests we might of course
also think of even more down-to-earth mechanisms, e.g. the power
mechanisms involved in conversation, such as interrupting or ques-
tioning, mechanisms much attended to in recent linguistic pragmatics.u
For Foucault, the choice between a conception of power as posses-
sion and a conception of power as interaction is closely connected to the
alternative of studying power via economic or via military models.
Indeed, Foucault attacks what he regards as the "economism of the
theory of power" (p/K 88). Foucault suggests that both classical political
theory and Marxism are representatives of an economism in the theory
of power.
In the classical political theory, economism is closely tied to the
traditional preoccupation of power theorists with questions of right, law
and legitimization. This link, Foucault holds (p/K 88-90), can be seen
especially clearly in the principal idea of contractualism according to
which presocietal individuals 'trade in' their powers for protection
provided by the state. Power is thus conceptualized as a commodity. On
the premisses of this model, i.e. "the juridical" or "liberal" model of
power, critical questions concerning, and analysis of, power, remain
confined to a reflecting on the conditions of the initial or ideal contract,

that is, confined to ever new justifications of either prevailing or ima-
gined bodies of law in terms of the initial trading-in of pre-societal,
individual powers. It is one of Foucault's most important claims in his
historical work that this model of power with its inevitable conceptual
link between power and law is impotent when applied to the study of
modem mechanisms of power. This is because these mechanisms - that
derive from medicine, the prison, and the army - operate not on the axis
of just/unjust or legitimate/illegitimate, but rather on the axis of nor-
mal/abnormal or healthy/sick.
In Marxism the economism figures somewhat differently. Even
though power is here taken as an internal relation and is not conceptual-
ized as a commodity, its interactional aspect remains underdeveloped
since Marxists are preoccupied with the allegedly "ultimate" causes of
power in economic structures rather than with the actual effects of power
mechanisms upon humans:
On this view, then, the historical raison d'etre of political power is

to be found in the economy.... Is it [Le. power) always in the
service of ... the economy? Or, on the contrary, do we need to
employ varying tools in its analysis ...? ... If that is the case, it is
not the models of functional subordination or formal isomorphism
that will characterize the interconnection between politics and the
economy (P/K 89).
Foucault proposes "Nietzsche's hypothesis" as the most promising

successor to an economically-minded study of power. According to
"Nietzsche's hypothesis", "the relationship of power lies in the hostile
engagement of forces" and is to be analyzed "primarily in terms of
struggle, conflict, and war" (P/K 91).
"Nietzsche's hypothesis" clearly brings out the interactional aspect
of power. Recall only that for Clausewitz in his classic Yom Kriege, war
is "an extended duel", "always the clashing of two living forces", or
"an act of violence in order to force the opponent to fulfill our wishes"
(Clausewitz 1918: 3,6). Foucault goes beyond Clausewitz, however, by
inverting the latter's notorious dictum that war is but the continuation of
politics by other means (P/K 90-92). To put the point more sharply than
134 PART II
Foucault does himself, whereas Clausewitz claimed that politics is the

underlying deep structure, or "logic", of the surface structures or
"grammars" of both diplomacy and warfare (1918: 641), Foucault holds
that it is the raison de la guerre that underlies peacetime politics as
much as military clashes: "Even when one writes the history of peace ...
it is always the history of [...] war that one is writing" (p/K 91).
In fact Foucault claims that even Clausewitz's own dictum was
already the inversion of an earlier principle of political thought. Interest-
ingly enough, Foucault claims that the proponents of that earlier prin-
ciple of politics as the continuation of war by other means 13 all mark
a countermovement to the long and still dominant liberal conception of
power. For them, war is never over, "law emerges from burnt cities and
devastated countries", society is always divided in two army-like sec-
tions, and all theoretical discourse is but part of this ongoing battle (LK
11_13).14
POWER AND PERSONAL IDENTITY
The fact that Foucault focusses on the question of what power relations
or mechanisms make of us, rather than on why individuals coerce or
manipulate one another, implies the possibility of a second way in which
the distinction between internal and external relations can be brought to
bear on an attempt to reconstruct his theory of power. The crucial point
of this second application is that the classical dichotomy of internal
versus external relations was usually connected to the opposition between
essential and accidental properties (Elster 1978: 22). Internal relations
were generally regarded as essential, and external relations were con-
ceived of as accidental; in other words, internal relations but not external
relations were treated as constitutive of what their relata are.
Applying this dichotomy between internal-essential and external-
accidental relations to relations of power, the second of Foucault's key
theses can be formulated as stating that relations ofpower are internal-
essential rather than external-accidental. That is to say, insofar as power
relations make us what we are, and insofar as relations of power are
constitutive of us as subjects or individuals, we cannot define ourselves,
nor understand who we are, without understanding the mechanisms of
power that shape us, without appreciating and analyzing our ability or
tendency to take up specific roles within power relations.
In line with this reconstruction, Foucault calls for the study of how
"peripheral subjects" are constituted "as a result of the effects of power"
(NGH 98). And in another passage we read:
[...] we should try to discover how it is that subjects are gradually,

progressively, really and materially constituted through a multipli-
city of organisms, forces [Le. powers], energies, materials, desires,
thoughts, etc. (p/K 97).
In still other terms, the central question is "how are we constituted as

subjects who exercise or submit to power relations?" (ENL 48).
In his historical studies on the prison, the asylum and the production
of the modern conceptions of health and sexuality, Foucault provides
plenty of material for deal ing with this question in terms of historical
analysis. Here, however, I shall confine myself to some comments from
a more topical perspective.
To start us off, note that Foucault's remark, according to which his
study of power is to be a "microphysics of power" (DP 29), can - to
some extent at least - be justified at this point. After all, modern physics
(microphysics or quantum mechanics and the theory of relativity) differs
from classical Newtonian physics most impressively in its abandoning of
the substance vs. attribute scheme, in its giving predominance to rela-
tions and structures rather than objects, and in being a science of effects
rather than a science of objects. For instance, whereas in the 19th
century it was still thought that science starts from "real" objects and
subsequently proceeds to conjectures as to their relations, microphysics
gives priority to relations and treats "real" everyday objects as mere
phenomena. The development of the theory of relativity has produced a
way of thinking that conceives of relations as independent of their
termini, and that conceptualizes objects "as strange functions of the
function that relates them" (Bachelard 1971: 34). The particle in mic-
rophysics is no longer a classically conceived "extremely small body";
rather, the particle has been stripped of almost all of those determina-
tions that classical physics attributed to physical bodies. The protons and
electrons of modern physics are not material objects with the property of
136 PART II
electricity because we know of electricity only as electrons and protons:

substance and attribute can no longer be kept apart. Moreover, the
particle in quantum mechanics has neither a definite absolute dimension,
nor a precise location. One speaks no more about its size but about its
"zone of influence" or "its power of opposition", Le. the distance one
electron can approach another when fired at it with a speed close to that
of light (ibid., 62).
Perhaps it was these characterizations of modem physics, characteri-
zations that loom large in Bachelard's epistemological writings, that
prompted Foucault to speak of his investigations as a "microphysics of
power". After all, Foucault, too, argues for the priority of relations (of
power) over individuals: he denies that power can ever be "localized
here or there" (p/K 98), and he tries to show that the individual is the
result of forces or powers, powers that - as we shall see later - cannot
be modelled upon the power relations between larger social, "macro-
physical", units.
Obviously notions of physical theory that resemble Foucault's
conception of power predate the twentieth century. Foucault's own talk
of "fields of forces" also invites reference to Faraday's conception of
fields of force with its revolutionary idea of forces as being primary with
respect to bodies}S Replace ~bodies' with 'subjects' and the analogy
works indeed. Furthermore, Faraday also sounds Foucauldian, and
Foucault Faradayan, because of Faraday's view that the amount and the
direction of force associated with each point in the field of force is
dependent upon, and determined by, the amount and the direction of
force in contiguous points. Foucault conceives similarly of the social
field of society. Not only are the power or force relations essential to the
points of the field, Le. the individuals, but he also holds that none of
them can be described and explained without taking into account its
power or force relations to others. Even Faraday's novel notion of
"action in the vicinity" or "near-action"l6 has a parallel in Foucault's
preoccupation with ecological mechanisms of power.
Turning from microphysics to objections, the major threat for the
conception of power relations as internal-essential might be seen to arise
from the observation that if all power relations in which we have been,
and in which we live now, are essential to our identity, then the hope of
us ever being free of those power relations must inevitably be vain: what
is essential to our identity cannot be made to disappear without that

identity disappearing. too.
It certainly makes no sense to claim that all power relation tokens -
relations in which we figure as relata - are essential to what we are. The
fact that we let ourselves be coerced into drinking a second beer by a
friend who uses some trivial. minor threat. can hardly as such be part of
what we are. Yet while it is unnatural to say that all of our power
relations are essential to our identity. it is still plausible to assume that
more lasting power relations. as well as our tendency and/or ability to
enter certain types of power relations as. say. coercer or coerced. do
have an intimate relation to what we are. In other words. some of our
more lasting power relations and tendencies of reacting to power rela-
tions are "productive": they shape us into what we are.
For instance. the power relations and mechanisms that we have
established with our students. spouses. long-term friends and colleagues.
parents and teachers. seem to be likely candidates for having to be
included into the rough definition of what we are. for having to be
included into the set of determinants of how we understand ourselves.
The case of tendencies and abilities is no less clear. Our ability to
act as. say. coercers. or our tendency to figure as likely victims of coer-
cion. does not stand isolated among our personality traits. Obviously.
since at least some of our more permanent traits are definitive of what
we are. and since personality traits predisposing us to being either good
or bad coercers can hardly be excluded from these central traits. our
tendency and/or ability to act as power exerciser or power subject is an
essential part of our identity.
It is true that. in some sense. the thesis of power relations as inter-
nal-essential. forces us to say that we do not remain what we are (or
were) when essential power relations. or tendencies to take up specific
positions in them. change. Yet rather than being a threat to Foucault's
view of power relations as essential. this implication might also be seen
as harmonizing with recent suggestions concerning the relativity of
personal identity!7 We all change over time. and over longer periods
of time small-scale changes add up to bigger ones. so much so that it is
reasonable to say that different time-slices of "our lives" are like dif-
ferent persons. If this idea is accepted nothing seems more natural than
to extend it explicitly to power relations. too: when ridding ourselves of
138 PART II
long-term and more fundamental power relations, or when becoming

entrapped in them, and when either developing, or freeing ourselves
from, tendencies to take up specific positions in specific types of power
relations, we become different, new persons.
NElWORKS OF POWER
As already indicated above, Faraday's idea that the force associated with
each point influences, and is influenced by, the forces associated with its
adjacent points, has a counterpart in Foucault. This counterpart is
Foucault's claim that every individual both exercises power and is the
target of the exercise of power by others. Just as for Faraday in the field
of force there is no point without force, and thus no point without rela-
tions to other points, so for .Foucault in the social field there is no
individual without his or her relations of power to some other individ-
uals. This claim can be regarded as the third key assumption of Fou-
cault's theory of power: Power is an omnipresent network in the social
field.
Foucault holds that "power is everywhere [...] because it comes
from everywhere" (p1K 93) and he also writes that ...
between every point of a social body, between a man and a woman,

between the members of a family, between a master and his pupil,
between every one who knows and every one who does not, there
exist relations of power ... (p1K 187).
Reminiscent of Faraday's notion of "networks of lines of force" and

his idea that forces can travel through a field without being the attribute
or possession of a body, Foucault also suggests that "power is employed
and exercised through a net-like organization", that "power circulates ...
in the form of a chain", and that individuals "circulate between ... [the]
threads [of power]; they are always in the position of simultaneously
undergoing and exercising this power" (p1K 98).
For Faraday the omnipresence of force does not of course imply that
force is evenly distributed over all points. Even though force is every-
where, Faraday allows for "clusters", "intensities" or "lines" of force,
FOUCAULDlAN GENEALOGY 139
Le. concentrations of force in various parts of the field. This idea can
also be expressed by saying that the density of lines of force is different
for different points in the field. 18 In an analogous way, Foucault pro-
poses that we should not take power to be "the best distributed thing in
the world, although in some sense that is indeed so" (p/K 99). While
power relations permeate the whole social body of society, power rela-
tions can be more dense in some regions and less dense in others.
Foucault comes close to Faraday's lines of force even terminologically
when speaking of control mechanisms as "lines of penetrations" (HS I
42).
While these parallels suggest some ways in which the discourse of
field theories of physics might provide heuristically useful concepts for
the study of power along Foucauldian lines, they do only limited work
in defending Foucault's omnipresence thesis against standard criticism.
All they achieve in this respect is to show that omnipresence does not
imply equal density, and that omnipresence need not preclude the
possibility of resistance. III To see this, we only need to remember that
Faraday's theory does not include the assumption that all vectors of
force have to point in the same direction.
I shall not risk pushing the parallel too far, however. Instead, I shall
make some topical remarks on the thesis of power as an omnipresent
network. More precisely I shall (1) comment on the question of whether
this amounts to a mystification of power, (2) take up the issue of wheth-
er this thesis is trivial, and (3) discuss the concepts linked to this thesis,
Le. the notions of chain and network.
(1) The first question that deserves attention is whether talk of power as
an omnipresent network does not render the category of power useless
as an analytical tool in the social sciences. Various writers - long before
Foucault formulated his own views - have voiced the suspicion that the
notion of power is of little use in our attempts to come to grips with
social and political life. These authors claim that explaining social
phenomena in terms of power "light[s] up very little of everything ... we
see a little of the side, but the face remains obscure" (Minogue 1959:
283). Things seem to get even worse when the concept of power is
enlarged so as to have as its extension the whole of society. It has been
suggested that in this all-embracingness, the notion of power is no longer
140 PART II
opposed to anything, that when taken in this way power, in fact, col-
lapses into the practice of the domain it was meant to illuminate (fumer
1989: 533-34). What remains after this radical expansion is a concept
that explains everything and nothing, in other words, a concept with a
mystical ring.
Admittedly, many more programmatic statements of Foucault -
when taken out of their context of historical studies - have something of
a mystifying ring. For instance, Foucault speaks as if power itself was
an actor ("Power can retreat here ... invest itself elsewhere ..... (p1K
56», says that "power is never localized here and there" and writes that
power has no origin (p1K 199). Actually it is precisely the latter claim
that leads him to reject the idea of developing any "theory" of power:
"If one tries to erect a theory of power one will always be obliged to
view it as emerging at a given place and time and hence to deduce it ... "
(ibid.)
Yet despite the implausibility of the latter position, it seems to me
that Foucault cannot be charged with collapsing the notion of power into
the notion of society. Even though ultimately only the whole of this
study can provide the evidence needed, be it said, nevertheless, already
here, that Foucault's detailed definition of power, his distinguishing
between relations of power, on the one hand, and relations of knowl-
edge, exchange, communication, production and kinship, on the other
hand, his more or less detailed descriptions of different forms, technolo-
gies, origins, structures and networks of power mechanisms, and finally
his attention to the opposition between vicious and non-vicious forms of
power, set him worlds apart from mystifications of power. After all,
Foucault does provide us with a conceptual framework that can be - and
has been - applied to detailed historical analyses.
(2) The Foucauldian version of the idea of power relations as penetrating

all through the social sphere might, from a different perspective, be
accused of triviality rather than mystification. After all, everyone knows
from their own experience that power relations can be identified all
around and between almost any interacting individuals.
Now, evaluating a thesis as to its degree of triviality is possible only
by placing it against the larger backgrounds first of the historical mo-
ment when it was advanced, and second of the overall theory to which
it belongs. Applying this to Foucault's omnipresence thesis, one cannot

but conclude that in the French context of the 1970's, a time when struc-
turalist Marxism loomed large, it was not trivial to advance the notion of
the omnipresence of power. In this period, the very term 'power' had
been reserved for the abstract level of class relations and structures, and
the Althusserian Nicos Poulantzas distinguished even conceptually
between "pouvoir" as the higher-level and genuine form of power of one
class over another, and "puissance" as the uninteresting and politically
irrelevant omnipresent relations of power between individuals (poulan-
tzas 1978a: 106). To state the omnipresence of pouvoir in this context
meant proposing that higher-level and lower-level powers are interrelated
enough to be referred to by the same concept. It was making a point, not
stating the obvious.20
(3) Finally, we need to examine Foucault's way of formulating the thesis

of the omnipresence of power in terms of chains and networks.
The fact that Foucault speaks of power relations as forming chains
or nets should not come as a surprise given his emphasis on relations
rather than attributes in the study and conceptualization of power. After
all, a network is but a set of relations holding between a set of points,
objects or individuals. A Foucauldian power network, however, has
some specific features that are worth mentioning. Since Foucault regards
power relations as internal-essential it is obvious that the individuals over
which the network is defined cannot be thought of as existing apart from
the network. In other words, the structural position of an individual, Le.
the sum of her (more basic) power relations to others, as well as the set
of positions she can take up within the network are (at least in part)
constitutive of the individual's identity.
It seems that a number of Foucault's suggestions concerning power
and knowledge can perhaps be made more transparent by formulating
them explicitly in the idiom of sociological network analysis. 21 For
instance, recall that Foucault speaks of power relations as being inter-
related, as being transformable into further new power relations by
"matrices of transformations", and that he allows for power relations to
"cross, .. sometimes cancel one another out, sometimes reinforce one
another". Translated into the idiom of network theo:y which represents
networks as directed graphs with nodes tied to one another by labelled
142 PART II
arrows, these ideas amount to ~aying that a given network of power ties
can be enlarged, reduced or transformed, where the latter possibility
implies the introduction of new types of (and thus new labels of) power
ties (or arrows).
Network concepts can perhaps also highlight Foucault's overall
interests as well as illuminate his notoriously obscure statement that
power "is the name that one attributes to a complex strategical situation
in a particular society" (HS I 93). According to this suggestion, to
describe power in a given society is not only to depict the different types
of networks of power relations and mechanisms existing in the society,
but also to portray the (types of) interrelation between various smaller
networks, to study the stability and instability of networks as well as the
causes for the stability or instability in question, and to sketch the ways
in which what might be called "the identity tie" (i.e. the power/control
of an individual over herself) is determined by its position in, or changes
in, the network. A Foucauldian study of these networks will, first, label
the ties according to the mechanisms involved, second, delineate how the
introduction of new types of ties is possible, and third, show how the
introduction of new types of ties changes already existing networks.
Furthermore, a Foucauldian study will also investigate the relations
between networks of social power and networks of knowledge, study
whether and how these networks are interlocked, mutually supportive or
isomorphic. Finally it will also address the question as to which macro-
structures are latent and implicit in the myriads of microlevel relations.
Thus genealogy performs a "reduction" of the multitude of individuals
and relations to equivalent classes; that is to say, it distributes individuals
into classes according to whether these individuals occupy an equivalent
or similar position in the network(s), and it distributes relation tokens
into classes according to their similarity and equivalence. In engaging in
this operation, this investigation of course performs an operation that is
also carried out by individuals and groups in society itself.
To be sure, this kind of investigation will not be able to answer in
any absolute sense the question as to where power originates from, since
its starting point will always be already existing networks of power. But
it will address the question as to how these networks change, how they
condition one another and how new types of networks emerge from
earlier ones.
NETWORKS OF POWER AND INVISIBLE HANDS
In order to further appreciate the fact that Foucault's own advocacy of

the omnipresence of power thesis leads to fruitful and interesting impli-
cations, we need to turn to his fourth thesis with respect to power:
Microlevel power relations and mechanisms are different, and thus to be
distinguished, from macrolevel power relations between classes, on the
one hand, and macrolevel power mechanisms, i.e. (systems oj) coercive
institutions, on the other hand. The emergence, persistence and effects of
macrolevel relations and mechanisms are often unintended and unfore-
seen.
Two distinctions appear in the first sentence of this thesis, the
distinction between a macro- and a microlevel, and the distinction
between relations and mechanisms. Neither of these distinctions have
been fully attended to in the literature, and regrettably Foucault does not
bring out clearly their importance to his enterprise.
Foucault speaks of "infinitesimal mechanisms" of power that be-
come "invested, colonized, utilized, involuted, transformed, displaced,
extended, etc. by ever more general mechanisms and by forces of global
domination" (p1K 99). He also refers to these mechanisms as "micro-
mechanisms" of power. "The more general mechanisms" of the last
quotation are but the coercive institutions of the prison, the asylum, the
clinic and the boarding school, institutions whose historical emergence
and development Foucault investigates in Madness and Civilization and
Discipline and Punish. Foucault calls these institutions and the systems
they form "blocks", "apparatuses" or "dispositifs" (p/K 197). The last,
not straightforwardly translatable word has several meanings, for in-
stance, "an ensemble of parts that constitute an apparatus or machine",
"an ensemble of measures that constitute an organization or plan", and
"an ensemble of (material) precautions and means for carrying out a
strategic, military operation"22. The last-mentioned meaning harmo-
nizes especially well with Foucault's liking for military terminology.
Power mechanisms do not have a life of their own. Rather they are
employed within micro- or macrorelations of power. The latter distinc-
tion is hinted at in the following methodological advice:
144 PART II
We need to identify the agents responsible [...] those which consti-

tuted the immediate social entourage, (the family, parents, doctors,
etc.) and not be content to lump them under the formula of a gene-
ralized bourgeoisie (p1K 101).
To be sure, Foucault does not rule out the possibility of speaking of

the bourgeoisie or the proletariat as classes: he writes, for instance, that
"the bourgeoisie was interested in power, not in madness" (p1K 102).
However, he holds that the attention given to the macrorelations between
classes must not lead us to disregard the microrelations with respect to
which macrorelations are but abstractions or reductions. Thus the oppo-
sition Foucault focusses on is one between microlevel power relations
between singular agents, or small groups, and macrolevel power rela-
tions between classes.
The distinction between the micro- and the macrolevel also surfaces
in Foucault's writings as the opposition between "tactics" and "strate-
gies". For Clausewitz tactics was "the theory of how to use forces in a
battle", and strategy "the theory of how to use battles for the purposes
of war" (1918: 69). In the same vain, Foucault refers to tactics as the
local employment of "infmitesimal mechanisms" in exercises of power
between individuals, and to strategy as the combining, using, and
influencing of these local relations of power for the purposes of social
war. The interdependence of these two levels is suggested in the follow-
ing passage:
... the great strategies of power encrust themselves and depend for
their conditions of exercise on the level of the micro-relations of
power. But there are always also movements in the opposite direc-
tion, whereby strategies which co-ordinate relations of power pro-
duce new effects and advance into hitherto unaffected domains (p/K
199-200).
Before going further it might be helpful to summarize the distinc-

tions just outlined in the form of a table (Figure 13).
relations mechanisms
between
(1) (2)
macrolevel classes dispositifs

(strategy)
(3) (4)
microlevel individuals, infinitesimal
(tactics) social
entourage
Figure 13
Foucault's central interest is with the birth and growth of (2). Thus
the question is: how is this birth and growth to be explained? Foucault
reproaches Marxism for trying to explain too much in terms of relations
between classes or the mode of production. For instance, he opposes
attempts to explain the prison or the asylum in terms of bourgeois inter-
ests. The bourgeoisie and its objectives, or the emergence of coercive
institutions, cannot be defined and understood without engaging in an
"ascending analysis" from the microlevel up (p1K 99). Put differently,
to reduce the multiplicity and variety of power networks to the opposi-
tion between two social classes is to bracket too many of precisely those
(types of) power ties that are needed to account for the emergence and
persistance of coercive institutions.
In line with this criticism Foucault's own historical analyses do not
seek to explain (2) in terms of (1), but rather to explain (2) via (3) and
(4), and to explain (3) in terms of (2) and (4). It would seem that
Foucault suggests a twofold explanation for the existence of dispositifs,
although it is hard to show that he has ever in fact presented such an
explanatory account in any detail. What a historical account of the
existence of a dispositif calls for is, first, a description of the structure,
profile or composition of the coercive institution, and second, an expla-
nation of the conditions of its possibility.
146 PART II
To tackle the first point is to move from (4) to (2). It is - as already

mentioned - to study how "infinitesimal mechanisms" of power are, or
have been, "invested, colonized, utilized, involuted, transformed,
displaced, extended, etc." by and within systems of dispositifs. One
investigates dispositifs both as social laboratories, and as the results of
this laboratory work: dispositifs are spaces where experimentation with
small, closed power networks, or social "microworlds" is possible,
spaces where new technologies of more efficient, more extensive, more
diverse forms of control, manipulation, authority, coercion, punishment
and deterrence are developed.
Suffice it here to mention only Foucault's favorite example, "pan-
opticism" as a "technological invention in the order of power, compa-
rable with the steam engine in 'the order of production" (p/K 112). This
example can also serve as an illustration of the more basic nature of
Foucauldian power mechanisms as compared with the traditional "spe-
cies" of power. As is easy to appreciate, the panopticon can be used to
strengthen or create relations of coercion, authority, etc. It can also be
used, say, to manipulate and coerce, both at the same time.
Developed and applied in prisons, hospitals, armies and factories,
the panopticon is a type of power network in which one individual, or a
few, can keep a much larger group under constant surveillance. Mem-
bers of the latter are isolated from one another yet are at all times visible
to the controller. (For instance, in the prison, cells are arranged in a
circle around a central tower, the cell side facing the tower having bars
rather than a wall.) The effectiveness of this technology - whose basic
structure can be adopted and modified for the needs and resources of dif-
ferent environments - does not simply lie in the extent of visibility it
provides or in the small number of individuals needed for control. Its
most intriguing feature is the effect the constant surveillance has upon
the controlled: they can be expected to internalize the controlling gaze
and thus to control themselves. And what is more, this technology is
truly 'closed' in that even the individual at the controlling end is herself
controlled, too: if she neglects her duty and the network breaks down,
she will be the first to suffer for it. As Foucault puts it:
This indeed is the diabolical aspect of the idea and all the applica-
tions of it. One doesn't have here a power which is wholly in the
hand of one person who can exercise it alone and totally over the
others. It's a machine in which everyone is caught, those who
exercise power just as much as those over whom it is exercised (p/K
156).
To address the second issue, the conditions of the possibility of a

coercive institution, is again a double task. First, it is to identify "the
immediate social entourage", that is, the (groups of) agents who con-
ceive of a given dispositif as enhancing the scope, intensity, amount and
efficiency of their own exercise of power over others. In other words,
these actors take the dispositif in question as changing prevailing net-
works of power in their favor, or as upholding an existing network of
power to their advantage, e.g. by providing these individuals with new,
more efficient mechanisms of power. (Step from (4) to (2).) Perhaps
Foucault's most convincing attempt regarding this task is his editing -
with Arlette Farge - of a selection of the famous Lettres de cachet
(1982), letters that in the seventeenth century could be sent directly to
the King asking for the imprisonment of one's family members, neigh-
bors, or employees. As the documents bring out impressively, people
from all social strata eagerly seized upon this possibility to enhance their
authority within their family, their neighborhood and their workplace.
Thus already existing power relations led people to "willingly invest
themselves"23 into this new technology of control and authority.
Second, to explain the conditions of the possibility of a coercive
institution is also to explain how individuals or groups become consti-
tuted as either inmates and victims, or as bottom level officials. (Here
we thus move from (2) and (4) to (3).) This question is of course natural
to ask given Foucault's thesis of power relations as internal-essential. In
his historical studies Foucault is more concerned with the inmates than
with officials, even though he mentions that the problem of training
officers, teachers and foremen became an acute one by the eighteenth
century and that the State long lagged behind the Church in acquiring the
lower cadres needed for control and supervision (p/K 157). As concerns
the issue of inmates, Foucault suggests, for instance, that the prison
functions in a self-defeating - or self-fulfilling - way: the networks of
power mechanisms within this dispositif constitute precisely those kinds
of personalities and identities that the prison is allegedly meant to re-
148 PART II
shape.
To complete the discussion of the interrelations between relations
and mechanisms, between the macro- and the microlevel, let us note that
Foucault is aware of the possibility that the emergence of (2) can be
unforeseen on the level of (3), Le. unforeseen by the individual actors
involved. 24 Foucault indicates his position in the following passage:
... the rationality of power is characterized by tactics that are often

quite explicit at the restricted level where they are inscribed (the
local cynicism of power), tactics which, becoming connected to one
another, attracting and propagating one another, but finding their
base of support and their condition elsewhere, end up forming
comprehensive systems: the logic is perfectly clear, the aim deci-
pherable, and yet it is often the case that no one is there to have
invented them ... an implicit characteristic of the great anonymous,
almost unspoken strategies ... (HS I 95).
Even though one might want to question whether the use of the
notion of "strategy" is a happy terminological choice here, the point
itself should not come as a surprise. After all I have already mentioned
Foucault's claim that the power networks within society change constant-
ly, and that the introduction of new elements at one point in the network
lead to re-arrangements throughout the whole network.
Unfortunately, over and above occasional hints like the quoted
passage, Foucault has not provided us with a theoretical account of
invisible hand or counterfinality explanations. Neither do his historical
studies include unambiguous cases of such explanations from which one
could infer his conception. This lacuna is unfortunate, particularly since
Foucault's vision of modern society as "panoptical" - Le. as a society
where everyone is controlled by someone, without there being someone
who controls everyone - obviously stands in need of an invisible hand
account. 25 It seems clear that any rational utility maximizer (or group
thereof) who can design such a society could also design a society where
only he, she or they is or are in command.
As Foucault's emphasis upon "ascending analysis" already suggests,
we may surmise that had Foucault developed his conception more fully,
he would have come to speak out in favor of aggregate rather than
functional mold invisible hand explanations. Whereas aggregate mold

explanations take intentions of actors as their input in order to yield, via
an assumed complex process, an unforeseen social phenomenon as their
output, the functional mold version explains the raison d'2tre of a social
phenomenon in terms of its function within a social system. U Further
evidence for this suggestion can be marshalled from an interview where
Foucault sides with methodological individualism by assenting to the
claim that "there are only ever transitory coalitions, some of which
immediately break up, ... others of which persist, but ... strictly speak-
ing individuals ... [are] the first and last components" (p/K 208).
Furthermore, in the same interview, Foucault fiercly denies wishing to
model his explanations of dispositifs and their effects upon biological
concepts like organism; thus he rejects precisely the notions upon which
functional mold invisible hand explanations are modelled:
... I have never used the metaphor of the organism. ... When I
speak of strategy, I am taking the term seriously: in order for a
certain relation of forces not only to maintain itself, but to accentu-
ate, stabilize and broaden itself, a certain kind of manoeuvre is
necessary. The psychiatrist had to manoeuvre in order to make
himself recognised as part of the public hygiene system. This isn't
an organism, any more than in the case of the magistrature, and I
can't see how what I'm saying can imply that these are organisms
(P/K 206).
11. THE GENEALOGICAL CONCEPTION OF POWER II:

SOCIAL POWER AND SCIENTIFIC KNOWLEDGE
THE PRODUCTIVITY OF POWER
Little has been said so far about the two topics that figure most centrally
in Foucault's genealogy, to wit, the productivity of power, and the
relation between power and knowledge.
Foucault never tires of reminding us that not all exercises of power
have merely the effect of restraining, limiting and restricting. In his
words, the effects of power are not in all, and perhaps not even in most,
150 PART II
cases, "negative". Rather, exercises of power, power mechanisms and

power networks, can also have productive or "positive" effects such as
the power subject's acquiring new skills, her adopting new beliefs, or
her being instrumental as an object of experimentation (and thus her
being involved in the growth of knowledge).
While being one of Foucault's most often cited genealogical ideas,
the thesis of the productivity of power, at least in its general form, is
neither original nor surprising. After all, manipulation has always been
conceptualized as the production of wishes or beliefs in the power
subject. However, even though the idea of the productivity of power is
uninformative in its general form, it gains in substance, when related to
Foucault's view of the relation between scientific knowledge and social
power. If it can be argued that the productivity of (social) power is
somehow internally related to the growth of scientific knowledge, we
obtain a thesis which is non-trivial. In this section I shall argue that such
conception naturally emerges on the basis of Foucault's suggestions.
Thus, as concerns the relation between social power and scientific
knowledge, the thesis that seems natural to attribute to Foucault can be
formulated as follows: The relation between scientific knowledge and
social power is internal-essential rather than external-accidental.
What makes it inviting to attribute this position to Foucault are his
statements to the effect that although social power and scientific knowl-
edge cannot be separated, their relation is nevertheless neither a concep-
tual one, nor one of identity; in other words, the relation between social
power and scientific knowledge is, at least in part, causal:
'Truth' is to be understood as a system of ordered procedures for

the production, regulation, distribution, circulation and operation of
statements. 'Truth' is linked in a circular relation with systems of
power which produce and sustain it, and to effects of power which
it induces and which extend it (P/K 133).
That Foucault does not conceive of the power/knowledge relation as

being merely accidental is obvious from statements like "truth is already
power" (P/K 133), or "there is no point in dreaming of a time when
knowledge will cease to depend on power" (P/K 52). Their identity is,
however, rejected in the following passage:
... when I read - and I know it has been attributed to me - the

thesis, "Knowledge is power", or "Power is knowledge", I begin to
laugh since studying their relation is precisely my problem. ... The
very fact that I pose the question of their relation proves clearly that
I do not identify them (pPHC 43).
In order to work our way towards a defense of an intemal-essential

relation between social power and scientific knowledge, let us start from
observing that all forms of successful power exercises presuppose or
involve knowledge and belief. For instance, in cases of coercion the
power exerciser is well advised to know - at least roughly - the value of
the booty to his victim; in cases of manipulation the coercer needs to
know something about human psychology; in cases of authority the
power subject relies on the allegedly superior knowledge of the author-
ity; and in cases of deterrence one agent wants the other to know what
will happen once the latter either starts or changes a specific course of
action.
However, when asking for the relation between knowledge and
power, we are obviously after something more, or different, from these
last-mentioned relations. What we are interested in, and what seems to
be the contested question, is not the occasional employment of knowl-
edge in exercises of coercion, nor - to mention another relation between
knowledge and power - the connections between knowledge and power-
to, i.e. power in the broad sense of capacity to bring about some state of
affairs. What we are interested in is rather the role that social power
plays in the shaping of scientific knowledge, the problem as to whether
all scientific knowledge is related to power, and the question of whether
science would be (or is) possible without the mechanisms and effects of
power exercises.
In order to bring out more clearly what is unsatisfactory about
approaches to the problem of power/knowledge that do not meet these
conditions, let us look briefly at one more recent attempt, by Barry
Barnes, to argue for an identity, or near-to-identity, between power and
knowledge. Pointing out the weakness of this proposal can pave the way
towards what I submit are more compelling perspectives or arguments:
the scientific community as a field of agonistic struggles, the interdepen-
dence between the emergence of dispositifs and some of the human
152 PART II
sciences, science as constituted by exclusion, and social power as a

determining factor of forms of scientific knowledge. I shall conclude by
asking how the internal relation between power and knowledge is to be
understood in modal terms, and what its implications are for the issue of
realism versus anti-realism.
BARNES' PROPOSAL
The line of reasoning of Barnes's The Nature of Power (1988) takes its
starting point from the "common sense concept of power" . According to
this concept, power is "the capacity to do work", or, more precisely,
since one and the same power usually can achieve more than one goal,
"the generalized capacity" to bring about (or hinder the coming about
of) states of affairs (ibid., 2-3).
To see how Barnes uses this general concept of power as capacity in
his definition of social power, we need to introduce two more of his
suggestions. On the one hand, Barnes argues that we can look upon a
society either as a set of routines and actions, or as a distribution of
knowledge: ". .. one characterization is as good as the other" since
"human beings act calculatively on the basis of what they know" (ibid.,
46). On the other hand, Barnes proposes that the power of a collective
is larger than the power of all its individuals taken singularly because the
capacity to act increases through coordination and exchange. Social
power is then the added capacity for action accruing from coordination,
Le. from forming a society. Thus, since the latter is eo ipso a distribu-
tion of knowledge, Barnes can write that ...
social power is the added capacity for action that accrues to individ-
uals through their constituting a distribution of knowledge and
thereby a society (ibid., 57).
However, this definition only tells us what social power is. From
this definition of "the nature" of social power, Barnes distinguishes the
issue of its possession. According to The Nature of Power, "social
power is possessed by those with discretion in the direction of social
action, and hence predominantly by those with discretion in the use of
routines" (ibid., 58). Furthermore, Barnes stresses that to be and to be

known to be socially powerful are one and the same thing. This is so,
since, for Barnes, society is not just any distribution of knowledge, but
rather a "self-referring" and "self-validating" one (ibid., 46). In other
words, the knowledge in question is knowledge about the roles of
various individuals in that society, knowledge about prevailing routines,
norms and practices in society. Moreover, this knowledge is true pre-
cisely because it is constitutive of society itself:
To come to believe something about the status of an individual is to

do two things at once: it is to accept a claim about his status and at
the same time to contribute to the constitution of his status (ibid.,
49).
For instance, to know that the chairperson in your department is power-

ful is to participate in making or keeping him powerful.
From these foundations Barnes advances three theses. According to
the first, (social) power is omnipresent. Even though social power is
possessed only by a minority, social power "lies outside" the power
holder, "right down there among the supposedly powerless". It is "only"
(!!) the discretion over this power that is concentrated (ibid., 62-63).
Capacities to act are everywhere; it is merely the discretion over them
that is unequally distributed.
Second, knowledge and social power "are the same things under
different forms of awareness, and a change in the one does not cause a
change in the other but rather entails it" (ibid., 169-70). Since the given
distribution of knowledge constitutes the powerful and provides them
with the social power qua discretion over the added capacities of their
polity, knowledge and social power are but two sides of the same coin.
To be socially powerful is to be a certain kind of referent of some given
distributed knowledge. In other words, to be socially powerful is to be
known to have, and thus have in fact, discretion over added capacities
accruing from the coordination of actions, that is, accruing from the
coordination or specific distribution of knowledge.
Third, and finally, the identity of (social) power and knowledge does
not imply a "simple correspondence" between specific singular "bits" of
knowledge and singular "bits" of social power. Rather:
154 PART II
The existence of a specific power, available at a specific point in the

social order, may be a matter of the existence of innumerable items
of knowledge, spread over every part of the social order. The
existence of a specific item of knowledge, at a specific point in the
social order, may be implicated in the constitution of numerous
powers available at many points (ibid., 169).
Barnes's ideas are instructive for our concerns in so far as his book
illustrates especially clearly that a one-sided attention to power-to rather
than power-over does only very limited work in helping us to understand
the relations between scientific knowledge and social power as domina-
tion.
Note, first of all, that Barnes indeed bases his argument crucially on
the revisionary redefinition of power as capacity. Phenomena like
coercion, authority, control, etc. thus remain inevitably bracketed in his
account. To be sure, in Barnes's case, power-over is preserved under the
new conceptualization of "discretion over social power". Yet it is not
difficult to point out that this suggestion is counterintuitive: not only
does this new conceptualization force Barnes to write that it is "only"
discretion rather than power which is hierarchized, but this conceptual-
ization also leads straightforwardly to his claim that every change in (the
distribution of) knowledge entails a change in (the distribution of)
power. Even if we grant him this claim for the case of capacity - and
even there it does not seem free of implausibility - it certainly does not
need to hold for discretion over social power(s) as added capacities.
Social power as domination (Le. Barnes's discretion) is certainly more
stable than this suggestion implies.
Secondly, even though Barnes explicitly includes scientific knowl-
edge of both the natural and the social sciences within his notion of
society-shaping knowledge, and even though he regards his investiga-
tions as closely linked to the sociology of (scientific) knowledge (ibid.,
170-71), his revisionary account of power does not even provide hints
as to how we should relate social power (in its standard sense as power-
over others) to the production and distribution of knowledge. All we can
say on the basis of Barnes's book on the relation between scientific
knowledge and power, is that scientific knowledge is part of the distrib-
uted knowledge of society, that it is in part constitutive of that society,
and that it is as such internally related, nay, identical, with capacities.

But these claims are either trivial - if one accepts Barnes's revisionary
definitions - or false - if one does not accept his leaving the concept of
power-over to the role of a 'poor relative'.
DEFENDING "POWERIKNOWLEDGE"
Turning from Barnes to Foucault, a first way to approach the issue of

science and power in a Foucauldian spirit is to place it within the context
of his overall theory of power. As we saw above, Foucault advances the
theses of the omnipresence, and of an ascending analysis of power. Since
science, or rather scientific communities, are part of the social field,
obviously both ideas also apply in its or their case. Unfortunately,
Foucault has not paid much attention to this application. He merely
informs us of his siding with Nietzsche in believing that the scientific
field is a field of struggles, and more precisely, that the "devotion to
truth" results from "the passion of scholars, their reciprocal hatred, their
fanatical and unending discussions and their spirit of competition" (NGH
78).
Yet while Foucault himself is disappointingly brief on this issue, it
seems easy enough to suggest questions for the social-historical study of
science that are but a transferring of Foucault's general ideas on power
to this special case. Thus taking our lead from the notion of power
networks and the distinctions between relations and mechanisms, the
micro- and the macrolevel, we might ask how the scientific community
as a power network relates to the scientific community as a knowledge
network (a network for the production and circulation of facts and
theories); what are the infinitesimal mechanisms of power in the scien-
tific community; what was the historical process by which they came to
be organized, combined and structured in institutions like the school, the
university and the laboratory; how was this organizing process possible,
Le. which - partly cynical - interests of which groups made these
organizations seem attractive; which - partly cynical - interests of which
groups and agents uphold and re-enforce these organizations today; and
how are individuals disciplined by power relations in this field for these
individuals to find these relations and mechanisms useful or inevitable.
156 PART II
To be sure. not all of these questions are new. Several of them have
already received at least partial answers in some more recent studies in
the philosophy and sociology of science. These enquiries study the use
of humans and animals as experimental ·material·. the agonistic struggle
among scientists for recognition. credit or credibility. identify the
disciplinary techniques employed. the mechanisms by which scientists
monopolize important resources like new instruments. mechanisms that
keep younger researchers obedient and loyal. and mechanisms by means
of which one prevents competitors receiving funds or gaining access to
important publication channels. However. while making for fascinating
reading. it can hardly be claimed that these investigations are more than
just a beginning. Thus. although a start has been made to increase our
understanding of power in the laboratory. not much work has yet been
done. for instance. on the mechanisms of power in university depart-
ments: the sociological anthropologists are not yet at large among
philosophy tribesmen. The perhaps best-known study that takes some
first steps in this direction. Pierre Bourdieu's masterly Homo Academi-
CUS. does not direct itself to the microsociological level even though it
makes many interesting observations relating to it. For example. Bour-
dieu is a perceptive observer of the ways in which French professors can
keep their pupils in a relation of dependence by determining their career
rhythm. their chances to publish. the time of their doctorate. and their
appointment to short-term positions. The character or attitude that these
mechanisms produce is. in Bourdieu's happy phrase. "the docile and
submissive. even somewhat infantile. attitude which characterizes the
good student of all eras" (Bourdieu 1988: 88).
However. treating science as an agonistic field of struggle among
scientists does not yet in itself establish that scientific knowledge qua
product of the scientific activity is internally-essentially interlocked with
social power. The defender of the received view might still hold on to
his or her position by distinguishing between the social context of
science and scientific knowledge itself. In order to undermine his or her
position more radically. we therefore need to turn to three other argu-
ments. 27
A disquieting suggestion
The first argument is little more than a disquieting suggestion that

applies perhaps only to some of the human sciences. Scientific knowl-
edge of the human sciences might be internally related to power mecha-
nisms because the very subject matter of those sciences is, at least
partly, constituted by mechanisms of power that those human sciences
help to justify, use as instruments, and have taken over from coercive
institutions. Put differently, scientific knowledge of the human sciences
might just be an enormous self-fulfilling prophecy because the mecha-
nisms upon which it is based are but versions of those mechanisms of
control, surveillance and manipulation that have constituted its objects.
This disquieting suggestion arises from Foucault's study of the
history of psychiatry and criminology. Foucault argues that the symp-
toms of mental disease or social deviance are diagnosed and treated by
the same mechanisms of surveillance, control, drill and categorization
that figure centrally in the mental clinic or the prison, on the one hand,
and that were created and refined in and for these coercive institutions
by psychiatrists and criminologists, on the other hand.
Science and exclusion
Another argument for an internal link between social power and scien-
tific knowledge is one that looms large in all of Foucault's genealogical
writings: scientific discourses constitute themselves only by exclusion
and prescription, Le. by drawing a line between scientific and non-
scientific knowledge. Foucault speaks of the excluded knowledge as
"subjugated knowledge", as "the historical contents that have been
buried and disguised in a functionalist coherence of formal systematiza-
tion" and that ...
have been disqualified as inadequate to their task or insufficiently

elaborated: naive knowledge ... beneath the required level of cogni-
tion or the scientific ... [e.g.] that of the psychiatric patient, of the
ill person, of the nurse, of the doctor ... of the delinquent ... (p/K
81).
158 PART II
Foucault also confronts those who seek to establish Marxism as a science

with the question, "what types of knowledge do you want to disqualify
...7" (P/K 85).
As Foucault sees it, exclusion functions through a discursive "or-
der", a "regime of truth" (P/K 133). Concerning the latter, Foucault
tells us that it is a "circular relation" between truth "as a system of
ordered procedures for the production, regulation, distribution, circula-
tion and operation of statements", on the one hand, and "systems of
power which produce and sustain it [Le. truth] ... and effects of power
which it induces and which extend it" (ibid.), on the other hand. Put
more succinctly, "one is within the [region of] the truth only when being
obedient to a discursive police" (LD 37).
In his L 'Ordre du Discours, his inaugural lecture at the College de
France in 1970, Foucault suggests a systematization of the different
forms of exclusions and prescriptions by which a field of science is
constituted. A first main category are "procedures of exclusion" (LD
11). Within this category Foucault distinguishes further between "inter-
dictions" concerning certain objects, styles of arguing, and individuals or
groups, a "rejection" of what is defined as irrational and mad, and a
"partition" between the true and the false - a partition that amounts to
an exclusion of what does not conform to some specific conception and
standard of truth (LD 11-15).
A second main category, "internal procedures", relates to prescrip-
tions on how statements have to be formed and formulated in order for
them to qualify as scientific within a given field of study. Such prescrip-
tions bear upon the manners according to which one is to produce "com-
mentaries" on the work of other scientists; the ways one is expected to
group statements and theories around the originators or "authors"; and
the list of objects or processes to be talked about, or conceptual and
technical instruments to be employed. Foucault refers to this last item as
"discipline" (LD 23-31).
Finally, the third main category refers to various procedures that
further define and regulate competent membership, that is, "conditions
of admission". These are, first, the "ritual", Le. the criteria of proper
social/discursive behavior; second, the "discursive societies", Le. visible
or invisible closed colleges all of which put specific demands upon their
members; third, the "doctrine", Le. the - at least verbal - adoption of,
or abiding by, specific religious, political and philosophical views, and

fourth and finally, "education" (LD 38-46).
That this categorization is a useful one, can easily be vindicated by
relating it to work done in the sociology or history of science, fields that
tend to be stronger in collecting empirical materials than in providing
theoretical frameworks. Steven Shapin's and Simon Schaffer's study of
Boyle's introduction of experimental science as a new language-game,
Leviathan and the Air-Pump. Hobbes, Boyle, and the Experimental Life
(1985), is a case in point. Shapin and Schaffer distinguish only vaguely
between "material", "literary" and "social technologies" employed by
Boyle (ibid., 25). Using Foucault's distinctions seems to give a more
fine-grained picture, however.
As concerns exclusion, Boyle's new science was defined by interdic-
tions with respect to objects like insubstantial beings and spirits, self-
moving matter, and metaphysics; interdictions of styles of arguing like
deductivism or ad hominem arguments typical of philosophical treatises
at the time (Hobbes); and interdictions of groups or individuals like
alchemists, rationalist philosophers and hermeticists. Panition between
the truth and the false took the specific form of accepting only the
probable as true, and of rejecting all claims to absolute truth as per se
unscientific (ibid., 58, 61-69, 72, 167, 176,203).
Internal procedures consisted of specific demands with respect to
commentary, the latter being required to be modest, and abstaining from
any sharp critical overtones; specific demands with respect to the catego-
ry of the author: experiments must be attributed to individuals not the
experimental community as a whole (Le. only individuals can fail, not
the scientific experimental community); and specific demands of the
discipline: use of the air-pump as the central instrument of experimenta-
tion, and acceptance of the air-pump as air-tight for all practical pur-
poses, rejection of all questions as illegitimate that cannot be handled
experimentally, abiding by the distinction between experimental matters
of fact and their ultimate physical causes or explanations, supporting a
sharp opposition between actual and thought experiments, regarding the
spring of air as the central object to be investigated, taking the reports of
eye-witnesses as decisive, and the employing naturalistic representations
in one's publications (ibid., 29,45,51,55,56,60-69).
Finally, conditions of admission included those of the ritual, Le. a
160 PART II
civil and liberal tone, an attitude of trust and modesty, a functional style
of writing, the abstention from preconceived expectations as regards
experimental outcomes, and the posing as a "priest of nature". Admis-
sion was also restricted by the Royal Society as a closed discursive
society: eye-witnesses were restricted to lawyers, priests and hand-picked
individuals, peasants were regarded as unqualified, testimony of experi-
mental outcomes by non-members were rejected, and only educated, or
disciplined witnesses were acceptable. Finally, doctrines to be adopted
included the rejection of both vacuism and plenism, the belief in the
reliability of sense-perception as a theory-neutral foundation of science,
fallibilism, a mechanistic conception of nature, and a religious concep-
tion of God as non-corporeal (ibid., 36-42,58-76, 171, 206, 218, 319,
336).
The fact that the interesting observations of Leviathan and the Air-
Pump can be reconstructed and put more succinctly in Foucauldian
terms, also suggests a reformulation of the book's main topical thesis. In
their rather brief dealings with the theoretical implications of their study,
Shapin and Schaffer conclude from their detailed account of Boyle's
program and its relation to the wider political situation of the time, that
science itself is to be analyzed in terms of politics. The authors hold that
in order to understand the possibility of the creation of scientific knowl-
edge we have to understand the working of scientific communities and
their structures. To deal with the latter, however, is to study the inter-
play of three factors: "(1) the polity of the intellectual community; (2)
the solution to the practical problem of making and justifying knowl-
edge; and (3) the polity of the wider society" (ibid., 342). These factors
are interconnected in that making and justifying knowledge is possible
only against the background, or within, an intellectual community with
specific rules of exclusion and prescription. Furthermore, scientific
knowledge usually spreads out from the scientific community into the
wider polity of society where 'it is used as a resource. And finally, the
introduction and maintaining of any specific scientific forms of life is
dependent on whether its advocates succeed in mobilizing support for it
in the wider society (ibid., 342).
One need not disagree with this conclusion to suggest that it is
perhaps brought into sharper focus by being conceptualized in terms of
social power rather than in terms of "politics". Relying on the notion of
power - regardless of whether in a Foucauldian or some other version -

at least immediately prompts us to pose a limited list of precise ques-
tions. These questions concern topics like responsibility, invisible hand
effects, networks, forms of sanctions, scopes, amounts, extensions, and
mechanisms of power, means, resources, distribution and intensity of
power, as well as the constitution of subjects in and by power mecha-
nisms.
Shifting from a vague notion of politics to a somewhat more elabo-
rated notion of social power, we can reformulate Schaffer's and Shapin's
program by saying that to study the possibility of knowledge is to
investigate the relations between the social power network of society and
the distribution of social power within and outside of science. The
production and justification of knowledge is possible only against the
background of, or within, an intellectual community as a network of
social power. Scientific knowledge usually spreads out from this com-
munity into the wider society, changing, enhancing and modifying
existing power relations by providing various individuals or groups with
new or improved power mechanisms, resources and means. Finally, the
introduction and maintaining of specific scientific forms of life - and the
power structures characteristic of them - depend upon whether or not
they succeed in convincing a sufficient number of (powerful) individuals
and groups of their usefulness, where this usefulness in part again is
measured by their production of, and improvement of power resources
and means.
Reformulated in this way, it is immediately clear not only that
"power is productive" but also that the relation between scientific
knowledge and social power is internal-essential rather than external-
accidental.
An argument from underdetermination
Even though the last argument seems sufficient to establish an essential

link between social power and scientific knowledge, it can perhaps
further be strengthened by another line of reasoning. This line of rea-
soning is implicit in the argument above - and actually implied in part
by Shapin and Schaffer's book. This third argument proceeds from the
162 PART II
existence of underdetermination in science to, first, the inevitable in-

volvement of social power in scientific decision making, and to, second,
the idea that social power shapes our forms of knowledge. One might
perhaps say that this argument is hinted at by Foucault when he stresses
the contingent character of scientific and social development, on the one
hand, and the pressure of power on science, on the other hand:
... things 'weren't as necessary as all that'; it wasn't as a matter of

course that mad people came to be regarded as mentally ill; it
wasn't self-evident that the only thing to be done with a criminal
was to lock him up; it wasn't self-evident that the causes of illness
were to be sought through the individual examination of bodies ...
(QM 104).
Power never ceases its interrogation, its inquisition, its registration
of the truth ... (p/K 93).
However, the argument from underdetermination is best presented

without trying to squeeze it out of Foucault's sketchy remarks.
The starting point, or first premiss, of the argument, Le. the exist-
ence of underdetermination in science, more specifically, the underdeter-
mination of theory by observation, and the underdetermination of the
interpretation of experiments, the choice of theory and the choice of
research programs by principles of rationality, is hardly controversial.
(Even though one might still debate the extent to which underdetermi-
nation plays a role.)
The second premiss states that underdetermination in science is
overcome by interests and that these interests are power interests. I
regard this premiss as highly plausible based on the number of case
studies that sociologists and historians have provided us with over the
last decade or so. It appears that, when faced with choices, scientists
tend to adopt interpretations, theories or lines of research that both
enable them to monopolize areas of investigation, and that offer them
either opportunities to dismiss the work of competitors as unfruitful,
unscientific, and unworthy of funding, or opportunities to answer to the
social demands of a particular class, the state, a political party or a
church (Shapin 1982).
The third premiss is the hypothetical sentence that if there is, or
were factors F, that determine what we regard as knowledge, good

science and scientific exemplars or paradigms, then the relation between
these Fs and knowledge is, or would be, internal-essential.28
Now given these three premisses, my conclusion should no longer
come as a surprise. Since underdetermination is overcome by non-
scientific interests, and since these non-scientific interests are power
interests, therefore power is one of the factors F that are internally-
essentially related to scientific knowledge. Since science qua social field
is, and always has been, penetrated by power relations, and since the
choice of theories and research programs has thus inevitably been
strongly influenced by the interest to increase power - be it the power of
individuals, classes or the military power of the state - then it becomes
indeed implausible to think of social power and scientific knowledge as
merely accidentally related.
REALISM AN MODALITY
To conclude this rather longish discussion of the internal-essential

relation between power and knowledge, let us, at least briefly, ask, first
what all of this amounts to for the issue of realism vs. anti-realism, and,
second, how, in more precise modal terms, this internal-essential relation
is to be understood.
I regard it as a distinct merit of the various arguments presented
above that they do not prejudge anything concerning the tricky problems
of realism or anti-realism in the philosophy of science.29 What Fou-
cault's position on this question might be is hard to decide on the basis
of his writings - at least as far as the natural sciences are concerned. (I
shall return to this topic below.) Be this as it may, both realist and anti-
realist philosophers can accept the thesis of an internal-essential relation
between scientific knowledge and social power. Nothing in the above
arguments shows that there is no true knowledge, that scientific realism
is wrong, or that in and through engaging in scientific work we do not
find out things about parts of reality. To be sure, if these proposals are
accepted, a study of social power cannot be excluded from the question
of why we model reality in one way rather than another, of what pieces
of reality we come to know, or of why science develops in one direction
164 PART II
rather than another. But all this does not amount to saying that science
is dealing - always and necessarily - with fictions.
I have already pointed out that an interesting and challenging argu-
ment for an essential relation between power and knowledge must not
stipulate this relation by fiat, Le. by providing revisionary definitions of
the two relata, definitions on the basis of which the relation between
social power and scientific knowledge becomes a conceptual one. This
demand, however, calls for an answer to the question of what the
internal-essential, but non-conceptual, relation between social power and
scientific knowledge amounts to in modal terms. It should be clear, in
any case, that since this relation is not a conceptual-logical or analytical
one, it cannot hold in all possible worlds. After all, we can imagine a
world of angels with scientific interests, or a world of humans in which
all knowledge is obtained through divine revelations. Yet the fact that
the relation between social power and scientific knowledge does not hold
in all possible worlds does not oblige us to treat this relation as an
accidental and contingent one. This is because we can preserve a notion
of necessity here by restricting the possible worlds to socially possible
ones. That is to say, the thesis of the internal-essential relation is tanta-
mount to the truth of the following counterfactual:
(*) If the scientific communities were not networks of social power

relations, and if they were not connected to various social power
networks of the wider society, scientific knowledge as we know it
would not exist.
This counterfactual (*) amounts to saying that in any possible state of

affairs in which scientific communities are not social power networks,
and which resemble our actual state of affairs as much as is permitted by
the absence of social power relations in science, scientific knowledge as
we know it does not exist. 30 Put differently, while not being a logically
necessary relation, the relation between social power and scientific
knowledge is socially necessary. In other words, the relation holds in all
those possible worlds which are socially similar to our world: they
contain humans or other intelligent creatures that resemble us in so far
as they do not obtain knowledge by revelations, they are not angels or
saints, they live in a world of scarce resources, and they face the prob-
lems of underdetermination of scientific theories by observation and

standards of rationality.
12. GENEALOGICAL RESEARCH STRATEGIES
INTRODUCTION
Proceeding from the question of what power is or how it works to the

question of how it is to be (or can be) investigated, it is clear that the
latter issue cannot but be closely related to the former. It is hardly
surprising, therefore, that in dealing with the Foucauldian theory of
power we already encountered a whole number of proposals on how
power might be studied. For instance, we saw Foucault attending to
power mechanisms and exercises; employing models of war and physics;
giving heed to mechanisms of ecological control by architectural mea-
sures; focussing on the training or disciplining of bodies; studying the
ways power relations and mechanisms shape the personal identity of the
coercer and the coerced; investigating power as a network of power
relations; examining the interconnections between microlevel and macro-
level power relations; analyzing coercive institutions as social labora-
tories for the development of power mechanisms; taking one's starting
point from individuals and groups rather than classes ("ascending analy-
sis"); relying on invisible hand or counterfinality explanations; making
use of concepts like "tactics" and "strategies"; considering the interrela-
tions between social power and scientific knowledge; taking into account
the involvement of interests; (re-)conceptualizing science as a social
process of exclusion and regimentation ("orders of discourse"); and
rehabilitating excluded, "subjugated" knowledge.
Many of these ideas are also alluded to in all of the four major
places where Foucault explicitly spells out his methodological "de-
mands", "rules" or "precautions".31 Since all of these rules flow fairly
straightforwardly from the theses discussed above as the Foucauldian
theory of power, there is little need for going over them in detail. What
calls for exploration and discussion, however, are some more general
methodological ideas (and the reasoning behind them), ideas that moti-
vate Foucault's choice of the label "genealogy" for his investigations
166 PART II
into power and knowledge. In other words, what needs to be investigated

are the central principles and suggestions that jointly constitute the
Nietzschean background, or the Nietzschean dimension, of Foucault's
enterprise.
Foucault himself describes the intention behind his genealogical
work as seeing "[...] to the extent that it is possible, with the aid of
Nietzsche's texts - but also with anti-Nietzschean theses (which are
nevertheless Nietzschean!) - what can be done in this or that domain"
(pPHC 251). He also suggests that "the genealogy of morals" might
stand as a title over his own work, and he praises Nietzsche for being
"the philosopher of power, a philosopher who managed to think of
power without having to confine himself within a political theory in
order to do so" (p1K 53). Foucault also confesses, however, to be
interested more in using Nietzsche's texts for his own purposes rather
than striving for 'correct' and 'faithful' interpretations of Nietzsche's
intentions:
For myself, I prefer to utilize the writers I like. The only valid
tribute to thought such as Nietzsche's is precisely to use it, to
deform it, to make it groan and protest. And if commentators then
say that I am being faithful or unfaithful to Nietzsche, that is of
absolutely no interest (p1K 54).
Statements like the last make it a spurious task to try to determine

whether Foucault is justified in calling his own enterprise "genealogy"
or not. (Or then, it is a task I am not especially interested in.) Thus I
shall not attempt here to match Foucault's own writings on Nietzsche
against the latter's own texts or against some widely-circulating inter-
pretations. Rather I want to point out briefly, by attending to Foucault's
essay "Nietzsche, Genealogy, History", what Foucault finds interesting
in Nietzsche's work. Even here, however, caution is called for: given
the above quotation, it would certainly be unwarranted to assume that
every Foucauldian reformulation of Nietzsche's aphorisms is eo ipso a
formulation of a thesis of Foucauldian genealogy.
As is obvious enough even from a superficial reading of "Nietzsche,
Genealogy, History", there are four main themes in Nietzsche's genealo-
gical writings, that Foucault p~ys special attention to.
The first theme is that of the contingency of the present, a topic that
centers around the notions of struggle and event. Let us call this first
theme the recovering ofagonistic events. Foucault describes Nietzschean
genealogy as an attempt to lay bare the contingency of the present and as
an opposition to explanations which rely on one exclusive principle of
explanation. To do genealogy is to "maintain passing events in their
proper dispersion; it is to identify the accidents" of history, it is to
understand that "truth or being does not lie at the roots of what we know
and what we are, but with the exteriority of accidents" (NGH 81). The
aim of genealogy is not to show how the present is necessitated by the
past. Rather the aim of the genealogist is to show that the present is the
product of a variety of contingent-accidental events of power struggles
that could have ended differently from the way they did. 32
The second theme, an attending to the human body, might be called
physicalism. Foucault points out that Nietzschean genealogy adopts not
only the point of view of the event and the struggle, but also the per-
spective of the human body as it is formed and molded in these strug-
gles.
The genealogical study of power does not give primacy to the effects
of power on the level of beliefs and convictions, but also gives heed to
the ways different power regimes control, confine, nourish and shape the
human body. Foucault refers approvingly to Nietzsche's idea that gene-
alogy "diagnose[s] the illnesses of the body, its conditions of weakness
and strength" (NGH 80). For the genealogist the body is "the inscribed
surface of events" (NGH 83), the genealogist does not believe that the
human body stands outside of history and that the body is a proper
subject matter only for the physiologists. The task Nietzsche sets for
genealogy is to "expose a body totally imprinted by history and the
process of history's destruction of the body" (NGH 87).33
Third, Foucault also shows sympathy for the attitude of distrust and
irony that informs all of Nietzsche's writings. Allegedly humanitarian
motives are to be unmasked as stemming from (and here Foucault quotes
Nietzsche) "detestable, narrowminded conclusions. Pudenda origo"
(NGH 77). The genealogist does not believe that humanity "gradually
progress[es] from combat to combat until it arrives at universal recipro-
city", rather humanity moves on "from domination to domination"
(NGH 85).
168 PART II
This distrust also extends to science and knowledge. Nietzschean

genealogy tears off the "the mask" which presents scientific or, more
specifically, historical consciousness as "neutral, devoid of passions, and
committed solely to truth" (NGH 95). Genealogy seeks to establish that
behind this mask all one finds are "the passion of scholars, their recipro-
cal hatred, their fanatical and unending discussions, and their spirit of
competition" (NGH 78), that behind that mask there are but "aspects of
the will to knowledge: instinct, passion, the inquisitor's devotion, cruel
subtlety, and malice" (NGH 95).
In this unmasking, genealogy takes up an ironic or parodic position.
It shows that 'high' motives can be redescribed by allegedly 'low' ones.
Thus Foucault can claim that genealogy makes a "parodic and farcial
use" of history or that genealogy is "history in the form of a concerted
carnival" (NGH 93-94). Genealogy ironicizes the traditional accounts
and scientific styles by disassociating itself from them and by showing
how they came about. In this sense, the function of genealogy is that of
a "curative science" (NGH 90). For instance, in breaking with the tradi-
tional view of unity and strong personal identity over time, it opts for
the "systematic dissociation of identity": "The study of history makes
one 'happy, unlike the metaphysicians, to possess in oneself not one
immortal soul but many mortal ones'" (NGH 94).
Fourth, and finally, Foucault also gives much attention to the
Nietzschean theme of perspectivism. Given the idea that everything is
relative to its historical conditions, and given the dissociation of identity,
this hardly comes as a surprise. Thus one of the most central traits of
genealogy is "its affirmation of knowledge as perspective" (NGH 90):
"Truth is undoubtedly the sort of error that cannot be refuted because it
was hardened into an unalterable form in the long baking process of
history" (NGH 79).
Of course, here Foucault could also have cited Nietzsche's noto-
riously famous sentences that "facts are precisely what they are not, only
interpretations", or that truth is "not something there, that might be
found or discovered - but something that must be created ... ,,34
In sum, "Nietzsche, Genealogy, History" suggests that the recovery
of agonistic events, physicalism, distrust and irony as well as perspec-
tivism are the four themes that Foucault regards as of special interest in
Nietzsche's genealogy. In the following I shall take up each of these
topics in more detail, discussing them from both a historical and a

topical point of view.
RECOVERING AGONISTIC EVENTS
The return of the event
In order to understand in greater detail what "recovering agonistic

events" amounts to as a genealogical research strategy, we have to return
once more to L~vi-Strauss and the French historical school around the
Annales. As seen in greater detail earlier, it was due to the combined
efforts of structuralism and Annales historians that the notion of event
had been a discredited one by the sixties. To begin with structuralism,
L~vi-Strauss draws a sharp dividing line between the particular and the
universal, and he seeks to leave behind the level of singular historical
events qua particulars in order to reach for invariant mental structures
qua universals. The latter are modelled upon, or depicted by, means or
techniques borrowed from structural linguistics and especially phonolo-
gy. While L6vi-Strauss leaves a role for history as "indispensible for
cataloguing the elements of any structure whatever", history, for him,
can be little more than an assistant to anthropology: " ... history leads to
everything, but on condition that it be left behind" (1966: 262). Anthro-
pology takes over from history in its effort to "grasp, beyond the con-
scious and always shifting images which men hold, the complete range
of unconscious possibilities" (1967: 23). These unconscious possibilities
provide us, L~vi-Strauss suggests, with "a logical framework for histori-
cal development" (ibid., 24).
In part, the success of the Annales school in postwar France was due
to its ability to present itself as an ally of structuralist concerns. L6vi-
Strauss himself granted that, for instance, the work of Lucien Febvre on
Rabelais was congenial to his own work (ibid., 25). Furthermore, a
leading exponent of the Annales school wrote in 1973, that "for half a
century, from Marc Bloch to Pierre Goubert, the best of French histo-
rians, systematically systematizing away, have in fact been structuralists.
Sometimes they realized it themselves; sometimes they did not; all too
often nobody else realized it at all" (Le Roy Ladurie 1981: 5). More-
170 PART II
over, a considerable number of Annales historians from Febvre and

Bloch on, conceived of anthropology as one of their most important
sister discipline. And even those who did not, could at least join forces
with structuralism in rejecting the category of event.
By the early 19708 however, several French historians began to raise
critical questions concerning the rejection of the category of the event.
Questions of priority aside, the work most responsible for are-evaluation
of the notion of event in France, is probably Paul Veyne's Writing
History, a book first published in 1971. Veyne's thesis is that "history
is an account of events: all else flows from that" (1984: 4). For Veyne,
historical study is essentially the production of narratives; and here it is
the narrative's "plot" that determines what is eventful. While Veyne
joins the Annalists in criticizing traditional history as a "history of
treaties and battles", his criticism turns not against the category of event
itself, but rather against the traditional tacit assumption of a set of
"'accepted' events", Le. events that are treated as given. From Veyne's
point of view this assumption is mistaken since events are not given but
constituted by the historian's ordering of the historical material according
to different routes ("itineraries") of inquirrs: "In short, the eventwor-
thy field does not comprise spots to be visited and that would be called
events; an event is not a being, but an intersection of possible itinera-
ries" (1984: 36).
Against Annales historians, on the other hand, Veyne points out that
attending to territories, prices and the growth of the population does not
inevitably imply a stepping beyond the domain of events. Rather it
amounts to a re-ordering of facts according to a new plot and thus to the
constitution of new (forms of) events:
Non-eventworthy history [Le. Annales histories] has been a sort of

telescope that, by letting us see in the sky millions of stars other
than those known to astronomers of old, would make us understand
that our dividing the starry sky into constellations was subjective
(ibid., 37).
Foucault's agrees with Veyne (against structuralists and Annalist

historians) that the category of the event is everything else but obsolete;
and, like Veyne, he also points out that rather than exorcising the
category of event from history, the Annales historians have "constantly

enlarged" the field of events by defining ever new series of events (LD
57). However, as in his earlier, archaeological, phase, Foucault ac-
knowledges his indebtedness to the work of Febvre, Bloch and their
followers. In fact, in his writings of the seventies he repeatedly expresses
his sympathy for, and indebtedness to, the Annales school's ideas. With
the latter he rejects a history of "those who held power - anecdotal
histories of kings and generals" (p/K 51), uses "the great eighteenth-
century demographic upswing in Western Europe", much studied by
Annalist historians, as an explanatory principle for the creation of the
modern problems of "population" , commends the Annales school for its
interest in the "'ignorable' materials" of popular culture and the body,
endorses their ideal of history as a means of distancing oneself from the
present, shares their distrust in the straightforward applicability of
linguistic-structuralist models to history, and proclaims himself indebted
to their study of spaces (p/K 3·7, 171; DP 25).
To take up the last two points in more detail, even though members
of the Annales school like Braudel or Le Roy Ladurie repeatedly express
their interest in structuralism and anthropology, they still deny that
structuralist methods could be straightforwardly applied to historical
materials (Braudel 1980: 44; Le Roy Ladurie 1981: 5). Foucault joins
them in this caution, but links it - here of course parting company - to
the rehabilitation of the event. For him "the history which bears and
determines us has the form of a war rather than that of a language:
relations of power, not relations of meaning" (p/K 114). And wars can
only be described by giving full heed to the notion of event:
One can agree that structuralism formed the most systematic effort
to evacuate the concept of the event, ... In that sense, I don't see
who could be more of an anti-structuralist than myself. ... The
problem is at once to distinguish among events, to differentiate the
networks and levels to which they belong, and to reconstitute the
lines along which they are connected and engender one another. ...
'semiology' is a way of avoiding its [Le. history's] violent, bloody
and lethal character by reducing it to the calm Platonic form of
language and dialogue (p/K 114-15).
172 PART II
Foucault's partly critical. partly sympathetic attitude towards "the

Annales paradigm" is even more clear with respect to spaces. Le.
geographical and architectural spaces. In one passage he claims that it
was Bloch's and Braudet's work on rural and maritime spaces and
Ari~'s work on architecture in the eighteenth century. that decisively
opened up the possibility of a study of spaces (p/K 148-49). Still.
Foucault suggests that one has to go beyond these studies in so far as the
latter are restricted by the conception that space "predetermines a history
which in turn reworks and sediments itself in it". In opposition to this
view. for the genealogist. space has to be studied as an "economic-
political form" (p/K 149). That is to say. architectural and natural space
have to be studied in their interrelations with power. with events and
processes of struggle.
Recovering contingency
In fact. Foucault's interest in mechanisms and effects of power and the

interrelations between power "and forms of knowledge is the deeper
motivation behind his rehabilitation of the event not only against struc-
turalism but also against the Annales school. Explaining his research
strategy of "recovering agonistic events" (eventalisation) to members of
that school in a discussion of Discipline and Punish (QM). Foucault
emphasizes above all three points of difference.
The first ingredient of "eventalisation" is the idea that attending to
events rather than structures can help to lay bare the contingency of
taken-for-granted self-evidences. The aim of genealogical enquiry is to
show that existing structures. like the categorizing of the mad as men-
tally ill. the hospital system or the prison institution. originate from
contingent events and that their existence was thus inevitable only after
various processes had taken place (QM 104). That is to say. even though
their existence became inevitable in some sense at a certain point in
history. say t. their existence is nevertheless still contingent not only in
so far as they owe their existence to a coincidence of various. originally
independent. factors. but also in so far as their emergence was unneces-
sary prior to t.
Foucault implies that this contingency of history is not properly
attended to by Annales historians. He charges them with tending towards

replacing historical contingencies by "the most unitary, necessary,
inevitable and (ultimately) extrahistorical mechanism or structure availa-
ble":
An economic mechanism, an anthropological structure or a demog-

raphic process which figure as the climactic stage in the investiga-
tion - these are the goals of a history that downgrades events (QM
106; cf. LD 61).
To put the central point more sharply than Foucault does himself,
we might want to say that his criticism of Annales historians is based on
a distinction between events that are "strongly" and events that are
"weakly" necessary. J6 An event is strongly necessary if its occurrence
at a specific point in the historical course of time had been necessary all
along; an event is weakly necessary if its occurrence merely became
necessary after certain other events had taken place. Using this distinc-
tion between strong and weak necessity, it is inviting to say that Fou-
cault charges the Annales historians with paying insufficient attention to
the turning points of history, those points in or after which a structure or
institution becomes necessary.
Recall also that in "Nietzsche, Genealogy, History" Foucault puts a
lot of emphasis on genealogy as the recovering of accidents and chance
in history, while in other places he speaks rather of the contingency of
the present. Putting these ideas together suggests going a step further and
distinguishing between two forms of contingency: contingency-proper or
weak contingency, and chance or strong contingency. To draw this
distinction, all we need is a way to drive a wedge between two cases:
contingency plus weak necessity, on the one hand, and contingency
without any necessity, on the other hand:
"p is weakly contingent at t" means: first, p obtains at t, and there

was some moment of time t', before t, such that at t' it was possible
that p would not obtain at t, and, second, there was some moment
of time t", identical with, or after t', such that at t" it became
necessary that p would obtain at t.
174 PART II
"p is strongly contingent at t" means: first, p obtains t, and there

was some moment of time t', before t, such that at t' it was possible
that p would not obtain at t, and, second, there was no moment of
time t", identical with, or after t', such that at t" it became neces-
sary that p would obtain at t.
The upshot of this exercise, I submit, is this: not only does Foucault
want to emphasize - unlike structuralism and the Annales School - weak
necessity and eo ipso weak contingency, but he also tends towards
depicting the origin of institutions like the prison as being due to strong
contingency qua chance. To appreciate this point we only have to keep
in mind the classical Aristotelian conception of chance as the concurring
of two independent causal chains (e.g. I walk to the water because of my
thirst, while a group of bandits goes by the same water on their way to
town), and Foucault's depiction of, for instance, the origin of the prison
as being due to concurring different, partially independent developments,
like the new interest in discipline, new architectural ideas, economic
problems, new methods of education and factory work.
In sum then, the first ingredient of Foucauldian eventalisation is to
suggest a description of those moments of chance or concurrences of
different series of events out of which subsequently weakly necessary
structures arise.
Caus(ll multiplication
The second ingredient of the Foucauldian research strategy of "recover-

ing the agonistic event", is "causal multiplication" (QM 104). Foucault
implies that the historical studies of the Annales School are based on a
far too limited repertoire of explanatory principles or causes. This
charge is expressed not only in the above quotation with its implication
that the Annalists try to reduce historical plurality to one mechanism or
structure; it is also implied in Foucault's explanation as to why his own
approach will seem inadequate in this quarter of the historical profession:
Clearly, viewed from the standpoint of this style of analysis, what

I am proposing is at once too much and too little. There are too
many diverse kinds of relations, too many lines of analysis, yet at

the same time there is too little necessary unity (QM 106).
Foucault explains that a singular event or process like the emergence of

the prison is a "'polyhedron' of intelligibilities" (QM 105), and that it
can be placed in an endless number of series or plots, Le. in an endless
number of relations to other events, none of which renders it necessary.
The agonistic dimension
Finally, the third ingredient of Foucauldian recovery of agonistic events

brings in the element of agon. In line with his reconstruction of Niet-
zschean genealogy, Foucault speaks of the importance of rediscovering
"the connections, encounters, supports, blockages, plays of forces,
strategies and so on which at a given moment establish what subsequent-
ly counts as being self-evident, universal and necessary" (QM 105). In
other words, the events to be selected by a genealogical study of his-
tories are struggles, conflicts, controversies, Le. events in and through
which new structures become established, existing structures either being
used as resources or being destroyed.
Foucault is not alone of course in believing that the study of net-
works of social power and scientific knowledge, and above all the
deconstruction of self-evidences, have to take their starting point from
controversies. Such modus operandi can be defended by reference to
other authors, too. For instance, Shapin and Schaffer17 in their Leviat-
han and the Air-Pump propose that one of the most effective ways for
making the self-evidences of our scientific experimental practices explicit
is to take up the position of a "stranger" to this practice. The stranger in
a culture becomes easily aware of what are self-evident assumptions for
its members simply because the stranger "is in a position to know that
there are alternatives to those beliefs and practices" (1985: 6). Now one
effective means of achieving a stranger's attitude towards our scientific
culture, Shapin and Schaffer argue, is to study scientific controversies of
the past, controversies out of which this culture emerged. On the one
hand, these controversies usually give rise to "disagreements over the
reality of entities or properties of practices whose existence or value are
176 PART II
subsequently taken to be unproblematic or settled", and, on the other

hand, the adversaries in such controversies usually try to "deconstruct
the taken-for-granted quality of their antagonists' preferred beliefs and
practices ... by trying to display the artifactual and conventional status
of those beliefs and practices" (ibid., 7).
However, unlike these authors, Foucault does not limit himself to a
study of how structures (e.g. dispositifs) and self-evidences (e.g. the
normal-pathological divide) emerge from a bundle of disparate contro-
versies and struggles. His call for a recovering of agonistic events, Le.
a conceptualization of history in terms of war, amounts to proposing -
in Veyne's terms - a somewhat broader, or longer "itinerary" through
the "eventworthy field". Events selected or constituted by this "plot"
are, in addition to those out of which (foreseen and unforeseen) struc-
tures result, also those events that are predetermined by these structures,
and even those events in which these structures and their (foreseen and
unforeseen) effects function in multiple, different ways as resources and
obstacles for individuals, groups and classes in their struggle for, or
against, mechanisms and effects of power and knowledge. For instance,
the creation of delinquency was, according to Foucault, an unforseen
effect of the penal structure that centrally relied on the prison. Yet
members of the bourgeoisie employed this side-effect in order to weaken
their primary enemy, the workers, by using the delinquents for the
"extraction of profit from pleasure through the organization of prostitu-
tion", as provocateurs and as a political-military resource. (Napoleon III
rose to power with the help of criminals (OP 280).YS
Put differently, on the genealogical plot, controversy does not end
with the emergence of structures. Since the structures in question are the
result of power struggles, they are always partly unstable and constantly
contested. Foucault therefore proposes that history be conceptualized as
an ongoing "war", in which "for each move by one adversary, there is
an answering one by the other" (p/K 57). In fact, even this formulation
is somewhat misleading, since rather than assuming a war between two
adversaries, Foucault proposes also that "there aren't immediately given
subjects of the struggle, one the proletariat, the other the bourgeoisie".
In the war in which ultimately "we all fight each other", there are only
ever transitory coalitions (p/K 209).
Foucault is well aware of the fact that his conceptualization of
history in terms of struggle is not new as such; foremost it figures of

course in Marxism. While acknowledging this, Foucault points out,
however, that Marxists have more often than not applied the vocabulary
of war and struggle somewhat half-heartedly.He writes, for instance,
that "'dialectics' is a way of evading the always open and hazardous
reality of conflict by reducing it to a Hegelian skeleton" (p/K 114-15),
that "the good old 'logic' of contradiction is no longer sufficient, far
from it, for the unravelling of actual processes" (p/K 164) and that even
though ...
the greatest of the Marxists (starting with Marx himself) insisted on

the 'military' problems (the army as an instrument of the state,
armed insurrection, revolutionary war) ... [when dealing with] ...
the 'class struggle' as the mainspring of history, they focus mainly
on defining class, its boundaries, its membership, but never con-
cretely on the nature of the struggle (pPHC 123).
This shortcoming of Marxism, Foucault suggests, is to be overcome

by replacing the vague notions of "struggle" and "historical contradic-
tion" by a whole network of military terms, and by applying these terms
to a wider variety of cases than just the economic sphere. Thus the
proponent of agonistic eventalisation does not speak merely, or pre-
dominantly, of class-struggle and he does not confine the application of
the concept of struggle to instances of military action among states or
parties in a civil war. Rather society is a battle-site for power-struggles
among individuals, a battle-site for their temporary alliances like interest
groups or classes. Each of these individuals or alliances have their own
unconscious or conscious strategies and tactics, they try to uphold or
bring about a (dis)equilibrium of forces, carry out offensives or counter-
offensives along different lines of penetration, retreat or resist or re-
organize forces, employ or build dispositifs (as ensembles of precautions
and measures for carrying out a strategic-military operation), engage in
investment (as the encircling of a fortress), and strive for their own
victory and the defeat of their antagonists.
In sum, then, Foucauldian recovery of the agonistic events suggests
- although only in broad outlines - a new plot for the study of social
history, the history of science and the history of institutions. To speak of
178 PART II
a plot here, and thus make use once more of a category that figures
centrally in Veyne's aforementioned study, seems useful because it
highlights the fact that this plot functions as a selecting principle within
the myriad of historical facts. Needless to say, there are many other
historical facts that will inevitably remain outside of the genealogical
scope. As a result, agonistic "eventalisation" in particular, and genealo-
gy in general, can make no claims to absolute truth or comprehensive-
ness. Unlike some other methodologists, Foucault is well aware of this
perspectival nature of his enterprise. Thus even though he proposes the
employment of this "itinerary" for the study of the history of society,
Foucault agrees that it is only one among several ways of cutting up this
history. For instance, he allows explicitly for "different strategies (...)
for integrating these unbalanced, heterogeneous, unstable, and tense
force relations" (HS I 93), and proposes that one is naturally driven
towards this plot only when taking the Marxian conception of "struggle"
as one's starting point (p/K 164).
PHYSICALISM
The second crucial research strategy arising from Foucault's Nietzsche-

anism is a focus upon the relations between power and the body. To
speak of Foucault's study of the history of the human body as "physical-
ism" can be justified by more than just his own use of a physicalist
idiom when writing about effects of power on the body. More important-
ly, a specific term for a Foucauldian history of the body seems to be
called for, in order to delimit his interest in, and methodology for, this
topic from the many other, already existing, "histories of the body".
Histories of the body
Indeed, the historical study of the human body is not as such a new or
original endeavour. As Foucault notes himself: "Historians have long
ago began to write the history of the body. They have studied the body
in the field of historical demography or pathology; they have considered
it as the seat of needs and appetites, as the focus of physiological pro-
cesses and metabolisms, as a target for the attacks of germs and viruses;
they have shown to what extent historical processes were involved in
what might seem to be the purely biological bases of existence ... " (DP
25).
One can even question whether this list covers the enormous variety
of studies that have been carried out, either directly under the heading of
"history of the body", or then with obvious relevance to the topic this
heading refers to. To give a more adequate overview, the following
figure (Figure 14) might be helpful.
Histories of the body investigate how ...

human bodies, their shapes, parts, organs, wants, desires or movements
were ...
sick or healthy, nourished, shaped, distributed, affected by diseases or events
in the natural world, treated by oneself or others (trained, punished, ob-
served, studied, distributed, fed, grown, stimulated, dressed, sacrilized,
desacrilized ...)
and/or categorized and/or evaluated and/or constituted as ...
normal or abnormal, useful or useless, efficient or inefficient, docile or
resisting, male or female, worthy or unworthy, sinful or nonsinful, sexually
attractive or unattractive, adult or infantile, physical or spiritual, young or
old, strong or weak, alive or dead, sick or healthy, according to race or
class, ...
and/or related to ...
other bodies, their shapes, parts, organs, wants, desires or movements, God,
the soul, the mind, intentions, ideology, tools, space, time, the state, society
and/or how the body or its pans and movements were used as symbols or
metaphors for, or interpreted by means of ...
social structures, the natural world ...
by/in ...
individuals, groups, classes, societies, professions, churches, sciences,
writers, artists, forms of art, literary genres, periods, social events (e.g.
festivals), political events, institutions ...
Figure 14
180 PART II
To flesh out this definition at least briefly, let us note that different
traditions and studies have dealt with different aspects or ingredients of
this facet sentence. For instance, Robert Mandrou's classic of Annales
scholarship, Introduction to modern France 1500-1640. An essay in
historical psychology (Mandrou 1976), opens with two chapters, "The
body: food and environment" and "The body: health, diseases and
'population'''. These chapters deal with the diet of different social
classes, the malnutrition of large parts of the rural population and the
diseases resulting from it, clothing, living accommodation, different
forms of widespread diseases and the ineffectivity of therapies, over-
population, death and birth rates, life expectancy and attitudes towards
violence and death. In later chapters, Mandrou also pays attention to
emotions, sexuality, conceptions of beauty, manual skills and the repre-
sentation of the human body in pictorial art. Enquiry into these matters,
that also feature repeatedly in the works of other authors of the Annales
School, are seen by Mandrou as contributing to a reconstruction of the
mentality of the period under investigation. In other studies, the Annales
historians have also attended to topics like attitudes towards the bodies
of the dead, physical education and corporeal punishment of children
(Ari~s 1960), or the belief in the Middle Ages according to which
touching the King's body can heal one's diseases (e.g. Bloch 1924).
The classical contributions to the history of the body by Norbert
Elias and Mikhail Bakhtin, must also be mentioned here. In his The
Civilization Process (Elias 1978), Elias describes the emergence of the
modern "closed" human being. This "homo clausus" emerges with the
arising of the feeling of embarrassment or the "Peinlichkeitsgefiihl" that
we experience when coming in close contact with other human bodies,
when seeing others spit, eat ·with their hands, or being nude. Elias
argues that the repression of immediate impulses, and the strict demarca-
tion of one's own body from others, was part and parcel of the individu-
alism and the respect for the other sex that slowly emerged in the
Renaissance. Elias also shows that this rise of "civilization" was linked
to the coming of the modern national state.
Bakhtin's book, Rabelais and His World (Bakhtin 1968), can be read
as depicting the kind of body that fell prey to the historical processes
described by Elias. Bakhtin investigates what he refers to as the "carne-
valesque" or "grotesque" body of the lower classes in medieval times
and the Renaissance, a body that is spontaneous, laughing, disrespectful,

and intimate with others. 39
Finally, and to bring this incomplete list to a provisional end, one
must also mention the numerous studies on the history of medicine and
on images of the body as involved in social and political thought. For
instance, studies carried out in and around the Edinburgh School in the
sociology of knowledge have frequently attended to the ways theorizing
in medicine, phrenology and eugenics related to political events and
class-struggles. For example, the 18th century assumption of feelings of
sympathy, an assumption made by physicians to explain the integration
of the body, was used by the Edinburgh elite as an analogy of society,
and as a resource for justifying the political status quo (Lawrance 1980).
Body, power, knowledge
While there certainly are links between these various studies of the
human body, on the one hand, and the Foucauldian "physicalism", on
the other hand, the latter also has several characteristic features that are
worth attending to in more detail.
Foucault himself sees the novelty of his project in its being based on
the observation that ...
the body is also directly involved in a political field; power relations

have an immediate hold on it; they invest it, mark it, train it, torture
it, force it to carry out tasks, to perform ceremonies, to emit signs
(DP 25).
Foucault's project in the study of the body is thus an investigation

into the relations between the body and power, or, more precisely, the
relations between social power, (scientific) knowledge, the human body
and personal identity.
This historical-political study of the body is motivated, above all, by
two ideas. On the one hand, it seems natural - given Foucault's activities
on behalf of prisoners, the mentally ill and hospitalized people - to
attribute to him the view that it is first and foremost the concrete physi-
cal suffering of human beings that a critical reflection of modern society
182 PART II
must attend to both politically and by way of historical study. On the

other hand, Foucault's interest in the triad of body-knowledge-power is
also motivated by a criticism of the prioritizing of "ideology" in Marxist
analyses of society. In one place, Foucault suggests that it might be
"more materialist" to study the effects of power on the body rather than
to focus exclusively on questions of ideology (p1K 58). In other places,
he goes further and questions the usefulness of the notion of ideology
altogether. Thus he claims, first, that the concept of ideology presup-
poses a problematical opposition with truth, second, that it is based on
a traditional notion of the subject, and third, that it is taken as distinct
from, and seen as determined by, an economic basis or infrastructure
(p1K 118). Rather than assuming naively ~ la Althusser the existence of
a non-ideological, scientific truth as the opposite of ideology, Foucault
holds that we have to study how "effects of truth are produced within
discourses"; rather than presupposing "a human subject on the lines of
the model provided by classical philosophy, endowed with a conscious-
ness which power is then thought to seize on" (p1K 58), we have to
study the creation or formation of this subject; and rather than trying to
explain all power as ultimately deriving from the economic infrastruc-
ture, we have to diversify our -explanatory models (p1K 118).
The deeper reason why Foucault finds an analysis in terms of
ideology questionable, and why he also denies the link between his
physicalism and the study of mentalities, is his belief that - especially in
the modem world - mechanisms of power have evolved that do not
necessarily involve conscious, mental representations in the power
subject. Instead, these new forms of manipulation make the body move
in specific ways without the power subject being able or forced to
deliberate on these movements. Moreover, given these new forms of
manipulation, even if the power subject deliberates on her actions, this
deliberation has already been influenced and shaped by systems of
control and surveillance over her body.
Since the details of this analysis belong more properly to a critical
review of Foucault's historical investigations, let us tum to his concep-
tual apparatus for investigating the relation between power and bodies.
Now, it is obvious that coercion has a rather intimate relation to the
body. Not only are many coercive threats directed against the body of
the power subject, but the power subject is also usually directly physi-
cally confronted with the coercer. Both coercer and coerced are typically
seen by one another, and of course the coercer must make the coerced
believe that any deviating from the demanded course of action will be
observed immediately. Furthermore, every modem society has a system
of coercive institutions, like the police or the prison, a system to guaran-
tee and enforce the punishment of lawbreakers. And, needless to say, a
considerable number of punishments are corporal, in that they confine
the body to a cell, threaten it with bodily pain, or destroy the body's
life.
Foucault's investigations into the penal systems of different histori-
cal periods suggest that these can be studied as so many "regimes" or
"modalities" of bodily confinement and the infliction of pain. Generaliz-
ing one of the notions that Foucault only applies to the modem form of
power over the body, we might say that every society has a specific
"political anatomy" (OP 138), a specific way of relating to the bodies of
different groups, sexes, classes, and age groups, and that this "anatomy"
comes out most clearly in the- way society punishes members of these
different categories.
Some passages of Discipline and Punish also suggest that these
regimes can be measured along various axes. A first such axis is the
distinction between two directions of visibility: some systems of power
rely more on the threatening visibility of the power holder, others
employ techniques of surveillance that leave the power holder by and
large invisible. 40 Second, systems of control also vary as to their
"scale" (OP 136) or degree of "resolution" (p1K 151): some force and
bind the body as a whole, as in the case of slavery, while others exercise
"upon it a subtle coercion, ... obtaining it at the level of the mechanism
itself - movements, gestures, attitudes, rapidity: an infinitesimal power
over the active body" (OP 137). Third, power over the body can be
exercised by inflicting pain, by training or by stimulating it. 41
For Foucault, various systems of "political anatomy" need to be
studied by attending to the kinds of knowledge or truth they presuppose
and produce. In fact, any study of the body-power relation that leaves
out this further variable of knowledge is insufficient from the standpoint
of genealogy, since it is, as will be remembered, one of the most basic
assumptions of the Foucauldian theory of power, that power is always
interrelated with knowledge. Neither can it be claimed that the interrela-
184 PART II
tion between power over the body and knowledge is merely a peculiarity
of modem societies with their specific forms of surveillance and control.
After all, to mention only one example that Discipline and Punish
discusses in some detail, judicial torture as the inflicting of extreme pain
on the suspect's body, obviously was, and was conceived of as, a search
for truth (OP 37).
The genealogical questions with respect to the triad of knowledge-
power-body fall, roughly, into three groups.
The first is concerned with the know-how or 'practical' knowledge
involved in systems of bodily control and punishment. This knowledge
is a "technological" knowledge of power mechanisms, a knowledge of
measures by which bodies are most effectively treated, by which pain is
most efficiently applied, and by which the body is best trained, disci-
plined, stimulated and controlled. This technological knowledge is
knowledge on how to intensify power over the body, how to lower the
costs of power exercises, how to standardize the means of power, and
how to increase the amount of power. It is knowledge of how the power
of the body qua capacity is increased and of how the power of the body
qua resource of resistance is diminished (OP 138). Of course, this
knowledge is not only presupposed by systems of bodily surveillance and
control, it is also refined and developed within such systems by trial and
error.
Foucault usually speaks of this "political technology of the body"
(OP 26) when dealing with the specific form of technological knowledge
involved in more recent forms of disciplining. However, his own study
of the knowledge of bodily pain involved in earlier systems of torture
suggests that this term can also be applied more widely. What makes this
generalization even more natural, is Foucault's view that this knowledge
is below the threshold of systematic science, to wit, that it is "diffuse"
... rarely formulated in continuous, systematic discourse; it is often

made up of bits and pieces; it implements a disparate set of tools or
methods. In spite of the coherence of its results, it is generally no
more than a multiform instrumentation. Moreover, it cannot be
localized in a particular type of institution or state apparatus. For
they have recourse to it; they use, select or impose certain of its
methods (OP 26).
The second group of genealogical question with respect to the body,

power and knowledge, focusses on scientific knowledge about the body,
its movements and functions. Genealogy asks how this scientific knowl-
edge is related to systems of control and the technological knowledge
behind them. Here Foucault's well-known thesis is that knowledge of the
human sciences as we know them, and especially their methods, were
made possible in part by "the great nineteenth century effort in discipline
and normalization" (p1K 61). This is especially clear in the case of
medical knowledge, one condition of which were new methods of
observation and control employed in the hospital. Here poor patients
were - and still are - used for medical experiments, and patients are
constantly put under surveillance. For instance, they are distributed
according to their symptoms and diseases in order to facilitate the
accumulation of medical truths. Furthermore, often scientific knowledge
of the body feeds back into the systems of control and technological
knowledge from which they emerge; cases in point are doctors working
as consultants in factories and prisons, as torture-instructors, or as
technical advisors in executions (pPHC 193).42
Adding to Foucault's suggestions, we can also note that again we
are not dealing with a totally new phenomenon: the interrelation between
power, body and knowledge is not confined to the modern era; after all,
cruel medical experiments on the living bodies of convicts were per-
formed way back in the third century B.C. in Alexandria by Herophilos
and Erasistratos - not even to mention the myriads of animals that have
been tortured to death over the last two millenniums in the name of
scientific progress. 43
Third, and finally, the genealogical study of the history of the body
as caught in the networks of science and social power, also questions the
body in its status as a prediscursive, natural object. That is to say,
Foucault seeks to lay bare the ways in which the body, its pleasures and
its sexuality have been constituted by systems of power and knowledge.
In other words, Foucauldian physicalism, here once more quite con-
genial to Faraday's enterprise, does not take bodies as given, and does
not allow for any clear line between the biological and the historical:
... what is needed is to make it [Le. the body] visible through an

analysis in which the biological and the historical are not consecu-
186 PART II
tive to one another, as in the evolutionism of the first sociologists,

but are bound together in an increasingly complex fashion in accord-
ance with the development of the modem technologies of power that
take life as their objective. Hence I do not envisage a 'history of
mentalities' that would take into account of bodies only through the
manner in which they have been perceived and given meaning and
value; but a 'history of bodies' and the manner in which what is
most material and most vital in them has been invested (p1K 152).
To tackle this problem is, for instance, to study the emergence of

the "sexualized bodies" (and the differences between male and female
bodies) in and through the interplay of various discourses (sexology,
medicine, eugenics, demography, pedagogy, jurisprudence, psychiatry,
obstetrics ...), methods of control and surveillance (in boarding schools,
birth clinics, army camps ...), as well as techniques of stimulation
(pornographic literature, films ...).
DISTRU~ AND IRONY
Turning finally to the third main genealogical research strategy, we can

start from the observation that skepticism with respect, for instance, to
the notions of man, progress or mental illness is of course a well-known
feature of Foucault's work throughout its different stages. More general-
Iy, Foucault cautions us against the tendency to assume that concepts or
words like 'madness' or 'deviance' denote natural objects; he asks us to
see these objects as resulting from prior objectivations. 44 I shall be
dealing with this skepticism or attitude of suspicion in the next chapter
under the heading of Foucault's "(methodological) perspectivism". Here
I rather want to take up Foucault's Nietzschean Leitmotifof distrusting,
parodizing or ironicising self-proclaimed purely scientific and humani-
tarian interests or motivations, i.e. his attention to non-scientific interests
in analyses of the historical emergence of the human sciences and
coercive institutions.
For instance, cases in point are Foucault's arguments to show that
the criticism of the prison reform movement in the 19th century was less
motivated by humanitarian concerns than by the interest to bring about
a more effective penal system45, and Foucault's attempts to demon-

strate that the control of children's sexuality was as much motivated by
doctors' and parents' interest in control and sex as it was motivated by
medical concerns. 4'
Interests
Foucault's emphasis on such non-humanitarian and non-scientific in-

terests and power structures is of course congenial with recent work in
the sociology of science, especially the studies carried out by the Edin-
burgh School. To focus more closely on these parallels, suffice it here to
take a closer look at the lecture "The dangerous individual" (01). In this
paper Foucault's strategy of explaining scientific and social development
via interests can clearly be seen. 47
In "The dangerous individual" Foucault investigates why crime in
the 19th century became increasingly medicalized or connected to mental
disease. He argues that a first step in this direction was taken when, in
the 1830's and 40's, psychiatrists began to publish widely on the phe-
nomena of "the ultimate crime", Le. cruel, terrible murders allegedly
without visible motivation or reason. Psychiatrists in part helped to
create, in part enrolled, the publics' as well as the judges', awareness for
these crimes, offering their own profession of psychiatry as promising to
provide the key to their solution.
Psychiatrists exploited the public fear of, and the judges' perplexity
over, these crimes by introducing the notion of "homicidal monomania" ,
Le. the notion of a form of insanity whose only visible expression or
symptom is homicide. As Foucault puts it: "Nineteenth-century psychi-
atry invented an entirely fictitious entity, a crime which is insanity, a
crime which is nothing but insanity, an insanity which is nothing but
crime" (01 132). More precisely, the nature of this form of insanity was
claimed to be such that only the psychiatrists could eventually come to
recognize its almost invisible symptoms.
It was by redefining the ultimate crime, a crime linked inextricably
to death and thus to fear, that psychiatrists went a long way in convinc-
ing the public as well as many legal experts of their competence in the
realm of crime and punishment. It was by redefining the cruel, motive-
188 PART II
less homicide in psychiatric terms that psychiatrists gained a position of

social power that they had not had before.
However, Foucault cautions us not to attend only to the professional
vested interests of the psychiatrists: "It is not enough to invoke some
sort of imperialism on the part of psychiatrists seeking a new domain for
themselves ... " (01 133). In order to understand why they could suc-
ceed, one has to take into account the earlier success of the medical
profession in convincing politically powerful groups, like industry, the
army and the police, of its own importance as an expert in public
hygiene, Le. as an expert in treating of the diseases of the newly invent-
ed category of the "population". Psychiatrists thus did not have to start
their own struggle for a larger share of social power from scratch. The
medical profession, from which psychiatrists had just evolved as a more
or less independent field, had provided the latter with a strategically
important basis. Psychiatrists merely needed to extend the new inter-
pretation of the doctors' socially crucial role to a new domain.
Moreover, the psychiatrists' interest in penetrating the legal appara-
tus was not independent of their other professional interests. The atten-
tion of psychiatrists to this whole issue was indeed connected to their
concern for justifying their right to impose a therapeutic confinement
upon the mentally ill. Homicidal mania merely worked as one case
where the public's fear and the judges' helplessness could be translated
into the need for such confinement.
Why then were the judges unable to defend their traditional power
of being the sole adjudicators of punishment? This too is an issue Fou-
cault proposes to address in terms of power. He argues that during the
18th century the traditional system of punishment as based on revenge
and deterrence had become inefficient as a system of power and had
been replaced by a new system based on panoptical surveillance and
reforming of the offender. The new focus on the individual, a focus that
came hand in hand with the new power mechanisms of surveillance,
confession and discipline, shifted awareness from the criminal act to the
criminal individual. In determining punishments, the judiciary was thus
forced to determine a punishment by considering how dangerous the
individual in question was and how he or she could be re-educated.
However, precisely in the case of the unmotivated, unreasonable mur-
derer this was impossible. In short, the psychiatrists had chosen the right
target for their attack.

The notion of the dangerous individual which was thus emerging
could then be exploited by the anthropology of criminal man in Italy and
the theory of social defense in Belgium. Once psychiatrists had suc-
ceeded in firmly establishing their hold on the legal apparatus towards
the end of the 19th century, they no longer needed the notion of homi-
cidal monomania. In fact, they could - and had to - give up this notion,
in order to further expand their area of influence. Relying on a more
refined picture of mental disease as developing by degrees, on the one
hand, and on the notion of degeneration, on the other hand, psychiatrists
could establish their professional competence with respect to ever more
forms of criminality. Psychiatry "enlarged, organized, and codified the
suspicion and the locating of dangerous individuals, from the rare and
monstrous figure of the monomaniac to the common everyday figure of
the degenerate, of the pervert, of the constitutionally unbalanced, of the
immature, etc." (01 149).
Even this brief sketch of a much more involved argument suffices
for us to see that Foucault's treatment of interests and power has strong
similarities with the use of interests as explanatory devices in the Edin-
burgh type of sociology of science. Interests focussed upon in the latter
are, foremost, (1) "professional nested interests" as scientists' interests
to (re-)define object areas for the purpose of establishing their (exclu-
sive) competence in dealing with them; (2) scientists' interests in keeping
laymen out of the scientific field; (3) scientists' interests in using concep-
tual tools drawn from the broader cultural or social context (in order to
arouse wider attention for their work); (4) interests in the wider society
to make scientific results effectively applicable; and (5) scientists'
interests in drawing support from classes or institutions (Shapin 1982).
All of these interests are also touched on by Foucault in various of
his works, even though not always under the same heading of interests.
It is especially interesting to note that Foucault agrees with these sociolo-
gists not only in claiming that straightforward explanations in terms of
classes are often rather questionable, but also in proposing that profes-
sional nested interests play a crucial role and that scientific work often
is dependent upon the support it manages to draw from state institutions,
industry or the Church.
More important than these parallels, however, is the fact that
190 PART II
Foucault's account summarized above is not open to one specific charge

that has been raised against the Edinburgh program. According to this
criticism, authors of the Edinburgh School neglect the fact that the
interests of the wider society which scientists enroll for their purposes
are often not pregiven, but rather produced by the scientists themselves.
This point has been especially convincingly treated by Bruno Latour in
his study of Pasteur (Latour 1983; cf. Latour 1987, 1988). For him, the
art of using interests is the art of "convincing others of what their
interests are and what they ought to want and to be. He who is able to
translate others' interests into his own language carries the day" (1983:
144). In Pasteur's case the interest of the public was "a consequence and
not a cause of Pasteur's efforts to translate what they want or what he
makes them want. They have no a priori reason to be interested at all,
but Pasteur has found them more than one reason" (ibid., 144).
Of course, a similar line is involved not only in Foucault's dealing
with the arising of the category of "the dangerous individual", but also
in Discipline or Punish or the History of Sexuality I. After all, the way
to success for psychiatrists was to translate the interest of the public to
be freed from the fear of death and crime into the interest both in the
confinement of various allegedly insane individuals and in the work of
psychiatrists themselves. The psychiatrists had of course helped to bring
about this fear by writing extensively on these special types of crime.
There is also a noteworthy parallel between Latour's rejection of the
distinction between the inside and the outside of the laboratory and
Foucault's attention in the two:-way borrowing of mechanisms of power
(and the knowledge coming with them) between coercive institutions and
the society at large. Latour deconstructs the border between the labora-
tory and the wider society by arguing, first, that experimentators typi-
cally translate "outside" problems into laboratory settings, and, second,
that the application of laboratory results to the outside world is, in fact,
an extension of laboratory conditions to the outside: " ... on the condition
that you respect a limited set of laboratory practices - disinfection,
cleanliness, conservation, inoculation gesture, timing and recording -
you can extend to every French farm a laboratory product made at
Pasteur's lab" (ibid., 152). In Foucault, the Latourian laboratory is
replaced by coercive institutions, but otherwise the line of thinking is
similar: coercive institutions are the social laboratories for new forms of
control and discipline (and the new forms of knowledge that support, or
arise from, these power mechanisms) and the results of these social
experiments are extended to ever new social domains. For instance,
Foucault argues that the effectiveness and docility of the worker is
achieved by transposing the power mechanisms of the prison and the
army to the workplace.
We might perhaps also claim that Foucault holds a counterpart of
Latour's thesis that "most of the time when we talk about the outside
world we are simply taking for granted the prior extension of a former
science built on the same principle as the one we are studying" (ibid.,
156). Translated into Foucault's framework this would amount to the
claim that the outside world is but a network of power, in part modelled
on some earlier coercive institutions. In any case, these two French
authors certainly concur that "in our modem societies most of the really
fresh power comes from sciences - no matter which - and not from the
classical political process" (ibid., 168).48
Ironic hypotheses
Replacing the standard picture of, say, psychiatry as motivated through

and through by a humanitarian concern, with a picture of psychiatrists'
as being motivated by interests in power and manipulation, has, of
course, something of the burlesque, the parodic or the ironic. After all,
what Foucault - and the sociologists - are, in the eyes of many scientists
and philosophers, engaged in, is an undue replacement of the dignified
by the 'low' and 'familiar'. Hardly surprising then that their writings
and ideas should meet with so -much resistance.
Foucault has been aware of this ironic or parodic element of his
work from the beginning; not unexpectedly for a student of Nietzsche
rather than of Mannheim, Merton or Habermas. 49 This element is
rather obvious not only in some of Foucault's asides or interviews ...
Why should so many people, including psychiatrists, believe that I

am an anti-psychiatrist? It's because they are not able to accept the
real history of their institutions which is, of course, a sign of psy-
chiatry being a pseudo-science. A real science is able to accept even
the shameful, dirty stories of its beginning (pPHC 15).
192 PART II
... but also in Foucault's research strategies. After all, Foucault asks us
to be aware of developments in and through which attempts to increase
the efficiency and amount of power backfire: the medical discourse on
homosexuality provides homosexuals with a vocabulary to articulate their
rights; the controlling of sexuality leads to its stimulation; and the
controllers become themselves controlled (the Panopticon). These
invisible hand effects, by themselves, provide a burlesque commentary
on the work of the architects of power.
In any case, it seems fair to say that Foucault favors the methodolo-
gical strategy of attempting to proceed by using, what one might term,
'ironic hypotheses'. When confronted with an institution whose openly
proclaimed purpose is to bring about, or hinder, or uphold a set of states
of affairs, say S, Foucault will suggest working on the opposite, ironic,
assumption that the real purpose of the institution was the hindering of
S (rather than bringing S about), or the bringing about of S (rather than
its hindering), or changing S (rather than upholding it). Just think here
of Foucault's hypotheses that the prison's function was to produce rather
than prevent crime, or that the medical classification of perversions was
an incentive rather than a rejection of these 'deviations'.
Finally, we might note also the reflexivity of Foucault's irony, Le.
his conscious attempt to make the reader aware of the transient and
provisional character of his own models and conceptions. Foucault's
writings do not employ a single framework, but from Madness and
Civilization up to the latter parts of the History of Sexuality, we are in
fact confronted with several vocabularies. Moreover, as seen earlier,
Foucault uses various different models for conceptualizing power, and
some of them are clearly incompatible (war, politics, physics, individ-
ualist). Thus, for Foucault, there is not one single framework for the
study of power, but rather there are "different strategies (...) for in-
tegrating these unbalanced, heterogeneous, unstable and tense force rela-
tions" (HS 193). Perhaps Foucault's reminder that "1 have never written
anything but fictions", a reminder that is qualified by the addition that "1
do not mean to say, however, that truth is therefore absent" (PIK 193),
is best understood in this sense, too.
13. GENEALOGICAL PERSPECTIVISM
INTRODUCTION
The fourth Nietzschean Leitmotif of Foucault's genealogical research,

Le. perspectivism, is important and complex enough to be best treated in
a separate chapter. Admittedly, perspectivism is a rather vague notion.
Even in the case of Nietzsche himself, interpreters use this concept in
different ways. While they agree that for Nietzsche all knowledge is
relative to perspectives, interpreters disagree over the implications of this
view. According to some Nietzsche scholars, his perspectivism does not
rule out the possibility of speaking meaningfully of truth and of evalu-
ating different perspectives in some non-arbitrary fashion (Nehamas
1985: 64-73). Others hold that Nietzsche rules out this possibility: for
them, perspectivism amounts to the doctrine that since everything is
relative to perspectives, we should not even use the concepts of truth and
falsehood (MacIntyre 1988: 352).
In this chapter, I shall use perspectivism as an umbrella-notion. I
outline Foucault's position with respect to such issues as constructivism,
skepticism and relativism, and then simply label his specific views on
these notions "perspectivism".
In dealing with relativism, skepticism, and constructivism, we enter
a rather complex and emotionally loaded area of philosophical debate.
What makes a discussion of these concepts difficult is not only the
vagueness of their meaning, but also, and following from this vagueness,
the fact that the positions these concepts denote seem praiseworthy to
some philosophers, but utterly distasteful, self-defeating and dangerous
to others. Often, for philosophers who take the latter view, there is not
much, for instance, that distinguishes relativism and skepticism from,
say, "total skepticism", "nihilism", or "irrationalism".
In suggesting that Foucault adopts some versions of relativism,
constructivism and skepticism, I thus prima facie discredit his work in
the eyes of more than just a few readers. However, rather than joining
them in rejecting Foucault's thought, I shall try to convince them that
relativism, constructivism and skepticism are not absurd positions. In
particular, I shall try to argue that epistemological relativism is not a
self-refuting doctrine, that methodological anti-realist constructivism is
194 PART II
not self-defeating, and that skepticism is coherent in both a limited and

an unlimited form.
SKEPTICISM
Foucault has been alternately praised and attacked for being a skeptic.
For instance, John Rajchman calls Foucault "the great skeptic of our
times" and endorses Foucault's thought as one that aims "for the free-
dom of withholding judgment on philosophical dogmas, and so of
acquiring relief from the restrictions they introduce into our lives and
our thought" (Rajchman 1985: 2). Some critics of Foucault, however,
see his skeptical position as self-refuting in so far as they regard his
alleged denial of all knowledge as pre-emptying the possibility of claim-
ing truth for his own work (e.g. Merquior 1985: 147; Merquior 1986:
259).
Against the latter position it is worth stressing that even if Foucault
were to be the universal skeptic these critics make him out to be, his
position need not be self-refuting. Indeed, it is possible to provide a
description of genealogical studies that avoids the charge of self-refuta-
tion with respect to those studies. Foucault could be a universal skeptic
and still write historical studies meant to undermine, case by case, the
central standards and categories of philosophy, psychiatry and the human
and social sciences. If we take his studies as being addressed to the
'dogmatic' proponent of these standards and categories, their employ-
ment by Foucault would not involve him in a contradiction: his use of
these standards and categories would only be a hypothetical and tentative
one. Foucault's arguments, premisses and conclusions, would be argu-
ments, premisses and conclusions for the proponent of these standards,
meant to convince these proponents that it is best to withhold judgement
on these standards.
However, charges of self-refutation aside, Foucault's skepticism is
not the universal skepticism that some of his critics accuse him of. To
see this, we only have to note, for example, that Foucault never, neither
directly nor implicitly, suggests that there is no knowledge, no progress,
and no truth. To be sure, Foucault does endorse skepticism in the
following passage:
1 adopt the methodological precaution and the radical but unaggres-

sive skepticism which makes it a principle not to regard the point in
time where we are now standing as the outcome of a teleological
progression which would be one's business to reconstruct histori-
cally: that skepticism regarding ourselves and what we are, our here
and now. which prevents one from assuming that what we have is
better than - or more than - in the past (p1K 49).
Note, however, that Foucault hastens to add: "1 don't say that humanity
doesn't progress. 1 say that it is a bad method to pose the problem as:
'How it is that we have progressed?''' (p1K 50).
Furthermore, Foucault's genealogical texts abound with defenses of
"subjugated knowledge" as well as with claims like "1 know that knowl-
edge can transform us". "knowledge is for me that which must function
as a protection of individual existence", and "I'm an empiricist" (pPHC
7, 14, 106).
Foucault's skepticism is thus at most a limited one. His skepticism
is first and foremost a skepticism with respect to the notions that have
been used for writing the history of the last three-hundred years. These
are the notions of an ever-increasing welfare and freedom. of the grow-
ing humanitarian care for the sick and the mad. of the education of ever
wider strata of the population. of our freeing ourselves from the suppres-
sion of sexuality, of the emergence of a scientific spirit that increasingly
frees itself from structures of power and domination. and of a liberating
philosophical discourse.
Foucault's limited skepticism in these mentioned cases does not
amount to denying that progress has. in fact. taken place. Rather his
skepticism grows from the disquieting awareness that, to put it in terms
ofWittgenstein's favorite quotation from Nestroy, "it is the nature of all
progress that it looks much greater than it really is"so. Foucault - like
e.g. Adorno, Horkheimer and Wittgenstein - stresses that each progress
has demanded a high price, and that often the progress has been mea-
sured one-sidedly. For instance. while it is true that the application of
the death penalty has been abandoned in some Western countries, to
praise this as a victory of humanitarianism is to overlook that the death
penalty is still applied on a large scale in the United States as well as in
more than one-hundred countries around the world. Furthermore, in the
196 PART II
West, the extensive application of capital punishment in the case of a

few major crimes has been replaced by a new system of punishment
which has enormously increased the range of punishable offenses and has
introduced large-scale imprisonment. As documented by Foucault and
others, these developments have led, and still lead, to the permanent
exclusion of large numbers of individuals from society.
To repeat, if Foucault can be said to suggest a withholding of
assent, this withholding does not amount to a withholding of assent with
respect to all knowledge claims. It would be more accurate to say that
Foucault wishes to withhold assent or unqualified approval with respect
to the standard narrative of progress, and, especially, with respect to the
normative conceptions that come with this plot.
Foucault's skepticism can be seen most clearly in his treatment, or
non-treatment, of normative ethical or political notions, a fact that
reflects itself in Foucault's approval of the label of a "skeptical ethics"
(1977: 67) for his work. Since he claims to have shown that many of the
notions that are used to justify progressive movements have actually been
shaped by the very "regimes of truth" and "regimes of practices" they
are meant to undermine, he wishes to abstain from using or endorsing
them in a systematic manner. Rather than building on these traditional
standards of ethics and politics, Foucault follows the classical skeptic in
turning to the immediate experience: in his case, to the suffering of the
allegedly mad, the prisoners, and the victims of medical tutelage.
CONSTRUCTIVISM
The second key ingredient of genealogical perspectivism, "constructiv-

ism", brings genealogy into close contact with more recent relativist
sociology of science. Two forms of constructivism, Le. the doctrine that
scientific facts result from human activity, can be distinguished. "Realist
constructivism" holds that even though scientific facts are "remade" in,
say, the laboratory, these facts are still "discovered" and determined by
reality. According to "irrealist constructivism", scientific facts are only
or merely the result of human activity. Facts are nothing but fabricated,
and they are not discovered. Nature is the result of construction, and
puts no constraints on beliefs and theories (Hacking 1988; cf. Knorr-
Cetina 1983). For the irrealist, it makes no sense to speak of nature or

reality as determining an experimental outcome, since the very categories
of nature or reality are themselves taken by the irrealist to be mere
products of scientific negotiations. As Bruno Latour and Steven WooI-
gar, two somewhat ambivalent proponents of this position, put it:
"Reality" cannot be used to explain why a statement becomes a fact,

since it is only after it has become a fact that the effect of reality is
obtained (1986: 180).
It is not simply that phenomena depend on certain material instru-
mentation; rather the phenomena are thoroughly constituted by the
material setting of the laboratory. The artificial reality, which
participants describe in terms of an objective entity, has in fact been
constructed ... (ibid., 64).
In other places, however, Latour and Woolgar bend more towards

realistic constructivism; e.g. at one point they write that "we do not
want to say that facts do not exist and that there is no such thing as
reality" (ibid., 180).
The leading figure of the Bath School in the sociology of scientific
knowledge, Harry Collins, urges sociologists of science to adopt the
methodological standpoint that "the natural world in no way constrains
what is believed to be" (1982: 140). Here nature has dropped out
completely as a determining factor of knowledge, and as the determining
factor of experimental outcomes.
Collins calls his own approach "relativism" . This is adequate, since
he, like Latour and Woolgar, and other sociologists of science more or
less loosely associated with the Edinburgh School, sees scientific knowl-
edge as relative to the partly differing standards of various scientific
cultures or groups. Be it noted, however, that irrealism, as formulated
by Collins above, need not go hand in hand with epistemological rela-
tivism. Even if reality does not determine our beliefs, it is still possible
that there are universal standards of rationality or ways of categorizing
which determine a singular "true" conception of the world. Analogously
to the distinction between universalist and relativist nominalism, one
might thus also distinguish between a universalist and a relativistic
irrealism. This move is inviting too because nominalism can be looked
198 PART II
upon as a special case of irrealism: irrealism with respect to categori-

zations.
To show that Foucault adopts irrealism (and thus eo ipso nomi-
nalism) at least as methodological points of view, is not especially
difficult. After all, already in his archaeology Foucault studied how
objects like madness are constituted in and through a number of contin-
gent processes. In other words, madness is not something waiting 'out
there' in nature or society to be discovered, but it can be discovered
only once it has been constituted in, by and through various practices
and struggles.
Foucault also states straightforwardly in several passages that his
interest is not to study how nature determines our knowledge, but rather
to investigate the processes of how certain 'natural' objects or categories
come into existence in and through both discursive and non-discursive
practices. Recall, for instance, his earlier quoted insistence that we
"should try to discover how it is that subjects [and by the same token,
the categorizations of subjects] are ... constituted through a multiplicity
of organisms, forces, energies, materials, desires, thoughts etc." (p/K
97). He also tells us that we must study scientific discourse as "a prac-
tice that we impose upon" nature (LD 55), and that truth is to be studied
not so much as a correspondence with reality but rather as "a system of
ordered procedures for the production, regulation, distribution, circula-
tion and operation of statements" (p1K 133). Thus where the sociologists
of science wish to study how statements put forward in the laboratory
are eventually turned into facts such that "all traces of production are
made extremely difficult to detect" (Latour and Woolgar 1986: 176),
Foucault is involved in a similar project with respect to central concepts
and "facts" of human "nature". Furthermore, where Latour and Wool-
gar seek to demonstrate "how a hard fact can be sociologically decon-
structed" (ibid., 107), Foucault attempts a similar feat by genealogical-
historical means.
While there thus can be little doubt about the question whether or
not irrealism is part and parcel of genealogy, it is worth pointing out that
Foucault is careful not to propose irrealism as a global standpoint with
respect to all sciences. He seems in fact to take irrealism foremost as a
methodological, heuristic device, Le. he studies practices and scientific
discourses as if in their constituting activity these were not constrained
by nature. Foucault neither claims that scientists and social practices are
free to construct any object or category they like, nor does he ever imply
that the construction of madness is not - in some way or other - related
to brain-damage and types of unusual behavior (cf. Veyne 1979: 47).
The reason why he need not take these 'objective facts' into account
from his starting point should be clear enough: in the discursive and
non-discursive interactions that Foucault studies, these 'objective'
features inevitably appear only as already interpreted.
To be sure, if Foucault were to endorse irrealism on a total scale,
then his own studies could be nothing but fairy-tales: if what is 'out
there' in nature and society does not determine our theories and beliefs
at all, we would have no reason to believe Foucault's own story of what
is 'out there' in society and history.
Fortunately, Foucault's position is, to repeat, more moderate. What
his methodological irrealism amounts to is merely the Canguilhemian
idea according to which "the object of the history of science [as a dis-
cipline] has nothing to do with the object of science". That is to say, the
object of the history of science is science as a social community or as a
succession of bundles of theories, whereas the object of science is nature
or society. While genealogy and the sociology of science bracket the
latter, they do not bracket the former. In other words, the genealogist
does not leave "the natural attitude" when studying the processes in and
by which science and apparatuses constitute objects. But he gives up this
natural stance and practices an epoche (in both the skeptical and Hus-
serI's sense) with respect to those objects that the scientist takes for real.
It is also possible, of course, to iterate this move from the natural
attitude to the genealogical one. We can not only study how sciences
constitute their objects; we could also investigate how, in turn, genealo-
gy constitutes its subject matters of disciplines and apparatuses. The only
move that is indeed a likely candidate for impossibility is the attempt to
do both investigations at one and the same time. Be it said also, to
conclude, that the infinite regress which opens up here can hardly be
used as an argument against the genealogical enterprise: after all, the
distinguishing feature of human action and thinking is precisely the fact
that it inevitably contains or implies reflexivity. 51
200 PART II
EPISTEMOLOGICAL RELATIVISM
Before taking up the question of what the third ingredient of genealogical

perspectivism, Foucault's epistemological relativism, amounts to, I want
to address the more basic question of whether charging a thinker with
being an epistemological relativist is to provide a decisive argument
against him or her. The reason why the vast majority of philosophers
and social scientists would answer in the positive is, obviously, the wide-
spread belief that epistemological relativism is a self-refuting, incoherent
standpoint. In this and the following section, I wish to suggest that this
received view is mistaken.
In order to establish this position, I shall, admittedly brief, take a
closer look at two recent books, one by Harvey Siegel and one by
Alaisdair MacIntyre, books that set out to refute epistemological rela-
tivism. Of course there are numerous other authors who have attempted
a similar feat. Restricting myself to these two writers can be justified,
however, by noting that since Plato there have been few innovations in
the reasoning against relativism and that the arguments presented by
Siegel and MacIntyre can be seen as culmination points of this long
tradition of 'refutations'.
"Relativism Refuted" - refuted
Siegel, in his book Relativism Refuted. A Critique of Contemporary

Epistemological Relativism (1987), takes epistemic relativism (= ER) to
be the following doctrine:
ER: For any knowledge-claim p, p can be evaluated (assessed,

established, etc.) only according to (with reference to) one or
another set of background principles and standards of evaluation
SI' ... Sll; and, given a different set (or sets) of background
principles and standards S'I' ... S '", there is no neutral (that is,
neutral with respect to the two (or more) alternative sets of
principles and standards) way of choosing between the two (or
more) alternative sets in evaluating p with respect to truth or
rational justification. p's truth and rational justifiability are
relative to the standards used in evaluating p (ibid., 6).
Let us look more closely at Siegel's three main arguments against

ER. The first two arguments reformulate, in more exact terms, ideas
already employed by Plato's Socrates against Protagorean relativism.
(1) Both Plato and Siegel hold that epistemic relativism is incoherent
"because if it is right, the very notion of rightness is undermined, in
which case relativism cannot be right" (ibid., 4). Siegel's point is that
the non-relativist is entitled to demand from the epistemic relativist
"right reasons" for the latter's position, where by "right reasons" Siegel
understands reasons that are not relative, Le. reasons based on a "neutral
(with respect to the presuppositions of relativists and non-relativists)
framework" (ibid., 8). Siegel's twist then is simplicity itself: if the
relativist provides these "right reasons", she has ipso facto transcended
his relativism and thus has refuted herself, since, according to epistemo-
logical relativism, there are no non-relative reasons. If, on the other
hand, the relativist does not provide any non-relative, but merely relative
reasons, the absolutist is not obliged to adopt relativism, since for her,
"'relative rightness' is no rightness at all" (ibid., 8). Siegel holds that
relativism, in taking all standards to be relative, makes it impossible to
provide any objective non-relative standards in terms of which epistemo-
logical positions - and among them relativism - can be evaluated.
It is fairly easy to show that Siegel's first argument fails to demon-
strate the incoherence of epistemological relativism. No relativist needs
to be impressed by Siegel's line of reasoning, for all it shows is that
relativism allows only for relative reasons, and that relativism is the
denial of the existence of a universally valid framework. But to show
this is not to refute relativism, it is merely to restate its content. To
answer the relativist challenge, it is insufficient to tell the relativists what
they already have been saying all along; if the absolutist wants to refute
this position she better point out to the relativists just what a list (pre-
ferably a longish list) of non-relative grounds for deciding epistemol-
ogical issues looks like. This, however, Siegel fails to do.
Let us note also that Siegel's claim, according to which relative
rightness is no rightness at all, is not very convincing. For instance,
right moves in the game of chess are certainly right only relative to the
game, yet they are right all the same.
Finally, we can also point out against Siegel that the giving of
relatively-right arguments for relativism is not as odd and insufficient as
202 PART II
he claims. Arguments for relativism are not per se "impotent", and

adopting the relativist position need not be the totally arbitrary decision
that Siegel makes it out to be. (" ... the relativist can adopt beliefs (and
act) only arbitrarily" (ibid., 21).) Given that both the framework or
culture relativist and her critic, Le. the absolutist, share the same frame-
work and culture, arguing for the relativist position can have force, and
adopting the relativist stand is not merely up to the individual. Given the
shared framework (shared with their absolutist opponent) the relativist
can, without contradicting herself, present arguments. Of course, the
consensus possibly emerging on this basis of a common framework will
be a local one, and perhaps only temporary. But this is not in itself an
argument against relativism, especially since no transcendental arguments
for necessary a priori truths and standards have so far gained general
acceptance.
(2) Siegel's second argument turns on the idea that "necessary some
beliefs are false": "relativism is incoherent because it holds that all
beliefs and opinions are true, yet, given conflicting beliefs, some beliefs
must necessarily be false - in which case relativism cannot be true"
(ibid., 6).
The details - and mistakes - of this line of reasoning can be spelled
out best by attending closely to Siegel's reformulation of Plato's way of
putting this argument against Protagorean relativism. The central pas-
sages from Siegel's text are the following (I have labelled the sentences
for subsequent reference):
[a] Protagorean relativism is false for all those who do not believe
it: [b] it is false for all if no one believes it; and [c] true only to the
extent that some number of people (perhaps only Protagoras him-
self) believe it ... [and Siegel quotes Plato:] [d] "it is more false
than true by just so much as the unbelievers outnumber the be-
lievers" ... [e] if opinions conflict about the truth of Protagoras'
relativism, then the Protagorean relativist must acknowledge the
truth of the opinion that the doctrine is false" (ibid., 5).
Applying this argument to his reformulation of ER, Siegel writes that ...
[f] it follows that, if according to some set ofstandards S1' ••. s" ER
is judged to be false, then, if ER is true, (at least according to that
set of standards 81 , .•. 8,,) ER is false. In this way, ER is self-refut-
ing, and so incoherent (ibid., 6).
To see why this line of reasoning is mistaken, be it noted first of all

with respect to Siegel's formulation of the second argument, that rela-
tivism need not hold that all beliefs and opinions are true. Within a
given framework, the distinction between opinion, belief and knowledge,
or the distinction between truth and falsehood can be maintained. Sie-
gel's way of formulating the relativist stand is thus inadequate, to say the
least.
Second, note that in moving from [a] to Cd], Siegel drops the rela-
tive notion of truth as truth-for-someone, and replaces it with the abso-
lute notion of truth. This is a question-begging move since the relativist
will precisely deny that truth as a non-relative, absolute notion, makes
sense. Siegel (and Plato) cannot, in order to avoid circularity, neglect
this position, since their aim is to derive an incoherence from premisses
accepted by the relativist.
To make the petitio principii of Siegel's reasoning explicit, let us
reformulate his and Plato's statements in more exact terms.
Siegel's [a] ("Protagorean relativism is false for all those who do not
believe it") is the claim that
[a] (x)("'B;cfPR] ... "'T;cfPR])

(For all persons x: if x does not believe in Protagorean rela-
tivism, then Protagorean relativism is not true-for-x.)
This statement can be accepted by the Protagorean relativist. However,

Siegel's [b] ("it is false for all if no one believes it") has two readings.
The reading that restates [a] would be [b']:
[b'] (..,Ex)(B;c[PR]) ... (x)('" T;c[PR])

(If there is no person x, such that x believes in Protagorean
relativism, then for all x: Protagorean relativism is not true-for-
x.)
204 PART II
On the second reading, a reading that must be at work in order to justify

Siegel's subsequent claim [c), truth is not relative:
[b"] (-,Ex)(Bx[PRj) .... (x)(Bx -'T[PRj)

(... then for all x: x believes that PR is not true.)
Only given the equivocation that exists between [b'] and [b"] does it
become understandable that Siegel can move on to the claim that [c) (PR
is "true only to the extent that some number of people [perhaps only
Protagoras himself] believe it") ...
[c'] (Ex)(Bx[PRj) - T[PRj

(PR is true if and only if there are some persons x such that
these x believe in PR.)
[c'], with its unindexed use of the truth operator, obviously flies in the
face of the basic contention of relativism, whether Protagorean or not.
Even worse, let us note that Siegel's choice of words invites yet another
equivocation. His expression "true only to the extent" might also be read
as an allusion to degrees of truth:
[c"] (Nx)(Bx[PRj) - d(I';N)[PRj

(protagorean relativism is true to a degree, if and only if there
is a number of people believing it; where any specific degree of
truth is dependent upon specific numbers of believers.)
That it is not a wild speculation to attribute this reading to Siegel is

obvious from the fact that he approvingly quotes Plato's [d) ("it is more
false than true by just so much as the unbelievers outnumber the be-
lievers"). Formally:
[d) (dm[-,PRj> dm[PRj) - ((Nx)(-,Bx[PRj) > (Nx)(BxfPRj))

(The degree of truth of not-PR is greater than the degree of
truth of PR, if and only if the number of x not believing PR is
greater than the number of people believing it.)
It hardly needs to be explicitly stated at all that nothing of this sort is

claimed by either Protagoras nor the modem relativist. The dropping of

the indices is also to be lamented upon in [e] ...
Given that a believes in PR (BafPRJ) and that b does not

(...,BbfPRJ), all that a is forced to acknowledge is that PR is not-
true-for-b (-,TbfPRj). a is not forced to admit, however, that PR is
not true (-'T[PRJ), and even less is a obliged to regard PR as not-
true-for-herself (..., TafPRJ). From' a believes that b does not believe
in PR' (Ba-,BbfPRJ), or from 'a believes that b believes PR to be
false' (BaBb[-' PRj), it does not follow that a must abstain from
believing PR (""BafPR]) .
... and [t]:
From 'epistemological relativism is true' (T[ERJ) and epistemologi-

cal relativism is not true for a framework frl' (-, TJI-J[ERJ) it does
not follow that 'epistemological relativism is not true' (""T[ERJ),
simply because the relativist does not accept the unindexed notion of
truth, Le. she does not accept T[ER] as a well-formed formula. All
that follows is, trivially, that 'epistemological relativism is not true
for frameworkfrl (...,Tjrl[ERJ). Again, given that two frameworks
differ in their truth value assignments with respect to ER (..., TJI-JfER]
and Tjr2[ERJ), it does not follow that ER is false (""T[ERJ) or that it
is false for both frameworks. At most, what might follow is that the
relativism-endorsing framework has to allow for the possibility of
ER being false when judged on the basis of other frameworks. But
that the assessment of epistemological theses varies with frame-
works, is part and parcel of the epistemological relativism anyway.
How could ER itself be exempt from it?
(3) Siegel's third argument seeks to establish that the very notion of
relative truth does not make sense and that relative truth presupposes an
absolute notion of truth. Siegel stresses this point when attacking Jack
Meiland who has proposed that truth-for-W (where W can range over
frameworks, cultures etc.) "denotes a special ... relation which does not
include the ... relation of absolute truth as a distinct part" (ibid., 13).
Now, Meiland puts this position somewhat clumsily by saying that "one
206 PART II
can no more reasonably ask what 'true' means in the expression 'true-
for-W' than one can ask what 'cat' means in the word 'cattle'" (ibid.).
Countering this example, Siegel has no difficulty in convincing his
readers that ...
'cat' is no more a meaningful part of 'cattle' than is 'ca', 'catt',

cattl', etc. 'True-for-W', on the other hand, is made up of the
hyphenization of distinct concepts ... (ibid., 13-14).
However, Siegel's suggestion for a better example, to wit, "action-at-a-

distance", will not do either since it prejudges the issue in his favor. For
given this example, it is easy to claim that truth-for is but a special
conception of truth (simpliciter). After all, action-at-a-distance falls
under the concept of action (simpliciter).
To make Meiland's position immune to Siegel's attack, it suffices to
say that what the proponent of "truth-for-W" is in fact doing is not so
much defining a new concept of truth as pointing out a hidden variable
which has been at work in the "absolute" concept of truth. Introducing
the hyphenization of "truth-for-W" is like pointing out to a child which
uses the words "left" and "right" in an absolute sense, that the real
meaning of "left" is "left-of", and the real meaning of "right", "right-
of". Obviously, correcting the child in this way does not presuppose any
"absolute" notion of leftness. ~n other words, the relativist holds that in
the absolute notion of truth one relatum has been overlooked, or that one
variable has been held constant, to wit, our ways of modelling, concep-
tualizing and conceiving of the world. When couched in these terms, it
is easy enough to appreciate that the concept of relative truth is not
completely different from the traditional notion of truth, and that Siegel's
criticism of Meiland's "truth-for- W" (viz. "he has provided no hint what
the string does mean or refer to" (ibid., 14) does not carry the day.
Telling a child that "left" really means "left-of" does not, of course,
mean that one has introduced something completely new. One is merely
pointing out to the child a feature of the concept that the child had
formerly overlooked. Mutatis mutandis, the relationship between "truth"
and "truth-for- W" is not such that the latter is parasitic upon the former.
The "not-to-be-negleeted truth"
MacIntyre's attitude towards relativism seems to be, at leastprimafacie,

more sympathetic than that of Siegel. For MacIntyre, relativism contains
a "not-to-be-neglected truth". An indication for the existence of this
truth, Macintyre suggests, is the fact that relativism has been refuted so
often in the history of philosophy, whereas "genuinely refutable doc-
trines only need to be refuted once" (1987: 385). This "not-to-be-neg-
lected truth" is assumed to reside in the feeling of incomprehension that
we experience when encountering the texts of other cultures (ibid., 404).
In his recent book, Whose Justice? Which Rationality? (1988),

MacIntyre also gives more general definitions of relativism and perspec-
tivism. Relativism is the doctrine that ...
to assert or to conclude this rather than that can be rational relative

to the standards of some particular tradition, but not rational as
such. There can be no rationality as such. Every set of standards,
has as much and as little claim to our allegiance as any other (ibid.,
352).
However, perspectivism goes further than relativism by drawing a

radical conclusion from the relativist stand: we should not use the
notions of truth and falsehood (at least in their traditional sense), and we
are to treat of different traditions or frameworks as so many perspectives
"for envisaging the realities about which they speak to us" (ibid., 352).
MacIntyre presents us with two arguments against relativism (and
thus ipso facto against perspectivism) and an additional criticism against
perspectivism in particular.
According to the first argument, relativism neglects the possibility
that within every tradition the debate over, or the development of, a
"problematic" can reach the point where members of the tradition begin
to question the standards of that tradition, and where ultimately the claim
of that tradition to truth can no longer be upheld:
And this by itself is enough to show that if part of the relativist's

thesis is that each tradition, since it provides its own standards of
208 PART II
rational justification, must always be vindicated in the light of those

standards, then on this at least the relativist is mistaken (ibid., 364).
MacIntyre's second argument continues along the same line. It

cannot be true that rationality equals the rationality of any particular
tradition, he argues, because within a tradition one can come to see that
the only way forward in dealing with its "problematic" is to tum to
other, rival, traditions and their standards (1987: 408; 1988: 364). This
undermines the relativist position, MacIntyre alleges, because relativism
claims that "there is no way in which each tradition can enter into
rational debate with any other, and no such tradition can therefore
vindicate its rational superiority over its rivals" (1988: 366).
Finally, and this is the argument specifically addressed to the
perspectivist, MacIntyre proposes that the perspectivist neglects the
importance and inevitability of the notion of truth for traditions of
enquiry. In other words, we cannot give up this notion because it is
"integral ... to tradition-constituted forms of enquiry" (ibid., 367).
As is easy enough to show, there is something suspicious not only
about MacIntyre's three arguments, but also about his definition of
relativism. As we already pointed out when criticizing Siegel's argu-
ments, for the relativist, the choice of standards and the adoption of
beliefs need not be totally arbitrary. Unfortunately, this is, nevertheless,
what MacIntyre assumes ("Every set of standards has as much and as
little claim to our allegiance as any other. ")
MacIntyre's first argument will work only if the relativist denies the
possibility of a development of standards within traditions or frame-
works, and if the relativist disallows for the possibility that one can
question some of a given tradition's standards in terms of other standards
of that same tradition. However, all the relativist needs to deny - in
order to be a relativist - is that there is some neutral ground of universal
standards, a ground in terms of which the standards of a given tradition
or framework, or then the failure of some of these standards, can be
assessed and evaluated.
Concerning MacIntyre's argument from encounters between tradi-
tions, three points must be stressed. First, MacIntyre himself concedes
that even though a tradition might 'objectively' need the help of another
tradition, "... it does not of course follow that there will be actual
acknowledgement" (ibid., 365). Seeking help from the other thus seems
to be a demand of "rationality ... qua rationality" (1987: 408). To see
that this will not do, one only needs to echo the title of MacIntyre's
book: Whose rationality? If this rationality is taken to be a universal one,
then the question has been begged. If it is the rationality of some tradi-
tion, however, the relativist position has not been undermined.
Second, if MacIntyre wants to claim that the possibility of one tradi-
tion's learning from another does presuppose some common criteria or
standards, then this, in itself, is not sufficient as a refutation of the
relativist. Even if common standards were needed in this case, all the
relativist needs to reject is that these criteria or standards can be univer-
sal, or a priori necessary. But this specific denial does not rule out the
possibility of there being standards which are relative, and common, to
at least two traditions.
Third, it also is not clear why the relativist would have to refuse to
acknowledge that the standards, the knowledge and the truths of some
given tradition, say tradition A, can be judged as superior with respect
to another tradition, say B. All the relativist has to contradict is that this
evaluation can be made from a ground which transcends all traditions.
But this leaves open the possibility of this evaluation being made from
within A or B, or both, or from within some other tradition.
Finally, MacIntyre's point against perspectivism (truth as inevitably
presupposed within traditions) does not succeed either, since - according
to his own definition of this stand - the perspectivist only rejects truth
and falsehood in their traditional sense, not in every sense.
To conclude this brief discussion of whether or not relativism is a
self-refuting doctrine, it remains to be said that even if the above argu-
ments against epistemological relativism as a theoretical position were to
succeed, this would still not endanger it as a heuristic device or met-
hodological stand. In other words, even the absolutist might accept the
idea that working on the basis of the relativistic hypothesis, Le. that
knowledge is nothing over and above what is accepted as such within a
given culture (be it scientific or not), is a more fruitful hypothesis for
historical and sociological research than working on the foregone conclu-
sion concerning the existence of universal principles of rationality,
common to all (possible) cultures. Perhaps, the absolutist might even
adopt this position as a methodological rule in the hope of collecting
210 PART II
material which, when day dawns, will undermine epistemological relati-

vism as a theoretical stand.
Foucault's relativism
In showing that epistemological relativism is not the self-defeating,

impotent and arbitrary view it is often made out to be, I hope to have
given the reader reason to believe that the epistemological relativism of
Foucauldian genealogy provides no compelling reason for rejecting the
Foucauldian project. What remains to be done here is merely to present
the details of genealogical relativism.
A first feature of genealogical epistemological relativism is that the
truth-defining frameworks and relative standards are essentially related
to social power. Foucault speaks of "truth" as a set of rules and pre-
scriptions for the production of scientific propositions, and he refers to
these sets as "regimes of truth" in so far as these knowledge-defining
frameworks are related to the systems of social power and sustain,
reinforce, and profit from, the knowledge produced (p1K 133).
Here it is worth emphasizing once more, and contradicting critics of
Foucaul~1, that the essential-internal link between truth and social
power does not as such tum truth into arbitrary opinion and illusion. On
the contrary, it is Foucault's very point that every socially organized
production of truth and knowledge is possible only against the back-
ground of, and within, a framework of rules, practices and standards, to
wit, that the production of knowledge is possible only against the back-
ground of, and within, systems of social power. (As seen earlier, how-
ever, Foucault considers that not all of these systems of social organiza-
tion are evaluatively on one par.)
Second, Foucault does not hold53 , that regimes of truth are homo-
geneous, that they cannot be changed from within, and that they cannot
be compared. That regimes of truth are not homogeneous is clear enough
from Foucault's constant reminders that within the same large-scale
system of social power, Le. within the same "strategy", several scientific
and philosophical discourses can exist, side by side, and that these
discourses abide, at least in part, by different, even incompatible stan-
dards:
Discourses are tactical elements or blocks operating in the field of

force relations; there can exist different and even contradictory
discourses within the same strategy; they can, on the contrary,
circulate without changing their form from one strategy to another,
opposing, strategy (HS I 102).
Foucault also writes that "relations of power-knowledge are not static

forms of distribution, they are 'matrices of transformations'" (HS I 99).
Furthermore, he suggests that even though different systems of domina-
tion have their own specific rationalities, "transmissions, transferences,
interferences" also occur (pPHC 37). Moreover, just as power in
general leads to effects which undermine its hold and encourage opposi-
tion to it, so do scientific discourses. Foucault proposes, e.g. that even
though the new discourses in the nineteenth century on "the species and
subspecies of homosexuality, inversion, pederasty, and 'psychic her-
maphrodism'" made possible and led to an increase in social control, the
very same discourses also led to the formation of an opposing discourse:
"homosexuality began to speak in its own behalf, to demand that its
legitimacy or 'naturality' be acknowledged, often in the same vocabu-
lary, using the same categories by which it was medically disqualified"
(HS I 101). Finally, Foucault also seeks to rescue "subjugated know-
ledge" as knowledge or beliefs that do not qualify as true by the stan-
dards of a given "regime of truth". In sum, it is hard to see how Taylor
can maintain that "Foucault tidies it up too much, makes it [Le. history]
into a series of hermetically sealed, monolithic truth-regimes" (faylor
1986: 98).
In connection with the problem of how existing regimes of truth can
be changed or questioned, Foucault maintains that we can "turn the will
to truth ... against truth" (LD 22). In other words, even though there is
no vantage point from which we could bystep our framework or regime
of truth, we still can investigate (on the basis of our form of knowledge)
into our framework's origin and historically contingent nature. The aim
of such inquiry, we are told is to "free thought from what it silently
thinks, and so enable it to think differently" (HS n 9):
What reason perceives as its necessity, or rather, what different

forms of rationality offer as their necessary being, can perfectly well
212 PART II
be shown to have a history; and the network of contingencies from

which it emerges can be traced. Which is not to say, however, that
these forms of rationality were irrational. It means that they reside
on a base of human practice and human history; and that since these
things have been made, they can be unmade, as long as we know
how it was that they were made (pPHC 37).
Naturally, this does not mean that we can come to know our own frame-
work in full from within itself, nor that we can find ourselves outside of
every framework. On the one.hand, the possibility of coming to know
our own framework completely is denied where Foucault writes that "we
have to give up hope of ever acceding to a point of view that could give
us access to any complete and definite knowledge of what may constitute
our historical limits" (ENL 47). On the other hand, Foucault explicitly
rejects the possibility of any universal standards; he speaks of rational-
ities in the plural, and suggests that there is no vantage point from which
our forms of rationality could be regarded as absolutely superior to
earlier ones (pPHC 36). He also proposes, for instance, that the cere-
mony of public torture was not irrational in its time, even though (most
would say) it is irrational for us today. In other words, there is no
"absolute against which [different systems of punishment] could be
evaluated as constituting more or less perfect forms of rationality" (QM
107).
Clearly, Foucault's suggestion that we study the origins of our
standards and frameworks historically as emerging from earlier ones
further undermines the claim that he conceives of different frameworks
as being completely incomprehensible to one another. For example, even
though he holds that the Greeks "had their own "regime of truth"
(pPHC 223), he still sees nothing strange about a study of Ancient
Greek morals, a study that is hoped to lead one "to explore what might
be changed, in [one's] ... own ·thought, through the practice of a knowl-
edge that is foreign to it" (HS n 9).
To conclude, while being relativistic in the sense of denying any
universal standards of rationality, and while holding that rational judg-
ments can only be made on the basis of historically contingent standards,
Foucault does not adopt any of the implausible ideas that MacIntyre and
Siegel see as definitive of relativism and perspectivism: for Foucault, the
historical nature of our standards of knowledge does not make our

standards - for us - impotent; he gives no "absolute" reasons for
relativism, confining himself instead to a case-by-case demonstration of
how allegedly timeless categories have had historically contingent
origins; he does not make the truth or validity of standard a function of
the number of frameworks abiding by them; he does not base his rela-
tivism on any claims of incommensurability and intranslatability; he does
not reject the use of the concepts 'true' and 'false', and he does not
exclude the possibility of incoherent standards within the same frame-
work.
14. GENEALOGICAL CRITICISM OF POWER

AND RATIONALITIES
Perhaps the most striking difference between Foucauldian genealogy and

research programs in the sociology of science is that genealogy does not
confine itself to descriptions and explanations of power and interest
structures in science. Rather, genealogy takes up a critical posture
towards the power of science in modem society. To analyse this posture
in a separate chapter seems imperative for two reasons. On the one
hand, it allows us to see genealogical methodology at work, and illus-
trates the critical accumen of genealogical relativism. On the other hand,
the overwhelming majority of critics of genealogy have concentrated
their attention on the critical posture of genealogy. To answer their
criticism will, I hope, turn the focus of debate over genealogy away
from this topic, and towards the issues dealt with in previous chapters.
GENEALOGY AS CRITICISM AND ANALYSIS OF RATIONALITIES
Let us start by reviewing Foucault's characterizations of the role of

genealogical analyses concerning specific political issues.
Foucault suggests that since the late sixties a series of oppositions
and struggles has emerged: women have begun to question male domi-
nance more radicaIly, children question the authority of their parents,
and psychiatry is criticized for its control over the mentally ill, medicine
214 PARTn
for its treatment of the sick, and administrations and bureaucracies for
their interventions in the realm of the individual (S&P 211). Foucault
proposes that these resistances have a number of common traits, such as
their international character, their being conflicts over power effects,
their immediacy (rather than being struggles against a class or the state,
they are struggles against specific effects), their being concerned with the
right of individuals to determine their own identity, and their being
struggles against privileges of knowledge (S&P 212). These resistances
are all similar in that they involve the rehabilitation of forms of subju-
gated knowledge and in that they are intertwined with theoretical-intel-
lectual work (p1K 81).
Obviously, the issues listed remind one of the topics that Foucault
deals with in his genealogical writings, especially if one remembers that
the problem of sexuality is linked to the male-female and the parents-
children question, and that the effects of discipline and stimulation (dealt
with in Discipline and Punish)' are related to the working of administra-
tions and bureaucracies.
But just what contribution can Foucauldian genealogy make to these
local struggles and resistances? To put it succinctly, Foucault believes
that the specific critical contribution of genealogy is not so much, or not
primarily, to criticize institutions and persons, but instead to make
criticizable the forms of knowledge, the standards of rationality, or prin-
ciples of reasoning not only on the basis of which such institutions arise
and tum out to be 'reasonable', 'useful' and 'self-evident', but also on
the basis of which the violence of these institutions becomes natural and
justifiable. Thus genealogy provides the groups opposing a given institu-
tion with an additional critical weapon: the debate over the institution
need no longer be carried out in terms of those forms of knowledge or
frameworks of reasoning which underlay the institution itself. 54
As Foucault points out time and again, his aim is first and foremost
the "wearing away [of] certain selfevidentnesses and commonplaces",
the changing and displacing of "ways of perceiving and doing things ...
or ... of forms of sensibility and thresholds of tolerance" (QM 109), the
grasping "why and how that-which-is might no longer be that-which-is",
the criticizing of "political rationality's very roots" and the making
intelligible in order to make criticizable (pPHC 36, 85, 101). That is to
say, rather than analyzing the working of factually existing institutions,
genealogy seeks to identify the network of knowledge, the standards of

rationality or rationalities that make these institutions possible and that
justify them. Since it is the central premiss of genealogy that knowledge
and power are internally-essentially related, genealogy cannot remain on
the level of violence and coercion in institutions like the prison; it has to
study the frameworks of thinking and knowledge that inform these
institutions.
The stress upon the analysis of rationality in the plural is crucial
here. Foucault tells us that to work with the simple opposition between
reason and unreason, or with the notion of a bifurcation of reason into
emancipatory and technical-strategical, is unhelpful. On the one hand,
relying on the dichotomy 'reason versus unreason' forces us to say either
that the penal system and its justification is based on reason or that it is
based on unreason. In the first case, we as critics place ourselves outside
of rationality and thus on the side of irrationality. In the second case, we
fail to realize the reasoning underlying the institution, and thus underesti-
mate its specific rationality and effectiveness. In brief, "such a trial
would trap us into playing the arbitrary and boring part of either the
rationalist or the irrationalist" (pPHC 59). On the other hand, relying on
the dichotomy of technical-strategic versus emancipatory reason, an idea
central for the Frankfurt school, tends to make us blind to the variety
("an endless, multiple bifurcation") of different forms of strategic
thinking, and reinstalls a simple-minded opposition between reason and
power. (pPHC 29, 35).
What genealogy attends to then is not whether people who rational-
ize something "conform to principles of rationality, but to discover
which kind of rationality they are using" (pPHC 59). Perhaps we might
say that Foucault's epistemic relativism shows its critical acumen here:
the idea is to show on the basis of which - historically contingent -
principles of reasoning certain strategies of control and punishment are
rationally defensible. Of course, to show this is to endorse neither these
strategies nor their principles. Instead it is to open up a new level of
debate, a level that lies deeper than the immediate criticism. What
genealogy strives to do is to demonstrate that even though certain judg-
ments and measures, say within the penal system, are defensible given
the underlying historically contingent principles and (local) frameworks,
these same judgments and measures are only defensible given these
216 PART II
historically contingent principles and (local) frameworks. To restate a

point stressed earlier, this is not to say that these frameworks or prin-
ciples can be questioned from some point beyond all frameworks, but it
is to show that these frameworks are not without alternatives, that we
can modify or enlarge them, that we can question them as merely local
frameworks from within the more global framework of our culture.
The task is thus, as Foucault puts it, "a liberation of thought", that
is meant to prepare the ground for "deep transformations" rather than
small corrections within a given system (pPHC ISS). To take the issue
of penal institutions, this liberation is one that allows us to question not
only specific insufficiencies in prisons, but rather to question the current
system of penal justice as a whole. By the same token, we acquire the
means "to avoid other institutions, with the same objectives and the same
effects, from taking their stead" (pPHC 84).
Foucault characterizes this radical form of criticism in a discussion
in 1971, a discussion in which he opposed his own form of work to what
he calls "the humanist's" critique. He claims that whereas a humanist
critique takes as given that, "the guilty are guilty and the innocent are
innocent" , and merely confines itself to demanding minor improvements
in the prison for the prisoner as a human being (" ... consequently, flush
toilets"), genealogy "isn't concerned with the soul or the man behind the
convict, but it seeks to obliterate the deep division that lies between
innocence and guilt" (LCMP 227).
After the publication of Surveiller et punir, Foucault writes in a
somewhat less radical tone. In 1977 he says that the aim of genealogy's
analysis of those rationalities in which the current penal system is
embedded is to make small, straightforward adjustments within the
prevailing system difficult:
... it's true that certain people, such as those who work in the
institutional setting of the prison - which is not quite the same as
being in prison - are not likely to find advice or instructions in my
books that tell them 'what is to be done'. But my project is precisely
to bring it about that they 'no longer know what to do', so that the
acts, gestures, discourses which up until then had seemed to go
without saying become problematic, difficult, dangerous (QM 110).
In other words, genealogy does not only abstain from making proposals
for limited reform, but even brings into question these very standards by
which we would be likely to evaluate proposals as to their efficiency,
humanity, or inevitability. For Foucault, to justify one's criticism with
the putting forward of a program for reform is to fall into "a trap in
which those who govern try to catch intellectuals" (pPHC 52); to fall
into this trap is to accept the validity of the idea "don't criticize, since
you're not capable of carrying out a reform" (QM 110). The reason why
this challenge can be rejected is not only that any such immediate and
straightforward proposal for reform must remain bound to the prevailing
standards, but also because those who govern usually withhold from
their critics the very information necessary for deciding which solutions
are possible (pPHC 52), and also, most importantly, because "'what
ought to be done' ought not to. be determined from above by reformers,
be they prophetic or legislative, but by a long work of comings and
goings, of exchanges, reflections, trials, different analyses" (QM 110).
Put differently, political decisions on the penal system, for example,
should be based on experimentation and a prolonged debate between not
only philosophers and the legislators, but with all the parties, including
the prisoners, who are involved. In this experimentation and in this
debate genealogical analysis participates only by laying bare the contin-
gency of the standards of those who see the present system as sufficient
and defensible.
Is RESISTANCE POSSIBLE?
The doubts that critics of Foucault have raised concerning the critical
posture of genealogy fall roughly into two groups. On the one hand,
many writers have claimed that, given Foucault's theory of power, all
criticism of modem science and society is futile. This is so, these writers
allege, because the latter theory leaves no room for resistance to power
structures. On the other hand, critics hold that the absence from Fou-
cault's writings of a normative theory for organizing a just society leaves
him without a normative basis for any criticism. In the reminder of this
chapter, I shall take up these two issues in tum.
To begin with the possibility of resistance, the thesis that I wish to
218 PART II
attribute to Foucault is the following: While power relations are an

inevitable fact of social life, not all power relations are evaluatively on
a par: we can distinguish - roughly - between hegemonial forms of
power which we can resist and non-hegemonialforms ofpower which we
need not resist.
While Foucault is notorious for his reluctance to make value judg-
ments, it is interesting that in several contexts he in fact draws a distinc-
tion between inevitable or harmless, and avoidable or harmful forms of
power. Thus, even though he says that we cannot ever reach a state
where truth is emancipated "from every system of power (which would
be a chimera, for truth is already power)", he calls for the "detaching
lot] the power of truth from the forms of hegemony, social, economic
and cultural, within which it operates at the present time" (p/K 133).
Foucault also tells us that genealogy fights for ...
the insurrection of forms of knowledge that are opposed primarily

not to the contents, methods or concepts of [...] science, but to the
institution and functioning of an organized scientific discourse within
a society such as ours (p/K 84).
The underlying premiss of these statements is formulated by Foucault

himself as: "Every power relation is not bad in itself, but it is a fact that
always involves danger" (pPHC 168).
The fact that Foucault believes it is possible to resist hegemonial
power is obvious not only from the task that he assigns to genealogy, but
it is also clear from many other contexts. For instance, Foucault holds
that we can speak of a power relation only where there is the possibility
of "refusal or revolt", even though the latter might consist only of
preferring death over submission. Foucault thus draws a distinction
between power and violence. Violence or force leave the victim no
choice between compliance or non-compliance; in these cases the victim
is treated as a body and not as someone whose deliberation is to be influ-
enced. (" ... [power] is a mode of action which does not act directly and
immediately on others. Instead it aets upon their actions.... A rela-
tionship of violence aets upon a body or upon things" (S&P 220).)
In one passage, Foucault calls the element of resistance inherent in
power relations "a certain plebeian quality or aspect" (p/K 138). While
doubting that the plebs is a genuine sociological entity, and while deny-
ing that the plebs can be identified with the proletariat, Foucault suggests
that the plebeian quality is "a centrifugal movement, an inverse energy,
a discharge" and that it can be found allover the social body "in a
diversity of forms and extensions, of energies and irreducibilities":
This measure of plebs is not so much what stands outside relations

of power as it is their limit, their underside, their counter-stroke,
that which responds to every advance of power by a movement of
disengagement (p/K 138).
While I certainly agree with Foucault's critics that Foucault's

conception of the resisting agent is underdeveloped, and that his histories
of the prison and the asylum give little attention to resistance, it still
seems to me that the often pronounced worry, according to which Fou-
cault's theory of power makes resistance inconceivable, is vastly exag-
gerated. 55
First, the claim that power is omnipresent clearly does not lead to
this result. From the fact that our daily lives are permeated both by
different, mostly harmless, forms of exercises of power, and by various
forms of power being exercised over us, it hardly follows that we cannot
oppose, e.g., the extensive power that professors exercise over unem-
ployed young scholars, or the extensive power that the Finnish police
has over foreigners living in Finland. Critics tend to misread the Fou-
cauldian idea of the inevitability of some or other forms of power as the
much stronger claim that all of the now-prevailing power relations are
inevitable and thus irresistible. But attributing the latter position to
Foucault is to fall prey to the fallacy of composition.5' It is to infer B
from A:
(A) For all times and all parts of society: there exist some (or other)
relations or mechanisms of power.
(B) For some (and the same) relations or mechanisms of power, these
relations and mechanisms of power exist for all times and all parts
of society.
220 PART II
The latter statement is contradicted not only by Foucault's stress on the

historical development of power mechanisms, but also by his allowing
for resistance against malicious forms of power.
Actually, one might even argue that the omnipresence of power
makes resistance more rather than less likely. After all, this thesis
includes the idea that all of us are exercisers as well as victims of power,
coercers as well as victims of coercion. And this can only mean that
none of us has fully adopted a submissive attitude that predisposes us to
be a willing victim of coercion: all of us harbour some of the rigidity
which makes us less likely to give in.
Furthermore, we also need to give proper heed both to Foucault's
notion of subjugated knowledge, and the contradictory effects, or unfor-
seen by-products, of mechanisms of power. Recall that Foucauldian
subjugated knowledge includes the knowledge of those - the prisoner,
the mentally ill, the patient - who more than any group in the Panoptical
society are the victims of power mechanisms. The fact that hegemonial
forms of power can only suppress but not destroy these forms of knowl-
edge shows that even the most massive manipulation cannot, on Fou-
cault's view, determine the victim's identity completely. In part, this
situation is due to the circumstance that no coercive institution, and no
system of coercive institutions, is ever immune to the danger of produc-
ing or using several, conflicting mechanisms of power such that they
either neutralize one another or have other unforeseen and resistance-
invoking counter-effects. Foucault explicitly allows for the possibility
that the combining of infinitesimal mechanisms leads to "numerous
phenomena of inertia, displacement and resistance" (p1K 142). He also
indicates, more concretely, that the preoccupation with the health of the
human body, a preoccupation central to the theme of bio-power, has in
fact led to such countereffects:
... once power produces this effect there inevitably emerge the
responding claims and affirmations, those of one's body against
power, of health against the economic system, of pleasure against
the moral norms of sexuality, marriage, decency. Suddenly, what
has made power strong becomes used to attack it (p/K 56).
Finally, the claim that Foucault does not conceive of human subjects
as being through and through determined by manipulation and control,

can also be backed by attending to a notion that surfaces in his last
writings, The History of Sexuality II and III. This notion is that of
"technologies of the self", a notion a discussion of which lies beyond the
scope and central interests of my investigations. Technologies of the self
permit individuals to effect by their own means or with the help of

others a certain number of operations on their own bodies and souls,
thoughts, conduct, and way of being, so as to transform themselves
in order to attain a certain state of happiness, purity, wisdom,
perfection, or immortality (Foucault 1988: 18).
Clearly, - and despite the fact that there is nothing especially new
or thrilling about Foucault's investigations into this topic - his allowing
for the possibility of such "technologies" rules out the possibility of us
being completely determined by interdictions and disciplinary mecha-
nisms. To repeat, these Foucauldian ideas on resistance do not amount
to more than hints or sketches. However, not only can the sketchiness of
these ideas be defended by the observation that resistance cannot usually
be predicted and deduced ("In the end, there is no explanation for the
man who revolts" (Foucault 1981c: 5).), Le. that it is unclear what more
can be said about resistance on a theoretical level, but it should also be
clear that even in their very rudimentary form, there ideas are sufficient
to show that, according to the Foucauldian theory of power, resistance
is not only possible, but even inevitable.
CRITICISM WITHOUT NORMATIVITY?
Commenting on the absence from Foucault's writings of a normative

theory for organizing a just society seems imperative for at least two
reasons. On the one hand, explaining this Foucauldian notion warrants
the vagueness of his distinction between hegemonial and non-hegemonial
power. On the other hand, there is hardly any ingredient of Foucault's
thought that has been denounced as often as the alleged contradiction
between his implicit criticism of modern forms of power and his unwil-
222 PARTll
lingness to spell out the normative basis of this criticism.

Nancy Fraser's way of putting this point can be regarded as repre-
sentative of many other commentators:
Why is struggle preferable to submission? Why ought domination to

be resisted? ... Only with the introduction of normative notions
could he [Foucault] begin to tell us what is wrong with the modem
power/knowledge regime and why we ought to oppose it (Fraser
1981: 284).
Like his stance on resistance, Foucault's version of a criticism

without an elaborated normative theory also seems to me to be much
more plausible than his critics allow for. Let me try to explain this
somewhat telegraphically, in part by drawing on Foucault's own argu-
ments, in part by submitting some further, non-Foucauldian arguments
in his support.
First, I find unconvincing Fraser's insistence that Foucault owes us
a normative theory explaining why we should prefer struggle to submis-
sion, Le. why we should resist domination. It seems arguable that the
onus of proof does not lie with those who are willing to resist but rather
with those who are not. It seems fair to attribute this position to Foucault
based on his answer to the question why he is interested in politics: " ...
why shouldn't I be interested? That is to say, what blindness, what
deafness, what density of ideology would have to weigh one down to
prevent one from being interested ... " (1974: 168). Perhaps it is also
worth adding that in one place Foucault suggests that it is precisely the
message of what he calls the humanist tradition that one should not resist
the exercises of power. In fact, Foucault even defines humanism as ...
the totality of discourse through which Western man is told: "Even

though you don't exercise power, you can still be a ruler. Better
yet, the more you deny yourself the exercise of power, the more
you submit to those in power, then the more this increases your
sovereignty" (LCMP 221).
Second, one need not be too impressed by the often-heard claim that
criticism needs to be based on firm criteria, and that to accept the
validity of a criticism is to accept the criteria upon which it is based.

After all, there are many good examples in the history of science and
philosophy where criticisms were not presented using firm criteria, and
many clear cases where the critique was accepted without the criteria. 57
For instance, nothing prevents one from accepting many of Habermas's
analyses and criticisms of modem society, while at the same time
rejecting his normative utopia of the ideal speech situation.
Third, and this too is an argument that can not be found - as far as
I can see - in Foucault's texts, one will usually manage to mobilize
much broader popular support for an initiative, say, against new mis-
siles, new nuclear plants, degrading conditions in hospitals, etc., if one
abstains from linking this initiative to some global conception of what
society should be like. It is enough that we share roughly the same
intuitions telling us, e.g. that maltreatment of prisoners or the mentally
ill is inhumane - we need not share the same moral or political theory.
Fourth, and here I come to a point that is central for Foucault
himself, one might argue that .the task of philosophers and intellectuals
is not to tell people how they should live their lives, but at most to par-
ticipate in the struggle for getting rid of vicious forms and mechanisms
of power. That is to say, "the intellectual's role is no longer to place
himself 'somewhat ahead and to the side' in order to express the stifled
truth of the collectivity ... " (LCMP 207-8), his or her task is not to
"deduce from the form of what we are what it is impossible for us to do
and know" (ENL 46), but rather the task is the more modest one of
pointing out that we need not be what we are, that is, showing "the
possibility of no longer being, doing, or thinking what we are, do, or
think" (ENL 46).
Fifth, one can also advance an argument against the preoccupation
with normative theories by reference to the question of priority, and the
current state of the left. On the one hand, arguing with Noam Chomsky
about political theory in 1974, Foucault held that the criticism of coer-
cive institutions is too important to be left alone until one has developed
"an ideal social model" (1974: 171). On the other hand, in 1977,
concerned by events in China, Foucault suggested that the progressive
groups and movements of the Western world can no longer refer to any
socialist country as a model. The left must "call in question the whole
important tradition of socialism in a fundamental way - since, after all,
224 PART II
everything which this socialist tradition has produced in history must be

condemned" (Foucault 1977: 68). Foucault believes that we are thrown
back to the year 1830 and that a leftist, progressive theory has to be
developed anew from the ground up. Even if one does not agree with
this position in all its radicality, its general drift is certainly even more
convincing in the 1990's than it was back in 1977.
Sixth, Foucault considers that utopian conceptions of what a just
society must look like are obstacles rather than accelerators of social
change. ("1 think that to imagine another system is to extend our partici-
pation in the present one." - (LCMP 230).) For instance, Foucault
regards as the central flaw in the development of Soviet society that it
tried to implement a conception of "an ultimately liberated human
nature" (1974: 174), a conception shaped by the bourgeois societies in
which it had been drawn up. More generally, criticizing modem society
in terms of justice or human nature is problematic because these are
"notions and concepts which have been formed within our culture,
within our type of knowledge and our form of philosophy, and that as a
result form part of our class system ... " (ibid., 187).
One noteworthy point about this sixth, and perhaps most central,
argument is that it expresses a position that has also frequently been
attributed to Marx. Several interpreters of Marx have suggested that
Marx cannot have wanted to criticize capitalist society for being unjust
since justice, "the juridical point of view", is but a function of the mode
of production. To measure society by the standard of justice is to remain
within the superstructure of society and thus to remain reformist rather
than become revolutionary.58
A further, rather obvious, observation is that the reasoning behind
this argument is of course compelling once we accept the thesis of an
internal link between power and knowledge. Clearly, there is no reason
why power interests, exclusion and manipulation should tum back at the
borders of the philosophical field.
Finally, a comment is called for on the alleged contradiction be-
tween the above-quoted Foucauldian rejection of the notions of justice
and human nature, and the ways these - or similar - notions appear in,
or are presupposed by, his own writings. Obviously, Foucault's own
interest in the history of modem forms of power is morally motivated,
- his own political activities are as good an indication of this as any. He
also accepts in the introduction to one of his articles in Le Nouvel

Observateur the comment that French institutions are "unjust". More-
over, he claims to feel "impatience for liberty"; he speaks out for
"bodies and pleasures"; and he calls for "a new form of right", "auto-
nomy" , "the right to knowledge" and the creation of "ourselves as a
work of art" (ENL 50; HS I 157; PIK 108; PPHC 51, 160; 1983: 236).
That these different pronouncements need not constitute a contradic-
tion or paradox is easily appreciated by returning once more to Rawlsian
terminology, to wit, Rawls' distinction between the concept and concep-
tions ofjustice, and his distinction between moral intuitions and theories.
It seems natural to suggest that Foucault's opposition to justice is not so
much an opposition to the concept of, and our moral intuitions concern-
ing, justice, as it is an opposition to modern theoretical conceptions
thereof. After all, what Foucault seeks to establish, in his historical
analyses, is that political-philosophical rationality either has its roots in
pre-modern conceptions of society - and is thus unable even to concep-
tualize modern power - or that it is intertwined with the rationality that
rules the Panoptical society. However, Foucault's criticism of that
rationality and the Panoptical society itself need not depend upon these
same conceptions of justice or human nature. Most of us that know
nothing of modern theories of justice or of humanism still probably feel
that there is something fundamentally unjust about the ways prisoners or
the mentally ill are treated, or that there is something inhumane about
many power mechanisms in science. Drawing, as Foucault does, on
these intuitions is not irrational, especially as long as the other factor of
the Rawlsian "reflective equilibrium" (1972: 20), Le. moral theories or
"moral geometry" (ibid., 121), stand in doubt. Perhaps most of us also
share Foucault's intuition that the construction of new moral or political
theories in the future - a possibility implied by Foucault's call for a
novel concept of right - has to be constrained by at least the following
restriction: "The search for a form of morality acceptable to everybody
in the sense that everybody should submit to it strikes me as catastro-
phic" (pPHC 253-54).
To conclude, let us note that even though the idea used in the last
paragraph to make sense of Foucault's drawing on moral intuitions
rather than elaborated ethical theories was not formulated by himself, a
similar stand is expressed by him with respect to the wider issue of his
226 PART II
relation to the Enlightenment tradition. On the one hand, Foucault

suggests that we can no longer adopt many of the central doctrines of the
Enlightenment, e.g. its humanism, i.e. its assumption of a permanent
human nature, its naive opposition between social power and knowledge,
or its utopianism. On the other hand, for Foucault, this rejection of
central doctrines of the Enlightenment does not imply that we can no
longer trust the intuition or attitude that underlies, and emerges from,
Kant's writings, and that continues to be at work from Hegel through
Nietzsche and Weber up to the Frankfurt School, namely the intuition,
ethos or attitude that our historical era calls for a "permanent critique",
a questioning of what is given to us as "universal, necessary, obliga-
tory", a readiness to experiment with new forms of social life, and a
constant attempt to disconnect the growth of our scientific and technical
capabilities from the "intensification of power relations" (ENL 48).
Foucault calls this attitude "the critical ontology of ourselves" and
explains that ...
it is to be conceived as an attitude, an ethos, a philosophical life in

which the critique of what we are is at one and the same time the
historical analysis of the limits that are imposed on us and an ex-
periment with the possibility of going beyond them (ENL 50).
NOTEs TO PART IT
1 For example, Robert Dahl and Stephen Lukes write as follows:

"A has power over B to the extent that he can get B to do some-
thing that B would not otherwise do" (Dahl 1969: SO).
" ... A exercises power over B when A affects B in a manner con-
trary to B's interests" (Lukes 1974: 27).
Z See Dahl (1969); Harsanyi (1969); Wrong (1979: 12-17).
3 "A person kept thoroughly confined and supervised, as an individual
in a strait-jacket, ... is not a participant in a reciprocal power-relation"
(Giddens 1979: 149).
4 E.g. Wrong (1979); Lukes (1974); Bachrach and Baratz (1969a) and
(1969b).
5 "In general, we understand by 'power' the chance of a man or a
number of men to realize their own will in a communal action even against
the resistance of others who are participating in the action" (Weber 1970:
SO).
Ii Interestingly enough, Lukes's characterization, which tends to push
Foucault's conception of power into the camp of (marxist-) structuralist
theories of power, can also be countered with the observation that the
Althusserian Nicos Poulantzas, whom Lukes considers rightly to be the
champion of a responsibility-denying conception of power, has in fact
claimed that Foucault retains too many ingredients of the Anglosaxon
sociological tradition. Little surprise that Poulantzas faults Foucault's
genealogy for this closeness (poulantzas 1978b: 41). In order to defend
Foucault against Poulantzas's charge of merely repeating old Anglosaxon
ideas, we need to turn to the features of Foucault's theory of power that set
him apart from this tradition.
7 In outlining the Leibnizian distinction below, I follow Rescher (1981:
56-83).
• This was already vaguely suggested by Veyne (1981: 63).
, The idea of applying the Leibnizian distinction to theories of power
has been elaborated most convincingly by Jon Elster, who incidentally
attributes the view that power relations are internal to Marx (Elster 1978:
20-27; Elster 1985: 92-95). Elster suggests that we conceive of power as
an external relation when we treat power as something that different individ-
uals or groups possess in different degrees, and when we distinguish
between groups in society on the basis of how much power they possess.
What this kind of approach leaves out, however, is what takes place in the
interaction between the groups or individuals.
227
228 NOTES TO PART II
10 Foucault's fairly wide concept of power mechanisms allows him to

speak of the exercise of power even in those cases in which other writers
speak of power as a disposition rather than an episode (or exercise). That is
to say, these writers wish to draw a distinction between the cases where
threats are actually made, and cases where the power subject submits in the
absence of direct interaction (e.g. Wrong 1979: 6). Foucault's claim that
power "exists only in action" (P/K 89) seems to blur this distinction. In
order to preserve this distinction within the Foucauldian framework, we
might suggest that regardless of whether the compliance follows the presen-
tation of threats or not, it is in any case the result of an exercise of power:
the two cases (of power as episode or disposition) are distinguishable by the
criteria either of whether the power exercise is temporally close or tempo-
rally remote from the power subject's action (or non-action), or by which
types of different mechanisms of power are involved (e.g. ecological control
versus threat).
11 The notion of "ecological control" was suggested by Cartwright
(1969: 140).
12 See, e.g. Kedar (1987).
13 Lawyers, historians, and political writers like Edward Coke (1552-
1634), John Lilburne (1614-1657), Nicolas Boulainvilliers (1688-1749),
Charles Hector Estaing (1729-1794), Augustin Thierry (1795-1856), and
later biologists, racists and proponents of eugenics (LK).
14 While the details of this history lie beyond the scope of this study, it
is perhaps interesting to note in passing that Foucault's turning around of
the Clausewitzian principle is not unique among political philosophers of
this century. For instance, in the thirties Antonio Gramsci sought to concep-
tualize the class struggle for hegemony in terms of the distinction between
war of movement and war of position. More precisely, Gramsci analyzed
economic crises as bombardments of heavy artillery that have to be employ-
ed by the left in order to move towards a frontal attack on the class oppo-
nent (Gramsci 1980: 268-71). Another example is of course Carl Schmitt,
one of the most eminent reactionary political thinkers of this century, who
again in the 1930's defined the essence of politics via the categories of
friend and enemy, taking these latter concepts as referring to friend and
enemy in a battle of war. Schmitt does not confine this opposition to
international politics but applies it to domestic politics as well: "The real
possibility of struggle that must always be present in order for us to speak
of politics, relates - when we take into account the 'primacy of domestic
politics' - not to the war between organized peoples (states or empires), but
to civil war" (Schmitt 1932: 20). Schmitt also faulted Clausewitz for having
NOTES TO PART II 229
neglected the importance of a study of war for an understanding of politics

(ibid., 21). Sutprisingly enough, these parallels between the arch-reac-
tionary Schmitt and Foucault seem not to have been exploited yet by
Foucault's intetpreters.
15 For good introductions to Faraday's thought, see Agassi (1971);
Berkson (1974); Sachs (1973).
16 On this notion, see Agassi (1971: 102-16).
17 See foremost Parfit (1984).
18 See, for a more detailed discussion, Berkson (1974: 105).
19 For instance, Jean Baudrillard argues that "if power were this
magnetization of the social field' then it would hardly any longer meet with
resistance" (Baudrillard 1977: 58).
20 Note also that even aside from the question as to whether or not
Foucault himself has made fruitful use of the omnipresence thesis, this idea
is in fact the basic starting point of one of the most elaborated theories of
power in Anglosaxon sociology. Talcott Parsons' theory of power is based
on the premiss that power is "a circulating medium, analogous to money"
(parsons 1969: 256). This medium is spread throughout society, even
though it can, at least in part, be "invested" or concentrated. Contrary to a
suggestion found in the literature (Kroker 1984), it seems to me that Fou-
cault shares little else with Parsons apart from this idea of the omnipresence
and circulating character of power. For example, the very parallel between
money and power, and Parsons' attempt to construe political institutions as
banks in which we invest our power-money are clearly diametrically
opposed to Foucault's anti-economism and his refusal to study power in
terms of legitimization. Yet the fact remains that Parsons' theory constitutes
a case where the omnipresence assumption finds a non-trivial application.
21 See, e.g. Anderson and Carlos (1976); Baumgartner, Buckley, Burns
and Schuster (1976); Lorrain and White (1971).
II Petit lArousse Illustre (1988: 320).
23 Deleuze (1988: 28).
Z4 This is an important point to take up, since Foucault has been
charged with leaving no room for this possibility. Charles Taylor has
accused Foucault of neglecting even a consideration of invisible hand
explanation, claiming that Foucault explains coercive institutions exclusively
in terms of "Power" as "a strange kind of Schopenhauerian will, unground-
ed in human action" (Taylor 1986: 88).
Z5 For an analysis of counterfinality phenomena, see Elster (1985).
Z6 See Ullmann-Margalit (1978) for a detailed discussion of this dis-
tinction.
230 NOTES TO PART II
27A fourth, highly interesting argument can be found in Rouse (1987).

21Since this idea draws a conclusion from observations by Ian Hack-
ing, a brief look at its original context might soften resistance to the subse-
quent conclusion.
At a time when the direction of the bulk of basic research in the natural
sciences and mathematics is determined by military funding, Hacking (1987)
asks us to consider whether it is not natural to assume that the very form of
our knowledge has been permanently shaped by this relation. Hacking takes
a form of knowledge of a branch of science to be "a structured set of
declarative sentences that stand for possibilities, that is, sentences that can
be true or false, together with techniques for finding out which ones are
true and which ones are false" (ibid., 243). Hacking suggests that as more
and more results of weapons research come to figure as exemplars or
paradigms of what constitutes good scientific work, it is reasonable to
assume that alternative ways of doing research drop out from our screen of
possibilities, and weapons research increasingly determines what we regard
as scientific knowledge, as meaningful areas of research or as possible
candidates for truth and falsehood (ibid., 260).
29 This point is also stressed by Hacking (1987: 250).
30 See Lewis (1973) for the analysis of counterfactuals employed here.
31 Thus, in L 'Ordre du discours, he formulates the following "metho-
dological demands": (1) "a principle of reversing", according to which the
seemingly positive figures of author, discipline, or will to truth are to be
treated as principles of rarification and exclusion; (2) "a principle of
discontinuity" , according to which discursive practices are to be attended to
as often excluding and ignoring one another; (3) "a principle of specificity",
according to which we are to regard discourses as practices that we impose
upon things; and (4) a principle.of "exteriority", according to which one is
to study the external conditions of the possibility of discourses. In the same
text Foucault also speaks of four "regulative principles of analyses", these
being "the event", "the series", "the regularity", and "condition of pos-
sibility". These notions are meant to replace notions of the traditional
history of ideas, concepts like creation, unity, originality and meaning, i.e.
notions connected to the idea of a subject that freely creates meaning (LD
53-56).
In Discipline and Punish, we are again given "four general rules" of
analysis: (1) attend to the "possible positive effects" of power and punish-
ment; (2) focus on "techniques" or "tactics" of power use and punishment;
(3) do not study "the history of penal law and the history of the human
sciences as two separate series"; and (4) investigate how control over the
body shapes "the soul" or personal identity (OP 23-24).

In "Two Lectures" (p1K 78-108) the four "methodological precau-
tions" are: (1) study power not on the level of the question of its legitimacy
but on the level of how different institutions work on individuals, Le. "try
to locate power at the extreme points of its exercise"; (2) study power in its
exercise rather than asking who has power and what are the intentions of
the power holder, Le. study how power shapes subjects; (3) analyze power
as "something which circulates", as "a chain" or as a "network"; (4)
"conduct an ascending analysis of power", and abstain from giving undue
importance to ideological effects (p1K 97-102).
Finally, in History of Sexuality J, Foucault formulates "four rules to
follow", telling us that these are "not intended as methodological impera-
tives; at most they are cautionary prescriptions": (1) the "rule of imma-
nence", according to which knowledge and power are to be studied as
internally related; (2) "rules of continual variations", according to which
"relations of power-knowledge are not static forms of distribution, they are
'matrices of transformations'''; (3) the "rule of double conditioning",
according to which macrolevel and microlevel power mechanisms are to be
regarded as interdependent; and (4) the "rule of the tactical polyvalence of
discourses", according to which discourses are to be studied as serving
different functions (HS 198-1(0).
32 "... genealogy must record the singularity of events outside of
any monotonous finality " (NGH 76).
"An event ... is not a decision, a treaty, a reign, or a battle, but the
reversal of a relationship of forces, the usurpation of power, ... " (NGH 88).
Genealogy is thus concerned "to reestablish the various systems of
subjection: not the anticipatory power of meaning, but the hazardous play
of dominations" (NOH 83).
33 Cf.: "The body is molded by a great many distinct regimes; it is
broken down by the rhythms of work, rest, and holidays; it is poisoned by
food or values, through eating habits or moral laws; it constructs resistan-
ces" (NOH 87).
34 Nietzsche (1968: 481, 552). Here quoted from Nehamas (1985: 42,
173).
3S Cf. Part I of this study, especially Foucault's notion of series.
36 Here and in the following I make use of ideas by G. H. von Wright.
See the papers in von Wright (1984).
37 Taking their lead from Collins (1985).
38 Foucault calls this ability of the dispositij of the prison to be adapta-
ble to different ends it strategic "saturation" (remplissement) (p/K 196).
232 NOTES TO PART n
39 Linking Elias's and Bakhtin's classical studies with more recent

feminist concerns, Dorinda Outram has suggested that Bakhtin's work is to
be taken as a corrective of Elias's, and that the important tum from the
"carnevalesque" to the homo clausus is situated around the French Revolu-
tion: "The victory of the homo clausus ... was not complete until, through
the French Revolution, a representative bourgeois culture of public domi-
nance could be constructed which functioned not only through institutional-
ized support in the new state, but also ... through the creation of a new
political culture defined precisely as that victory of homo clausus over both
the feminine and the carnivalesque" (Outram 1989: 16). Outram shows
convincingly that the French Revolution led to the creation of a political
culture from which women were excluded and that this exclusion reflected
itself in conceptions of the male and the female body.
40 The shift from the first to the second is one of the central themes of
DP.
4. The latter alternative is of course the central topic of HS I.
e Cf. Amnesty International (1987: 139).
43 See e.g. Osnowski (1988) .
.... This feature of Foucault's work is convincingly stressed by Veyne
(1981).
45 DP, part two.
46 HS I, passim.
47 Even though Foucault claims at one point that he is not concerned
with interests, I shall ignore this disclaimer and follow his interlocutors in
holding that he does in fact provide explanations in terms of interests (p/K
205-6).
... There are, of course, many other interesting parallels between
Foucauldian genealogy and Latour's work (1987, 1988). See Rouse (1987)
for further elaboration on these parallels.
4' Sociologists of science have only recently attended to this feature of
their work. Thus Steve Woolgar has interpreted the irony of the sociological
enterprise concerning science as arising from the fact that sociology replaces
the account presupposed by the natural scientists (nature determines know-
ledge) with an account that sees knowledge and nature as determined by
scientists' negotiations, quarrels and contingent laboratory settings. Woolgar
also distinguishes between "stable" and "dynamic irony": the former is
nothing but the replacement of an actor's account with an allegedly more
adequate one; the latter signals, additionally, its own transient and pro-
visional character: "dynamic irony asks the reader constantly to recognize
the fragility of the ironicist's own account ... " (Woolgar 1983: 260).
50 Quoted from Baker and Hacker (1983: 4).

51 See also Hesse (1980: 44-48).
52 Like MacIntyre (1987) and Taylor (1986).
53 Pace Taylor (1986: 94).
54 Even though Foucault claims that direct criticism of moral wrong-
doing within institutions is not the primary task of genealogical criticism, it
is by no means true that he has abstained from criticizing persons and
institutions directly. This much has to be firmly said in answer to those
many critics who claim that Foucault always depicts power as a "Schopen-
hauerian will" (Taylor 1986: 88) for which no one is responsible. Fou-
cault's numerous articles and interviews for widely-circulating French daily
and weekly papers speak a totally different language. For instance, in an
article published in Le nouvel observateur in 1972 (Foucault 1972), Fou-
cault attacks the French president, Pompidou, for not having pardoned two
men who were subsequently executed with the guillotine. He refers to the
executor by name, mentions his hometown and even reveals his salary for
carrying out these executions. Foucault also speaks directly of those who
are responsible for the prison system and the constant threat of death its
inmates are faced with: "the legal administrative body, the Church, the
armed police, and, in the shadow, the president of the Republic ... [and] the
prison personnel". Furthermore, in the same article Foucault criticizes the
prison sentence imposed upon a woman for having had an abortion, and
deplores the case of a woman who was imprisoned because she could not
pay back a loan. To take another example, in an article for Liberation
published in 1981 (Foucault 1981), Foucault speaks explicitly of "the
political class", quotes prisoners' reports on their hunger strike and the
unwillingness of the prison warden to negotiate, and he claims that for the
ruling political elite the ideals of the French Revolution mean preciously
little.
55 Concerning the possibility of resistance on the basis of Foucault's
theory of power, I have already referred to Baudrillard's argument that
power qua magnetization of the social field simply leaves no room for
resistance. Michel Pecheux has claimed that Foucault's theory of power is
marked by a "masked biologism which ... does indeed make revolt strictly
unthinkable". Pecheux believes that on the basis of Foucauldian premisses
it is impossible to distinguish between "subjection of human individuals and
the procedures of animal domestication" (Pecheux 1982: 219). Moreover,
Nicos Poulantzas has argued that Foucault lacks a subject or agent of re-
sistance, suggesting that only a social class like the proletariat can take this
position (poulantzas 1978a: 137).
234 NOTES TO PART n
S4i The role of this fallacy in sociological and political thought has been
studied by Elster (1978).
57 Agassi (1988b: 213) makes this point in a different context.
51 On this issue see the papers in Cohen, Nagel and Scanlon (1980).
Cf. Lukes (1985).
BmLIOGRAPHY
1. WRITINGS, EDmONS, AND INTERVIEWS BY FOUCAULT
1954 Maladie tnentale et personnalite, P.U.F., Paris.

1960 "Introduction a L' Anthropologie de Kant". These comptementaire
pour Ie doctorat des lettres, Universite de Paris, Faculte des lettres
et des sciences humaines. This unpublished thesis is preserved at the
Centre Michel Foucault in Paris.
MC Madness and Civilization, translated by J. Barchilon, Pantheon,
New York, 1965. Originally published in French as Folie et de-
a
raison. Histoire de lafolie l'age classique, PIon, Paris, 1961.
1966 "Une histoire restee muette", La Quinzaine litteraire, July 1st.
1968 "Foucault repond a Sartre", La Quinzaine litteraire, March lst-
15th.
1969 "Michel Foucault explique son dernier livre", Magazine litteraire
28 (April-May): 23-25.
1971a "Entretien avec Michel Foucault", in Bellour 1971: 135-44.
1971b "Foucault responds/2" , Diacritics vol. 1: 60.
UJ L'ordre du discours, Gallimard, Paris, 1971.
1972 "Les deux morts de Pompidou", Le nouvel observateur, Dec. 4.
BC The Birth of the Clinic: An Archaeology of Medical Perception
translated by A. Sheridan Smith, Pantheon, New York, 1973.
Originally published in French as Naissance de la clinique. Une
archeologie du regard mMical, PUP, Paris, 1963.
OT The Order of Things. An Archaeology of the Human Sciences,
Tavistock, London, 1974. Originally published in French as Les
mots et les choses. Une archeologie des sciences humaines, Galli-
mard, Paris, 1966.
AK The Archaeology ofKnowledge, translated by A.M. Sheridan Smith,
Tavistock, London, 1974. Originally published in French as L 'ar-
cheologie du savoir, Gallimard, Paris, 1969.
1974 (with N. Chomsky), "Human Nature: Justice versus Power", in
Elders 1974: 133-98.
1975 I, Pierre Riviere, having slaughtered my mother, my sister and my
brother, edited by M. Foucault, translated by F. Jellinek, Pantheon,
New York. Originally published in French as Moi, Pierre Riviere,
ayant egorge ma mere, ma soeur et mon frere, Gallimard, Paris,
1973.
235
236 BIBLIOGRAPHY
MIP Mental Illness and Psychology, translated by A. Sheridan, Harper

Colophon, New York.
1976 Uber Strajjustiz, Psychiatrie und Medizin, translated by W. Seitter,
Merve, Berlin.
LCMP Language, Counter-Memory, Practice: Selected Essays and Inter-
views, edited by D.F. Bouchard, Cornell University Press, Ithaca,
N.Y., 1977.
1977 "Die Folter, das ist die Vernunft. Ein Gesprich Knut Boesers mit
Michel Foucault", Uteraturmagazin, vol. 8: 60-68.
DP Discipline and Punish: The Birth of the Prison, translated by A.
Sheridan Smith, Pantheon, New York, 1977. Originally published
in French as Surveiller et punir. Naissance de la prison. Gallimard,
Paris, 1975.
1978 "Politics and the Study of Discourse" , Ideology and Consciousness,
vol. 3: 7-26.
P/K Power/Knowledge: Selected Interviews &: Other Writings 1972-
1977, edited by C. Gordon, Pantheon, New York, 1980.
HS 1 The History ofSexuality 1: An Introduction, translated by R. Hurley,
Pantheon, New York, 1978. Originally published in French as
Histoire de la sexualitel: La volonte de savoir, Gallimard, Paris,
1976.
1980 Herculine Barbin, Being the Recently Discovered Memoirs of a
Nineteenth Century French Hermaphrodite, edited by M. Foucault,
translated by R. McDougall, Pantheon, New York. Originally
published in French as Herculine Barbin dite Alexina B. , Gallimard,
Paris, 1978.
1981a "Foucault at the ColI~ge de France II: A Course Summary", Phi-
losophy and Social Criticism, vol. VIII: 349-59.
1981b "II faut tout repenser la lois et la prison", Liberation, July 6th.
1981c "Is it useless to revolt?", Philosophy and Social Criticism, vol. 8:
1-9.
1982 (Edited with A. Farge), Le desordre des families: Lettres de cachet
des Archives de la Bastille au XVme si~le, Gallimard, Paris.
1983 "On the Genealogy of Ethics: An Overview of Work in Progress",
in Dreyfus and Rabinow 1983: 229-52.
S&:P "The Subject and Power", in Dreyfus and Rabinow 1983: 208-26.
HS 11 The Use of Pleasure, translated by R. Hurley, Pantheon, New
York, 1985. Originally published in French as Histoire de la sexu-
alite 2: L'usage de plaisirs, Gallimard, Paris, 1984.
BIBLIOGRAPHY 237
HS III The Care ofthe Self, translated by R. Hurley, Pantheon, New York,
1985. Originally published in French as Histoire de la sexualitl 3:
Le souci de soi, Gallimard, Paris, 1984.
FR The Foucault Reader, edited by P. Rabinow, Penguin Books,
Harmondsworth, 1984.
ENL "What is Enlightenment?" , in FR 32-50.
NGH "Nietzsche, Genealogy, History", in FR 76-100.
LK VomLicht des Krieges zur Geburt der Geschichte, translated by W.
Seitter, Merve, Berlin, 1986.
QM "Questions of Method: An Interview with Michel Foucault", in
Baynes et aI. 1987: 100-17.
1987 (and B. Blanchot), Foucault. Blanchot, Zone Books, New York.
1988 "Technologies of the Self", in Martin et al. 1988: 16-49.
PPHC Politics. Philosophy. Culture. Interviews and Other Writings 1977-
1984, edited by L.D. Kritzman, Routledge, New York and London,
1988.
DI "The Dangerous Individual", in PPHC 125-51.
2. OTHER SOURCES
Ackermann, R. (1989), "The New Experimentalism", British Journalfor

the Philosophy of Science, vol. 40: 185-90.
Agassi, J. (1963), Towards an Historiography of Science, History and
Theory, Beiheft 2.
Agassi, J. (1971), Faraday as a Natural Philosopher, The University of
Chicago Press, Chicago and London.
Agassi, J. (1988a), "Cohen contra Kuhn - clues hidden in many details", in
Agassi 1988b: 123-30.
Agassi, J. (1988b), The Gentle Art of Philosophical Polemics, Open Court,
La Salle, Ill..
Althusser, L. (1970a), "From Capital to Marx's Philosophy", in Althusser
and Balibar 1970: 11-69.
Althusser, L. (1970b), "The Errors of Classical Economics: Outline of a
Concept of Historical Time", in Althusser and Balibar 1970: 91-118.
Althusser, L. and E. Balibar (1970), Reading Capital, translated by B.
Brewster, NLB, London. (The French original was published in 1968.)
Althusser, L. (1977), For Marx, translated by B. Brewster, Verso, London.
(The French original was published in 1965.)
Amnesty International (1987), United States ofAmerica. The Death Penalty,
238 BIBLIOGRAPHY
Amnesty International Publications, London.

Anderson, B. and M.L. Carlos (1976), "What is social network theory?",
in Burns and Buckley 1976: 27-51.
Bachelard, G. (1927), Essm sur 10 connaissance approch~e, Vrin, Paris.
Bachelard, G. (1940), La philosophie du Non, P.U.F., Paris.
Bachelard, G. (1951), L 'aetivitl rationaliste de 10 physique contemporaine,
P.U.F., Paris.
Bachelard, G. (1964), The Psychoanalysis ofFire, translated by A.C. Ross,
Routledge at Kegan Paul, London. (The French original was published
in 1938.)
Bachelard, G. (1971), Epistemologie. Ausgewithlte Tate, edited by D.
Lecourt, Ullstein, Fnnkfurt am Main.
Bachelard, G. (1978), Die Philosophie des Nein, translated by G. Schmidt
and M. Tietz, B. Heymann, Wiesbaden.
Bachelard, G. (1984), The New Scientific Spirit, translated by A. Gold-
hammer, Beacon Press, Boston. (The French original was published in
1934.)
Bachrach, P. and M.S. Baratz (1969a), "Decisions and Non-decisions: An
Analytical Framework", in Bell et al. 1969: 100-109.
Bachrach, P. and M.S. Baratz (1969), "Two faces of power" , in Bell et al.
1969: 94-99.
Baker, P.M.S. and G.P. Hacker (1983), An Analytical Co~ntary on Witt-
genstein's Philosophical Investigations, Basil Blackwell, Oxford.
Bakhtin, M. (1968), Rabe10is and His World, translated by H. Iswolsky,
Harvard University Press, Cambridge, Mass. (The Russian original was
first published in 1965.)
Barnes, B. (1977), Interests and the Growth of Knowledge, Routledge and
Kegan Paul, London.
Barnes, B., The Nature of Power (1988), Polity Press, Cambridge.
Barnes, B. and S. Shapin (cds.) (1979), Natural Order. Historical Studies
of Scientific Culture, SAGE, London.
Baudrillard, J. (1977), Oublier Foucault, editions galilee, Paris.
Baumgartner, T., W. Buckley, T.R. Bums and P. Schuster (1976), "Meta-
power and the structuring of social hierarchies", in Bums and Buckley
1976: 215-88.
Baynes, K., J. Bohman, and T. McCarthy (oos.) (1987), After Philosophy.
End or Transformation?, Massachusetts Institute of Technology, Cam-
bridge, Mass.
Bell, R., D.V. Edwards and at R.H. Wagner (cds.) (1969), Political
Power. A Reader in Theory and Research, The Free Press, New York,
BIBLIOGRAPHY 239
Collier-Macmillan Ltd., London.

Bellour, R. (1971), Le livre des aUlres, Editions de I'Heme, Paris.
Benton, T. (1984), The Rise and Fall of Structuralist Marxism. Althusser
and his Influence, MacMillan, Houndsmills, Basingstoke.
Berkson, W. (1974), Fields of Force. The Development of a World View
from Faraday to Einstein, Routledge and Kegan, London.
Blanchot, M. (1987), "Michel Foucault as I Imagine Him", translated by J.
Mehlman, in Foucault and Blanchot 1987: 61-109.
Bloch, M. (1924), Les rois thaumaturges, Paris.
Bloch, M. (1953), The historian's craft, translated by P. Putnam, Vintage
Books, New York. (The French original was first published in 1941.)
Bloch, M. (1962), Feudal Society, translated by L.A. Manyon, Routledge
& Kegan Paul, London.
Bloor, D. (1976), Knowledge and Socia/Imagery, Routledge and Kegan
Paul, London.
Boucher, D. (1985), Texts in Context. Revisionist Methodsfor Studying the
History of Ideas, Martinus Nijhoff, Dordrecht.
a
Bouligand, G. et al. (eds.) (1959), Hommage Gaston Bachelard, P.U.F.,
Paris.
Bourde G. and H. Martin (1983), Les koles historiques, Editions du Seuil,
Paris.
Bourdieu, P. (1988), Homo Academicus, translated by P. Collier, Polity
Press, Cambridge. (The French original was published in 1984.)
Braudel, F. (1972-74), The Mediterranean and the Mediterranean World in
the Age of Philip 11, translated by S. Reynolds, 2 vols., Harper and
Row, New York. (The French original was first published in 1949.)
Braudel, F. (1980), On history, translated by S. Matthews, The University
of Chicago Press, Chicago.
Briihmann, H. (1980), 'Der Begriffdes Hundes bellt nicht'. Das ObjekJ der
Geschichte der Wissenschaften bei Bachelard und Althusser, B. Hey-
mann, Wiesbaden.
Burguiere, A. (1982), "The Fate of the History of Mentalites in the Anna-
les", Comparative Studies in Society and History, vol. 24: 424-37.
Burguiere, A. (ed.) (1986), Dictionnaire des sciences historiques, P.U.F.,
Paris.
Burke, P. (1986), "Strength and Weaknesses oftbe History of Mentalities" ,
History of European Ideas, vol. 7: 439-51.
Bums, T. and W. Buckley (oos.) (1976), Power and Control. Social Struc-
tures and Their Transformations, SAGE Studies in International So-
ciology 6, SAGE Publicatio~s, London and Beverly Hills, Cal.
240 BIBLIOGRAPHY
Canguilhem, G. (1955), La formation du concept de rejlexe aux XVIle et

XVIlle siecles, P.U.F., Paris.
Canguilhem, G. (1959), "Sur une epistemologie concordataire", in Bouli-
gand et al. 1959: 3-12.
Canguilhem, G. (1968), Etudes d'histoire et de philosophie des sciences,
Vrin, Paris.
Canguilhem, G. (1978), On the Normal and the Pathological, translated by
C.R. Fawcett, Reidel, Dordrecht. (The French original was first pub-
lished in 1943.)
Canguilhem, G. (1988), Ideology and Rationality in the History of the Life
Sciences, translated by A. Goldhammer, The MIT Press, Cambridge,
Mass. (The French original was first published in 1977.)
Cartwright, D. (1969), "Influence, Leadership, Control", in Bell et al.
1969: 123-65.
Cartwright, N. (1983), How the Laws of Physics Lie, Oxford University
Press, Oxford.
Cassirer, E. (1932), Die Philosophie der Aujkll1rung, second edition, Mohr,
Tiibingen.
Cavaill~, J. (1987), Sur la logique et la theorie de la science, Librairie
Philosophique J. Vrin, fourth edition, Paris.
Chartier, R. (1982), "Intellectual History or Sociocultural History? The
French Trajectories", in LaCapra and Kaplan 1982: 13-46.
Chaunu, P. (1970), "L'histoire serielle. Bilan et perspectives", Revue
historUjue, vol. 494: 297-320.
Childe, V.G. (1956), Piecing Together the Past. The Interpretation of
Archaeological Data, Routledge and Kegan Paul, London.
Cbilde, V.G. (1962), A Short Introduction to Archaeology, Collier Books,
New York.
Clausewitz, K. von (1918), Yom Kriege, B. Behr's Verlag, 13th edition,
Berlin und Leipzig.
Clegg, S. (1979), The Theory of Power and Organization, Routledge and
Kegan Paul, London.
Cohen, LB. (1980), The Newtonian Revolution, Cambridge University
Press, Cambridge.
Cohen, LB. (1985), Revolution in Science, Harvard University Press,
Cambridge, Mass.
Cohen, M., T. Nagel, and T. Scanlon (008.) (1980), Marx, Justice, and
History, Princeton University Press, Princeton.
Collins, H.M. (1982), "Special Relativism", Social Studies of Science, vol.
12: 139-43.
BIBLIOGRAPHY 241
Collins, H.M. (1985), Changing Order: Replication and Induction in

Scientific Practice, SAGE, London.
Condren, C. (1985), The Status and Appraisal of Classical Texts. An Essay
on Political Theory, Its Inheritance, and the History ofIdeas, Princeton
University Press, Princeton.
Cousins M. and A. Hussain (1984), Michel Foucault, MacMillan, Hounds-
mills.
Crombie, A.C. (1981), "Philosophical Presuppositions and Shifting Inter-
pretations of Galileo" , in Hintikka et at. 1981: 271-86.
Crombie, A.C. (1986), "What is the history of science?", History of
European Ideas, vol. 7: 21-31.
Crombie, A.C. (1988), "Designed in the Mind: Western Visions of Scien
ce, Nature and Humankind", History of Science, vol. 26: 1-12.
Dahl, R.A. (1969), "The Concept of Power", in Bell et al. 1969: 79-93.
Deleuze, G. (1988), Foucault, translated by S. Hand, The Athlone Press,
London. (The French original was published in 1986.)
Dreyfus, H. and P. Rabinow (1983), Beyond Structuralism and Hermeneu-
tics, The Harvester Press, second edition, Brighton.
Duhem, P. (1905), Les origines de la statique, vol. 1, A. Herrmann, Paris.
Duhem, P. (19063), Etudes sur Uonard de Vinci, vol. 1, A. Herrmann,
Paris.
Duhem, P. (1906b),1.A Theorie physique. Son objet et sa structure, Cheva-
lier & Riviere, Paris.
Duhem, P. (1914), Le Systeme du monde. Histoire des doctrines cosmologi-
a
ques de Platon Copernic, A. Herrmann, Paris.
Duhem, P. (1969), To save the phenomena. An essay on the idea ofphysi-
cal theory from Plato to Galileo, translated by E. Doland and C. Masch-
ler, The University of Chicago Press, Chicago.
Dunn, J. (1968), "The Identity of the History of Ideas", Philosophy, vol.
43: 85-104.
Elders, F. (ed.) (1974), Reflexive Water. The Basic Concerns of Mankind,
Condor, London.
Elias, N. (1978), Uber den Prozej der Zivilisation. Soziogenetische und
psychogenetische Untersuchungen, Suhrkamp, Frankfurt am Main. (The
German original was first published in 1939.)
Elkana, Y. (1981), "A Programmatic Attempt at an Anthropology ofKnow-
ledge", in Mendelsohn and Elkana 1981: 1-76.
Elster, J. (1978), Logic and Society. Contradictions and Possible Worlds,
John Wiley & Sons, Chichester.
Elster, J. (1985), Making Sense of Marx, Cambridge University Press,
242 BIBLIOGRAPHY
Cambridge.
Eribon, D. (1989), Michel Foucault, Flammarion, Paris.
Febvre, L. (1955), "Sur Einstein et sur l'histoire", Annales E.S. C., vol. 10:
305-12.
Febvre, L. (1982), The Problem of Unbelief in the Sixteenth Century. The
Religion of Rabelais, translated by B. Gottlieb, Harvard University
Press, Cambridge, Mass. (The French original was first published in
1942.)
Fekete, J. (ed.) (1984), The Structural Allegory. Reconstructive Encounters
with the French Thought, University of Minnesota Press, Minneapolis.
Feyerabend, P. (1984), Wissenschaft ais Kunst, Suhrkamp, Frankfurt am
Main.
Feyerabend, P. (1987), Farewell to Reason, Verso, London and New York.
Fichant, M. (1969), "L'idee d'une histoire des science", in Pecheux and
Fichant 1969: 49-139.
Fleck, L. (1980), Entstehung und Entwicklung einer wissenschaftlichen
Tatsache. Einftlhrung in die Lehre vom Denkstil und Denkkollektiv,
edited and introduced by L. Schafer and T. Schnelle, Suhrkamp, Frank-
furt am Main. (First published in 1935.)
Fraser, N. (1981), "Foucault on Modem Power: Empirical Insights and
Normative Confusions", Praxis International, vol. 1: 272-87.
Fraser, N. (1985), "Michel Foucault: A 'Young Conservative'?", Ethics,
vol. 96: 165-84.
Furet, F. (1985), "Quantitative methods in history", translated by D.
Denby, in Le Goff and Nora 1985: 12-27.
Galison, P. (1987), How Experiments End, The University of Chicago
Press, Chicago.
Galison, P. (1988), "History, Philosophy, and the Central Metaphor",
Science in Context, vol. 2: 197-212.
Giddens, A. (1979), Central Problems in Social Theory. Action, Structure
and Contradiction in Social Analysis, The Macmillan Press, London and
Basingstoke.
Giere, R.N. (1988), Explaining Science. A Cognitive Approach, The Uni-
versity of Chicago Press, Chicago and London.
Goldmann, L. (1969), The Humiln Sciences and Philosophy, translated by
H.V. White and R. Anchor, Cape, London. (The French original was
first published in 1952.)
Gramsci, A. (1980), Zu Politik, Geschichte und Kultur, edited by G. Zamis,
ROderberg, Frankfurt am Main.
Greenleaf, W.F. (1966), Oakeshott's Philosophical Politics, Longmans,
BIBLIOGRAPHY 243
London.
Gutting, G. (1989), Michel Foucault's archaeology of scientific reason,
Cambridge University Press, Cambridge.
Hacking, I. (197S), The Emergence of Probability. A Philosophical Study of
Early Ideas about Probability, Induction and Statistical Inference,
Hacking, I. (1982), "Language, Troth, and Reason", in Hollis and Lukes
1982: 48-66.
Hacking, I. (1983a), "The Accumulation of Styles of Scientific Reasoning" ,
in Henrich 1983: 4S3-65.
Hacking, I. (1983b), Representing and intervening. Introductory topics in
the philosophy of natural science, Cambridge University Press, Cam-
bridge.
Hacking, I. (1984), "Five parables", in Rorty et al. 1984: 103-24.
Hacking, I. (198S), "Styles of Scientific Reasoning", in Rajchman and West
1985: 14S-6S.
Hacking, I. (1987), "Weapons Research and the Form of Scientific Knowl-
edge", Canadian Journal of Philosophy, Supplementary Volume 12:
237-61.
Hacking, I. (1988), "The Participant Irrealist At Large in the Laboratory",
British Journalfor the Philosophy of Science, vol. 39: 277-94.
Hakosalo, H. (1991), Bio-Power and Pathology: Science and Power in the
Foucauldian Histories of Medicine, Psychiatry and Sexuality, Reports
from the Department of History, University of Oulu (Finland).
Harris, E.C. (1979), Principles of Archaeological Stratigraphy, Academic
Press, London.
Harsanyi, J.C. (1969), "Measurement of Social Power, Opportunity Costs,
and the Theory of Two-person Bargaining Games", in Bell et al. 1969:
226-38.
Henrich, D. (ed.) (1983), Kant oder Hegel, Klett-Cotta, Stuttgart.
Hesse, M. (1980), Revolutions and Reconstructions in the Philosophy of
Science, Harvester, Brighton.
Hintikka, J., C.D. Groender, E. Agazzi (eds.) (1981), Theory Change,
Ancient Axiomatics, and Galileo's Methodology, Reidel, Dordrecht.
Hirsh, A. (1981), The French New Left: An Intellectual History from Same
to Gorz, South End Press, Boston.
Hollis, M. and S. Lukes (eds.) (1982), Rationality and Relativism, Black-
well, Oxford.
Holton, G. (1988), Thematic Origins of Scientific Thought. Kepler to
Einstein, Harvard University Press, Cambridge, Mass. (First published
244 BIBLIOGRAPHY
in 1973.)
Hoy, D.C. (ed.) (1986), Foucault: A Critical Reader, Basil Blackwell, New
York.
Hughes G.E. and M.J. Cresswell (1968), An Introduction to Modal Logic,
Methuen, London.
Huizinga, J. (1970), The Waning ofthe Middle Ages. A Study ofthe Forms
ofLife, Thought and Art in France and the Netherlands in the XlVth and
XVth Centuries, translated by F. Hopman, Arnold, London. (First
published in 1924.)
Jacob, F. (1974), The Logic of liVing Systems. A History of Heredity,
translated by B.E. Spillmann, Allan Lane, London. (The French original
was published in 1970.)
Kedar, L. (ed.) (1987), Power through discourse, Ablex, Norwood, N.J.
Knorr-Cetina, K. (1981), The Manufacture of Knowledge. An Essay on the
Constructivist and Contextual Nature of Science, Pergamon, Oxford.
Knorr-Cetina, K. (1983), "The Ethnographic Study of Scientific Work:
Towards a Constructivist Interpretation of Science", Knorr-Cetina and
Mulkay 1983: 115-40.
Knorr-Cetina, K. and M. Mulkay (eds.) (1983), Science Observed, SAGE
Publication, London and Beverly Hills.
Koyre, A. (1957), From the Closed World to the Infinite Universe, The
Johns Hopkins University Press, Baltimore and London.
Koyre, A. (1973), Etudes d'histoire de la pensee scientifique, Gallimard,
Paris.
Koyre, A. (1978), Galileo Studies, translated by J. Mepham, The Harvester
Press, Hassocks. (The French original was first published in 1939.)
Kragh, H. (1987), An Introduction to the Historiography of Science, Cam-
bridge University Press, Cambridge.
Kroker, A. (1984), "Modem Power in Reverse Image: The Paradigm Shift
of Michel Foucault and Talcott Parsons", in Fekete 1984: 74-103.
Kuhn, T. (1970), The Structure of Scientific Revolutions, second edition,
The University of Chicago Press, Chicago.
Kusch, M. (1989), lAnguage as Calculus vs. lAnguage as Universal
Medium. A Study in Husserl, Heidegger and Gadamer, Kluwer, Dor-
drecht.
Komer, S. (1984), Metaphysics: its structure and function, Cambridge
University Press, Cambridge.
LaCapra, D. and St.L. Kaplan (eds.) (1982), Modern European Intellectual
History. Reappraisals and New Perspectives, Cornell University Press,
Ithaca and London.
BIBLIOGRAPHY 245
Lakatos, I. (1974), "Falsification and the Methodology of Scientific Re-

search Programmes", in Lakatos and Musgrave 1974: 91-196.
Lakatos, I. and A. Musgrave (eds.) (1974), Criticism and the Growth of
Knowledge, Cambridge University Press, Cambridge.
Lakatos, I. (1978), The methodology of scientific research programmes,
Philosophical Papers Volume I, edited by J. Worrall and G. Currie,
Latour, B. (1983), "Give me a Laboratory and I will Raise the World", in
Knorr-Cetina and Mulkay 1983: 141-70.
Latour, B. (1987), Science in Action, Open University Press, Milton Key-
nes.
Latour, B. (1988), The Pasteurization ofFrance, Harvard University Press,
Cambridge, Mass.
Latour, B. and S. Woolgar (1986), Laboratory Life. The Construction of
Scientific Facts, Introduction by J. Salk, with a new postscript by the
authors, Princeton University Press, Princeton, N.J. (First published in
1979.)
Laudan, L. (1977), Progress and Its Problems. Towards a Theory of
Scientific Growth, Routledge and Kegan Paul, London and Henley.
Laudan, L. (1984), Science and Values. The Aims ofScience and Their Role
in Scientific Debate, University of California Press, Berkeley.
Laudan, L., A. Donovan, R. Laudan, P. Barker, H. Brown, J. Leplin, P.
Thagard, S. Wykstra (1986), "Scientific Change: Philosophical Models
and Historical Research", Synthese, vol. 69: 141-223.
Lawrance, C. (1980), "The Nervous System and Society in the Scottish
Enlightenment", in Rousseau and Porter 1980: 19-40.
Le Roy Ladurie, E. (1972), Times of Feast, Tunes of Famine: A History of
the ClimlJle since the Year ]000, translated by B. Bray, George Allen &
Unwin, London. (The French original was published in 1969.)
Le Roy Ladurie, E. (1981), Tht: Mind and Method of the Historian, trans-
lated by S. Reynolds and B. Reynolds, The University of Chicago Press,
Chicago. (The French original was published in 1978.)
Lecourt, D. (1975), Marxism and Epistemology. Bachelard, Canguilhem
and Foucault, translated by B. Brewster, NLB, London. (The book
contains two parts, "Part one: Gaston Bachelard's historical epistemolo-
gy" (pp. 23-118, first published in French in 1969), and "Part two: For
a critique of epistemology" (pp. 119-216, first published in French in
1972).)
Le Goff, J. (1985), "Mentalities: a history of ambiguities" ,translated by D.
Denby, in Le Goff and Nora 1985: 166-80.
246 BmuOGRAPHY
Le Goff, J. and P. Nora (eds.) (1985), Constructing the past. Essays in

historical methodology, Cambridge University Press, Cambridge.
Lemert, C. and G. Gillan (1982), Michel Foucault. Social Theory as
Transgression, Columbia University Press, New York.
Lepenies, W. (1978), "Wissenschaftsgeschichte und Disziplingeschichte",
Geschichte und Gesellschajt, vol. 4: 437-51.
Leslie, M. (1970), "In defense of Anachronism", Political Studies, vol.
XVIn: 433-47.
Uvi-Strauss, C. (1966), The Savage Mind, Weidenfold and Nicolson,
London. (The French original was published in 1962.)
Uvi-Strauss, C. (1967), Structural Anthropology, translated by C. Jacobson
and B.G. Schoepf, Anchor, Garden City, N.Y. (The French original was
published in 1958.)
Uvi-Strauss, C. (1976), Tristes Tropiques, translated by J. and D. Weight-
man, Penguin Books, New York. (The French original was published in
1955.)
Lewis, D. (1973), Counterfactuals, Harvard University Press, Cambridge,
Mass.
Lewis, D. (1983a), "Counterpart Theory and Quantified Modal Logic", in
Lewis 1983b: 26-46.
Lewis, D. (1983b), Philosophical Papers, Vol. I, Oxford University Press,
Oxford.
Lorrain, F. and H. C. White (1971), "Stroctural Equivalence ofIndividuals
in Social Networks", Journal ofMathematical Sociology, vol. 1: 49-80.
Lovejoy, A. (1960), Essays in the History ofIdeas, Capricorn Books, New
York. (First published in 1948.)
Lovejoy, A. (1973), The Great Chain ofBeing. A Study ofthe History ofan
Idea, eleventh printing, Harvard University Press, Cambridge, Mass.
(First published in 1936.)
Lukes, S. (1974), Power: A Radical View, Macmillan, Houndmills.
Lukes, S. (1985), Marxism and Morality, Clarendon Press, Oxford.
Lukes, S. (1986a), "Introduction", in Lukes 1986b: 1-18.
Lukes, S. (ed.) (1986b), Power, Blackwell, Oxford.
MacIntyre, A. (1987), "Relativism, Power, and Philosophy", in Baynes et
al. 1987: 385-411.
MacIntyre, A. (1988), Mose Justice? Which Rationality?, Duckworth,
London.
Mandrou, R. (1975), Introduction to modern France 1500-1640. An essay
in historical psychology, translated by R.E. Hallmark, Holmes & Meier,
New York. (The French original was published in 1961.)
BIBLIOGRAPHY 247
Martin, L.H., H. Gutman & P.H. Hutton (eds.) (1988), Technologies ofthe
Self. A Seminar with Michel Foucault, Tavistock, London.
Mendelsohn, E. and Y. EIkana (eds.) (1981), Sciences and Cultures.
Anthropological and Historical Studies of the Sciences, Reidel, Dor-
drecht.
Merquior, J.G. (1985), Foucault, Montana, London.
Merquior, J.G. (1986), From Prague to Paris. A Critique of Structuralist
and Post-Structuralist Thought, Verso, London.
Minogue, K. (1959), "Power in Politics", Political Studies, vol. 7: 269-83.
Nehamas, A. (1985), Nietzsche. Life as Literature, Harvard University
Press, Cambridge, Mass.
Nietzsche, F. (1968), The Will to Power, translated by W. Kaufmann and
R.J. Hollingdale, Vintage Press, New York.
O'Farrell, C. (1989), Foucault. Historian or Philosopher, MacMillan,
Houndsmills.
Osnowski, R. (ed.) (1988), Menschenversuche, KaIner Volksblatt Verlag,
Kaln.
Outram, D. (1989), The Body and the French Revolution. Sex, Class, and
Political Culture, Yale University Press, New Haven and London.
Parfit, D. (1984), Reasons and Persons, Clarendon Press, Oxford.
Parsons, T. (1969), "On the Concept of Political Power", in Bell et al.
1969: 251-84.
P&heux, M. (1969), "Ideologie et histoire des sciences. Les effets de la
coupure galil6enne en physique et en biologie", in Pecheux and Fichant
1969: 13-47.
Pecheux, M. and M. Fichant (1969), Sur l'histoire des sciences, Fran~is
Masperio, Paris.
P&heux, M. (1982), Language, Semantics and Ideology, translated by H.
Nagpal, Macmillan, London. (The French original was published in
1975.)
Petit Larousse IUustr~ 1988, Larousse, Paris.
Pocock, J.G.A. (1972), Politics, Language and Time, Methuen, London.
Popper, K. (1974), "Normal Science and its Dangers", in Lakatos and
Musgrave 1974: 51-58.
Poulantzas, N. (1978a), Political Power and Social Classes, translated by
T. O'Hagan, Verso, London. (The French original was published in
1968.)
Poulantzas, N. (1978b), Staatstheorie. Politischer Uberbau, Ideologie,
Sozialistische Demokratie, translated by H. Arenz, T. Brackmann, H.
Friedhoff and R. LOper, VSA, Hamburg. (The French original was
248 BIBLIOGRAPHY
published in 1977.)
Rajchman, J. (1984), Michel Foucault: thefreedom ofphilosophy, Columbia
University Press, New York.
Rajchman, J. and C. West (eds.) (1985), Post-Analytic Philosophy, Colum-
bia University Press, New York.
Rawls, J. (1972), A Theory ofJustice, Clarendon Press, Oxford.
Reingold, N. (1980), "Through Paradigm-Land to a Normal History of
Science", Social Studies of Science, vol. 10: 475-96.
Rescher, N. (1981), Leibniz's Metaphysics of Nature, Reidel, Dordrecht.
Revel, J. (1986a), "Foucault, Michel, 1926-1984", in Burgui~re 1986:
290-92.
Revel, J. (I986b), "Mentalites", in Burgui~re 1986: 451.
Rorty, R., J.B. Schneewind and Q. Skinner (eds.) (1984), Philosophy in
History, Cambridge University Press, Cambridge.
Rouse, J. (1987), Knowledge and Power. Toward a Political Philosophy of
Science, Cornell University Press, Ithaca and London.
Rousseau, O.S. and R. Porter (eds.) (1980), The Ferment of Knowledge.
Studies in the historiography of eighteenth-century science, Cambridge
University Press, Cambridge.
Sachs, M. (1973), The Field Concept in Contemporary Science, Charles C.
Thomas, Springfield, Ill.
Sarton, o. (1975), Introduction to the History of Science. Volume 1: From
Homer to Omar Khayyam, Krieger, Huntington, N.Y.
Sartre, J.-P. (1969), Being and Nothingness, translated by H. Barnes,
Methuen & Co., London. (The French original was published in 1943.)
Schmitt, K. (1922), Der Begriff des Politischen, Duncker & Humblot,
Munchen und Leipzig.
Shapin, S. (1982), "History of Science and Its Sociological Reconstruc-
tion", History of Science, vol. 20: 157-211.
Shapin S. and S. Schaffer (1985), Leviathan and the Air-Pump. Hobbes,
Boyle, and the Experimental Life, Princeton University Press, Princeton.
Shapin, S. (1988), "Robert Boyle and Mathematics: Reality, Representa-
tion, and Experimental Practice", Science in Context, vol. 2: 23-58.
Sheridan, A. (1980), Michel Foucault. The Will to Truth, Tavistock Publi-
cations, London.
Siegel, H. (1987), Relativism Refuted, Reidel, Dordrecht.
Skinner, Q. (1969), "Meaning and Understanding in the History of Ideas",
History and Theory, vol. 8: 3-53.
Smart, B. (1985), Michel Foucault, Tavistock, London.
Steiner, o. (1971), "The mandarin of the hour - Michel Foucault", New
BmLIOGRAPHY 249
York Review of Books, February 28: 8 & 28-31.

Stoianovich, T. (1976), French Historical Method. The Annales Paradigm,
Cornell University Press, Ithaca and London.
Taylor, C. (1986), "Foucault on Freedom and Truth", in Hoy 1986:
69-102.
Toulmin, S. (1972), Human Understanding, Vol. I. General Introduction
and Part I, Clarendon Press, Oxford.
Turner, S. (1989), "Depolicizing Power", Social Studies of Science, vol.
19: 533-60.
Ullmann-Margalit, E. (1978), "Invisible-Hand Explanations" , Synthese, vol.
39: 263-91.
Veyne, P. (1981), Der Eisberg der Geschichte. Foucault revolutioniert die
Historie, translated by K. Tholen-Struthoff, Berlin. (The French original
was published in 1979.)
Veyne, P. (1984), Writing History. Essay on Epistemology, translated by
M. Moore-Rinvolucri, Manchester University Press, Manchester. (The
French original was published in 1971.)
Vilar, P. (1976), A History of Gold and Money 1450-1920, translated by J.
White, NLB, London. (The Spanish original was published in 1969.)
Vilar, P. (1985), "Constructing Marxist history", translated by I. Patterson,
in Le Goff and Nora 1985: 47-80.
Waldenfels, B. (1983), Phlinomenologie in Frankreich, Suhrkamp, Frank-
furt am Main.
Weber, M. (1970), From Max Weber, translated, edited and with an Intro-
duction by H.H. Gerth and C. Wright Mills, Routledge and Kegan Paul,
London.
Wilson, E.O. (1977), "Biology and the Social Sciences", Daedalus, vol. 2:
127-40.
Wisan, W.L. (1981), "Galileo and the emergence ofa new scientific style",
in Hintikka et al. 1981: 311-39.
Woolgar, S. (1983), "Irony in the Social Sciences", in Knorr-Cetina and
Mulkay 1983: 239-66.
Woolgar, S. (ed.) (1988), Knowledge and Reflexivity. New Frontiers in the
Sociology of Knowledge, SAGE, London.
Wootton, D. (1988), "Lucien Febvre and the Problem of Unbelief in the
Early Modem Period" , JOUT7UJ1 ofModern History, vol. 60: 695-730.
Wright, G.H. von (1984), Truth, Knowledge, and Modality, Blackwell,
Oxford.
Wrong, D.H. (1979), Power. . Its Forms, Bases and Uses, Blackwell,
Oxford.
INDEX OF NAMES
Ackermann, R. 114
Adorno, T.W. 195
Agassi, J. 49, 88, 112, 229, 234
Althusser, L. 1, 10, 12,35-41,54,57,58, 103, 105, 112, 141, 182
Anderson, B. 229
Archimedes 26
Aries, P. 41, 112, 180
Aristotle 49, 53, 92
Austin, J.L. 59
Bachelard, G. xii, 1-6, 10, 12, 27-37, 40, 54, 57, 86, 89, 90, 103, 105,
106, 110, 112, 135, 136
Bachrach, P. 227
Bacon, R. 114
Baker, P.M.S. 233
Bakhtin, M. 180, 232
Baratz, M.S. 227
Barnes, B. 117, 118, lSI-ISS
Baudrillard, J. 229, 233
Baumgartner, T. 229
Benton, T. 36
Berkson, W. 229
Berr, H. 13
Binet, A. 34
Blanchot, M. 1
Bloch, M. 6, 13-21, 112, 169-172, 180
Bloor, D. 117
Bohr, N. 31,49,53
Boulainvilliers, N. 228
Bourde, G. 13, 112
Bourdieu, P. 156
Boyle, R. 65, 67, 68, 70, 71, 74, 109, 159, 160
Braudcl,F. 6,14,18,20,37,38,41,112,171,172
Briihmann, H. 30, 32
Buckley, T.W. 229
Burckhardt, J. 15,39, 55
Burguiere, A. 16
Burke, P. 18
Bums, T.R. 229
250
INDEX OF NAMES 251
Canguilhem, G. xii, 1, 3, 6, 12,32-37,40,41,53, 85, 86, 90, 106, 110,

112, 113, 199
Carlos, M.L. 229
Carnap, R. 89
Cartwright, D. 228
Cartwright, N. 128
Cassirer, E. 1, 55
Cavailles, J. 1, 6, 110, 112
Chadwick, J. 27
Chartier, R. 15, 18, 40, 112
Chaunu, P. 21
Childe, V.G. 7,8,42
Chomsky, N. 75, 223
Clausewitz, K. von 129, 133, 134, 144, 228
Clement, N. 34
Cohen, I.B. 51-54, 83, 84, 113
Cohen, M. 234
Coke, E. 228
Collins, H.M. 117, 197, 231
Comte, A. 34
Condren, C. 49, 50, 54, 112
Copernicus, N. 27, 64, 76, 91
Cresswell, M.J. 113
Crombie, A. 94,95, 106, 110
Dahl, R. 227
Darwin, C. 33,34,51,64,92
Deleuze, G. 2, 229
Democritos 26, 53
Derrida, J. xi
Desargues, G. 54
Descartes, R. 35, 46
Desormes, C.B. 34
Diderot, D. 46
Dilthey, W. 55
Dirac, P. 28
Dreyfus, H. 2, 3, 59
Duhem, P. 1,24-26,88
Dunn, J. 112
Durkheim, E. 13
Einstein, A. 27, SO, 51, 107
252 INDEX OF NAMES
Elias, N. 180, 232

Elkana, Y. 114
Elster, J. 134, 227, 229, 234
Erasistratos 185
Estaing, C.H. 228
Euclid 31, 32
Faraday, M. 107, 136, 138, 139, 185, 229
Farge, A. 147
Febvre, L. 13-20, 38, 40, 112, 169-171
Fechner, G.T. 34
Feyerabend, P. 89, 90, 103, 104, 113
Fichant, M. 25, 88, 112, 113
Fleck, L. 113
Fontenelle, B. 46
Fraser, N. 128,221,222
Freud, S. 64, 84, 92
Furet, F. 20-24, 112
Galilei, G. 24, 25, 89
Galison, P. 86, 87, 90, 95-97, 102, 103
Gazdar, G. 75
Giddens, A. 227
Gillan, G. 2, 4
Goethe, J.W. von 92
Goldmann, L. 1, 55, 56
Goubert, P. 169
Gramsci, A. 228
Greenleaf, W.F. 112
Gutting, G. 2, 35, 112
Habermas, J. xi, 191,223
Hacker, G.P. 233
Hacking, I. 8, 85, 90, 94-96, 106, 108, 113, 128, 129, 196, 230
Harris, E. 9-12
Harsanyi, J.C. 227
Hegel, G.W.F. 31, 37-39,43, 117,226
Heidegger, M. 2
Herophilos 185
Hesse, M. 233
Hobbes, T. 70, 71, 109, 159, 160
Holton, G. 50, 51
Honegger, C. 112
INDEX OF NAMES 253
Horkheimer, M. 195
Hugbes, G.E. 113
Huizinga, J. 15
Husserl, E. 2, 6, 199
Jacob, F. 112
Joyce, J. 4
Kant, I. 6, 128, 226
Kedar, L. 228
Kepler,J. 25,26, SO
Knigbt, D. 10
Knorr-Cetina, K. 196
Korhonen-Kuscb, R. 112
Koyre, A. 1,24,26,27,40,85,89,90
Kragb, H. 9, 49
Kroker, A. 229
Kubn, T. 2,8, 12, 77, 85, 86, 89-91,93-99, 102-106, 113
Komer, S. 112
Labrousse, C.-E. 20
Lakatos, I. 90, 103, 104, 110
Langlois, C.-v. 13
Laplace, P.S. 34
Latour, B. 62, 112, 117, 190, 191, 197, 198, 232
Laudan, L. 85, 86, 90-98, 102-104, 110, 112-114
Lawrance, C. 181
Le Goff, J. 15, 18, 19
Le Roy Ladurie, E. 18,20,21, 169, 171
Lecourt, D. 27, 31, 34, 35, 85, 86
Leibniz, G.W. 46, 76, 130, 131,227
Lemert, C. 2, 4
Lepenies, W. 8, 34, 112
Leslie, M. 114
UVi-Strauss, C. 5, 9, 169
Uvy-Brubl, L. 15, 112
Lewis, D. 78-81,230
Lilbume, J. 228
Locke, I. 39
Lorrain, F. 229
Lovejoy, A. 1,44-47, SO, 51, 54, 112
Lukacs, G. 55
Lukes, S. 117-122, 127, 129,227,234
254 INDEX OF NAMES
Lyell, C. 51
Mach, E. 52
MacIntyre, A. 193, 200, 207-209, 212, 233
Mallarme, S. 55
Mandrou, R. 19, 20, 180
Mannheim, K. 191
Martin, H. 13, 112
Mux,K. 1,36,37,39, 128, 129, 132, 133, 141, 145, 158,177,178,182,
224,227
Maxwell, I.C. 31, 107
Meiland, 1. 205, 206
Mendel, G. 33, 84, 92
Merleau-Ponty, M. 6, 112
Merquoir, I.G. 194
Merton, R.K. 191
Michelet, 1. 54
Minogue, K. 139
Nagel, T. 234
Nehamas, A. 193,231
Newton, I. 26, 28,49,51,92,107,135
Nietzsche, F. xiii, 133, 166-168, 173, 178, 186, 191, 193,226,231
O'Farrell, C. 1, 2
Osnowski, R. 232
Outram, D. 232
Palingenius 27
Paracelsus 113
Parfit, D. 229
Parsons, T. 229
Pasteur, L. 190
Pecheux, M. 112, 113, 233
Plato 26, 52, 64, 171, 200-204
Pocock, I.G.A. 112
Poe, E. 55
Popper, K. 90, 113
Poulantzas, N. 141, 227, 233
Prochaska, G. 35
Protagoras 201-205
Quetelet, A. 34
Rabinow, P. 2, 3, 59
Rajchman, 1. 194
INDEX OF NAMES 255
Raspail, F.-V. 31
Rawls, J. 117,225
Reingold, N. 113
Rescher, N. 227
Revel, J.-F. 1, 16,41, 112
Ricoeur, P. 6, 112
Rouse, J. 230, 232
Sachs, M. 229
Sarton, G. 11, 88
Sartre, J.-P. 56, 124
Scanlon, T. 234
Schaffer, S. 65-71, 159-161, 175
Schmitt, C. 228, 229
Schopenhauer, A. 229
Schuster, P. 229
Searle, J. 2
Seignobos, C. 13
Shapin, S. 65-71, 159-161, 175, 189
Sheridan, A. 112
Siegel, H. 200-207, 212
Simiand, F. 20
Skinner, Q. 49, 50, 54, 55, 112, 114
Smart, B. 3
Spinoza, B. 39,49
Stoianovich, T. 112
Taylor, C. 211,229, 233
Thierry, A. 228
Toulmin, S. 90-92, 106, 107, 113
Turner, S. 140
Tylor, E.B. 34
Ullmann-Margalit, E. 229
Veyne, P. 170, 176, 178, 199, 227, 232
Vilar, P. 112
Waldenfels, B. 112
Wallace, A.R. 33
Wallon, H. 15
Weber, M. 124, 226, 227
White, H.C. 229
Willis, T. 35
Wilson, E.O. 112
256 INDEX OF NAMES
Wisan, W.L. 113

Wittgenstein, L. 195
Woolgar, S. 62, 197, 198,232
Wootton, D. 17, 18
Wright, G.H. von 231
Wrong, D.H. 227,228
INDEX OF SUBJECfS
anachronism 3, 5, 33, 100, 110, 111

animism 28-30
anthropology xi, 12, 45, 110, 156, 170-173
- structural 5, 171
anticipation 31, 49, 88
antipositivist revisionism 90
apparatus 36, 143, 199
archaeological isotopia 79, 81
archaeological isomorphism 81
archaeological model 81
archaeology
- of knowledge xiii, xiv, 1-115
- proper 6-14
- stratigraphy 7-12
archive 22
associated domain 61-63
author xiv, 55, 62, 158, 159
authority 120, 121, 127, 146, 1:47, 151, 154
chance 173, 174
coercion 120, 121, 127, 146, 151, 154, 165, 182, 183, 215, 220
coercive institution 145, 147, 165, 182, 183,215,220
coherence 54-58
commentary 158, 159
concept 8, 30, 35, 37, 48, 60, 62, 64, 65, 68, 71-73, 80, 81, 100, 101,
108, 109, 111
condition of admission 158, 159
constructivism xi, 8
- historical xii, 20-23, 38, 41, 43, 111
- realist vs. irrealist 193, 196-199
contingency 167, 172-174,212
continuity vs. discontinuity xi, xii, xiv, 5, 11, 12, 17, 25-27, 30-32,
37-39,42,43,47,51,58,83-111
contradiction 46, 54-58, 71, 74, 177
control, ecological 132, 146, 147, 154, 165,228
controversity 175, 176
conventionalism 29
correlate 61, 62
counterfinality explanation 148, 165, 229
counterpart 74, 75, 78, 79, 82
257
258 INDEX OF SUBJECTS
derivation tree 71, 73

deterrence 120, 146, 151
discourse 3, 8, 11, 49, 66, 71, -75, 77, 78, 105-109, 115, 182, 198, 210,
211,222
- discourse analysis III
- orders of 165
discursive formation xiv, 58-83, 97,99-102, 105-110, 115
dispositif 143, 145, 147, 149, 151, 177,231
distrust 167, 168, 186
document 22,41,54
empirism 26, 27, 29, 86
Enlightenment 55, 225, 226
enunciation 63, 101
enunciative modality 64, 65, 67,71-73,81, 100, 101, 108-111
epistemological act 29-33
epistemological obstacle 29-33
epistemological profile 29, 32
epistemology, historical xii, 1, 5, 6, 12, 24, 27-35, 40-43, 50, 57, 105,
136
epoche 199
event 14, 20-23, 31, 41, 43,58,59, 101, 170-172,175, 176, 230, 231
exclusion 36, 100, 115, 151, 157-159, 165,224
experimentalism xi, 65, 96, 114
field of force 128, 129, 136, 138
force 120, 121, 123, 127, 218
formalism 29
form of knowledge 94, 152,214,218
formulation 63
genius 25, 33, 48, 52-54
genealogy
- Foucauldian xii, 115-234
- Nietzschean 166-168
geology 5,9
guiding assumption 103, 104
historical psychology 19, 45, 180
history
- accepted 31, 32, 110, 111
- Annalist xiii, 1, 4, 5, 6, 12-24,37,40-43,50,94, 111, 112, 169-
175, 180, 186
- general 43
- narrative 14, 23, 31, 86, 170, 196
- obsolete 31
- of ideas xii, 1,3,5,9, 15, 18,40,43-58,94, 111,230
- of mentalities 15-21,45,94, 180, 186
- of the body 178-186
- quantitative 20, 21
- recurrent 31, 110, 111
- serial 12,20-24, 38, 41-43, 54, 58, 112
- total 43
- traditional 13-15,21-24,42,55, 170
- vertical 9, 10
humanism 37, 167, 186, 195, 216, 222, 225, 226
idealism 29
ideas (themata) 9, 14,25,44-58
- unit-ideas 45, 47, 48, SO
ideology 33, 37, 86, 103, 182
incommensurability 17,97, 102,213
influence 31,33,49,59, 121
interest 115, 119, 121, 124, 145, 155, 162, 163, 165, 186-191,213,231
interface 11, 77
internal procedure 158, 159
interrelation (between discursive formations) 64, 71-77, 100, 108
invisible hand explanation 143, 148, 149, 161, 165, 229
irony 167, 168, 186, 191, 192, 232
itinerary 170, 176, 178
justice 224, 225
line of force 138, 139
manipulation 120, 121, 127, 146, 151, 191, 220, 224
marxism 5, 128, 129, 145, 158, 177, 178
- structural 1, 35-40, 227
mentality (mentalit6) 5, 12, 15-21,28,44,45,48,54,94, 110, 112, 180,
186
microphysics 27, 128
necessity 173, 174
nominalism 197, 198
object 56,60-67, 71-73, 76, 100, 101, 108, 109, 111, 115, 186
oeuvre xiv, 8, 54, 55, 59
origin 14, 34, 46, 48, 212, 213
originality SO, 54
panopticism 146, 148, 188, 192, 225
paradigm 90, 92-94, 97, 103, 105-109, 163
personal identity 126, 134-137, 165, 168, 181,231
perspectivism 168, 186, 193, 207-209, 212

phenomenology 2,5,6, 199
philosophy of history 9, 43
physicalism 167, 168, 178-186
plot 176-178
political anatomy 183
positivism 25, 26, 29, 86, 88, 95
possible worlds 76, SO, 81
postmodernism (in the philosophy of science) 86,95, 102
power
- amount of 120, 161, 192
- as capacity 125, 154
- as domination 154, 167
- as interaction 130-132
- as possession 130, 132
- base of 119, 126
- concept of 117, 118, 122., 139, 152
- conception of 117, 118, 126, 130, 132
- costs of 120
- counterfactual analysis of 118, 119, 123
- definition of 122, 129, 140
- economism of 132, 133
- effects of 158, 167, 172,214,230
- exercise of 130, 131, 138, 149-151,219,228
- extension of 120, 161
- hegemonial vs. non-hegemonial forms of 218
- holder 119, 123, 124
- intensity of 120, 161
- interests lIS, 121, 162, 224
- /knowledge 118, 126, 149-165,211,233
- macrolevel 130, 143-148, ISS, 165
- means of 120, 126, 161
- mechanism 115, 116, 118, 127, 131-134, 137, 143, 146, 147, ISO,
156,157,161,165,172,182-184,190,191,210,220,222,228
- methodology of (for study of) 117, 165-192
- microlevel 126, 130, 143-148, ISS, 165
- microphysics of 128, 135, 136
- models of 118, 127-129, 132, 133, 192
- mystification of 139, 140
- network: of xii, lIS, 116", 126, 127, 138-148, ISO, ISS, 161, 164,
165, 175, 185, 231
- omnipresence of 118, 138-142, 153, ISS, 219, 220, 229

- physical 118, 119
- power-over 154, 155
- power-to 125, lSI, 154
- productivity of 149, ISO, 161
- relation of 117-120, 124-140, 144, 147, ISS, 163-165, 217-219,
226,227
- scope of 120, 161
- social xiv, 118, 119, 153, 156, 160, 162, 181,210
- species of 118, 120, 127, 146
- strength of 120
- structures 130
- subject 119, 120, 123-125, 131, 135
- theory of xiii, 116, 117, 130, 134, 155, 165,217,221,233
- vicious forms of 117, 140
- zone of acceptance of 120
precursor, predecessor 25,31, 33, 49, 53, 88
prehistory 5, 8, 37, 105, 107, 111
problematic 30, 31, 33, 36, 58, 207
progress 23, 25, 28, 30-32, 48, 88, 90, 185, 186, 195, 196
psychoanalysis 5, 6, 40
rationality, scientific 90, 102, 162, 207-209, 211, 212, 214-216, 225
- reticulated model of xi, 97, 98
realism 28, 29, 152, 163
- complex 28
recovering of agonistic events 167-178
reduction 142
referential 61, 62
region of interpositivity 80, 82
regularity 64, 70, 72, 73, 75, 81, 100, 230
relation 100
- accessibility 75, 80-82
- accidental 134-136, 150, 161
- discursive 66, 67, 73, 74, 81
- essential 134-136, 147, 150, 152, 161, 163,215
- external 130-136, 150, 161
- internal 130-136, 147, 150, 152, 161, 163,215,227,231
- primary 66
- reflexive 66
relativism 94, 193, 196, 197,200-213,215
repeatability 61, 63
resistance 124, 125, 139, 177,217-223

revolution(s), scientific 3, 24, 26, 27, 51, 83-85,91-93, 104
rules 73, 75, 100
sections (in archeaology) 10, 87
series 5,7,8,23,41-43,58,59,64,66, 71, 73,75, 100, 101, 109-111,
115,230
sign 61, 64, 70
skepticism 193-196
sociology of science xi-xii, 3, 4, 65-71, 175, 176, 181, 187, 189, 196,
198,232
statement 8, 12,50,54,57-64,68-80,82, 100, 108, 111, 115, 158
strategy 64, 70-72, 100, 101, 108, 111, 124, 126, 144, 145, 148, 149,
165, 178, 192, 210
structuralism 9, 85, 169-172
style 93, 94, 113
- of reasoning 94,95, 107
subjective position 60, 61, 67
subjugated knowledge iS7, 165, 195, 211, 214, 220
tactics 144, 145, 148, 165, 230
technologies of the self 221
teleology 31, 43, 48
threshold 22, 23, 32,43, 57, 103, 105, 106, 107, 113
time 5, 23, 32, 37,40
- archaeological 8
- economic 38, 39
- ideological 37-39
- temporality 23
- time scales 6, 8, 14, 19, 21, 28
totalization 41, 46, 47
tradition 25, 48, 59, 77,92,93, 103, 110,207-209
transformation 12,51-53,59,99, 141, 142,216,231
truth 150, 148, 168, 182, 183, 192, 197,203,207, 213, 218
- absolute notion of 178, 203, 205, 206
- regime of 158, 196, 210, 211, 212
- relative notion of 203, 205, 206
underdetermination 161-165
violence 120, 123-125, 215, 218
Weltanschauung 23,43,54,55
Weltbild 42, 56, 59
Zeitgeist 15, 23, 39, 42, 43, 54, 59, 85
SYNTHESE LffiRARY
Studies in Epistemology, Logic, Methodology, and Philosophy ofScience
1. J. M. Bochenski, A Precis of Mathematical Logic. Translated from French and
Genoan by O. Bird. 1959 ISBN 90-277-0073-7
2. P. Guiraud, Problemes et methodes de la statistique iinguistique. 1959
ISBN 90-277-0025-7
3. H. Freudenthal (ed.), The Concept and the Role of the Model in Mathematics and
Natural and Social Sciences. 1961 ISBN 90-277-0017-6
4. E. W. Beth, Formal Methods. An Introduction to Symbolic Logic and to the Study
of Effective Operations in Arithmetic and Logic. 1962 ISBN 90-277-0069-9
5. B. H. Kazemier and D. Vuysje (eds.), Logic and Language. Studies dedicated to
Professor Rudolf Carnap on the Occasion of His 70th Birthday. 1962
ISBN 90-277-0019-2
6. M. W. Wartofsky (ed.), Proceedings of the Boston Colloquium for the Philosophy of
Science, 1961-1962. [Boston Studies in the Philosophy of Science, Vol. I] 1963
ISBN 90-277-0021-4
7. A. A. Zinov'ev, Philosophical Problems of Many-valued Logic. A revised edition,
edited and translated (from Russian) by G. KUng and D.O. Comey. 1963
ISBN 90-277-0091-5
8. G. Gurvitch, The Spectrum ofSocial Time. Translated from French and edited by M.
Korenbaum and P. Bossenoan. 1964 ISBN 90-277-0006-0
9. P. Lorenzen, Formal Logic. Translated from Genoan by FJ. Crosson. 1965
ISBN 90-277-OO80-X
10. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings of the Boston Colloquiumfor
the Philosophy of Science, 1962-1964. In Honor of Philipp Frank. [Boston Studies
in the Philosophy of Science, Vol. II] 1965 ISBN 90-277-9004-0
11. E. W. Beth, Mathematical Thought. An Introduction to the Philosophy of Mathe-
matics.1965 ISBN 90-277-0070-2
12. E. W. Beth and J. Piaget, Mathematical Epistemalogy and Psychology. Translated
from French by W. Mays. 1966 ISBN 90-277-0071-0
13. G. KUng, Ontology and the Logistic Analysis of Language. An Enquiry into the
Contemporary Views on Universals. Revised ed., translated from Genoan. 1967
ISBN 90-277-0028-1
14. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings of the Boston Colloquium for
the Philosophy of Sciences, 1964-1966. In Memory of Norwood Russell Hanson.
[Boston Studies in the Philosophy of Science, Vol. III] 1967 ISBN 90-277-0013-3
15. C. D. Broad,Induction, Probability, and Causation. Selected Papers. 1968
ISBN 90-277-0012-5
16. G. Patzig, Aristotle's Theory of the Syllogism. A Logical-philosophical Study of
Book A of the Prior Analytics. Translated from Genoan by J. Barnes. 1968·
ISBN 90-277-0030-3
17. N. Rescher, Topics in Philosophical Logic. 1968 ISBN 90-277-0084-2
18. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings of the Boston Colloquiumfor
the Philosophy of Science, 1966-1968, Part I. [Boston Studies in the Philosophy of
Science, Vol. IV] 1969 ISBN 90-277-0014-1
19. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings ofthe Boston Colloquiumfor
the Philosophy ofScience, 1966-1968, Part /I. [Boston Studies in the Philosophy of
Science, Vol. V] 1969 ISBN 9O-277-OO15-X
SYNTHESE LffiRARY
20. J. W. Davis, D. J. Hockney and W. K. Wilson (eds.), Philosophical Logic. 1969

ISBN 90-277-0075-3
21. D. Davidson and J. Hintikka (eds.), Words and Objections. Essays on the Work of
W. V. Quine. 1969, rev. ed. 1975 ISBN 90-277-0074-5; Ph 90-277-0602-6
22. P. Suppes, Studies in the Methodology and Foundations ofScience. Selected Papers
from 1951 to 1969.1969 ISBN 90-277-0020-6
23. J. Hintikka, Modelsfor Modalities. Selected Essays. 1969
ISBN 90-277-0078-8; Ph 90-277-0598-4
24. N. Rescher et al. (eds.), Essays in Honor of Carl G. Hempel. A Tribute on the
Occasion of His 65th Birthday. 1969 ISBN 90-277-0085-0
25. P. V. Tavanec (ed.), Problems of the Logic ofScientific Knowledge. Translated from
Russian. 1970 ISBN 90-277-0087-7
26. M. Swain (ed.), Induction, Acceptance, and Rational Belief. 1970
ISBN 90-277-0086-9
27. R. S. Cohen and R. J. Seeger (eds.), Ernst Mach: Physicist and Philosopher. [Boston
Studies in the Philosophy of Science, Vol. VI]. 1970 ISBN 90-277-0016-8
28. J. Hintikka and P. Suppes, Information and Inference. 1970 ISBN 90-277-0155-5
29. K. Lambert, Philosophical Problems in Logic. Some Recent Developments. 1970
ISBN 90-277-0079-6
30. R. A. Eberle, Nominalistic Systems. 1970 ISBN 9O-277-0161-X
31. P. Weingartner and G. Zecha (eds.),lnduction, Physics. and Ethics. 1970
ISBN 9O-277-0158-X
32. E. W. Beth, Aspects ofModern Logic. Translated from Dutch. 1970
ISBN 90-277-0173-3
33. R. Hilpinen (cd.), Deontic Logic. Introductory and Systematic Readings. 1971
See also No. 152. ISBN Ph (1981 rev.) 90-277-1302-2
34. J.-L. Krivine,lntroduction to Axiomatic Set Theory. Translated from French. 1971
ISBN 90-277-0169-5; Ph 90-277-0411-2
35. J. D. Sneed, The Logical Structure ofMathematical Physics. 2nd rev. ed., 1979
ISBN 90-277-1056-2; Pb 90-277-1059-7
36. C. R. Kordig, The Justification ofScientific Change. 1971
ISBN 90-277-0181-4; Pb 90-277-0475-9
37. M. Capek, Bergson and Modern Physics. A Reinterpretation and Re-evaluation.
[Boston Studies in the Philosophy of Science, Vol. VII] 1971 ISBN 90-277-0186-5
38. N. R. Hanson, What I Do Not Believe, and Other Essays. Ed. by S. Toulmin and H.
Woolf. 1971 ISBN 90-277-0191-1
39. R. C. Buck and R. S. Cohen (eds.), PSA 1970. Proceedings of the Second Biennial
Meeting of the Philosophy of Science Association, Boston, Fall 1970. In Memory of
Rudolf Carnap. [Boston Studies in the Philosophy of Science, Vol. VIII] 1971
ISBN 90-277-0187-3; Pb 90-277-0309-4
40. D. Davidson and G. Harman (eds.), Semantics ofNatural Language. 1972
ISBN 90-277-0304-3; Ph 90-277-0310-8
41. Y. Bar-Hillel (ed.), Pragmatics ofNatural Languages. 1971
ISBN 90-277-0194-6; Ph 90-277-0599-2
42. S. Stenlund, Combinators, A-Terms and ProofTheory. 1972 ISBN 90-277-0305-1
43. M. Strauss, Modern Physics and Its Philosophy. Selected Paper in the Logic,
History, and Philosophy of Science. 1972 ISBN 90-277-0230-6
44. M. Bunge, Method, Model and Matter. 1973 ISBN 90-277-0252-7
45. M. Bunge, Philosophy of Physics. 1973 ISBN 90-277-0253-5
SYNTHESE LmRARY
46. A. A. Zinov'ev, Foundations of the Logical Theory of Scientific Knowledge

(Complex Logic). Revised and enlarged English edition with an appendix by G. A.
Smirnov, E. A. Sidorenka, A. M. Fedina and L. A. Bobrova. [Boston Studies in the
Philosophy of Science, Vol. IX] 1973 ISBN 90-277-0193-8; Ph 90-277-0324-8
47. L. Tondl, Scientific Procedures. A Contribution concerning the Methodological
Problems of Scientific Concepts and Scientific Explanation. Translated from Czech
by D. Short. Edited by R.S. Cohen and M.W. Wartofsky. [Boston Studies in the
Philosophy of Science, Vol. X] 1973 ISBN 90-277-0147-4; Pb 90-277-0323-X
48. N. R. Hanson, Constellations and Conjectures. 1913 ISBN 90-277-0192-X
49. K. J. J. Hintikka, J. M. E. Moravcsik and P. Suppes (eds.), Approaches to Natural
Language. 1973 ISBN 90-277-0220-9; Pb 90-277-0233-0
50. M. Bunge (ed.), Exact Philosophy. Problems, Tools and Goals. 1973
ISBN 90-277-0251-9
51. R. J. Bogdan and I. Niiniluoto (eds.), Logic, Language and Probability. 1973
ISBN 90-277-0312-4
52. G. Pearce and P. Maynard (eds.), Conceptual Change. 1973
ISBN 90-277-0287-X; Ph 90-277-0339-6
53. I. Niiniluoto and R. Tuomela, Theoretical Concepts and Hypothetico-inductive
Inference. 1973 ISBN 90-277-0343-4
54. R. Fraisse, Course of Mathematical Logic - Volume 1: Relation and Logical
Formula. Translated from French. 1973 ISBN 90-277-0268-3; Ph 90-277-0403-1
(For Volume 2 see under No. 69).
55. A. Grilnbaum, Philosophical Problems of Space and Time. Edited by R.S. Cohen
and M.W. Wartofsky. 2nd enlarged ed. [Boston Studies in the Philosophy of
Science, Vol. XII] 1973 ISBN 90-277-0357-4; Ph 90-277-0358-2
56. P. Suppes (ed.), Space. Time and Geometry. 1973
ISBN 90-277-0386-8; Pb 90-277-0442-2
57. H. Kelsen, Essays in Legal and Moral Philosophy. Selected and introduced by O.
Weinberger. Translated from German by P. Heath. 1973 ISBN 90-277-0388-4
58. R. J. Seeger and R. S. Cohen (eds.), Philosophical Foundations of Science. [Boston
Studies in the Philosophy of Science, Vol. XI] 1974
ISBN 90-277-0390-6; Pb 90-277-0376-0
59. R. S. Cohen and M. W. Wartofsky (eds.), Logical and Epistemological Studies in
Contemporary Physics. [Boston Studies in the Philosophy of Science, Vol. XIII]
1973 ISBN 90-277-0391-4; Pb 90-277-0377-9
60. 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, /969-/972. [Boston Studies in the Philosophy of Science,
Vol. XIV] 1974 ISBN 90-277-0392-2; Ph 90-277-0378-7
61. R. S. Cohen, J. J. Stachel and M. W. Wartofsky (eds.), For Dirk Struik. Scientific.
Historical and Political Essays. [Boston Studies in the Philosophy of Science,
Vol. XV] 1974 ISBN 90-277-0393-0; Ph 90-277-0379-5
62. K. Ajdukiewicz, Pragmatic Logic. Translated from Polish by O. Wojtasiewicz. 1974
ISBN 90-277-0326-4
63. S. Stenlund (ed.), Logical Theory and Semantic Analysis. Essays dedicated to Stig
Kanger on His 50th Birthday. 1974 ISBN 90-277-0438-4
64. K. F. Schaffner and R. S. Cohen (eds.), PSA /972. Proceedings of the Third
Biennial Meeting of the Philosophy of Science Association. [Boston Studies in the
Philosophy of Science, Vol. XX] 1974 ISBN 90-277-0408-2; Pb 90-277-0409-0
SYNTHESE LmRARY
65. H. E. Kyburg, Jr., The Logical Foundations ofStatistical Inference. 1974

ISBN ~277-0330-2; Ph ~277-0430-9
66. M. Grene, The Understanding of Nature. Essays in the Philosophy of Biology.
[Boston Studies in the Philosophy of Science, Vol. XXIII] 1974
ISBN ~277-0462-7; Ph ~277-0463-5
67. 1. M. Broekman, Structuralism: Moscow, Prague, Paris. Translated from German.
1974 ISBN ~277-0478-3
68. N. Geschwind, Selected Papers on Longuage and the Brain. [Boston Studies in the
Philosophy of Science, Vol. XVI] 1974 ISBN 90-277-0262-4; Ph 90-277-0263-2
69. R. Fraisse, Course of Mathematical Logic - Volume 2: Model Theory. Translated
from French. 1974 ISBN 90-277-0269-1; Pb 90-277-0510-0
(For Volume 1 see under No. 54)
70. A. Grzegorczyk, An Outline of Mathematical Logic. Fundamental Results and
Notions explained with all Details. Translated from Polish. 1974
ISBN 90-277-0359-0; Ph 90-277-0447-3
71. F. von Kutschera, Philosophy ofLanguage. 1975 ISBN 90-277-0591-7
72. J. Manninen and R. Thomela (eds.), Essays on Explanation and Understanding.
Studies in the Foundations of Humanities and Social Sciences. 1976
ISBN 90-277-0592-5
73. J. Hintikka (ed.), Rudolf Carnap, Logical Empiricist. Materials and Perspectives.
1975 ISBN ~277-0583-6
74. M. Capek (ed.), The Concepts of Space and Time. Their Structure and Their
Development. [Boston Studies in the Philosophy of Science, Vol. XXII] 1976
ISBN 90-277-0355-8; Pb ~277-0375-2
75. J. Hintikka and U. Remes, The Method of Analysis. Its Geometrical Origin and Its
General Significance. [Boston Studies in the Philosophy of Science, Vol. XXV]
1974 ISBN 90-277-0532-1; Ph ~277-0543-7
76. J. E. Murdoch and E. D. Sy11a (eds.), The Cultural Context of Medieval Learning.
[Boston Studies in the Philosophy of Science, Vol. XXVI] 1975
ISBN 90-277-0560-7; Ph ~277-0587-9
77. S. Amsterdamski, Between Experience and Metaphysics. Philosophical Problems of
the Evolution of Science. [Boston Studies in the Philosophy of Science,
Vol. XXXV] 1975 ISBN 90-277-0568-2; Ph 90-277-0580-1
78. P. Suppes (00.), Logic and Probability in Quantum Mechanics. 1976
ISBN 90-277-0570-4; Pb 90-277-1200-X
79. H. von Helmholtz: Epistemological Writings. The Paul Hertz / Moritz Schlick
Centenary Edition of 1921 with Notes and Commentary by the Editors. Newly
translated from German by M. F. Lowe. Edited, with an Introduction and Bibliog-
raphy, by R. S. Cohen 8'ld Y. Elkana. [Boston Studies in the Philosophy of Science,
Vol. XXXVII] 1975 ISBN 9O-277-029O-X; Ph ~277-0582-8
80. 1. Agassi, Science in Flux. [Boston Studies in the Philosophy of Science,
Vol. XXVIII] 1975 ISBN 90-277-0584-4; Ph 90-277-0612-2
81. S. G. Harding (ed.), Can Theories Be Refuted? Essays on the Duhem-Quine Thesis.
1976 ISBN 90-277-0629-8; Pb 90-277-0630-1
82. S. Nowak, Methodology ofSociological Research. General Problems. 1977
ISBN ~277-0486-4
83. J. Piaget, J.-B. Grize, A. Szeminska and V. Bang, Epistemology and Psychology of
Functions. Translated from French. 1977 ISBN 90-277-0804-5
SYNTHESE LIBRARY
84. M. Grene and E. Mendelsohn (eds.), Topics in the Philosophy of Biology. [Boston
Studies in the Philosophy of Science, Vol. XXVll] 1976
ISBN 9O-277-0595-X; Pb 90-277-0596-8
85. E. Fischbein, The Intuitive Sources ofProbabilistic Thinking in Children. 1975
ISBN 90-277-0626-3; Pb 90-277-1190-9
86. E. W. Adams, The Logic of Conditionals. An Application of Probability to
Deductive Logic. 1975 ISBN 9O-277-0631-X
87. M. PrzeI~ki and R. W6jcicki (eds.), Twenty-Five Years of Logical Methodology in
Poland. Translated from Polish. 1976 ISBN 90-277-0601-8
88. J. Topolski, The Methodology of History. Translated from Polish by O. Woj-
tasiewicz.1976 ISBN 9O-277-0550-X
89. A. Kasher (ed.), Language in Focus: Foundations, Methods and Systems. Essays
dedicated to Yehoshua Bar-Hillel. [Boston Studies in the Philosophy of Science,
Vol. XLIII] 1976 ISBN 90-277-0644-1; Pb 90-277-0645-X
90. J. Hintikka, The Intentions of Intentionality and Other New Models for Modalities.
1975 ISBN 90-277-0633-6; Pb 90-277-0634-4
91. W. StegmUller, Collected Papers on Epistemology, Philosophy of Science and
History ofPhilosophy. 2 Volumes. 1977 Set ISBN 90-277-0767-7
92. D. M. Gabbay, Investigations in Modal and Tense Logics with Applications to
Problems in Philosophy and Linguistics. 1976 ISBN 90-277-0656-5
93. R. J. Bogdan, Local Induction. 1976 ISBN 90-277-0649-2
94. S. Nowak, Understanding and Prediction. Essays in the Methodology of Social and
Behavioral Theories. 1976 ISBN 90-277-0558-5; Ph 90-277-1199-2
95. P. Mittelstaedt, Philosophical Problems of Modern Physics. [Boston Studies in the
Philosophy of Science, Vol. XVIll] 1976 ISBN 90-277-0285-3; Ph 90-277-0506-2
96. G. Holton and W. A. Blanpied (eds.), Science and Its Public: The Changing
Relationship. [Boston Studies in the Philosophy of Science, Vol. XXXIII] 1976
ISBN 90-277-0657-3; Ph 90-277-0658-1
97. M. Brand and D. Walton (eds.), Action Theory. 1976 ISBN 90-277-0671-9
98. P. Goehel. Outline of a Nominalist Theory of Propositions. An Essay in the Theory
of Meaning and in the Philosophy of Logic. 1980 ISBN 90-277-1031-7
99. R. S. Cohen, P. K. Feyerabencl, and M. W. Wartofsky (eds.), Essays in Memory of
Imre Lakatos. [Boston Studies in the Philosophy of Science, Vol. XXXIX] 1976
ISBN 90-277-0654-9; Ph 90-277-0655-7
100. R. S. Cohen and J. J. Stachel (eds.), Selected Papers of Uon Rosenfield. [Boston
Studies in the Philosophy of Science, Vol. XXI] 1979
ISBN 90-277-0651-4; Ph 90-277-0652-2
101. R. S. Cohen, C. A. Hooker, A. C. Michalos and J. W. van Evra (eds.), PSA 1974.
Proceedings of the 1974 Biennial Meeting of the Philosophy ofScience Association.
[Boston Studies in the Philosophy of Science, Vol. XXXII] 1976
ISBN 90-277-0647-6; Pb 90-277-0648-4
102. Y. Fried and J. Agassi, Paranoia. A Study in Diagnosis. [Boston Studies in the
Philosophy of Science, Vol. L] 1976 ISBN 90-277-0704-9; Ph 90-277-0705-7
103. M. PrzeI~ki, K. Szaniawski and R. W6jcicki (eds.), Formal Methods in the
Methodology ofEmpirical Sciences. 1976 ISBN 90-277-0698-0
104. J. M. Vickers, Beliefand Probability. 1976 ISBN 90-277-0744-8
105. K. H. Wolff, Surrender and Catch. Experience and Inquiry Today. [Boston Studies
in the Philosophy of Science, Vol. LI] 1976
ISBN 90-277-0758-8; Pb 90-277-0765-0
SYNTHESE LIBRARY
106. K. Kosik, Dialectics of the Concrete. A Study on Problems of Man and World.
[Boston Studies in the Philosophy of Science, Vol. LII] 1976
ISBN 90-277-0761-8; Ph 90-277-0764-2
107. N. Goodman, The Structure of Appearance. 3rd ed. with an Introduction by G.
Hellman. [Boston Studies in the Philosophy of Science, Vol. LIII] 1977
ISBN 90-277-0773-1; Pb 9O-277-0774-X
108. K. Ajdukiewicz, The Scientific World-Perspective and Other Essays, 1931-1963.
Translated from Polish. Edited and with an Introduction by 1. Giedymin. 1978
ISBN 90-277-0527-5
109. R. L. Causey, Unity ofScience. 1977 ISBN 90-277-0779-0
110. R. E. Grandy, Advanced Logic for Applications. 1977 ISBN 90-277-0781-2
111. R. P. McArthur, Tense Logic. 1976 ISBN 90-277-0697-2
112. L. Lindahl, Position and Change. A Study in Law and Logic. Translated from
Swedish by P. Needham. 1977 ISBN 90-277-0787-1
113. R. Tuomela, Dispositions. 1978 ISBN 9O-277-0810-X
114. H. A. Simon, Models of Discovery and Other Topics in the Methods of Science.
[Boston Studies in the Philosophy of Science, Vol. LIV] 1977
ISBN 90-277-0812-6; Ph 90-277-0858-4
115. R. D. Rosenkrantz, Inference, Method and Decision. Towards a Bayesian
Philosophy of Science. 1977 ISBN 90-277-0817-7; Ph 90-277-0818-5
116. R. Tuomela, Human Action and Its Explanation. A Study on the Philosophical
Foundations of Psychology. 1977 ISBN 9O-277-0824-X
117. M. Lazerowitz, The Language of Philosophy. Freud and Wittgenstein. [Boston
Studies in the Philosophy of Science, Vol. LV] 1977
ISBN 90-277-0826-6; Pb 90-277-0862-2
118. Not published
119. J. Pelc (ed.), Semiotics in Poland, 1894-1969. Translated from Polish. 1979
ISBN 90-277-0811-8
120. I. Ptlrn, Action Theory and Social Science. Some Formal Models. 1977
fSBN 90-277-0846-0
121. J. Margolis, Persons and Mind. The Prospects of Nonreductive Materialism. [Boston
Studies in the Philosophy of Science, Vol. LVII] 1977
ISBN 90-277-0854-1; Ph 90-277-0863-0
122. 1. Hintikka, I. Niiniluoto, and E. Saarinen (eds.), Essays on Mathematical and
Philosophical Logic. 1979 ISBN 90-277-0879-7
123. T. A. F. Kuipers, Studies in Inductive Probability and Rational Expectation. 1978
ISBN 90-277-0882-7
124. E. Saarinen, R. Hilpinen, I. Niiniluoto and M. P. Hintikka (eds.), Essays in Honour
ofJaakko Hintikka on the Occasion ofHis 50th Birthday. 1979
ISBN 90-277-0916-5
125. G. Radnitzky and G. Andersson (eds.), Progress and Rationality in Science. [Boston
Studies in the Philosophy of Science, Vol. LVIII] 1978
ISBN 90-277-0921-1; Pb 9O·277-0922-X
126. P. Mittelstaedt, Quantum Logic. 1978 ISBN 90-277-0925-4
127. K. A. Bowen, Model Theory for Modal Logic. Kripke Models for Modal Predicate
Calculi. 1979 ISBN 90-277-0929-7
128. H. A. Bursen, Dismantling the Memorv Machine. A Philosophical Investigation of
Machine Theories of Memory. 1978 ISBN 90-277-0933-5
SYNTHESE LffiRARY
129. M. W. Wartofsky, Models. Representation and the Scientific Understanding.

[Boston Studies in the Philosophy of Science, Vol. XLVIm 1979
ISBN 9~277-0736-7; Pb 90-277-0947-5
130. D. lhde, Technics and Praxis. A Philosophy of Technology. [Boston Studies in the
Philosophy of Science, Vol. XXIV] 1979 ISBN 9~277-0953-X; Pb ~277-0954-8
131. 1.1. Wiatt (ed.), Polish Essays in the Methodology of the Social Sciences. [Boston
Studies in the Philosophy of Science, Vol. XXIX] 1979
ISBN 9~277-0723-5; Pb ~277-0956-4
132. W. C. Salmon (ed.), Hans Reichenbach: Logical Empiricist. 1979
ISBN ~277-0958-0
133. P. Bieri, R.-P. Horsunann and L. KrUger (cds.), Transcendental Arguments in
Science. Essays in Epistemology. 1979 ISBN 9~277-0963-7; Pb ~277-0964-5
134. M. Markovi~ and G. Petrovi~ (cds.), Praxis. Yugoslav Essays in the Philosophy and
Methodology of the Social Sciences. [Boston Studies in the Philosophy of Science,
Vol. XXXVI] 1979 ISBN 9~277-0727-8; Pb 9~277-0968-8
135. R. W6jcicki, Topics in the Formal Methodology of Empirical Sciences. Translated
from Polish. 1979 ISBN ~277-1004-X
136. G. Radnitzky and G. Andersson (cds.), The Structure and Development of Science.
[Boston Studies in the Philosophy of Science, Vol. LIX] 1979
ISBN ~277-0994-7; Pb 90-277-0995-5
137. 1. C. Webb, Mechanism, Mentalism and Metamathematics. An Essay on Finitism.
1980 ISBN 90-277-1046-5
138. D. F. Gustafson and B. L. Tapscott (cds.), Body, Mind and Method. Essays in Honor
of Virgil C. Aldrich. 1979 ISBN 90-277-1013-9
139. L. Nowak, The Structure ofIdealization. Towards a Systematic Interpretation of the
Marxian Idea of Science. 1980 ISBN 90-277-1014-7
140. C.Perelman, The New Rhetoric and the Humanities. Essays on Rhetoric and Its
Applications. Translated from French and German. With an Introduction by H.
Zysldnd.1979 ISBN 9~277-1018-X; Pb 90-277-1019-8
141. W. Rabinowicz, Universalizability. A Study in Morals and Metaphysics. 1979
ISBN ~277-102~2
142. C. Perelman, Justice, Law and Argument. Essays on Moral and Legal Reasoning.
Translated from French and German. With an Introduction by H.l. Berman. 1980
ISBN 9~277-1089-9; Pb ~277-109~2
143. S. Kanger and S. Ohman (cds.), Philosophy and Grammar. Papers on the Occasion
of the Quincentennial ofUppsala University. 1981 ISBN ~277-1091-0
144. T. Pawlowski, Concept Formation in the Humanities and the Social Sciences. 1980
ISBN ~277-1096-1
145. J. Hintikka, D. Gruender and E. Agazzi (cds.), Theory Change, Ancient Axiomatics
and Galilea's Methodology. Proceedings of the 1978 Pisa Conference on the History
and Philosophy of Science, Volume I. 1981 ISBN 90-277-1126-7
146. 1. Hintikka, D. Gruender and E. Agazzi (eds.), Probabilistic Thinking, Ther-
modynamics, and the Interaction of the History and Philosophy of Science.
Proceedings of the 1978 Pisa Conference on the History and Philosophy of Science,
Volume II. 1981 ISBN 90-277-1127-5
147. U. Monnich (cd.), Aspects ofPhilosophical Logic. Some Logical Forays into Central
Notions of Linguistics and Philosophy. 1981 ISBN 90-277-1201-8
148. D. M. Gabbay, Semantieal Investigations in Heyting's Intuitionistie Logic. 1981
ISBN 90-277-1202-6
SYNTHESE LIBRARY
149. E. Agazzi (ed.), Modern Logic -A Survey. Historical, Philosophical, and Mathemati-
cal Aspects of Modem Logic and Its Applications. 1981 ISBN 90-277-1137-2
150. A. F. Parker-Rhodes, The Theory of Indistinguishables. A Search for Explanatory
Principles below the Level of Physics. 1981 ISBN 9O-277-1214-X
151. J. C. Pitt, Pictures, Images, and Conceptual Change. An Analysis of Wilfrid Sellars'
Philosophy of Science. 1981 ISBN 90-277-1276-X; Pb 90-277-1277-8
152. R. Hilpinen (ed.), New Studies in Deontic Logic. Norms, Actions, and the Founda-
tions of Ethics. 1981 ISBN 90-277-1278-6; Pb 90-277-1346-4
153. C. Dilworth, Scientific Progress. A Study Concerning the Nature of the Relation
between Successive Scientific Theories. 2nd, rev. and augmented ed, 1986
ISBN 90-277-2215-3; Ph 90-277-2216-1
154. D. Woodruff Smith and R. Mcintyre, Husserl and Intentionality. A Study of Mind,
Meaning, and Language. 1982 ISBN 90-277-1392-8; Ph 90-277-1730-3
155. R.1. Nelson, The Logic ofMind. 2nd. ed., 1989
ISBN 90-277-2819-4; Ph 90-277-2822-4
156. J. F. A. K. van Benthem, The Logic of Time. A Model-Theoretic Investigation into
the Varieties of Temporal Ontology, and Temporal Discourse. 1983; 2nd ed., 1991
ISBN 0-7923-1081-0
157. R. Swinburne (ed.), Space, Time and Causality. 1983 ISBN 90-277-1437-1
158. E. T. Jaynes, Papers on Probability, Statistics and Statistical Physics. Ed. by R. D.
Rozenkrantz.1983 ISBN 90-277-1448-7; Ph (1989) 0-7923-0213-3
159. T. Chapman, Time: A Philosophical Analysis. 1982 ISBN 90-277-1465-7
160. E. N. Zalta, Abstract Objects. An Introduction to Axiomatic Metaphysics. 1983
ISBN 90-277-1474-6
161. S. Harding and M. B. Hintikka (eds.), Discovering Reality. Feminist Perspectives on
Epistemology, Metaphysics, Methodology, and Philosophy of Science. 1983
ISBN 90-277-1496-7; Ph 90-277-1538-6
162. M. A. Stewart (ed.), Law, Morality and Rights. 1983 ISBN 9O-277-1519-X
163. D. Mayr and G. SUssmann (eds.), Space, Time. and Mechanics. Basic Structures of a
Physical Theory. 1983 ISBN 90-277-1525-4
164. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. I:
Elements of Classical Logic. 1983 ISBN 90-277-1542-4
165. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. IT:
Extensions of Classical Logic. 1984 ISBN 90-277-1604-8
166. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. 1lI:
Alternative to Classical Logic. 1986 ISBN 90-277-1605-6
167. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. IV:
Topics in the Philosophy of Language. 1989 ISBN 90-277-1606-4
168. A. J. I. Jones, Communication and Meaning. An Essay in Applied Modal Logic.
1983 ISBN 90-277-1543-2
169. M. Fitting, Proof Methods for Modal and Intuitionistic Logics. 1983
ISBN 90-277-1573-4
170. J. Margolis, Culture and Cultural Entities. Toward a New Unity of Science. 1984
ISBN 90-277-1574-2
171. R. Tuomela, A Theory ofSocial Action. 1984 ISBN 90-277-1703-6
172. J. J. E. Gracia, E. Rabossi, E. Villanueva and M. Dascal (eds.), Philosophical
Analysis in LAtin America. 1984 ISBN 90-277-1749-4
173. P. Ziff, Epistemic Analysis. A Coherence Theory of Knowledge. 1984
ISBN 90-277-1751-7
SYNTHESE LmRARY
174. P. Ziff, Antiaesthetics. An Appreciation of the Cow with the Subtile Nose. 1984
. ISBN 90-277-1773-7
175. W. Balzer, D. A. Pearce, and H.-I. Schmidt (eds.), Red~tion in Science. Structure,
Examples, Philosophical Problems. 1984 ISBN 90-277-1811-3
176. A. Peczenik, L. Lindahl and B. van Roennund (eds.), Theory of Legal Science.
Proceedings of the Conference on Legal Theory and Philosophy of Science (Lund,
Sweden, December 1983). 1984 ISBN 90-277-1834-2
177. I. Niiniluoto, Is Science Progressive? 1984 ISBN 90-277-1835-0
178. B. K. Matilal and J. L. Shaw (eds.), Analytical Philosophy in Comparative
Perspective. Exploratory Essays in Current Theories and Classical Indian Theories
of Meaning and Reference. 1985 ISBN 90-277-1870-9
179. P. Kroes, Time: Its Structure and Role in Physical Theories. 1985
ISBN 90-277-1894-6
180. J. H. Fetzer, Sociobiology and Epistemology. 1985
ISBN 90-277-2005-3; Pb 90-277-2006-1
181. L. Haaparanta and J. Hintikka (eds.), Frege Synthesized. Essays on the Philosophical
and Foundational Work of GonIob Frege. 1986 ISBN 90-277-2126-2
182. M. Detlefsen, Hilbert's Program. An Essay on Mathematical Instrumentalism. 1986
ISBN 90-277-2151-3
183. J. L. Golden and J. J. Pilotta (eds.), Practical Reasoning in Human Affairs. Studies
in HonorofChaim Perelman. 1986 ISBN 90-277-2255-2
184. H. Zandvoort, Models of Scientific Development and the Case of N~lear Magnetic
Resonance. 1986 ISBN 90-277-2351-6
185. I. Niiniluoto, Truthlikeness. 1987 ISBN 90-277-2354-0
186. W. Balzer, C. U. Mouiines and J. D. Sneed, An Architectonic for Science. The
Structuralist Program. 1987 ISBN 90-277-2403-2
187. D. Pearce, Roads to Commensurability. 1987 ISBN 90-277-2414-8
188. L. M. Vaina (ed.), Matters ofIntelligence. Conceptual Structures in Cognitive Neuro-
science. 1987 ISBN 90-277-2460-1
189. H. Siegel, Relativism Refuted. A Critique of Contemporary Epistemological
Relativism. 1987 ISBN 90-277-2469-5
190. W. Callebaut and R. Pinxten, Evolutionary Epistemology. A Muitiparadigm
Program, with a Complete Evolutionary Epistemology Bibliograph. 1987
ISBN 90-277-2582-9
191. J. Kmita, Problems in Historical Epistemology. 1988 ISBN 90-277-2199-8
192. J. H. Fetzer (ed.), Probability and Causality. Essays in Honor of Wesley C. Salmon,
with an Annotated Bibliography. 1988 ISBN 90-277-2607-8; Pb 1-5560-8052-2
193. A. Donovan, L. Laudan and R. Laudan (eds.), Scrutinizing Science. Empirical
Studies of Scientific Change. 1988 ISBN 90-277-2608-6
194. H.R. Otto and J.A. Tuedio (eds.), Perspectives on Mind. 1988
ISBN 90-277-2640-X
195. D. Batens and J.P. van Bendegem (eds.), Theory and Experiment. Recent Insights
and New Perspectives on Their Relation. 1988 ISBN 90-277-2645-0
196. J. Osterberg, Self and Others. A Study of Ethical Egoism. 1988
ISBN 90-277-2648-5
197. D.H. Helman (ed.), Analogical Reasoning. Perspectives of Artificial Intelligence,
Cognitive Science, and Philosophy. 19&8 ISBN 90-277-2711-2
198. J. Wolenski, Logic and Philosophy in the Lvov-Warsaw School. 1989
ISBN 90-277-2749-X
SYNTHESE LIBRARY
199. R. W6jcicki, Theory of Logical Calculi. Basic Theory of Consequence Operations.

1988 ISBN 90-277-2785-6
200. J. Hintikka and M.B. Hintikka, The Logic of Epistemology and the Epistemology of
Logic. Selected Essays. 1989 ISBN 0-7923-0040-8; Pb 0-7923-0041-6
201. E. Agazzi (ed.), Probability in the Sciences. 1988 ISBN 90-277-2808-9
202. M. Meyer (ed.), From Metaphysics to Rhetoric. 1989 ISBN 90-277-2814-3
203. RL. Tieszen, Mathematical Intuition. Phenomenology and Mathematical
Knowledge. 1989 ISBN 0-7923-0131-5
204. A. Melnick, Space, Time, and Thought in Kant. 1989 ISBN 0-7923-0135-8
205. D.W. Smith, The Circle ofAcquaintance. Perception, Consciousness, and Empathy.
1989 ISBN 0-7923-0252-4
206. M.H. Salmon (ed.), The Philosophy of Logical Mechanism. Essays in Honor of
Arthur W. Burks. With his Responses, and with a Bibliography of Burk's Work.
1990 ISBN 0-7923-0325-3
207. M. Kusch, Language as Calculus vs. Language as Universal Medium. A Study in
Husserl, Heidegger, and Gadamer. 1989 ISBN 0-7923-0333-4
208. T.C. Meyering, Historical Roots of Cognitive Science. The Rise of a Cognitive
Theory of Perception from AntiqUity to the Nineteenth Century. 1989
ISBN 0-7923-0349-0
209. P. Kosso, Observability and Observation in Physical Science. 1989
ISBN 0-7923-0389-X
210. J. Kmita, Essays on the Theory ofScientific Cognition. 1990 ISBN 0-7923-0441-1
211. W. Sieg (00.), Acting and Reflecting. The Interdisciplinary Tum in Philosophy. 1990
ISBN 0-7923-0512-4
212. J. Karpinski, Causality in Sociological Research. 1990 ISBN 0-7923-0546-9
213. H.A. Lewis (ed.), Peter Geach: Philosophical Encounters. 1991
ISBN 0-7923-0823-9
214. M. Ter Hark, Beyond the Inner and the Outer. Wittgenstein's Philosophy of
Psychology. 1990 ISBN 0-7923-0850-6
215. M. Gosselin, Nominalism and Contemporary Nominalism. Ontological and
Epistemological Implications of the Work of W.V.O. Quine and of N. Goodman.
1990 ISBN 0-7923-0904-9
216. J.H. Fetzer, D. Shatz and G. Schlesinger (eds.), Definitions and Definability.
Philosophical Perspectives. 1991 ISBN 0-7923-1046-2
217. E. Agazzi and A. Cordero (oos.), Philosophy and the Origin and Evolution of the
Universe. 1991 ISBN 0-7923-1322-4
218. M. Kusch, Foueaul!'s Strata and Fields. An Investigation into Archaeological and
Genealogical Science Studies. 1991 ISBN 0-7923-1462-X
Previous volumes are still available.
KLUWER ACADEMIC PUBLISHERS - DORDRECHT / BOSTON / LONDON

Kusch, Martin - Foucault's Strata and Fields

Uploaded by

Copyright:

Available Formats

Kusch, Martin - Foucault's Strata and Fields

Uploaded by

Document Information

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

Kusch, Martin - Foucault's Strata and Fields

Uploaded by

Copyright:

Available Formats

FOUCAULT'S STRATA AND FIELDS

LOGIC, METHODOLOGY, AND PHILOSOPHY OF SCIENCE

JAAKKO HINTIKKA, Boston University

DONALD DAVIDSON, University ofCalifornia, Berkeley

Printed оп acid-free paper

PART I: FOUCAULDIAN ARCHAEOWGY . . . . . . . . . . . 1

2. ON THE VERY NOTION OF 'ARCHAEOLOGY' . . . . . 5

3. THE NEW HISTORIES IN FRANCE 12

4. ARCHAEOLOGY, THE NEW HISTORIES, AND THE

5. THE ARCHAEOLOGICAL MODEL I: IDENTIFYING

6. THE ARCHAEOLOGICAL MODEL II: BEYOND CON-

7. ARCHAEOLOGY OF KNOWLEDGE AND OTHER HIS-

NOTES TO PART I 112

PART ll: FOUCAULDIAN GENEAWGY . . . . . . . . . . . .. 115

9. THE CONCEPT OF POWER 117

10. THE GENEALOGICAL CONCEPTION OF POWER I:

11. THE GENEALOGICAL CONCEPTION OF POWER II:

12. GENEALOGICAL RESEARCH STRATEGIES 165

DISTRUST AND IRONY . . . . . . . . . . . . . . . . . . . . . .. 186

13. GENEALOGICAL PERSPECTIVISM 193

14. GENEALOGICAL CRITICISM OF POWER AND RA-

NOTES TO PART II . . . . . . . . . . . . . . . . . . . . . . . . . . .. 227

INDEX OF NAMES . . . . . . . . . . . . . . . . . . . . . . . . . . .. 250

INDEX OF SUBJECTS 257

I am indebted, most of all, to my colleagues and students in the Depart-

have spent many hours on the phone and in Chinese restaurants in

Qulu, June 1991

In recent years, a large number of books and articles on Foucault has

niques within a scientific culture.

archaeological and the genealogical projects are without links and

Foucault's archaeological discours de la methode, The Archaeology of

Unfortunately, existing book-size presentations of Foucault reflect

labors of the most important school in historical scholarship of this

historical epistemology, focussing especially on notions like mentality,

2. ON THE VERY NOTION OF 'ARCHAEOLOGY'

In choosing the term 'archaeology' as a label for his approach to the

Psychoanalysis was also regarded as an essential archaeological

In Folie et deraison: Histoire de la folie a l'age classique (1954)

It is also characteristic of the archaeologist's work, that she must

change, and different relative ages of contemporaneous sciences. In

Knowledge demands that the historian attend to discontinuities in the

3. THE NEW HISTORIES IN FRANCE

THE ANNALES EXEMPLARS

In order to see the importance of the Annales School for Foucauldian

Against traditional history

Marc Bloch, in his Apologie pour I'histoire, identifies another theme

If the best-known theorists of our methods had not shown such an

Lucien Febvre criticizes not only traditional history in general, but

The history of mentalities

However, Febvre distances himself from the Burckhardtian way of

prevailing attitudes and beliefs of the sixteenth century in terms of these

era has no assurance that it will be able to transmit these mental

psychology ("archaeopsychology") and ethnology as the proper ap-

The history of mentalities must also be distinguished from the

Finally, in a book that otherwise closely follows the outline of

Serial history and the construction of the historical fact

Mandrou's distinction between different time scales within a mentality is

Ladurie no longer underwrites Marc Bloch's famous claim, according to

... today's historian must give up methodological naivety and think