Hacking - Scientific Reason

SCIENTIFIC REASON
Ian Hacking
Scientific Reason
Copyright © 2009 by the Institute for Advanced Studies in Humanities and
Social Sciences, National Taiwan University, Taipei, Taiwan (R.O.C.)
All rights reserved.
Ian HACKING
Edited by Der-Lan YEH and Jeu-Jenq YUANN
Published by Si-Chen LEE

Institute fot Advanced Studies in Humanities and Social Sciences,
National Taiwan University
www.ihs.ntu.edu.tw
No.1, Sec. 4, Roosevelt Rd., Taipei, 10617 Taiwan (R.O.C.)
886-2-3366-3750 FAX 886-2-3366-3751
ntuihs@ntu.edu.tw (General Office)
newsletter@mail.ihs.ntu.edu.tw (Editorial Office)
Distributed by National Taiwan University Press

www.press.ntu.edu.tw
886-2- 3366- 3993 FAX 886-2-2363-6905
ntuprs@ntu.edu.tw
Copy editing by Bookman Books Ltd.

www.bookman.com.tw
Copy editor Lynn SAUVÉ
Editoral liaison Emily CHANG
Interior and cover design Iris HSU
GPN 1009703212
ISBN 978-986-01-6115-1 (Hardcover)
Retail Price NT$ 350

Sales restricted to Taiwan (R.O.C.) only.
Printed in Taiwan (R.O.C.)
This book may not be reproduced, in whole or in part, or transmitted in any form or by any means,
electronic or mechanical, including photocopying and recording, or in any information storage or
retrieval system without the prior written permission of the Institute for Advanced Studies in
Humanities and Social Sciences, National Taiwan University.
Contents
Editors' Foreword IX
Preface XV
Lecture 1 ON THE HISTORICAL ROOTS OF

SCIENTIFIC REASON 1
1. Learning how to learn 3
2. Styles of scientific thinking 6
3. "The European Tradition" 8
4. Styles as constituted by methods and objects 9
5. A crude template for organizing the past 10
6. Crystallization 13
7. Regrets about the word "style" 17
8. Self-authenticating styles 21
9. Objects 22
10. Leibniz and Bourdieu 24
11. Society 26
12. Bernard Williams' "Truthfulness" 29
13. Silence about truth 30
14. Truthfulness about the past 33
15. Mathematics 36
16. Caution about "X " 40
17. The Galilean style 41
vi / SCIENTIFIC REASON
18. The laboratory style 42

19. Logic 45
20. Three radical propositions 46
I. Styles are self-authenticating 46
II. Ontological debates 47
III. Cognitive foundations and cultural
history 48
Lecture 2 WHERE DO MATHEMATICAL OBJECTS

COME FROM? 49
1. Philosophical anthropology 49
2. Mathematics as a motley 51
3. Nature 53
4. Folkloric history of a discovery of a human
potential 56
5. Truthfulness about geometrical objects 58
6. Argumentative Greeks 59
7. Mathematics has mattered to some and only
some philosophers. 61
8. The obsession with mathematics 62
9. Two visions of a remarkable book on Greek
mathematics 66
10. An aside on Science Studies 67
11. Return to the two-faced paradox 68
12. Plato for better or worse 69
13. The shift in conceptions about telling the
truth 71
14. Proof, not axioms, and not calculation 72
15. Two conceptions of proof 74
16. A second crystallization within the
mathematical style? 75
Contents / vii
17. Mathematical objects 78

18. Broaden your horizons! 80
19. Comparative anthropology of reason 81
20. Ancient Chinese mathematics 84
21. The origin of the objects 86
Lecture 3 THE LABORATORY STYLE OF THINKING

AND DOING 89
1. Recapitulation 89
2. Experimental exploration and hypothetical
modelling 92
3. Recapitulation of 3 cautions 93
4. Exploration and analogy: styles (2) and (3) 94
5. Measurement 95
6. The hypothetico-deductive method 97
7. Architectonic reasoning 99
8. The Galilean style 100
9. A new "Form of Life" 104
10. A new actor: not a person but a piece of
apparatus 107
11. The new place: The laboratory 109
12. The creation of phenomena 111
13. What Thomas Hobbes saw dearly 112
14. The decisive conceptual shift 116
15. Ontology: theoretical entities 118
16. The vacuum 119
17. The biotechnicallaboratory 121
18. Two kinds of laboratory? 127
viii / SCIENTIFIC REASON
Lecture 4 REALISMS AND ANTIREALISMS 129

1. Plurals 129
2. Not realism versus nominalism 130
3. Universals 132
4. "Ontology recapitulates philology." 134
5. Against too much linguistic relativity:
Nietzsche and Chuang Tzu 136
6. General ontology and the special sciences 140
7. Dummett's type of antirealism 143
8. An allusion to Richard Rorty 143
9. Arthur Fine's Natural Ontological Attitude 145
10. So what about my experimental argument for
entity realism? 146
11. What was the experimental argument? 147
(a) If, (NOT "only if") 147
12. What was the experimental argument? 148
(b) The strongest argument (NOT a
conclusive one) 148
13. A strong argument can always have a false
conclusion 149
14. Best but not conclusive 152
15. When we can not interfere 154
16. The self-vindication of the laboratory sciences 155
17. The stability of the sciences 156
Bibliography 158
Index of Subjects 168
Index of Names 173
Editors' Foreword
In November 2007, the philosophical community of Taiwan

was privileged by a series of lectures delivered by Professor Ian
Hacking. This book, which compiles the lecture content,
manifests clearly Professor Hacking's brilliance and erudition in
philosophy, but also transmits covertly an inspiring message to
the people of this part of the world. In order to fully appreciate
the significance of Professor Hacking's visit to Taiwan, we will
attempt to outline his message in what follows.
The focus of Hacking's thought is revealed in the title,

Scientific Reason, which traditionally relates to the concept of
truth. Rather than dealing with scientific truth from a static and
ahistorical point of view, Professor Hacking, with the notion of
truthfulness, conceives of it as coming into being dynamically
and historically. By connecting B. William's genealogical inquiry
of truthfulness with A. C. Crombie's styles of scientific thinking,
he removes science from the broadly positivistic European
tradition and reinstates it on the world stage by emphasizing two
ideas: scientific objectivity and the cognitive capacities of human
beings. This represents a challenge to conventional thinking
about the nature of science and is greatly inspiring to us in
Taiwan.
x / SCIENTIFIC REASON
Far from conceiving scientific reason in accordance with its

positivistic sense of being logical and prescriptive, it is essentially
characterized in these lectures as something historical and
descriptive. According to Hacking, this characterization helps to
better capture the development of science and consequently the
embedded nature of reason. Although positivistic scientific
thinking has recently dwindled in popularity in Taiwan, it
nonetheless still finds adherents. Hacking illuminates an old
topic by stressing its historical and descriptive roots.
Describing science from a historical dimension is by no

means an easy task unless scientific reason itself is accounted for.
Professor Hacking explicitly claims that his approach is at once
"radical and simple". It is simple because it relies on truisms and
it is radical for its conclusions are unarguably persuasive.
Hacking begins by drawing attention to an apparent
contradiction manifest in the distinction between objects made
by nature, and objects nurtured by humans. They are different
indeed, but according to Hacking, not incompatible. In fact, their
difference dissolves as soon as we ask the question: "What
cultural elements are needed to sustain a discovery about our
cognitive capacities?" Take the emergence of mathematics as an
example. Mathematics emerged and developed in various places
around the ancient world (Greece, Persia, China) and at different
times. This fact has to be explained culturally. Without this
essential component, no understanding of the intellectual
evolution of mathematics would be intelligible. Obviously, this
opens the gate for "philosophical anthropology" to which these
lectures contribute.
The two aspects of innate human ability and the development

of social institutions together constitute a desirable intellectual
Editors' Foreword / xi
orientation through which scientific reason can be understood.

This understanding is achieved by demonstrating philosophically
how "objects" are discovered in history, a process Hacking calls
"styles". Professor Hacking admits right in the beginning that the
notion of style is a "crude template", even a regrettable one. The
notion is taken mainly from Crombie's magnum opus on the
global history of science from which six styles of scientific
thinking are distinguished. Professor Hacking also discloses that
the notion of style is unsatisfactory for being too connotative and
volatile to offer a clear sense. Besides, the notion's contextual
implications might also prevent us from seeing its usefulness in
characterizing scientific reason across great stretches of time.
Nonetheless, the notion is adopted by Hacking because of its
presence in all of the scientific disciplines, and also because of its
connection to what Hacking calls 'continual evolution' and
'discontinuous crystallization'. Let us briefly describe this central
idea.
While taking Crombie's historical survey of science as a

useful road map for the long-standing development of European
science, Professor Hacking contributes his idea of discontinuity,
i.e., crystallization. The conception, originally inspired by M.
Foucault, is further developed by Professor Hacking to
demonstrate how objects become perceptible and thus at some
point, available to perceivers through a style. According to
Hacking, this undermines traditional debates of realism versus
anti-realism lingering in the philosophy of science, treating them
as mere by-products of particular styles of (scientific) reasoning.
However, having said this, concerns of relativism automatically
arise. Would styles turn out to be relativistic? Would they be
ontological horizons within which objects and facts are
xii / SCIENTIFIC REASON
discovered, and even phenomena created? Would what is within

a style become "non-existent" as soon as it is taken out of the
style? Would Professor Hacking fall prey to relativism by his use
of "styles"? To these questions, his answer is a resounding "no".
The act of identifying philosophical objects within styles of

scientific reasoning reveals simultaneously our cognitive abilities
as well as human indebtedness to culture. People use their
cognitive abilities to cope with life's challenges. This is a truism.
Our intellectual achievements are impressive but diverge along
cultural lines. However, the multiplicity of various achievements
in relation to cultural relativity does not undermine objectivity in
science. For, scientific facts are independent of perceiving
subjects, even when we grant culturally dependent origin.
Scientific achievements are justified on the basis of objectivity.
Professor Hacking labels the acquirement of objectivity through
specific historical conditions "Lebnizian", which is characterized
as being both historicist and rationalistic.
Styles of scientific reasoning exist naturally and culturally

and they are, according to Professor Hacking, self-authenticating.
However, they alone do not warrant access to scientific
objectivity. Styles evolve dynamically within human social
interactions. They thrive and are popular for a time, and
eventually are abandoned, become moribund and even extinct.
Styles evolve naturally and culturally in many parts of the world,
and are unpredictable or invisible until they crystallize and
become available to cognitive agents. But, each style is not
explicit unless a genealogical depiction of "what happened?",
"when was that?" and "who did it?" is reconstructed
retrospectively. The reconstruction itself would never be able to
ensure its truth, yet this does no harm since it is definitely a
Editors' Foreword / xiii
convincing way of conveying the "truthfulness" of the past, the

time and the person. As there can certainly be more than one way
of telling the truth about the past, the time and the person, there
is no reason to limit the reinstatement of truthfulness within the
realm of the Western tradition. Professor Hacking's message is
further tailored for us by referring to the Chinese tradition of
historiography started by Sima Qian in China. There is
absolutely no doubt about the fact that more forms of
truthfulness can be told, given one knows well how to do it.
What is even more important from this message is the reasoning:
giving truthfull descriptions of the past applies not only to
history, but also to science in general and reason in particular.
Professor Hacking's message gives us the opportunity to

undertake an academic adventure within one of the most crucial
topics in human history. Far from being a monolithic mode of
thinking, reason itself essentially characterizes science, i.e., as the
objective truth. Though the message is derived from a historically
retrospective point of view, it applies equally to the future. As a
search from the known into the unknown, science will definitely
fuel its global expansion by incorporating all available elements.
Insofar as the process of incorporation prevails globally, all
peoples stand cognitively and culturally as unified parts of the
scientific undertaking so conceived. All the peoples of the world
therefore have a share in the bounty of scientific achievement as
well as responsibility to respect it.
Finally, at the end of this long foreword, we wish to express
our sincere gratitude to Professor Hacking for his insightful
lectures and his willingness to let us publish them in book form.
Professor Hacking's visit was organized jointly by the Institute
for Advanced Studies in Humanities and Social Sciences,
xiv / SCIENTIFIC REASON
National Taiwan University and the Department of Philosophy,

National Taiwan University with partial sponsorship by Taiwan's
National Science Council. On behalf of the organizers, we would
like to take this opportunity to thank all those who helped to
make this visit a success and those who attended the lectures.
Last, but not least, we are very appreciative to Simon van
Rysewyk for proofreading this foreword and Lynn Sauvé for
proofreading everything.
Jeu-Jenq Yuann,
Associate Professor, Department of Philosophy,
National Taiwan University and
Der-Lan Yeh,
Associate Professor, Department of Foreign Languages and Literatures,
National Taiwan University
Taipei, Taiwan,
July 2008
Preface
These chapters are rewritten versions of a series of four

lectures given in Taiwan during November 2007. They are an
exposition of a single large theme, but as the talks were given in
different places, I spoke to different audiences that overlapped
but were not identical. Hence each lecture had to be fairly
self-contained. I have eliminated most repetition, but not all, for I
believe that some recapitulation in the course of four chapters is
no bad thing. The venues were as follows:
Lectures I and II were given at the Humanities Forum, Institute
for Advanced Studies in Humanities and Social Sciences,
National Taiwan University, 9th and 10th of November 2007.
Lecture III was given at a Science, Technology and Society
Workshop, held at National Tsing Hua University, 12th
November 2007.
Lecture IV was given to the Philosophy Department, Soochow
University, 14th November 2007.
The audiences were wonderful, and significant changes have

been made for this version, thanks to probing questions that
were asked. It was a memorable visit to Taiwan, of little more
than two weeks. I learned much; on the academic and scholarly
side, in addition to meeting with people interested in philosophy,
I had the opportunity to meet with scientists in cold atom
laboratories at National Tsing Hua University and National
xvi / SCIENTIFIC REASON
Chung Cheng University. I also was able to meet clinicians and

researchers into autism in the National Taiwan University
Hospital in Taipei, and teachers, clinicians, and parents of autistic
children in Taichung. I well remember a long afternoon meeting
with philosophy students and teachers at National Chung Cheng
University. All my hosts were incredibly generous. I will mention
only special thanks to Chin-Mu Yang and Ruey-Lin Chen, and
above all to the mastermind and originator of my visit, Jeu-Jenq
Yuann.
December 2007
Lecture 1
ON THE HISTORICAL ROOTS

OF SCIENTIFIC REASON
These lectures address very traditional subjects of philosophical

discussion: truth, reason, knowledge and the sciences. These are
ancient topics. My spiritual ancestor is more Leibniz, and hence
by indirection Plato, than it is Aristotle. What we add, in our
historicist times, is that we are aware that nothing is eternally
fixed, unless it is purely formal. Everything evolves; most things
decay. Even the constants of nature, such as Planck's constant and
the speed of light, may not be quite as constant as they are taken
to be, in simple down-home physics.1
Many thinkers who are not analytic philosophers take it to be
obvious, that even truth has a history. I shall say the opposite (for
I take truth to be a formal concept), yet my colleagues, the
analytic philosophers, may fear that I set my sails too much in
that direction. Caution is called for. In order to connect truth with
history, I shall invoke Bernard Williams' recent idea about what
1 A standard recent survey paper is, Jean-Philippe Uzan, "The Fundamental

Constants and Their Variation: Observational and Theoretical Status," Reviews of
Modern Physics, 75 (2003): 403-455.
2 I SCIENTIFIC REASON
he calls truthfulness. Ways to tell the truth have a history, or what

Williams called a genealogy.2 In order to connect reason and
practice I shall speak about reasoning, rather than some abstract
ahistorical presence, the noble reason of more traditional
philosophizing. These are not mere cosmetic changes: they
represent new ways to pursue the Leibnizian project of
understanding truth and reason.
These topics are entirely impersonal, but I shall take the
liberty of making the occasional personal remark. People keep on
asking me how I can identify as an analytic philosopher and yet
be so comfortable in making use of the past in order to
understand the present. An occasional aside will suggest answers
to that question.
My approach is at once radical and simple. I shall keep on
coming back to that. On the one hand I rely on truisms. On the
other hand I make them yield conclusions about reasoning that
are rather radical, so radical that sometimes I shall have to say,
yes, I really do mean what I am saying.
Among the many oddities of my approach is that I shall often
cite, out of context, a well-known saying by a well-known dead
philosopher—Schlick or Husserl, for example, as well as the
canonical great ones—and say, that is very close to what I am
saying here. I am not arguing from authority when I do this, for I
do not go on to argue that the philosopher meant his words in
the way in which I take them. I want rather to say, it looks as if
these seemingly unfamiliar ideas have been around for a long
time, in slightly different guises. I shall also use folklore about
the history of the sciences more than a respectable person should.
That is because I think there is a lot of wisdom in some folklore
2 Bernard Williams, Truth and Truthfulness: An Essay in Genealogy, Princeton:

Princeton University Press, 2002. See my review in Canadian Journal of Philosophy,
34 (2004): 137-148.
ON THE HISTORICAL ROOTS OF SCIENTIFIC REASON / 3
that we are ill-advised to forget, even when it violates the rules of

rigour.
1. Learning how to learn

My fundamental observation is that reasoning, finding out,
and techniques of discovery have a history. It is not just the
history of facts discovered, theories proposed, and technologies
invented. We have not only learned an amazing amount about
the world and how to change it: we have also had to learn how to
find out.
There are two sides to that. We had to bring to the surface
various kinds of innate ability that human beings may have had
forever, but whose exercise does not come naturally. And we had
to evolve social organizations within which those abilities could
be fostered.
On the one side, then, I wish to point to cognitive science, the
study of mental capacities. On the other side, I refer to the history
of civilizations and of their institutions. I deliberately say "point"
and "refer." Cognitive science is thriving but still in its infancy.
There are many confident assertions about the brain and its
abilities, but in all honesty our knowledge is sketchy and
conjectural. Hence we can only point to current guesses about
cognition.
In contrast we know an enormous amount about the history
of civilizations, but only recently have historians begun to think
seriously about the role of the sciences within them. We are
beginning to know a great deal of the microhistory and
microsociology of this or that scientific incident, but I am
interested in a larger view, philosophical and anthropological, of
scientific reason in the life of our species. Thus I shall refer to
stories about the past, but I shall not engage in historical writing.
4 / SCIENTIFIC REASON
I should emphasize at the start that I am a philosopher who

exploits facts about the past, but I am in no way a historian of the
sciences. Moreover, I use the past in a retrograde way, to try to
understand the present. This can be dressed up as brilliantly
up-to-date, using a phrase taken from Michel Foucault, "history
of the present," or it can be dismissed with Herbert Butterfield's
phrase, "whig history."3 But in truth I am not doing history at all;
I shall only be using it.
Cognition and culture are, then, two dimensions that provide
the space in which to understand scientific reason. We have
many different cognitive abilities, and human history runs on
many paths. Not surprisingly, there are many ways to conduct
scientific research. For example,
• Mathematicians construct deductive proofs (among other
things).
• We make theoretical models of aspects of nature in order to
understand them or to alter them.
• The laboratory sciences demand not just "experiment," but
also the building of apparatus to elicit, and often to create,
phenomena.
• Taxonomists classify living things according to principles of
hierarchic structure, although what those principles are,
continue to be matters of intense debate.
• Decision under uncertainty, thinking in probabilities, is yet
another distinct style of scientific thinking.
3 Sir Herbert Butterfield wrote one of the first textbooks of history of early modem
European science, based on lectures he got up for science undergraduates in
Cambridge University in 1948. The Origins of Modern Science, 1300-1800, London
and Cambridge: Deighton Bell, 1949. His chief authority was Alexandre Koyré. But
his fame is for political and social history of England; he despised "Whigs"—of
whom the philosopher cum historian David Hume is the best example —who wrote
about the past as a series of events that made sense chiefly in that they led on to and
validated the glorious present.
• There may also be a genetic way of understanding, most

successful in such evolutionary theories as Darwin's theory
of natural selection, but also tried out in enterprises as
diverse as Freudian analysis and Marxist historiography.
These are distinct ways to find things out, practiced in what we
call the sciences. They have histories that are to some extent
independent. They are grounded in cognitive capacities about
which, at present, we can only speculate. These are distinct styles
of scientific thinking, each of which has been developed in its
own way, in its own time frame, and each of which contributes to
the larger fabric of scientific imagination and action.
There is nothing peculiar to the sciences about the
uncovering of innate cognitive (and other) human potentialities,
and the development of social institutions within which they
may thrive. A most obvious comparison is with music. Analogies
between music and what we now call sciences struck ancient
Greek philosophers, and were preserved in the standard course
of education in mediaeval universities in Europe. The
"quadrivium" consisted of arithmetic, geometry, music and
astronomy. We now file three of those as sciences, and have
moved music to one side. This is in part because although we
have "music theory," we do not think of music itself as
propositional, in the way in which we view arithmetic, geometry,
and astronomy as being sciences whose knowledge can be
expressed by sentences. In these lectures I shall take for granted
that music is not a style of scientific thinking, yet that distinction
may itself be a contingent fact about the history of thought. As it
happens, after the new learning of the 17th century, the sciences
evolved in one direction and music in another. Perhaps things
could have gone differently.
The early European organization, which put music beside
astronomy, does not seem ever to have been on the cards in
6/ SCIENTIFIC REASON
China. There, if we wanted to shake up our preconceptions, we

might offer calligraphy as a science alongside astronomy.
2. Styles of scientific thinking

Mathematical deduction, taxonomic inquiry, hypothetical
modelling, experimental exploration, life in the laboratory,
probabilistic reasoning, historic-genetic thinking— I got the idea
of a small manifold of distinct styles of scientific thinking from
the historian of science A.C Crombie. I encountered his ideas at
a conference in 1978 in Pisa, and have never looked back.4 He
thought in traditional terms of the more or less continual
evolution and growth of methods of scientific reasoning, all the
way back to ancient times. He wanted to organize a global
history of science on an encyclopaedic scale. He had an ambition
to produce a "historical anthropology" of European science. That
is a valuable phrase that I suppose he took from the French
historian of the ancient Mediterranean world, Jean-Pierre
Vernant.5 I do not really agree with Crombie's project, but it
launched my own. Indeed you will see at the start of my second
lecture that I regard my enterprise as, in a sense, anthropological.
But not as historical anthropology; rather as what can be called
philosophical anthropology, whose direct ancestor is of course
Kant.
4 The proceedings were published in J. Hintikka, D. Gruender, and F. Agazzi (eds.)

Theory Change, Ancient Axiomatics, and Galileo's Methodology: Proceedings of the
1978 Pisa Conference on the History and Philosophy of Science, 2 vols., Dordrecht:
D. Reidel, 1981.
5 "Historical anthropology" has no standard usage in English, but the French
expression, "anthropologie historique," is well known to scholars who work in that
language. It started with historians of the so-called Annales school, and in particular
with Vemant's new way of studying ancient Greece and other Mediterranean
civilizations. The term is now used to name university courses, research groups, and
so forth, and is a standard entry in French encyclopedias, e.g. the Universalis.
In 1994 Crombie finally produced his life work, the

three-volume Styles of Scientific Thinking in the European Tradition:
The History of Argument and Explanation Especially in the
Mathematical and Biomedical Sciences and Arts.6 These are three
curious and obsessive tomes of which no one other than I has
made much use, but which perhaps unwittingly conveyed to me
a new vision of truth and reason. Crombie spoke of exactly six
styles of scientific thinking in the European tradition, each
developing according to its own trajectory and time scale. I have
no commitment to his precise classification of six, as a definitive
analysis of the history of Western science, but I find it a very
useful template. Here are brief tags for each of his styles:
mathematical, hypothetical, experimental, taxonomic, statistical,
and genetic.
None of these six defines a scientific discipline. It is important
to avoid a possible misunderstanding. Styles (in Crombie's sense)
are not sciences, at least not in the present usage in which we talk
about chemistry or palaeontology as sciences. Most of the
sciences use most of the six styles of thinking. Take an extreme
example. Taxonomic reasoning must seem wholly removed from
mathematics—until you reflect that some of the most profound
theorems are about classification, say the exhaustive
classification of the finite groups. Such theorems go back to the
five regular solids that so impressed Plato and his heirs.
Conversely, systematic biologists construct phylogenetic trees
based on fossil and now molecular-genetic evidence. A standard
tool of analysis is the method of maximum likelihood, developed
in applied statistics and using fairly elementary yet deep
mathematical principles. Most sciences use most of Crombie's six
styles of scientific reasoning. Arguably any well developed
6 London: Duckworth, 1994.

science uses all of them. That goes for relatively "lowbrow"

sciences, such as meteorology or mineralogy, as well as for those
that are usually ranked higher up on the pecking order.
3. "TheEuropean Tradition"
Two aspects of Crombie's title should be noticed. There is the

subtitle: The History of Argument and Explanation especially in the
Mathematical and Biomedical Sciences and Arts. I shall not dwell on
this, but it represents an unusual pairing of mathematics and
biotechnology, not to mention the very old fashioned but
fascinating juxtaposition of the sciences and the practical arts.
Crombie had a great deal to say about medicine, and I shall say
nothing. I regard that as my loss.
A second feature of his title is that he is writing about styles
of scientific thinking in the European tradition. Crombie did in fact
imagine a "comparative historical anthropology" of the sciences
which, in particular, would compare what happened in Europe
with what happened in Asia. A recent series of studies by
Geoffrey Lloyd does exactly that, with ancient Greece and
ancient China being the two civilizations up for comparison.7
But Crombie said nothing about science in East or South Asia.
Worse, he took only the most cursory glance at West Asia and
North Africa, which were the sources of so much Greek thought.
Although I shall take up questions of Chinese mathematics
briefly in the second lecture, my own work is thoroughly
Eurocentric. If I were to tie my presentation more closely to
Chinese science, I would take a page from Crombie and write
about mathematical and medical traditions in East and West.
Instead, my own predilection is to emphasize what I shall call
7 Geoffrey Lloyd, Ancient Worlds, Modern Reflections: Philosophical Perspectives

on Greek and Chinese Science and Culture, Oxford: Oxford University Press, 2004.
laboratory science, the laboratory style of thinking and doing.

That is the topic of my third lecture. I shall argue that it is a very
specific cultural invention, a crystallization of a very general
feature of human nature, namely curious exploration and
tinkering with the world as we found it. That event, which
happened in Europe in mid-seventeenth century, has made the
human race a parasite on the face of the earth, gradually
devouring the planet and all that is to be found there.
Of course the cultural invention, the laboratory, has long
passed from Europe. The most important laboratories of the
present day do biotechnology, and the most important labs in
that field are the venture capital laboratories of California and the
state enterprise laboratories of Shanghai. I shall touch on this
geopolitical observation at the end of the third lecture.
4. Styles as constituted by methods and objects

I am engaged in turning Crombie's odd history into an even
odder philosophy. A single sentence of Crombie's serves as a sort
of pivot for making the turn:
We can establish in the classic scientific movement a taxonomy of

six styles of scientific thinking, distinguished by their objects and
by their methods of reasoning.8
The words here, with which I do the turning, are objects and
methods of reasoning. They are, in themselves, anodyne. The
objects with which the mathematical style concerns itself are
often called, by analytic philosophers, abstract objects, such as
numbers, shapes, and groups. The objects with which the
taxonomic style concerns itself are, for example, the species and
genera of systematic biology, not mere classifications of living
8 A. C. Crombie, Styles of Scientific Thinking in the European Tradition, vol. I, p.

83.
things, which are found in all languages, but objects bearing a

definite role of sub- and super-ordination to other objects of the
same sort.
5. A crude template for organizing the past

Crombie's initial list of six styles, characterized by the objects
they examine and the methods they use, had, for me, an initial
plausibility. In the beginning, it was convenient to have a
historian hand me a simple catalogue on a plate. Nowadays the
community of historians regards Crombie as a relic, a strange
fellow who knew a lot but who worked outside the current
sphere of historical practices. Nevertheless, when I began, I was
able to blame any problems about a list of styles of reasoning on
an eminent historian! Here is pretty much the first exposition
that I encountered:
The active promotion and diversification of the scientific methods

of late medieval and early modem Europe reflected the general
growth of a research mentality in European society, a mentality
conditioned and increasingly committed by its circumstances to
expect and to look actively for problems to formulate and solve,
rather than for an accepted consensus without argument. The
varieties of scientific method so brought into play may be
distinguished as:
(a) the simple postulation established in the mathematical sciences,

(b) the experimental exploration and measurement of more complex
observable relations,
(c) the hypothetical construction of analogical models,
(d) the ordering of variety by comparison and taxonomy,
(e) the statistical analysis of regularities of populations and the
calculus of probabilities, and
(f) the historical derivation of genetic development.
The first three of these methods concern essentially the science of

individual regularities, and the second three the science of the

regularities of populations ordered in space and time.9
Thus Crombie in that talk I heard in 1978. He had not quite fixed
on his terminology of "styles of scientific thinking"—as you see,
he at first spoke of methods, rather than styles, but I shall keep
the two words distinct, in line with his later thought that styles of
scientific thinking are constituted by methods of reasoning and
objects of inquiry.
There is a more profound difference between this inaugural
paragraph of 1978, and Crombie's big book of 1994. Here he
speaks of late medieval and early modem Europe—roughly
speaking the time leading up to what other historians of
European science have long called the scientific revolution. In the
final three volume magnum opus, we have a tale going back into
ancient Greece where, for Crombie, a set of social institutions
created a cast of mind in which what we call the sciences began.
The massive work of 1994 is organized around the six styles, each
of which is presented as stemming from ancient antecedents. In
1978, as the above passage shows, he really thought of his styles
of scientific thinking as taking definitive shape much later than
that.
My chief innovation, in connection with Crombie's list, will
be to urge that his long-term styles of 1994 are punctuated with
what I call crystallizations, in such a way as to emphasize the
early modem period in European history. This may diminish a
certain tension between his earlier and his later discussion of
styles of scientific thinking.
Before becoming analytic, or worrying about the long or the
short term, we should honour our immediate take on the list of
9 A.C. Crombie, "Philosophical Perspectives and Shifting Interpretations of

Galileo", in Hintikka et al, Theory Change (op. cit. note 4), pp. 271-86 on p. 284.
six. Each item simply feels different, looks different, and may well
be different. Each may reflect the use of a different group of skills
for which the human mind is well suited.
Crombie's list of six is a template, that is, a convenient pattern
to follow. There are innumerable ways to carve up something so
extended as thousands of years of many civilizations of evolving
"sciences." There are many possible frameworks of analysis:
think only of Thomas Kuhn's paradigms, Imre Lakatos's research
programmes, Gerald Holton's themata-—he list goes on. The
value of any division must lie in how one uses it. None is an
ultimate and definitive framework in its own right.
I have sometimes tried to construct precise definitions of a
style of scientific thinking, which would validate Crombie's
template. That is a misguided enterprise. It is better to see what
one can do with his idea. A form of Occam's maxim provides a
good rule of thumb: the list of styles of scientific thinking should
not be enlarged beyond necessity.
Every style enumerated by Crombie is a living, evolving,
organism. He used his template as a way to organize the history
of science. He has been amply criticized, and some of the
criticisms strike at the heart of his historiographic project.10 I
refrain from criticism because I am not interested in his catalogue
for its own sake. My aim is rethinking the entire structure of
scientific reasoning from what I call a Leibnizian point of view. It
is a view that, as we shall see, can also be called anthropological
in something like the original sense of Kant's Anthropologie. I
mean a study of the human species, its innate powers, and its
10 R. Iliffe, "Rational Artistry," History of Science, 36 (1998): 329-357, an extended

critical study review of Crombie's Styles of Scientific Thinking in the European
Tradition. For shorter reviews, R. Ariew, isis 86: 82-83; A.J Meadows, New
Scientist, 23 July 1994, p. 38; K. Magruder, Sixteenth Century Journal 26 (1995):
406-410.
ON THE HISTORICAL ROOTS OF SCIENTIFIC REASON /13
current practices. It helps immeasurably to have a template to

hand, albeit furnished by a historian whom his juniors have
come to dismiss. Readers will keep on saying, how on earth can
you take Crombie seriously? Answer: because he offered so
useful a road map—and by way of honestly owning an
intellectual debt. The template makes sense as a list of our ways of
finding out about the world that we call scientific—makes sense
as a starter kit.
Crombie's descriptions are sometimes dated. He did his
thinking when mathematics was thought of as the axiomatic
method. He thought of Euclid as a matter of postulates. Hence
his characterization of the first style in his list as postulational.
He was a man of his time. I am a man of my time; I came of age
when proof theory flourished. For me, Greek geometry is a
matter of proof, not postulates. I can summon historians who
will tell me that the standard Euclid of axioms is more a matter of
European mediaeval organization of texts than what the old
Greeks actually did. So I am right and Crombie is wrong? That is
a matter of scholarship, which is not so important in the biggest
picture. Whether we focus on postulates or proof, mathematical
reasoning and the ability to do it is something all of us recognize,
even if some of us are good at it and others are not. We know
when something demands mathematics. That is an amazing fact
on which I shall dwell in the second lecture. Why do we call all
that mathematics at all? It is so many things, a motley, axioms,
proofs, pictures, insight, calculation. But there it is, a distinct style
of thinking, no matter how we label it.
6. Crystallization
Crombie's vision of the history of the European sciences
favours continuity. My instinct is exactly the opposite. I like to
tell the story of each style of scientific thinking as having at least

one sharp moment of crystallization, a fixing of how to go on in
the future, usually after centuries, perhaps millennia, of inchoate
precursors. I acquired this habit early, in The Emergence of
Probability, first published in 1975, some years before I had ever
heard of Alistair Crombie, let alone his styles. My book was an
archaeology of knowledge, in the sense of Michel Foucault. It
was very much the first such archaeology to be written in English,
and it was the first in any language to deploy an analytic idiom. I
explain this in the new Introduction to the 2006 edition of the
book.11
I argued in 1975 that anything we can readily recognize as
probable reasoning began around 1650. New kinds of statements
began to be made by people all over Europe and we began to
enter a new world, a world of chance, the world we inhabit today,
of which Ulrich Beck's "risk society" is only one among many
prominences.12 Yes, you can find anticipations of things that
were said after 1650 in many earlier times, in many places and
climes. In the original book I mentioned a striking mastery of
probability-type reasoning in a Sanskrit classic, composed over
two millennia ago in India. There must be parallel examples in
China to which no one has yet drawn attention. For thirty years
scholars have produced examples of anticipations of probability
in order to refute my claim. Nevertheless, as I explain in the new
Introduction to Emergence, I stick to my claim. Many writers have
11 Ian Hacking, The Emergence of Probability, Cambridge University Press, 1975;

second edition, with a new "Introduction 2006,"~ Cambridge University Press, 2006.
Thanks to the gross incompetence of the New York Branch of the Press, the new
Introduction is neither paginated nor listed in the Table of Contents.
12 This term, "risk society,"' has become almost a cliché of our times, thanks to Ulrih
Bech, Risk Society: towards a New Modernity, London: Sage, 1992, translated from
Risikogesellschaft: auf dem Weg in eine andere Moderne, Frankfurt am Main:
Suhrkamp, 1986.
found rich elements of other systems of thought, in which we can

recognize hesitant precursors of our own probabilities. Crombie
himself, in presenting his fifth style of scientific thinking, takes
probabilities back into the mists of time. One can tell such a story,
but in my opinion it teaches little, because only after 1650 did
human beings begin to put the possible pieces together, and see
that confidence in diverse opinions, and the frequencies with
which things happen, have the same underlying structure. That
is, the structure of the laws of mathematical probability.
I shall not dwell on probability for more than another
moment—I have devoted too much of my life to it. In a parody
"Philosophical Dictionary" composed by Dan Dennett a long
time ago, the verb "to hack" is defined as obsessive attention to
detail, with this example, "he spent years hacking his way
through the statistical jungle." But for just a moment let me hack
at the jungle one more time. Lots of truths about probability are
not temporal at all: they are mathematical theorems of
probability calculus. Some other statements using probability
may be true at all times and places, say facts about the half-life of
radium. Many more are tensed, true or false at a time. But, to
exaggerate in order to display the point, no such statements, of
any of the three sorts, could be made before 1650. There was no
way to assert those truths; no conditions for their truth or
verification were in place. Methods of reasoning about them had
not yet come into being. There was, as I shall put it later in this
lecture, no truthfulness about what we call probability, until my
arbitrary date of 1650.
I have, evidently, a certain enthusiasm for mutations in
systems of thought, inherited from Gaston Bachelard.13 But
13 I say "inherited", For Anglophone philosophers of my generation, it was of

course Kuhn's Structure of 1962 that dominated all discussion. But my specific
enthusiasm for breaks in systems of thought was got from Michel Foucault's The
notice that I do not now call the emergence of probability in 1650

a mutation, or even a revolution.14 I have just called it a
crystallization. When water freezes, it becomes a completely new
substance, ice. Such phase transitions, as the physicists now call
them, are reversible. But there is no going back from the
probabilistic world we now inhabit. A crystallization in the
evolution of a style of scientific thinking is in effect irreversible.
I shall, in every case, cut Crombie's styles to my own pattern,
noting at least one crystallization. I feel no tension between
crystallization and continuity. My attitude to the past may be the
very opposite of Crombie's, but opposites can be complementary.
Here again we may usefully recall Foucault, from whom I
probably acquired this stance. At the beginning of the Archaeology
of Knowledge- his not very successful (in my opinion) attempt to
describe his own early methodology - he recalls the power of
Braudelien longue durée accounts of the past. Climates, not kings,
rule. The big events in Mediterranean history are for example the
cutting down of most of the trees in Greece to build ships. That
may have changed the world, but the great naval battle of
Salamis is a mere incident in the history of tree-felling. Turning
Greece into a barren rock changed the weather, and so the
fortunes of peoples. Foucault made plain that his tale of systems
of thought in terms of radical and almost instantaneous
mutations was completely consistent with such magisterial
versions of slowly moving centuries.
I have absolutely no commitment to any definitive
Order of Things (Les mots et les choses: une archéologie des sciences humaines,
Paris: Gallimard, 1966.) That in tum goes back to the influential writings of Gaston
Bachelard (1884-1962), who took up a lot of space in Paris bookstores during the
1960s.
14 On the specific question of revolutions in the history of probability, see my paper,
"Was There a Probabilistic Revolution 1800-1930?", in Lorenz Krüger et aI., (eds.),
The Probabilistic Revolution, Cambridge: M.LT. Press, 1987, vol. I, pp. 45-58.
classification of styles of scientific thinking, or to Crombie's

continualism as against my punctuated past. But I shall be
conservative. There is no point in creating artificial arguments
about how to classify fundamental styles of scientific reasoning. I
shall be parsimonious, imitating Occam's maxim. I shall stick
with Crombie's six styles, but I shall punctuate them. The
punctuations, I shall urge, are more important for understanding
scientific reasoning than are the long drawn-out stories of
predecessors. It is only when a style crystallizes that people
really understand how to find out things using that style.
Hence in the case of each style I shall note at least one
discontinuity, such as what I call the emergence of probability. I
do not invent these points: on the contrary, they are already
legendary, and each has its own trailblazing figure, or, at least a
household name. There is Pascal. There is Linnaeus, in biological
classification. Inevitably there is Galileo in hypothetical
modelling. These names happen to name real people. But each
stands for, symbolizes, a profoundly new beginning in which a
host of real human beings in concert collaborated to produce a
new way of going on. Far from resorting to old-time history of
science, with its tales of the hero in history, I invoke legendary
heroes with names that mayor may not correspond to the names
of historical personages. They serve as the names of
crystallizations of styles.
7. Regrets about the word "style"

I much regret having started to use the word "style" after
listening to Crombie, for it has acquired ever so many
connotations since Heinrich Wölfflin introduced it into art
history early in the 20th century.l5 There was a great vogue for
the word in Germany, including its compound Denkstil,
thought-style, or style of thinking. It figures in Oswald Spengler's
analysis of the decline of the West, published before and after the
First World War. It was very important, for example, to Karl
Mannheim's approach to sociology.16 Among the positivist
philosophers, we find Rudolf Carnap speaking of styles of
thinking, but it is not central to his analysis.17 The Nazis made a
great play with a specifically Jewish thought-style.l8
The crescendo of talk about thought-styles—in Mannheim,
Carnap, and the Nazis alike—occurred in the 1930s. That was the
decade of what we now regard as an enormously important book
by Ludwik Fleck, translated as Genesis and Development of a
Scientific Fact. That title omits the German subtitle, Introduction to
the theory of the thought-style and the thought-collective.19 Fleck's
idea of a Denkstil differs profoundly from Crombie's notion of a
style of thinking. Fleck is concerned with a way of thinking, a
way of finding out that is current in a specific community, a
thought-collective, at a definite time, and which, evolves, mutates,
or dies in a brief span of time. The thought-collective is a good
15 Heinrich Wölfflin, Kunstgeschichliche Grundbegriffe: Das Problem der

Stilentwicklung in der neueren Kunst, Munich: Bruckmann, 1915. (Principles of Art
History: The Problem of the Development of Style in Later Art, New York: Dover,
1929.)
16 For references see my "'Style' for Historians and Philosophers," Studies in
History and Philosophy of Science 23 (1992), 1-20.
17 Rudolf Carnap, The Logical Structure of the World: Pseudoproblems in
Philosophy, Berkeley: University of California Press, 1969, p. xvi.
18 1 have not kept a record of the sources which I recall here, but I well remember
the shock with which I encountered them when skimming German science
journalism published in the 1930s.
19 Ludwik Fleck, Entstehung und Entwicklung einer wissenschaftlichen Tatsache:
Einführung in die Lehre von Denkstil und Denkkollektiv, Basel: Benno Schwabe,
1935. Reprinted, Frankfurt am Main: Suhrkamp, 1980; Genesis and Development of
a Scientific Fact, Chicago: the University of Chicago Press, 1979.
name for what Thomas Kuhn less aptly named a "disciplinary

matrix," a body of core workers and their appendages in some
field of inquiry. Fleck's Denkstile are relatively short-lived, while
Crombie's styles of scientific thinking are very much a matter of
the longue dureé. They evolve and crystallize, yes, but they persist
over a long period of time, and are used to varying degrees in all
of the scientific disciplines and more.
One reason for disquiet about the word "style" is, then, that it
already had an entrenched usage in work of the highest calibre.
And at exactly the same time, call it 1935, it was also used in
utterly despicable ways. It was a vogue word of German culture,
put to uses of many types. Happily we remember only the best
and have suppressed the worst.
A second reason for unhappiness is that the word "style" is
used widely in many contexts. In the sciences it is altogether
natural to contrast the style of one famous researcher to that of
another. A whole series of papers to that effect were published
after I first wrote about styles of reasoning.20 It is a happy
accident here, that I used to write of "styles of reasoning." I have
abandoned that locution, and have gone back to Crombie's
original "styles of scientific thinking." This leaves "styles of
reasoning" as an expression for others to continue using in their
own ways. For example Arnold Davidson appropriated my
"styles of reasoning" in his brilliant study of perversion, but
meant something wholly different from what I meant.21 Most
recently, we have Styles of Reasoning in the British Life Sciences in
20 The work I am describing goes back to my "Language, Truth and Reason,"

Rationality and Relativism, ed. M. Hollis and S. Lukes, Oxford: Blackwell, 1982,
48-66. The next stage of development was my "style" paper of 1992, note 16 above.
21 Arnold Davidson, The Emergence of Sexuality: Historical Epistemology and the
Formation of Concepts, Harvard University Press, 2001.
the first part of the nineteenth century.22 Perhaps I shall mention

the phrase "style of reasoning" again, but I shall never use it
henceforth. By accident it is in the public domain and is there for
others to use. No one, so far as I know, writes about styles of
scientific thinking except Crombie and, after him, myself, so we
own the expression, for a little bit longer.
Styles of scientific thinking, in the sense of Crombie's
definitive six, are to be sharply distinguished from a more
generic concept of style. Recently Geoffrey Lloyd has been using
"style of inquiry" as a fundamental analytical tool in his
persuasive comparative studies of ancient Greek and Chinese
science.23 It is likely that in the future there will be many more
uses of the word "style" for conjuring up this or that aspect of the
sciences.
"Style" might seem, then, to have become an overused word,
both in the 1930s and now. Is it not a worn-out duck that should
be put out of its misery? Sometimes I think so, but, after much
word-searching and soul-searching, I feel stuck with Crombie's
expression. I shall continue to speak of long-standing styles of
scientific thinking, punctuated by moments of crystallization.
Occam's maxim acquires a special utility. If we resolve not to
enlarge the list of styles of scientific thinking beyond necessity,
we avoid tedious debates about how to define styles of
thinking—as if they had an essence to be discovered—and thus
can get down to work.
22 James Elwick, Styles of Reasoning in the British Life Sciences: Shared

Assumptions, 1820-1858, London: Pickering & Chatto, 2007.
23 He actually started out using my old term "style of reasoning" for just this
purpose, but then quietly moved on to his own terminology. "Deux styles
radicalement different de raisonnement mathématique." G. Lloyd, "Preface" to
Karine Chemla et Guo Shuchun, Les neuf chapitres: Le classique mathématique de
la Chine ancienne et ses commentaires, Paris: Dunod, 2004, p. xi.
8. Self-authenticating styles
So far, so simple. Now I forewarn of the radical. I pass from
the historical to the metaphysical, from historical anthropology
to philosophical anthropology, and from bland description to an
attempt to undo ontological debates.
I contend that when it crystallizes, a style of scientific
thinking introduces new objects, and new criteria for the truth or
falsehood of statements about those objects. A style, with its
specific methods of reasoning, does not answer to any criteria
except its own. It is not good because it helps find out the truth in
some domain. It itself defines the criteria for truth—telling in its
domain. Thus in a certain sense each style is autonomous and
"self-authenticating". That certainly sounds radical, and that is
what I mean.
Few such surprising doctrines are all that new. In the heyday
of logical positivism, Moritz Schlick coined the slogan, "the
meaning of a statement is its method of verification." That was
quickly revised and then abandoned, but has more merits than
are usually assigned to it. Schlick meant a method of verification
appropriate to an individual statement, rather than very general
kinds of method. Without embracing his strict verificationism,
we assert that until there are methods of reasoning that bear on
the truth or falsehood of a scientific statement, the question of its
truth or falsehood does not arise. Meaning, as simplistic forms of
analytic philosophy have insisted, demands the possibility of
applying truth conditions. Taking advantage of Bernard
Williams' "truthfulness" that I shall soon explain, it can be said
that scientific statements come up for grabs, as true or false, only
when there comes into play a method for reasoning about their
truth.
Most ordinary statements are not like this. They have truth
conditions, methods of verification and so forth, independently

of styles of scientific thinking. The cat is on the mat. Likewise
most objects are not introduced into discourse alongside styles of
scientific thinking. Sticks and stones.
One is obliged to make pedantic disclaimers that on reflection
are trivial. For example, it would never occur to me to suggest
that statements about dinosaurs in the distant past were false in
that distant past, just because we could not assert them.24 That
would be absurd. The statement, "Dinosaurs roamed the earth
during the Jurassic epoch" is true, but it makes no sense to say
that it is true now, or that it was true a million years ago. It is a
statement that English grammar— I shall speak just for my own
first language—throws into the past tense, but it is not true or
false at a time at all. Likewise I do not mean that the Pythagorean
Theorem became true only when geometry came into being as a
body of knowledge. The Pythagorean Theorem is untensed, it is
not true or false at a time.
9. Objects
Styles of scientific thinking introduce their own distinct class
of objects. Think of the abstract mathematical objects ('platonist'),
of the unobservable theoretical entities at the centre of the recent
debates about scientific realism, or of systematic biology with its
taxa. Each style is specific to it its own domain, but only because
it introduces the objects peculiar to that domain. It does not
create them—to say that would be foolish—but they have no
24 About 1985 Nelson Goodman said in public, "Ian, do you mean to suggest. .. "
[that the dinosaurs etc.] I said, no! I like it that the very first time this question was
posed to me, it was by the greatest pragmatist of his day, the greatest nominalist of
all time, the very man who is often taken to suggest, what he himself thought was
manifestly absurd, namely the ultimate reductio ad absurdum of any foolish version
of his own sound ideas.
place in thought outside the styles.

As a corollary, the new classes of objects make possible the
interminable ontological debates in those domains, for example
between platonism and antirealism in mathematics, or scientific
realism versus the various types of instrumentalism and
empiricism about physics. Ontological debates within the
sciences result from (a) the introduction of objects by styles of
thinking (b) the fact that we talk about these objects using
sentences in which names for the objects serve as grammatical
subjects, and (c), as already emphasized by Nietzsche, most
languages demand an existential presupposition for terms in the
subject position. This is certainly true of European languages, so
it is not surprising that these ontological debates are above all
European in nature. I shall return to this question in the second
lecture. Is ontology simply a facet of European languages and
culture, which Chinese thinkers ought to ignore?
Just as most common statements are independent of scientific
styles of thinking, so the most deep-seated ontological debates
have nothing to do with styles. Is there an external world? Are
there other minds but mine? Or (a vexed problem in various
guises for philosophers of logic) are there universals, classes,
properties, or do only individuals exist? Ontological issues all,
but they have nothing to do with styles of scientific thinking.
I hold these doctrines, about statements, objects, and styles of
scientific thinking, to be profoundly rationalist in character, very
much in the spirit of my mentor Leibniz. Far from implying some
sort of relativism, the doctrine that styles of thinking are
self-authenticating is part of an explanation of what we call
objectivity.
One must be wary when it comes to objectivity, self-conscious.
As Lorraine Daston and Peter Galison have done so much to
show, the epistemological concept of objectivity has a chequered
past, protean and polysemic. 25 The words "objective" and

"objectivity," together with their mates "subjective" and
"subjectivity," acquired their present meanings in European
languages only at the time of Kant.26 When I speak of objectivity
I mean chiefly to affirm that the truths discovered in the sciences
are true, independent of what we think, or of how we discover
them. That is wholly consistent with saying that their truth
conditions are products of the styles of thinking in whose
domain they fall.
10. Leibniz and Bourdieu

Allow me an unexpected comparison here. Pierre Bourdieu's
Pascalian Meditations of 1997 was, in effect, his philosophical
testament.27 He too was much concerned with the objectivity of
the sciences, including of course his own sociology. In my
opinion he was unduly hostile to most social studies of science,
whereas I learn much from the best practitioners of science
studies, even when some of them think I am an old
stick-in-the-mud who has learned nothing.28
Bourdieu was deeply committed to a rationalist vision—it is
too often forgotten that his early work is both about Leibniz, and
that it is truly Leibnizian. Since he was a significant man of our
times, he was historicist. In the chapter called "The Historicity of
25 Lorraine Daston and Peter Galison, Objectivity, New York: Zone Books, 2007.
26 Or a little before, for example Isaac Watts, Logic, London, 1724: "Objective
certainty, is when the proposition is certainly true in itself; and subjective, when we
are certain of the truth of it" (II. ii. § 8.)
27 Pierre Bourdieu, Méditations pascaliennes, Paris: Seuil, 1997; Pascalian
Meditations, Stanford University Press, 2000.
28 For my own reflections on science studies in connection with Bourdieu, see Ian
Hacking, "La science de la science chez Pierre Bourdieu," in J. Bouveresse et Daniel
Roche (eds.), La Liberté par la connaissance. Pierre Bourdieu (/930-2002), Paris:
Odile Jacob, 2004, 147-162.
Reason" in his Meditations, he wrote that "We have to acknowledge

that reason did not fall from heaven as a mysterious and forever
inexplicable gift, and that is therefore historical through and
through; but we are not forced to conclude, as is often supposed,
that it is reducible to history." He went on to insist that, "It is in
history, and only in history, that we must seek the principle of the
relative independence of reason with respect to history."
Moreover—and here I paraphrase an overly French sentence—he
thought that "the singular history of reason takes place" in
"completely specific," "strictly historical," but entirely
29
exceptional conditions. I could present my use of styles of
scientific thinking as a long gloss of that passage of Bourdieu's.
For it is in completely specific conditions that new styles of
scientific thinking come into being and flourish.
This kind of historicism, which I cheerfully call Leibnizian, is
very close to my own. It is the exact opposite of Husserl's final
project in the Crisis of European Civilization. Husserl's historical
and scientific insights are enormously important, but my project
is wholly different from his (insofar as I understand it). He
thought that we had fundamentally lost touch with what we
were really doing in the sciences, because we had buried the
original intuitions under innumerable layers of sediment. That
was the deep source of the crisis, of which the evident and
superficial signs were the state of Germany, heartland of
European culture and science, in 1936. The task of a
transcendental phenomenology was to recover the grounding
intuitions and rework up to the present from there. The noble
aim was to rescue European civilization from disaster. Much as I
respect what Husserl wanted to do, I believe that all ideas of
recovering a first understanding are fundamentally mistaken,
29 Meditations pascaliennes, 130; Pascalian Meditations, 109.

quite aside from the imagined political benefits.30 I shall

nevertheless choose to name one crystallization of a style of
scientific thinking "Galilean," and that name goes almost directly
back to Husserl, who, it seems to me, had a remarkable grasp of
the profound changes effected in scientific reason at the time of
Galileo.
11. Society
A style of scientific thinking answers to no criteria but its own.
This does not mean that it needs no support, as if it were some
sort of free-standing inhabitant of what Popper called the Third
World, a reincarnation of the headier postulates of Plato and
Frege. It is people who think. People reason. People find out.
People need to eat, and even if they are rich, they need people to
listen to them. Styles are enabled by institutions. Geoffrey Lloyd's
comparative study of ancient science in Greece and China,
mentioned earlier, attends closely to the institutional settings,
that allowed various of what he calls "styles of inquiry" to
flourish in one or the other society.
Styles of scientific thinking in the European tradition are just
as much social practices as Fleck's Denkstile. They are less local
and more enduring in part because they are built on
fundamental cognitive capacities. If they become extinct it will
not be by refutation, but by being abandoned. Or they can
become moribund. There is no better example than Greek
mathematics, which came to a full stop after Archimedes. People
wrote commentaries in Greek, Arabic and Latin, but no new
maths was created for a millennium. It was not so different in
30 I explain this in connection with mathematics and Husserl's "Ursprung" in

"Husserl on the origins of geometry," in David Hyder, Husserl's Epistemology,
Stanford University Press, forthcoming.
China.
I have asserted that styles of scientific thinking are in a
certain sense self-authenticating. Perhaps this implies that they
cannot be undone "internally"; they cannot be called into
question from within, for they furnish their own canons of
correctness. But we can just lose interest, for all sorts of reasons.
Maybe people quit creating mathematics after Archimedes
because no one had any new ideas. Or the next generation just
got bored with those old fogeys—there was no pay-off anyway.
The abandonment of a style of scientific thinking could arise for
the most banal of reasons.
Herman Hesse's The Glass Bead Game is a parable for a society
built around an intellectual activity, wholly viable, maintained by
complex institutions.31 It is not unlike some aspects of the
mathematical sciences, but the core is music and it begins with
the art of fugue. In the end, the Magister Ludi, the veritable
Archimedes of the game, decides he must, in all humanity, leave
the institutions that make the Glass Bead Game possible. From
our point of view, such a system, described by a great and
inventive novelist, is just human enough to be fascinating, and
yet has the air of absurdity.
Hesse well knew that some of his contemporaries regarded
mathematics as a purely formal game. Nowadays we pay not
only pure mathematicians but also string theorists out of the
public purse. So far, the stunning conceptual innovations of
string theory have no empirical or testable consequences
whatsoever. There is always a little worry, in such fields, that the
purse will dry up, and in a generation or two the activity will
31 Das Glasperlenspiel, Zurich: Fretz and Wasmuth, 1943. The Glass Bead Game:
A Tentative Sketch of the Life of Magister Ludi Joseph Knecht, together with
Knecht's Posthumous Writings edited by Herman Hesse, London: Jonathan Cape,
1970.
come to an end.
Here are a few sentences from another source that you would
not expect. J. M. Coetzee, as it happens like Hermann Hesse, won
a Nobel Prize for literature. His Tanner Lectures on Human
Values present a lecture about animals by the Australian novelist,
Elizabeth Costello. She in turn supplements a lecture imagined
by Kafka in 1917. It is by Red Peter, an ape who has learned to
talk. We read that "seen from outside, from a being who is alien
to it, reason is simply a vast tautology. Of course reason will
validate reason as the first principle of the universe—what else
should it do. Dethrone itself? Reasoning systems, as systems of
totality, do not have that power."32 But maybe Costello could
persuade everyone to stop thinking that way.
From time to time there is, in Richard Bernstein's phrase, a
rage against reason.33 One such time was the late 1960s, when
science was seen as a tool of capitalist colonial war. Young rebels
tried to sack places of research. That tactic was foolish but right.
If you want to criticize a style of thinking, you cannot do it on its
own terms (for then you are merely thinking according to that
style, trying to do it better). You must destroy its institutional
base. For very different reasons, the cultural revolution of the
same epoch successfully destroyed a generation of Chinese
scientists. Reason was hated in China and America at about the
same time. Perhaps science could have come to a full stop,
forever.
I do not myself believe that is an option, because I suspect
that when people have uncovered ways of finding out, they will
continue to use them so long as they can do so. Only coming to
think, e.g., that there is nothing to do, after Archimedes, might
32 J. M. Coetzee, The Lives of Animals, Princeton: Princeton University Press, 1999.

33 Richard Bernstein, "The Rage Against Reason", in E. McMullen (ed.),
Construction and Constraint, Notre Dame: Notre Dame University Press, 1988,216.
ON THE HISTORICAL ROOTS OF SCIENTIFIC REASON/ 29
terminate a way of thinking "voluntarily." I have the optimistic

view that coercion, whether in the form of a cultural revolution,
or in the form of American "political correctness", will exhaust
itself. But I grant that is a certain type of optimism.
The most pessimistic view would attend to the thought that
the styles of scientific thinking have gradually created a parasite
upon the face of the earth: the human race. The sciences have
turned us into an organism that consumes more and more of its
environment. Now a parasite must in the end be weaker than its
host, for if it kills its host, the parasite itself dies. A fear of
man-as-parasite-upon— the-face-of-the-earth, gradually killing its
host, might lead our race to stop engaging in the sciences, for fear
of putting an end to ourselves. Scientific thinking might just
stop ...
12. Bernard Williams' "Truthfulness"

My kind of historical specificity, to repeat Bourdieu's phrase,
is directed not at the experience of grounding reason but at
Leibnizian conceptions. I make a change that looks like casuistry,
a play on words. I turn from truth and reason to truthfulness and
reasoning. Now "truthfulness" is not a common English word. It
does not translate easily into Chinese. I take it from Bernard
Williams' last book, Truth and Truthfulness: An Essay in
Genealogy.34 It is the most interesting book in living memory
about truth and its value. It is about many things, and my use of
it is selective. I want to extend its notion of a genealogy of
truthfulness. Williams' metaphorical use of this idea of genealogy
is strictly taken from Nietzsche, with a somewhat different
emphasis from that of Foucault, who of course was inspired by
the same source.
34 Op. cit. note 2.

Williams offered two sub-genealogies, one of telling the truth

about the historical past, and the other of authenticity about
oneself. The time frame of his chapters on these topics is quite
different: the Athens of Thucydides and the Paris of Rousseau
and Diderot. He chose them for several reasons. Their intrinsic
interest. Their importance to his own philosophical work. But
also because these were among the two most contested domains
of truth-telling during the recent so-called culture wars. One of
his aims was to show that the possibility of telling the truth about
the past (a) has a history (contrary to dogmatic thinkers who
conceive of history as the relating of facts, facts that could have
been recounted any time, anywhere). And (b) this is a history of
how objectivity about the past came into being (contrary to
dogmatic thinkers who say that there is no objective historical
truth). Likewise for the self and authenticity: he opposed (a),
those who insist that there just are truths about individual
human beings, independent of the context in which they are
understood, and (b), those who say there is no objective truth
about a person, only the stories that people tell about themselves.
I shall try to extend Williams' idea to the sciences, where,
notoriously, a few years ago there was a similar (a)/(b)
debate—timeless fact versus no objectivity. I do so by connecting
Williams's genealogical approach with Crombie's idea of a
handful of evolving styles of scientific thinking.
13. Silence about truth

Williams most emphatically did not write a genealogy of
truth.
One thing I shall not consider, however, is the history of the concept
of truth, because I do not believe there is any such history. The
concept of truth itself—that is to say, the quite basic role that truth
plays in relation to language, meaning and belief—is not culturally

various, but always and everywhere the same. (p. 61)
Truth, then, has no history, beyond the fact that it is coeval with
the emergence of linguistic structures to convey information.
This conception is Aristotelian and Tarskian. The adjective 'true'
has many uses, but truth is a formal concept, essential to
semantics but with no semantics of its own. Thus it spans all
informative discourse, and has no genealogy.
Aristotle: "To say that that which is the case, is the case, and
that which is not the case, is not the case, is to say the truth."35
That is a blank, formal, assertion, conveying in passing the
fundamental fact that the adjective "true" primarily applies to
what is said, was said, or can be said. There is undoubtedly a
history of when human beings began to talk, to say things
informatively, to make what we can recognize as assertions. But
there is no further history of truth than that.
I read Aristotle's maxim as an early version of Tarski's equally
formal semantic theory of truth. Its scheme, " s is true if and only
if p", makes as plain as Aristotle did that the adjective applies to
sentences. The somewhat overblown theory of metalanguages
derives from that trifling grammatical fact. Tarski himself wrote
that his semantic theory appeared to be consistent with, and even
to express the core motivation for, every substantive "theory" of
truth, every theory which says what truth "is"—correspondence,
coherence or whatever. That is one way of saying that his own
theory is formal, and content-free. That means that, strictly
speaking, it is not a "theory" at all.
No philosopher dedicated more of his career to the concept of
35 Aristotle Metaphysics ~ (Book IV), 1011 b25. In translation of Cristopher Kirwan,

Oxford: Clarendon Aristotle, 1971: "To say that which is is, and that which is not, is
not, is true". In W. D. Ross, Oxford: Clarendon Aristotle, 1908: "To say that what is,
is, and that what is not, is not, is true".
truth than did Donald Davidson. He published at least a dozen

distinct papers with "truth" (or "true") in the title. They run from
"Truth and Meaning," in 1967, to "Truth Rehabilitated," in 2000.36
An unfinished volume was published after his death, with the
title Truth and Predication.37 Nothing would be more foolish than
a summary, in a few words, of his thoughts about truth. I shall
mention only the obvious: that he was ever loyal to Tarski, that
he rejected substantive theories of truth, and that he thought
truth is a bedrock idea that cannot be defined. He would not say
that truth is a purely formal notion; he would not have said that
truth has no semantics, because he would not have believed that
was the right way to put the matter. Yet I am sure that my
abrasive, "truth is a merely formal concept," owes much to
thinking about (and conversations with) Donald Davidson.38
As a matter of both principle and convenience I shall say
nothing substantive about truth.39 Why "convenience"? Because
truth has recently been such a contested notion among analytic
philosophers, with their competing theories of truth, that it
would be imprudent to dabble in waters held to be very deep,
certainly over my head. I have views about those debates, as is
shown by my doctrine that truth is a purely formal notion, but
those views are not important to my investigation of styles of
scientific thinking.
36 I refer only to papers in the five volumes of collected essays published by Oxford
University Press.
37 Cambridge, Mass.: Belknap.
38 The longest chapter in my Why Does Language Matter in Philosophy?
(Cambridge University Press, 1975), is called "Donald Davidson's Truth".
39 Thus I re-express most of what I said about truth, in the two previous papers
mentioned in footnote 6, in terms of truthfulness.
14. Truthfulness about the past

By truthfulness, Williams meant telling the truth about
something. This has two aspects. The truthful person is both
quite accurate and quite sincere. These are fairly independent
virtues. Williams was concerned with how and when it became
possible to be accurate and sincere about some subject matter. A
genealogy of truthfulness about X, will have two branches,
whose rhythms may be very different. Thus Williams's history
chapter focuses on accuracy and his chapter on self-knowledge
focuses on sincerity. Since I am interested in the sciences, I focus
on accuracy, although it is important also to see what happens to
sincerity in the scientific context.
Williams spumed vague generalities. The transformation in
the Western conception of telling the truth about the past
happened at a definite moment, at the time of the work of one
man. And that man, in the European tradition, was Thucydides.
Thucydides has become emblematic, a sort of icon, standing for
the beginning of the writing of history in the Mediterranean
world. No man stands alone: it is Thucydides, his hearers, and
his readers who count, in that highly specific social universe
which was Athens. There is another specific social universe in
which history writing came into being in much the same way as
it did in Greece: ancient China. The invention of writing is a
precondition for history in Williams's sense, so China is the place
to look for a decisive conceptual change like that which occurred
in the Mediterranean at the time of Thucydides.
The standard candidate is Sima Qian (145-90 BCE) or perhaps
his father, Sima Tan. One should not think that exactly the same
change in conceptions of the past occurred in China and Greece.
The Simas, father and son, had high official status at court for
long periods, a status which had no counterpart in Greece at the
time of Thucydides. Sima Qian's history of two millennia became

the paradigm, quite literally, of Chinese historiography—the
model of historiography that was followed for centuries. One
could argue that it was a social and cultural history unknown in
the West until the twentieth century.
It is not news that Thucydides is the first real Western
historian. Hume said as much. So have generations of scholars,
each with their own explanation of what makes this moment in
historiography something new. Williams's version is in terms of
truthfulness: there was, "most basically, a shift in conceptions of
what it is to tell the truth about the past." This is the fundamental
move in the historical parts of Williams' genealogy of
truthfulness. It is a logical operator whose form is this schema,
call it (*)
(*) A shift in conceptions of what it is to tell the truth about X.
That sounds as if X is a given, a timeless given, X = the past, or, in

the case of the emergence of authenticity, X = the self. No. New
ways to tell the truth about X change our conceptions of X itself.
Williams talks of a shift from "a 'local' to an 'objective' view of
the past." (p. 163). Here are some more snippets about the new
idea of history:
1) Historical time provides a rigid and determinate structure for
the past. (po162)
2) this significant change that took place in the fifth century B.
000
C., the invention in the West of historical time.

3) Did the change bring with it an increase in explanatory power?
Surely, yes; and this was so in terms of anyone's conception of
explanation. (p. 170)
4) Does that mean that those who operate in the new style, who
have the "objective" conception of time, are more rational or
again better informed than the others? No, if that means (as it
is usually meant to imply) that those in the traditional practice
were confused or believed something else. (p. 170)

5) The invention of historical time was an intellectual advance,
but not every intellectual advance consists of refuting error or
uncovering confusion. (p. 171)
I won't go on, you will see the point. We could add to Williams'
remarks that there can be completely different conceptions of the
past, and not merely ones that evolve in our Western notions.
The anthropologist Michael Lambek has provided a splendid
example from Madagascar, in which a people there express their
sense of their past by ceremonies involving complex role playing
by spirits.40 This is so alien that the staunchly rationalist among
us will hold that we "are more rational or again better informed
than" these "others." I cannot agree, although if their chronology
completely differed from the Western one, I might have to say
that they were not talking about the past (as we understand it).
Lambek's genius is to give us an understanding of the rationality
of this other conception of telling the truth about the past.
There was a new conception of the past, and correspondingly
new kinds of statements to make about it. There is the demand
that every event should happen before, after, or overlap with,
every other event. Williams claims that earlier writers, including
Herodotus, were not constrained by such an idea. But obviously
he does not suggest the silly idea that events themselves got into
a dateable ordering only in the time of Thucydides!
The schema extracted from Williams is altogether formal and
can be applied across the board. To repeat,
(*) a shift in conceptions of what it is to tell the truth about X.
Snippet (2) above suggests a 2nd schema:

(**) This significant change took place in the Y century, and its
40 Michael Lambek, The Weight of Past: Living with History in Matanga,

Madagascar, London: Palgrave, 2002.
emblemis Z.
When X= the past, Y = the fifth century BCE, and Z = Thucydides.

Using a terminology suggested above, Thucydides is a
legendary trailblazer, an icon around whom crystallized new ways
of telling the truth about the past. Here we have a sharp
discontinuity, embedded within a longer and more continuous
practice of talking about collective knowledge of times gone by.
15. Mathematics
Williams's schema can be applied to the sciences. He might
not have welcomed this. His Ethics and the Limits of Philosophy
kept the sciences distinct from humane questions about values.41
Let us see, however, how it can be done. Start with geometrical
relations. Who shall stand in for Thucydides? Here is the legend:
"A new light flashed upon the mind of the first man (be he Thales
or some other) who demonstrated the properties of the isosceles
triangle"—thus Kant.42 In that paragraph Kant waxes poetical
about entering "upon the sure path of science," "that royal road."
Kant calls this "the intellectual revolution" which was the
coming into being of mathematical demonstration. It was the
discovery of our capacity for mathematical proof, or rather proof
in geometry.
The iconic trailblazer of the crystallization of geometrical
proof is Thales, which is not to say that there was for sure any
such historical figure. The legend is that when X = geometrical
relations, Y = early in the sixth century BCE, and Z = Thales.
We need two things to understand styles of scientific thinking:
on the one hand, the study of mental capacities, and on the other,
the history of civilizations and of their institutions. What cultural
41 Harvard University Press, 1985.

42 Critique of Pure Reason, B. xi.
elements were needed to sustain a discovery about our cognitive

capacities? In answer, ask why proof loomed so large in ancient
Greece. There was lots of mathematics in Babylonia, and
certainly in China. Chinese had splendid computational devices,
but no proofs such as matured in Euclid. Why? The answer
favoured by Reviel Netz starts with the familiar fact that
Athenians were the most argumentative bunch of people ever
known.43 They countenanced no higher authority than themselves,
when it came to settling an argument. On looking back at Crombie,
one finds a remarkable, similar account of the importance of
argument, even though he did not fix on proof and deduction as
I will do in the second lecture.
China provides a contrast. To parody, an issue could be
settled by edict, so compelling proof— Wittgenstein's "hardness
of the logical must"— had no special interest. But in Athens,
proofs seemed to have the strange power of establishing truths, of
themselves, for those who could study them. That mattered,
because of the practices of settling arguments current in Greece.
It did not carry the same weight in China.
There is a lot of evidence that spatial, geometrical, thinking
involves cognitive capacities different from arithmetical,
combinatorial and algorithmic reasoning. Let us introduce
another legend for the algorithmic or combinatorial style of
thinking, the legend of Al-Khwarizmi (about 780-850) working in
the House of Wisdom at Baghdad, after whom the very word
"algorithm" is named. The title of one of his books gave us the
word "algebra." So now we have a new X, Y, and Z: a shift in
conceptions of what it is to tell the truth about numbers and
other quantities; this significant change took place in the ninth
43 Reviel Netz, The Shaping of Deduction in Greek Mathematics: A Cognitive

History, Cambridge University Press, 1999. Geoffrey Lloyd also supports this
account in Ancient Worlds.
century and its emblem was AI-Khwarizmi. This is a different

crystallization from the one that happened in the time of the
legendary (imaginary?) Thales.
Notice that in both cases, geometrical and combinatorial
thinking, the new methods of reasoning, including proofs,
provide a wholly new level of "explanatory power." See (3)
above. And to repeat (4), this does not mean that those who
operated in the new style, who attained the "objective"
conception of space or of computation, were more rational or
again better informed than their predecessors.
Kant may have got it right, in the Transcendental Aesthetic,
when he located arithmetic and geometry in separate
compartments of the mind. He linked one to the experience of
the succession of discrete unities in time, the other to the
experience of spatial configurations. However we account for
them, it is quite plausible to invoke the idea of "mental modules"
here—fundamentally distinct capacities inherent in the human
brain. Of course there is a big debate about modularity among
cognitive scientists, ranging from Dan Sperber's "modularity all
the way down,"44 to that of Jerry Fodor, who pioneered the
transfer of modularity from Chomsky on grammar to a wider
range of applications.45 "Modularity gone mad," writes Fodor of
Sperber. 46 It may nevertheless appear that module—talk in
moderation is a good parable, while we are still finding out about
44 Dan Sperber, "In Defense of Massive Modularity", in Dupoux, E., Language,

Brain and Cognitive Development: Essays in Honor of Jacques Mehler, Cambridge,
Mass: MIT Press, 2002, pp. 47-57, on p. 48.
45 Jerry Fodor, The Modularity of Mind: An Essay on Faculty Psychology,
Cambridge, Mass: MIT Press, 1983.
46 Jerry Fodor, "Modules, Frames, Fridgeons, Sleeping Dogs, and the Music of the
Spheres." In J. Garfield (Ed.), Modularity in Knowledge Representation and
Natural-language Understanding, Cambridge, Mass: MIT Press, 1987, pp. 26-36,
on p. 27.
the brain. As Bernard Williams says about modules somewhere

in Truth and Truthfulness, "Why not?" Why not, especially if there
has been a long-standing awareness of what appear to be distinct
faculties, long preceding Kant.
Advocates of parallel distributed processing will say that
there are very good answers to the question "why not?" Thus
in the case of taxonomic reasoning there is an entrenched
tradition that favours the idea of innate classificatory modules,
especially when it comes to classifying living things. Thus the
picture for the taxonomic style of thinking is that it is
back-grounded by modules. But alternative models of cognition
are very much to hand. A recent paper illustrates the claim that
all the phenomena evidenced for innate classificatory modules
are better accounted by a general processor of the parallel
distributed type.47
But suppose the research programme of modules does part
out. No single module will correspond to exactly one style of
thinking. Each demands many, and modules of different types.
What will soon be called the laboratory style requires a
combination of hand and eye coordination, of deductive skills,
and a lot else.
Recall the paired virtues associated with truthfulness. There
is a wholly new standard of accuracy about geometrical relations,
to the extent that "accuracy" no longer seems quite the right
word. On the other hand sincerity seems to drop out because
proof becomes the sole criterion of correctness in this new
domain. Or from a social point of view, what matters is not that
the geometer is sincere, but that initiates see that the argument of
the geometer is indeed a proof. We reject the Kantian
genius-in-history lote that "a new light flashed upon the mind of
47 Timothy T. Rogers and James L. McClelland, "Semantic Cognition: A Parallel

Distributed Processing Approach," Behavioral and Brain Sciences, 2007.
40I SCIENTIFIC REASON
the first man who demonstrated ... " As if one man did the trick! It
takes a community for there to be a practice of proving. But there
is no other standard, of what is a correct proof, than proof itself.
Proofs are self-authenticating. Many a proposed proof has
proven to be fallacious, but the standard of validity or fallacy is
proof itself.
In the case of geometry, we have a new way of telling the
truth, and of ensuring that it is the truth-namely proof. In
quotation (4), Williams spoke of "operating in a new style." It is
probably fortuitous, but it is handy that Crombie and Williams
used that same mischievous word, "style".
16. Caution about "X"

Williams's two examples have a plausible topic concerning
which there is a change in conceptions: the past, or the self. In my
first extension to the sciences, once again there was not difficulty
in producing a shorthand label for the "X" about which there
was a change in conception: geometrical objects, or geometrical
relations—or just plain geometry, if chosen. A change in
conception could be so radical that after the change in conception,
we do not have a ready tag, referring indiscriminately to the past
and to the future, to indicate the change. I follow a host of
historians, scientists and philosophers who single out Galileo as
marking a radical change going on in his time. God wrote the
Book of the World in mathematics, and it was the task of natural
philosophy to decode that book. It seems clear that a pretty
radical shift in conception occurred early in the seventeenth
century. But a shift in the conception of what? One thing that
happened in the time of, and partly through the work of, the
historical Galileo was a shift in the conception of motion. But
something more radical was crystallizing, namely a shift in the
conception of what it is to tell the truth about Nature herself. The

truth is to be told in the language of mathematics, statements that
are to be controlled (to use Crombie's word) by observation. And
not only observation. Alexandre Koyré and by derivation
Crombie favoured a platonistic Galileo who did not make
experiments. Telescopes yes, to observe the heavens, but
experiments, no. It took a devoted amateur and book-collector to
establish that Galileo really did build an enormous inclined plane,
in order to check his wonderfully simple mathematical theory of
falling bodies. Drake's critics replied that no times of the fall from
start to finish on the enormous plane could be measured with
such accuracy in those days. Think again, said Drake: Plainsong,
the Gregorian chant, provides a remarkably accurate
48
chronometer.
17. The Galilean style

In the case of the crystallization of each style of finding out,
there is a "radical shift in conceptions of what it is to tell the truth
about something." Husserl can be read as making exactly that
claim about the mathematization of the world that he attributed
to Galileo. Later writers, whose expertise is as seemingly
different as the cosmologist Steven Weinberg and the
grammarian Noam Chomsky, have attributed talk of the
"Galilean style" to Husserl. The fields of expertise of these two
famous men are not so different as may at first appear, for
cosmology and grammar are sciences where experiment (in any
48 Stillman Drake, "Galileo's Discovery of the Law of Free Fall". Scientific

American 228, #5 (1973), pp. 84-92. Galileo At Work. Chicago: University of
Chicago Press, 1978. Drake presents Galileo as an inductivist, discovering the law of
free fall on the basis of how a ball fell in his apparatus. I cannot support that reading.
I suggest however, that he used more than mere observation to control his
mathematics just by observation: he used what happened with his laboratory
instrument. A classic case, perhaps, of the hypothetico-deductive method.
strict understanding of the word) is not possible: one must make

theoretical models and test them against observation. The
Galilean style is the crystallization of what Crombie called the
style of hypothetical modelling, and of course its iconic
trailblazer is Galileo. Crombie too gave Galileo pride of place in
the history of that style, though not as originator, because he
rightly saw a long chain of precursors who used "the method of
hypotheses" and the making of models, both intellectual and
analogical.
I shall follow the maxim of not enlarging the catalogue of
styles of scientific thinking. I shall not say that what Husserl and
then Weinberg and then Chomsky called the Galilean style is a
new style of scientific thinking. Better to say that it is a definitive
crystallization of what Crombie called (c) the hypothetical
construction of analogical models. I shall in the future often
speak of the Galilean style, because that mode of reference
attends to something very specific about (c), the hypothetical
construction of models.
Galileo then serves as trailblazer. Notice how easy it is to pass
the buck. Neither Husserl (for Galileo) nor Kant (for Thales) is
used as an authority. My rhetorical trick is rather: look here—we
have been saying this all along forever and anon, only we have
not noticed that we were saying it.
18. The laboratory style

Then there is Crombie's experimental style: (b) the experimental
exploration and measurement of more complex observable
relations. This probably has no strict beginning. Human beings
have been curious, looking, tinkering, exploring, even measuring,
forever. A new kind of truth-telling did begin when a community,
call it that of Robert Boyle, not merely studied hypothetical
models, controlled by observation and measurement, but also

created instruments and apparatus for interfering with the
course of nature in order to extract its deepest secrets. Here X =
(let us say) the minute unobservable parts of material
nature—what in quite recent times philosophers came to call
theoretical entities. Laboratory science involves a distinctive
practice of truthfulness that begins with creating devices that
work; one can even begin to speak of truthful apparatus. In
mechanical matters we more commonly, and rightly, call the
instruments reliable, an oddly mechanical mixture of Williams'
virtues of sincerity and accuracy.
The classic account of this event is to be found in Steven
Shapin and Simon Schaffer's Leviathan and the Air Pump: Hobbes,
Boyle and the Experimental Life.49 One of the many wonderful
features of that book is that the main protagonist is not a person
but an instrument, the air pump. The other players, Boyle and
Hobbes, are perfectly cast. The book includes a translation of a
disquisition by Hobbes, published in 1662. Hobbes was aware
that Boyle was shifting the conception of what it is to tell the
truth about matter (and the void), by using instruments to create
new phenomena, and making our statements answer to the
phenomena produced by apparatus. I regard this as a, or perhaps
the, fundamental crystallization of (b) the experimental style. In
the future, especially in the third lecture, when I speak of the
laboratory style of thinking and doing, I am not enlarging
Crombie's list of six styles, but attending to a particular
crystallization of Crombie's style (b).
What exactly was this crystallization? Here Shapin and
Schaffer are invaluable in presenting the story of Hobbes vainly
fighting the new laboratory of Robert Boyle. There are quite
49 Princeton University Press, 1986.

enough public phenomena in the world already, Hobbes says

huffily. We do not need more, that are produced in a closet, and
checked only by the members of a secret society (the Royal
Society of London). Hobbes knew that a new style of thinking
was in the making, and hated it. His primary objection was,
perhaps, that whereas geometry was open to all who worked at it,
and had a distinctive certainty, the new physics was a matter for
initiates working in private, and produced merely probable
knowledge, opinio, not scientia. But he also says very clearly that
God has given us enough phenomena, and we do not need you
to give us more dubious ones, Mr Boyle.
One need not choose Boyle as the icon of the crystallization of
experiment into the laboratory style. The same new criteria of
truth checked by artificial phenomena were emerging all over
Europe. Boyle serves well because a great and articulate old man,
Hobbes, saw the writing on the wall, and spat at it.
Schaffer and Shapin, and Shapin in his book, A Social History
of Truth50 urge that what made the new laboratory science
possible was the custom of trust among the gentry who formed
the Royal Society. A gentleman could be trusted to report truly
the phenomena that had been produced and witnessed. Whether
Shapin is historically correct or not, he nicely draws attention to
the (defeasible) assumption that whoever writes up an
experimental inquiry is telling the truth.
Our practice of peer review has absorbed much of the
assumption of sincerity right across the board, in all the sciences,
including mathematics. Now we ask peers to review, whereas
once we relied on peers, that is, English aristocrats and landed
gentlemen.
As you find other "shifts in the conception in what it is to tell
50 A Social History of Truth: Civility and Science in Seventeenth century England,

University of Chicago Press, 1994.
the truth about X," at rather sharp points in time, you quickly hit
upon a figure of almost legendary proportions— Thucydides.
Rousseau/Diderot, Thales, Galileo, Boyle. For X = the species and
the higher taxa of systematics, legend congeals around Linnaeus.
For X = probabilities it is Pascal.
The doctrine of the hero in history has been totally debunked
by every historian of science in living memory. But there is still
truth in folklore. The fact that we have a legend, and a handy
emblematic trailblazer for a crystallization of each style of
scientific thinking, shows how tradition already recognizes that
each crystallized style is a radical innovation that can, for ease of
story telling, be associated with a giant. Each giant, to repeat, is
only a figurehead. I do not need reminding of this. An essential
part of the argument of The Emergence of Probability was precisely
that a new conception of probability seemed to spring up, almost
spontaneously, after 1650, in most regions of Europe. Pascal is
both the traditional emblem of this fact, and at the same time
merely a bit player.
19. Logic
Logic has come to mean deductive logic, perhaps as
formalized by first-order logic. There is an older tradition, in
which logic names the canons of reasoning. Peirce divided logic
into three parts, deduction, induction and abduction, or what in
the nineteenth century was called the method of hypothesis.
Peirce liked triads, and he was well aware how this triadic logic
itself fell under the mediaeval triad, the trivium, the bare
essentials of an education, namely grammar, logic, and rhetoric.
Deduction, induction, and abduction are not styles of
scientific thinking, in the sense of the present lectures. They are
surely founded on human cognitive capacities, although their
expression varies from language to language, and from culture to

culture. Peirce could well have added classification as a fourth
branch of logic, equally founded on some cognitive capacities.
All four are human universals, not simply because they are
accessible to all, but because they are the background within
which human discourse can take place. They have no beginning.
They are part of our animal nature. Popper liked to say that
amoeba make inductions: so be it, if you will.
None of the four branches of logic is among the "methods of
reasoning" which characterize styles of scientific thinking. They
are rather presupposed, in varying degrees, by every style.
Evidently the geometrical and algorithmic styles are more closely
linked to deductive logic than are other styles. The Galilean style,
Crombie's style of hypothetical modelling, must be an outgrowth
of abduction. The probabilistic style of thinking builds upon our
inductive capacities. The taxonomic style builds on naive
classification. What Bourdieu called "the singular history of
reason," namely cultural discoveries that take place in highly
specific and local situations, take for granted a background of
what Peirce called logic.
20. Three radical propositions

This is not a philosophy to expound briefly. The work must
be done in looking at specific styles of scientific thinking. In these
lectures I shall examine only styles (a) and (b) in some detail, the
mathematical style in the second lecture, and the experimental
style, crystallized as the laboratory style, in the third. But certain
aspects of the project can now be highlighted. Here are some of
the philosophical implications of this approach.
I. Styles are self-authenticating

Each way of finding out introduces its own criteria of
evidence, proof and demonstration. Each determines the criteria

of truth-telling that applies in its own domain. That leads to a
somewhat radical thesis about truthfulness and objectivity.
Scientific styles are in a certain sense self-authenticating. For
each style there is a class of sentences that are candidates for
truth or for falsehood only in the context of the style. The only
way to find out whether they are true or false is by using the
relevant style. The criteria of truthfulness are determined by the
style. All individual propositions are fallible. In reasoning
according to a style, one can always make mistakes. But it is in
the framework of the given style that one establishes that an error
has been made.
A style of scientific thinking could wither, fail, or become
extinct. In that case we are disinclined to call it scientific any
more. The disappearance of a style is always caused by external
forces. On the present theory, a style of thinking begins with the
discovery of how innate capacities can be used in new ways to
find out about something. That is what has traditionally been
called part of the "internal" history of science. But what
maintains some way of finding out must be its use in a cultural
context. That is part of an "external" history. The
self-authentication is internal, the perseverance, external.
II. Ontological debates

Each new style in the sciences introduces a new class of
objects to study, in effect, a new X in Williams's schema (*). But
matters do not rest there. Each new class of objects invites an
ontological debate, often described as realism versus some kind
of anti-realism. These debates are mere by-products of the styles
of thinking.
Take the controversies about mathematics between so-called
platonists and constructivists. Or take the confrontation between
those who say that unobservable theoretical entities are real, and
those, from Auguste Comte to Bas van Fraassen, who deny it. In
systematics, some contend that the species are real, but not the
higher taxa. Others insist that genera, classes, orders are real, are
part of the natural order. And so on. Each ontological debate
takes place within its own scientific style. That is because every
style of finding out creates its own objects. We are on the verge of
a genealogical theory about the nature of the classic ontological
debates in the sciences.
III Cognitive foundations and cultural history

Third, each of these ways of finding out is grounded in
typically human capacities, including cognitive ones and
physiological ones. Undoubtedly these capacities are the product
of evolution by natural selection. They are universal.
But scientific styles are themselves the product of cultural
innovation and evolution. Much of this has happened in the
Mediterranean regions—North Africa, West Asia and Greece—
and later in Europe. Each has a beginning in history, which
sometimes exists chiefly in the form of legend, and each has its
own trajectory of development.
Cognitive science and neuroscience are proper attempts to
understand the capacities underlying this history, and help
explain how what evolved at certain specific moments in
unimportant parts of the world, has spread to become part of the
universal human heritage.
From a different perspective, we should see a study of styles
of scientific thinking as part of "the natural history of human
beings" (Wittgenstein, Philosophical Investigations, § 415), or as
part of a philosophical anthropology.
Lecture 2
.........
WHERE DO MATHEMATICAL
OBJECTS COME FROM?
1. Philosophical anthropology
I think of these lectures as being a contribution to

anthropology, in a rather old fashioned sense. In English,
"anthropology" has come to mean, primarily, the study of
particular peoples or social groups, often delimited by language
and, or, region. It has become synonymous with ethnography.
But when Kant published his book Anthropologie in pragmatischer
Hinsicht—Anthropology from a pragmatic point of view—he really
thought of it as a study of human nature from, among other
things, a social point of view.51 It includes chapters on how to
have a good dinner party! Michel Foucault took this book as
marking a decisive break in European thought. Man, he said,
began to reflect on what it is to be Man. Hence a title of the
chapter in one of his books, "Man and his Doubles", man as both
subject and object of inquiry. 52 This way of thinking of our
51 German original, 1796. Translated and edited by R. Louden, Anthropology from a

Pragmatic Point of View, Cambridge University Press, 2006.
52 Michel Foucault, The Order of Things, chapter 8.
50 I SCIENTIFIC REASON
species, he claimed, became possible only towards the end of the

18th century in Europe, and in a way different from earlier
epochs and other places. It was not (Foucault claimed) what
Hume was doing in his great Treatise on Human Nature published
some 80 years earlier.
I have no interest, here, in defending claims about the date or
character of this alleged radical break. To continue for a moment
using words no longer in fashion, I rather like Francis Galton's
pair, "nature" and "nurture" as dividing up the two fundamental
aspects of anthropology. Galton himself wryly said that his
couplet is just a "convenient jingle of words."53 What he meant
by nurture includes more than just mother's breast and knee; it is
on the streets, it is television too. Galton was mostly concerned
with the nature and nurture of individuals, a topic for the various
disciplines loosely grouped under psychology. When we tum to
human beings in groups, some would prefer to speak of nature
and culture. To paraphrase Galton, nature is what people are
born with, what they bring with themselves into the world, while
culture is everything that affects them as persons in society after
their conception and birth. Obviously it is a feeble distinction,
but it is a handy pointer. It is a jolly jingle but also a useful one if
treated as a jingle rather than as a known and well-understood
fact.
Because of the subsequent uses of the name, anthropology, let
us call one kind of successor to Kant's enterprise philosophical
anthropology. I thus call these lectures a contribution to
philosophical anthropology, because they reflect on some aspects
of human nature as they have been discovered and nurtured by
groups of people and encoded in societies, even in civilizations. I
mean in particular, those capacities that are innate to human
53 Francis Galton, English Men of Science: Their Nature and Nurture, London:
Macmillan, 1874, p. 12.
WHERE DO MATHEMATICAL OBJECTS COME FROM / 51
beings and apparently only to human beings, and which have

been cultivated at quite specific times and places into the range
of talents, activities, techniques and bodies of knowledge that we
call the sciences.
2. Mathematics as a motley
I resolutely use the plural, "sciences" when I talk about their
philosophy, rather than talking about the philosophy of science.54
This practice is customary in French, and was common in 19th
century English, for example when William Whewell wrote three
volumes of History of the Inductive Sciences (1837) and two
volumes of Philosophy of the Inductive Sciences (1840). It is never to
be taken for granted that there is one definite thing, science, or
even one definite thing, technoscience. There is not even one
definite thing, mathematics; there is rather what Wittgenstein
called the motley of mathematics.55 Think how different is the
arithmetic that all of us learned when we were children, and the
proof of Pythagoras' theorem which many of us learned as
adolescents, or, what is tantamount to that theorem, the proof in
Plato's dialogue, the Meno, of how to construct a square double
the size of a given square. Think then of the mathematical idea
that Fermat had, when he wrote down what came to be called his
last theorem, and think of the proof ideas that lie behind Andrew
Wiles's discovery of a way to prove Fermat's last theorem.
Think next of the mathematics used in modelling the
behaviour of atoms. I recently became fascinated by very cold
54 As explained in my piece with a doubly plural title, "The Disunities of the

Sciences," in The Disunity of Science. Boundaries, Contexts and Power, edited by P.
Galison and D.J Stump, Stanford University Press, 1996,37-74.
55 Ludwig Wittgenstein, Remarks on the Foundations of Mathematics, edited by G.
H. von Wright, R. Rhees and G. E. M. Anscombe, translated by G. E. M. Anscombe,
3rd and revised ed., Oxford: Blackwell, 1978, III-§48, p. 182.
atoms, basically at absolute zero. I try to read some of the

mathematics that enables us to understand why they behave in
such extraordinary ways. This mathematics is old-fashioned,
taken from a physicist's toolbox, much of which has been around
for a century; the chief difference is that we can now use very
powerful computers to make approximate solutions to complex
equations that cannot be solved exactly, as well as to construct
simulations that enable us to establish intimate relations between
theory and experiment.
Think next of economists, also making models, who are as
incapable of understanding the reasoning of the physicist as most
physicists are of making sense of modem econometrics. Then ask,
why on earth do we call all of these "mathematics"? I suspect
there are several different types of useful answers to that
question. Some will themselves be mathematical in character,
while others may draw on other resources.
I would be content if any of you, who remember this lecture
five years from now, were to think to yourselves, "All I can
remember is that he made me ask, why are so many different
types of activities, of thinking and doing, all called
mathematics?" I would be even happier if you then thought,
"You know, that is a really serious philosophical question in its
own right!"
I begin with these remarks because I shall start where
philosophers, at least since Plato, have always started, with a
particular aspect of mathematics, namely geometrical theorems
established by short, perspicuous proofs. I want to warn in
advance that this hallowed starting point is on the one hand
essential to my philosophical anthropology, but is on the other
hand only one little bit of the amazing motley that is
mathematics.
3. Nature
We have good reason to think that the styles of scientific
thinking, which led me to speak of "scientific reason", are
grounded on "built-in" cognitive capacities. These are part of our
collective human inheritance. There is nothing relative or
historical or culture-dependent about them. (Or so I assume
without detailed argument.) Doubtless there have been
evolutionary pressures to favour these potentials, but do not take
for granted that they have any adaptive value for natural
selection. The ways in which they have come into being may well
be not for their immediate usefulness, but only as by-products of
genuinely functional adaptations in the development of humans
and animals. That is to suggest that the various cognitive
capacities essential to scientific reason may have emerged as
what Stephen Jay Gould and Richard Lewontin called
spandrels.56 Indeed there are scholars who urge some
explanation of the development of mathematical skills in terms
of evolutionary psychology, but I simply do not believe it. I do
not believe that the ability to do what we call mathematics had,
in the first instance, any adaptive value at all. Undoubtedly the
ability to navigate our three dimensional world is essential for
survival, and the ability to tell more from less, and even to count
a bit, would help birds, animals, and then people to get around,
but that takes us almost no way into the world of mathematics.57
56 Stephen Jay Gould and Richard C. Lewontin, "The Spandrels of San Marco and
the Panglossian Paradigm: A Critique of the Panglossian Paradigm," Proceedings of
the Royal Society of London, B, Biological Sciences 205 (1979): 581-98. Stephen
Jay Gould, "The Exaptive Excellence of Spandrels as a Term and Prototype,"
Proceedings of the National Academy of Sciences, 94 (1997): 10750-5. A spandrel is
an architectural feature, often decorative, that developed, as it were unintentionally,
out of a functional feature of the design of the building.
57 Thus despite the interest of the work by the cognitive scientist, Stephen Dahaene,
I cannot agree with the title of his book, The Number Sense: How the Mind Creates
It is nevertheless, a plausible hypothesis that if there is a part of

the brain that does mathematical reasoning, it will have
developed in that part which enabled earlier animals to count.
Chomsky has long been suggesting that, just as there is an
innate species—specific faculty for the acquisition of grammar,
there are also innate human capacities to reason mathematically.
He conjectures, as I understand him, that the mathematical
capacities may be by—products (spandrels?) of linguistic
capacities. Other cognitive scientists hold that there is a quite
specific module of mathematics, that the human brain is, in their
words, "hard-wired" for mathematical, or at any rate arithmetical,
thinking.58
I do want to emphasize that when I speak of the capacity to
reason mathematically, I refer to far more than the ability to
count or to tell a few shapes from each other. There is of course
tremendous regional variety in people around the globe, and lots
of cognitive variability within a single family. The ability to
reason mathematically is surely more variable from person to
person than is the child's ability to pick up the language spoken
around it, but it is also, in this picture, part of the human mind. It
appears that many people cannot do much more than count, do
rote arithmetic, and to recognize a few shapes. Whether this is a
matter of choice or intrinsic limitation I have no idea.
It also seems plausible to imagine that there is a module or
group of modules dedicated primarily to getting around
spatially and to imagining spatial configurations, and which is
connected with our ability to reason geometrically. It seems
plausible then to imagine that there is another module, or group
Mathematics, Oxford University Press, 1987. In my opinion, he teaches us about the

innateness of counting, but not about the innateness of mathematical reasoning.
58 Brian Butterworth, What Counts: How Every Brain is Hardwired for Math. New
York: Free Press, 1999.
of them, dedicated to numerical and combinatorial reasoning,

which in its more advanced stages becomes an algebraic skill. We
can understand Kant as recognizing this fact, even if we no
longer describe the phenomena in his way. He distinguished
arithmetic from geometry, and argued that what he called the
synthetic a priori knowledge in arithmetic derives from a
transcendental apperception of discrete units in time, while those
of geometry arise from a transcendental apperception of space.
Indeed some people seem to be particularly good at combinatorial
reasoning, while others are gifted at geometrical reasoning, as if
one group is better endowed with one group of modules, and the
other with another.
Some researchers claim that females process spatial
information differently from males. That is a story that could fit
in to the narrative of modules. But be cautious: these are
illustrative tales, just—so stories. These furnish ways of looking at
the facts, but they are only pictures. They might even be helpful
pictures, so long as we do not fantasize that we actually know
anything much about such modules or capacities. I find these
illustrative fables about innate abilities useful as way-points to
frame our ignorance. They are in the same league as Kant's way
of presenting his insights.
In this picture, such capacities are part of our universal
human inheritance, within which there is lots of individual
variation. The discovery and cultivation of human capacities is
something else. It happens in specific human histories, in
contingent historical circumstances, and is part of the history of
civilizations. There is, moreover, no reason to think that
capacities, even when discovered, will be exploited in the same
way by different cultures, although there is every reason to
believe that if one society wants to learn the cognitive skills of
another, it can.
4. Folkloric history of a discovery of a human potential

Now let us focus on the discovery of our capacity for
geometrical thinking, starting with geometry. I have already
mentioned Kant, but he had more to say. Being a philosopher, I
have always known the dramatic passage about mathematics
that Kant wrote for the introduction to the 2nd edition of his
Critique of Pure Reason. I have already repeated a few of those
words in the first lecture, and used them to introduce the idea of
a new way of telling the truth about geometrical shapes. Now I
shall quote the paragraph in full. It is good to remember that our
most abstruse philosopher, the sage of Konigsberg, loved a good
story. And let us not forget that the fact that he told the story,
does not imply that it is literally true.
Kant was embarrassingly Eurocentric. I mean that literally, as
well as figuratively. He thinks that the example of a
revolutionary discovery which will be uppermost in the mind of
his readers is the discovery of the sea—route around Africa to
Asia—where Europeans will, for a while, be able to exploit the
civilizations of the East. Kant does write of Egyptians preceding
Greeks, but he knows nothing of ancient Chinese mathematics.
He is writing, remember, in 1787, two years before what Europe
regards as its fundamental revolution, the French one. Hence
revolution, political, geographical, and intellectual, was much in
the air. He writes:
In the earliest times to which the history of human reason extends,
mathematics, among that wonderful people, the Greeks, had
already entered upon the sure path of science. But it must not be
supposed that it was as easy for mathematics as it was for
logic—in which reason has to deal with itself alone—to light upon,
or rather construct for itself, that royal road. On the contrary, I
believe that it long remained, especially among the Egyptians, in
the groping stage, and that the transformations must have been
due to a revolution brought about by the happy thought of a single

man, the experiment which he devised marking out the path upon
which the science must enter, and by following which, secure
progress throughout all time and in endless expansion is infallibly
secured. The history of this intellectual revolution— far more
important than the discovery of the passage round the celebrated
Cape of Good Hope—and of the fortunate author, has not been
preserved. But the fact that Diogenes Laertius, in handing down
an account of these matters, names the reputed author of even the
least important among the geometrical demonstrations, even of
those which, form ordinary consciousness, stand in need of no
such proof, does at least show that the memory of the revolution,
brought about by the first glimpse of this new path, must have
seemed to mathematicians of such outstanding importance as to
cause it to survive the tide of oblivion. A new light flashed upon
the mind of the first man (be he Thales or some other) who
demonstrated the properties of the isosceles triangle. The true
method, so he found, was not to inspect what he discerned in the
figure, or in the bare concept of it, and from this, as it were, to read
off its properties; but to bring out what was necessarily implied in
the concepts that he had himself formed a priori, and had put into
the figure in the construction of by which he presented it to
himself. If he is to know anything with a priori certainty he must
not ascribe to the figure anything save what necessarily follows
from what he has himself set into it in accordance with his
concept.59
To parody Kant, he is imagining the first mathematical light bulb

going on over someone's head. Thales is no more than a legend,
and what is required is not "one man", but rather a community,
at least a small network of collaborators, teachers and pupils,
who grasp the possibility of making a proof, and who are able to
share stunning arguments.
59 Critique of Pure Reason, B x-xii, trans. N. Kemp Smith, London: Macmillan,

1930, p. 19.
Like everyone else nowadays, I totally reject Kant's

individualism, but there is something in it of profound importance.
The cognitive capacity is in the first instance individual, and
human beings had to learn how to make use of that capacity of
individuals. What is missing in Kant's statement is the explicit
acknowledgement that the cultivation of the capacity is
communal. What matters is the nurture of the insight that legend
attributes to Thales, the nurture of the practices of proof.
Kant waxes poetical about entering "upon the sure path of
science," "that royal road." Kant calls this "the intellectual
revolution" which was the coming into being of mathematical
demonstration. I say that it was, first, the discovery of our
capacity for mathematical proof, or rather proof in geometry.
And secondly it was the recognition of the importance of this
discovery, and the development of a community that honoured
proofs, that mattered. But also, there was a shift in conception of
geometrical objects. We acquired a new way of telling the truth
about them— by proving facts about them.
5. Truthfulness about geometrical objects

My first lecture presented a genealogical schema for telling
the truth about something. The crystallization of a style of
thinking can produce:
Following Williams' example of Thucydides, the trailblazer of the

new way of telling the truth about the past, I suggested a second
schema:
(**) The significant change took place in the Y century, and its
emblem is Z.
Evidently the legendary trailblazer of geometry is Thales, an icon

who quite possibly was a historical figure, but perhaps the name
is simply a point about which legends have accrued. What is
striking is the extent to which folklore seems to have recognized
the need for such an accretion of myth. Kant's story is about:
(*) A shift in conceptions of what it is to tell the truth about
geometrical objects.
(**) This significant change took place in the 6th century BC, and its
emblem is Thales.
The new way to tell the truth about mathematical objects was to
prove relations between them, prove them a priori. Now why
should this ability be honoured, institutionalized?
6. Argumentative Greeks
Geoffrey Lloyd drew attention to an aspect of life in the
Greek city-state; Reviel Netz has elaborated on that theme.60
City-states were organized in many ways, but Athens is of
central importance. It was a democracy of citizens, who were all
male and none of whom were slaves. It was a democracy for the
few; not what today we call a democracy at all. But within those
few, there was no ruler. Argument ruled.
Athenians were the most consistently argumentative bunch
of self-governors of whom we have any knowledge. We read
Aristotle for his logic and not for his rhetoric. Greeks read him
for his rhetoric; his logic was strictly for the Academy. The
trouble with arguments about how to administer the city and
fight its battles is that no arguments are decisive. Or they are
decisive only thanks to the skill of the orator, or the cupidity of
the audience. But there was one kind of argument to which
60 Reviel Netz, The Shaping of Deduction in Greek Mathematics: A Study in

Cognitive History. Cambridge University Press, 1999.
oratory seemed irrelevant. Any citizen, and indeed any slave

who was encouraged to take the time, and to think under critical
guidance, could follow an argument in geometry. He could come
to see for himself, perhaps with a little instruction, that an
argument was sound. He could even create the argument, find it
out for himself. In geometry, arguments speak for themselves to
the inquisitive mind.
Cynics will say that this is a lie from the start. A "little"
instruction? The instruction is just a kind of rhetoric, a kind of
oratory. Look at the classic, the demonstration to be found in the
Meno, of how to double a square. The slave boy is said to
discover the technique by himself, unaided. Indeed Socrates is so
taken by knowledge not acquired by observation and experience,
that he makes the fantastic proposal that the boy must have
known already, in some previous existence. The proof of a
theorem is evidence for an immortal soul?!! That is what I mean
by a philosopher being bowled over by the experience of
understanding or making a proof.
Yes, the slave boy is coached by leading questions from
Socrates. But then there is this extraordinary phenomenon,
accessible to almost any thoughtful reader of the Meno, of seeing
that the square on the diagonal is twice that of the given square.
And a diagram is at least psychologically as well as historically
essential to this perception. But in company with the diagram,
and talk about the diagram, there is a new kind of experience, of
conviction based solely on the perception of a new truth.
So what? Lloyd remarks that this phenomenon is truly
impressive to members of an argumentative society that has no
recourse to a ruler, and whose final criterion is nothing more
than talk and persuasion. In this respect, Ancient Greek society
was perhaps unique among the civilizations of which we have
any knowledge. I shall return to the theme of comparative
studies, and make comparisons to mathematics in ancient China.

First, however, let us look at the impact of the discovery of proof
of European philosophy.
7. Mathematics has mattered to some and only some

philosophers.
A deep commitment, to reflection on the nature of
mathematics is found in some, but not all, Western philosophy.
For me, as a Westerner, the obsession with mathematics has
always felt "natural". I have taken it for granted. Indeed when I
was young I thought that it was the very mark of the profound
philosophers!
The implication runs only one way. The profound (I
imagined) were obsessed by mathematics, but many who were
obsessed by mathematics are not profound. Indeed a great many
philosophers of mathematics are boring, arid and sterile,
pursuing degenerating research programmes. Nevertheless, I
thought, with the enthusiasms of youth, that every really deep
philosopher is moved by the phenomena of mathematics, by the
experience of doing mathematics. Needless to say, I was wrong,
but it is worth thinking about why, and how, mathematics has
mattered to some but only some philosophers in the European
tradition.
I made his the topic of a lecture in 1998, under the title,
"What mathematics has done to some and only some
philosophers."61 One thing it has done is to make so many
philosophers ask where mathematical objects come from—the
title question of the present lecture. Here are some of the famous
names I mentioned in that lecture nine years ago: Plato,
61 A Dawes Hicks Lecture to the British Academy, published in Mathematics and

Necessity, edited by T.J. Smiley, London: British Academy, 2000, pp. 83-138.
Descartes, Leibniz, Kant, Russell, and Wittgenstein, to which I

now add Husserl.62
8. The obsession with mathematics

According to Bertrand Russell, 'the question which Kant put
at the beginning of his philosophy, namely "How is pure
mathematics possible?' is an interesting and difficult one, to
which every philosophy which is not purely sceptical must find
an answer."63
Russell exaggerated. Many philosophies that are not purely
sceptical have had no interest in Kant's question. It never even
occurred to them, much less struck them as important. I shall not
be invidious, but we quickly think of canonical Western
philosophers of all periods who have not troubled themselves
with mathematics at all. Hence "some and only some". But the
spirit of Russell's remark is right. A great many of the
philosophers whom we still read have been deeply impressed by
mathematics, and have gone so far as to tailor much of their
philosophy to their vision of mathematical knowledge,
mathematical reality, mathematical objects, or, what I think is
crucial, mathematical proof.
Why do so many philosophers try to answer Kant's question?
And, incidentally, why do a great many not address it? I think
that the answer will have something to do with the experience of
doing some kinds of mathematics. What is there about the
immediate feel of this or that piece of mathematics that has
62 I omitted Husserl because the lecture focused on issues central to analytic

philosophy. I have partially made up for the omission by a paper, "Husserl on the
Origins of Geometry", to be published in a volume on Husserl, edited by David
Hyder, and published by Stanford University Press.
63 Bertrand Russell, The Problems of Philosophy, London: The Home University
Library, 1946, 84. Kant's question is stated in The Critique of Pure Reason,
translated by Norman Kemp Smith, London: Macmillan, 1929, 56 (B 20).
fascinated this or that philosopher? Recall the motley. It is only

some mathematical reasoning, I suggest, that invites the idea of
mathematical objects and of a priori truths. It is chiefly the
experience of proof and of proof ideas that moves the
philosophers to take mathematics seriously.
I speak of this or that 'piece of mathematics', because we ought
to look at mathematics in action—proofs more than theorems—vital
understanding rather than quiescent truths-discovery as much as
knowledge—and for the creative mathematician, we need to look
at proof-ideas that expand into new domains, more than mere
proofs that settle a question. Experiences connected with live
mathematics have driven the philosophers who have built
cornerstones out of it. This is true not only of a Descartes or a
Leibniz, mathematician-philosophers, but also of astonished
onlookers, a Plato or a Wittgenstein.
What struck the philosophers? Bertrand Russell told us what
troubled him in 1912. "The apparent power of anticipating facts
about things of which we have no experience is certainly
surprising."64 Our name for the phenomenon that surprised
Russell is "a priori knowledge." That, and words like "necessary
truth"—or Wittgenstein's "the hardness of the logical
must"—express not so much a property of mathematics as an
experience some of us have in proving or understanding the
proofs of some propositions. These fancy philosophical notions
are derived from the experience of proof.
Some philosophers have drawn quite extraordinary
inferences from the possibility of mathematics. Socrates and the
immortality of the soul speaks for such wild inference. We
should reflect in an immediate and almost childlike way on the
elementary phenomena that fascinated them, in order to grasp
64 Russell, ibid., 85.

that question, "How is pure mathematics possible?" Too often we

are pleased to fly off into subtlety or technicality without asking
what worries us.
Observe that there are two groups of philosophers who have
been deeply perturbed by some aspect of mathematics. Some are
inflationary, some deflationary. Two of the inflationary
philosophers, Plato and Leibniz, draw remarkable conclusions
about, well, everything, from their experience of mathematics.
Inflation leads philosophers to make grotesque claims.
Deflation, on the other hand, leads philosophers to say things
about mathematics itself that are widely regarded as absurd.
John Stuart Mill saw himself as doing mighty battle against
inflation. He argued that there is no difference in principle or
practice between mathematical truths and other empirical
generalizations. Many mathematicians find that absurd, although
in modern times it has had his defenders.65
Descartes, is a very complex figure. He appears to be a
hyperbolic deflationist about mathematics. He apparently held
that God could make two plus two equal to five. Most
philosophers find that so incomprehensible that they suppose
that he cannot literally have meant just that. (I think he did: I take
philosophers blown away by mathematics very seriously indeed.)
Yet he attached hyperbolic importance to perspicuous proof: you
had to take a proof in all at once, to grasp it with the mind.
Wittgenstein is just as complex, and perhaps more
misunderstood than Descartes. One thinks of some of the more
curious of his Remarks on the Foundations of Mathematics, which all
too many people read as saying that mathematical truths are not
fixed or made determinate by anything other than social practice.
Russell was too distinguished a nominalist to take direct
65 Philip Kitcher, The Nature of Mathematical Knowledge, Oxford: Oxford

University Press, 1983.
WHERE DO MATHEMATICAL OBJECTS COME FROM /65
questions about mathematical objects seriously. The most

innovative part of his life as a philosopher was dedicated to an
indirect approach, by showing that they are "logical
constructions". Hence his demonstrably unsuccessful deflationary
attempt to prove that mathematical objects exist only in the sense
of being logical constructions: the amazing project of Principia
Mathematica. One of the virtues of Russell's dedication to
"logicism" was that it cast the doctrine into a form precise
enough for Kurt Gödel to refute it definitively.66
Husserl was inflationary in quite a different way from Plato.
He thought that the 1936 crisis in European civilization, of which
the Nazi regime was only the most gross manifestation, could be
overcome if we made direct contact with the original experience,
the Ursprung, of geometry.67 (Another part of his project was the
attempt to restore contact with the Galilean style of thinking, one
of my topics in the third lecture in this series, "The Laboratory
Style of Thinking and Doing.")
I speak ironically of these great men because they are my
personal heroes. Russell was the hero of my intellectual boyhood.
In my intellectual adolescence, Wittgenstein took an iron grip of
me that has never been relinquished. I mentioned in the first
lecture that I call my attitude to reason Leibnizian; Leibniz was a
youthfullove affair about which I still fantasize. Only in maturity
did I understand that Descartes was the deeper thinker. In
doddering old age I am sometimes fascinated by Husserl. Above
all, these are the philosophers whom I take seriously, to say
nothing of the gods, Plato, Aristotle, and Kant.
66 My own ideas about the essence of the logicist programme, and its motivation,
are found in "What Is Logic?" Journal of Philosophy, 86 (1 979): 285-319
67 Der Krisis der europäischen Wissenschaften und die transzendentale
Phänomenologie, completed 1936, and translated as The Crisis of European Sciences
and Transcendental Phenomenology; An Introduction to Phenomenological
Philosophy, Evanston, Illinois: Northwestern University Press, 1970.
9. Tho visions of a remarkable book on Greek

mathematics
Some but only some Western philosophers have been so
deeply impressed by mathematics that it has left a permanent
stamp on some or all of their philosophy. I do not want to single
out those philosophers who are completely unmoved by
mathematics, who philosophize as if mathematics were of no
importance to any philosophical reflection worth the name. They
are far more than deflationary: they are dismissive of the
philosophical pretensions of mathematics, and do not think them
worth deflating. I admire some of those philosophers, but not so
many, and I shall not speak of the dead ones who cannot tum
upon me.
I have the good fortune to be able to single out a living one, a
friend, and I come to praise him, not so much to disagree with
him as to face in a direction opposite to his. I refer to Bruno
Latour, a founding contributor to modem science studies. He
would prefer, I think, not to be called a philosopher, in English,
although he is certainly a philosophe, as the word is used in
contemporary French. He wrote the first notable ethnography of
the laboratory, Laboratory Life, co-authored in 1979 with Steve
Woolgar. Perhaps he would not disdain the title of
philosophically-minded anthropologist, much as I would be
proud of the label, anthropologically-minded philosopher.
He has recently written a brilliant critical study68 of a book I
have already mentioned: The Shaping of Deduction in Greek
Mathematics: A Study in Cognitive History. The author, Reviel Netz,
is a fairly young historian of ancient mathematics from Israel,
trained at Cambridge University, and now working at Stanford.
68 Bruno Latour, "The Netz-Works of Greek Deductions," Social Studies of Science,

38 (2008): 441 - 459.
The book was published in 1999. Why is Latour reviewing

the book eight years later? The excuse is that he is reviewing the
paperback edition of the book. The fact is that he was introduced
to the book only recently, and was enormously impressed.69 The
opening sentence reads,
This is, without contest, the most important book of science studies
to appear since Shapin and Schaffer's Leviathan and the Air Pump.
He is referring to the book subtitled Hobbes, Boyle and the

Experimental Life, published in 1986.70 It will assume a central
role in my third lecture, on the laboratory style of thinking and
doing.
Now here is a remarkable fact: I completely agree with
Latour's statement. If we exclude Latour's own work from the
competition, then I say "without contest" that these two books,
one about deduction, and one about experiment, are the two
most important contributions to science studies, period. But I say
so for reasons exactly opposite to Latour's!
There is a common point of agreement. Like all readers of
Netz, both Latour and I are astonished by his tour de force in
reconstructing the missing diagrams from surviving Greek texts,
and his demonstration that they are integral to the textbooks of
Euclid and the creativity of Archimedes alike. But the two of us
focus on different fundamental insights of Netz. Both of us are
right in our foci, but are led along different paths by them.
10. An aside on Science Studies

First a word of caution. I have learned immensely from the
69 I happen to be the person who told him about the book, so he very kindly
dedicated his review essay to me.
70 Steven Shapin and Simon Schaffer, Leviathan and the Air Pump: Hobbes, Boyle
and the Experimental Life, Princeton University Press, 1986.
field of science studies that came into being in the 1970s, but I
have never claimed to contribute to it, neither under the older
name of SSK, Social Studies of Knowledge, or a current name
such as SST, Society, Science and Technology. Latour was one of
the founding fathers of the field, inaugurating a basic on-going
programme of research, roughly speaking, the theory of scientific
networks. It is parallel to Barry Barnes's and David Bloor's work,
often referred to as the Edinburgh school or the strong
programme of scientific knowledge. Most workers in the field of
science studies regard me as at best an eccentric outlier, and I am
quite happy with that perception. Thus whereas it is entirely
fitting for Latour to pick the two most important books of science
studies of the past 21 years, I am not qualified to do so.
Nonetheless I would pick the same two books as Latour has.
What I should say is that, for my purposes, in developing a
philosophical anthropology of scientific reason, Netz's book is
(without contest) the most important book to come out of science
studies since Shapin and Schaffer's Leviathan and the Air Pump.
Latour is expressing the parallel thought, that for his purposes
these are the two most important books to come of science
studies in the past 21 years. Among his purposes is network
theory, and he takes full advantage of a pun; he calls his critical
study, "The Netz-Works of Greek Deductions."
11. Return to the two-faced paradox

After that necessary qualification about my being a
philosopher who does not practice that specialist discipline
properly called science studies, let me return to the paradox:
Latour and I think these two books, Shaping and Leviathan, are
the two most important, the two most useful, books to come out
of science studies. But we think so for diametrically opposite
reasons. Yet I can add that I well understand Latour's point of

view. He is fascinated by Netz's account of Greek mathematics in
action. Archimedes' network of correspondents around the
Mediterranean is a beautifully delimited network within which
mathematics thrived, within which it was created, and within
which it was stabilized. In much the same way, the Royal Society
of London was an evolving network of participants in the
creation and stabilization of the new science of the
mid-seventeenth century. But for me what matters in Netz is
summed up in his subtitle, A Study in Cognitive History. Latour
altogether ignores precisely what delights me, the "cognitive
history".
My philosophical anthropology of scientific reason holds that
styles of scientific thinking are grounded in innate potentials,
many of which are cognitive, which have to be discovered in the
course of human history. Netz has written exactly the study that I
want, an account of how the potential for deductive proof in
geometry emerged, not from some legendary Thales, but from
what we can infer from existing texts.
12. Plato for better or worse

My list of Western philosophers who were bowled over by
mathematics began with Plato. Latour puts the matter the other
way. I do not think we should describe Plato as bowled over by
mathematics, that is, he was not a victim, or a patient. He was an
exploiter, an agent. He kidnapped mathematics. He wilfully
misconstrued the new science. Latour says that "a strange
operation of channelling (not to say kidnapping) by platonist
philosophers of a narrowly specialized set of skills nurtured
inside tiny networks of cosmopolitan practitioners of Greek
geometry," enabled them to create a fascination with, to quote
again, "notions like 'demonstration', 'modelling', 'proving',

'calculating', 'formalism', 'abstraction'." He writes:
To the great surprise of those who believe in the Greek Miracle, the
striking feature of Greek mathematics, according to Netz, is that it
is completely peripheral to the culture, even to the highly literate
one. Medicine, law, rhetoric, political sciences, ethics, history, yes,
mathematics no. [... ] With one exception: the Plato-Aristotelian
tradition. But what did this tradition (itself very small at the time)
take from mathematicians? [... ] only one crucial feature: that there
might exist one way to convince which is apodictic and not rhetoric
or sophistic. Philosophy extracted from mathematicians, not a
full-fledged practice, but only a way to radically differentiate itself
through the right manner of achieving persuasion. (p. 445)
Latour takes Netz's analysis as a beautiful illustration of

knowledge sustained by a network of creators and distributors of
that knowledge. Nowhere is that better illustrated than by
Archimedes, who, working out of Syracuse in Italy, created and
maintained an unparalleled body of new understanding. In a
book written for a popular audience, Netz powerfully compares
him to Galileo, but goes further.
Archimedes is the most important scientist who ever lived.71
Mimicking what Whitehead said about Plato, Netz continues,

The safest general characterization of the European scientific
tradition is that it consists of a series of footnotes to Archimedes.
For starters, he asserts that Galileo's statics borrowed, always

with admiration, from what he knew of Archimedes, and that his
dynamics derives directly from the Greek mathematician. I will
not quarrel with Latour's use of the Archimedes circle—a handful
71 Reviel Netz and William Noel, The Archimedes Codex: Revealing the Secrets of
the World's Greatest Palimpsest, London: Wiedenfield and Nicholson, 2007, p. 26.
of mathematicians around the shores of the Mediterranean Sea,

whose focus is Alexandria—as a perfect exemplar for Network
Theory.
My interest in these two great books, Leviathan and Shaping, is
entirely different. In the case of Netz, my reading is closer to the
author's intentions than is Latour's. In the case of Shapin and
Schaffer, as I shall say in the next lecture, Latour's reading is
certainly closer to the intentions of the authors than mine. Latour
very much plays down the aspect of Netz indicated in the
subtitle, Cognitive History. So, indeed, do most readers, who tend
to notice only that Netz's book is a brilliant reconstruction of the
missing diagrams in ancient texts. For me, it is also the first
account of the discovery of a fundamental human cognitive
capacity, the ability to make demonstrative proofs. The story of
the diagrams is an absolutely essential part of that account.
13. The shift in conceptions about telling the truth

Is Plato to blame for the Western philosophical obsession
with mathematics? Yes and no. Yes, he is to blame for suggesting
the idea inculcated by Aristotelians, that all philosophical
reasoning should acquire the apodictic character of mathematical
proof. I claimed, following the lead of Lloyd and Netz, that
argumentative Athenians were impressed by the apparent ability
to settle some questions outside of rhetoric. This was what
appeared to be a new way of settling some questions, namely
proof. Thus the (*), the shift in conceptions of what it is to tell the
truth about geometrical objects, not only made a dramatic impact
on the debaters, but also on the man who despised democratic
debate, namely Plato.
Bernard Williams emphasized the new way of telling the
truth about the past. It seems to me that the new way of telling
the truth about geometrical objects had a far more profound

effect on Western civilization, by no means all for the good.
People became convinced that the only right and definitive way
to reason was deductively. Arguably it was only during the
"scientific revolution" that it was possible finally to shake off this
terrible idea.
Nevertheless, I do not believe that the Western philosophical
fascination with mathematical objects and mathematical truth
was due solely to Plato and the platonists. The experience of
perspicuous proof is itself a remarkable fact, and it is that which
so impressed the philosophers.
14. Proof, not axioms, and not calculation

There was a lot of talk, beginning in the 19th century and
taking up a lot of the 20th century, to the effect that mathematics
is axiomatic, and that the core of mathematics is something called
"the axiomatic method". But even during the heyday of an
emphasis on formal methods, there were always thinkers who
urged that mathematical ideas, including what we may call proof
ideas, are of the essence. This notion has now made it into
popular science, with science writers telling the world that what
was important about Andrew Wiles and Fermat's conjecture was
that a very general new way of doing mathematics had been
introduced. Wiles solved the classic problem, but that is just an
anecdote in history compared to the way that he opened a lot of
new doors, and connected previously unrelated domains. The
same is said more recently for the proof by Perelman (and others)
of the Poincaré conjecture.
It is important to repeat that there is not just one thing:
mathematics. People do not often enough ask, of some practice or
problem, "Why do I immediately recognize this as a mathematical
question?" Wittgenstein had that wonderful phrase: "the motley of

mathematics". Old-fashioned shopkeeper's arithmetic, before they
had calculators and lost the old skills, was "mathematics". So are
the famous proof-discoveries. What on earth do they have in
common? There is no "mathematical must" in calculation.
Despite my awareness of the "motley", the aspect of
mathematics that produces most of the underlying puzzles and
disagreements, is the phenomenon of perspicuous proof, proof
that you can take in, it seems, all at once, and see as a whole. The
type of proof that Descartes thought was the ultimate or perhaps
the only type of proof. It somehow has never seemed right, that
we could settle questions just by thinking. Or, to put matters the
other way around, it has seemed to some philosophers that
special kinds of thinking are the glory of the human mind and
the core of human nature. They have thought that just because
we can (apparently) settle some questions just by thinking.
By proof I do not mean any old calculation. I mean those
proofs such that we have the feeling that we not only see that the
theorem must be true, but also that we understand why it is true.
There is the familiar ah-ha experience that Martin Gardner used
to emphasize, the feeling of seeing, of grasping, what follows from
what. Recently I had to sum up a bunch of expenses in order to
get some money back— adding a list of some 12 numbers. I got 3
different results with my pocket calculator—I am clumsy. So I
did the sums by hand, "carry 3" and all that. In the end, I was
pretty sure I had got the right answer, but I certainly did not see
why it was the right answer. It was just what comes out of
mechanical operations that I learned as a child. But on those now
rare occasions when I take in a new proof idea, the experience is
totally different. Wittgenstein was especially interested in what
he called perspicuous proofs. Perhaps Descartes thought they
were the only proofs worthy of the name.
There is no "must" in most calculation. You just get that

result by following some procedures in which you were trained
as a child. Make the comparison with using a subway system. I
am just learning my way around Taipei. One fine day I am near
the Shipei stop and feel I would like to go to the zoo. So I look on
the map, and see that first I must take the Hsintien line, change at
Taipei Main Station, proceed on the Bannan line, get off at
Chunghsio Fuhsing, and get on the Mucha line going south,
finally I will reach the stop for Taipei Zoo. Any number of things
may go wrong, since I am a stranger, but if I do it right, and there
are no problems on the subway, then I will arrive at the zoo.
Calculating is like that, and there is no necessity to it. No one
would ever get the idea of a priori truth, logical necessity, or
mathematical objects, from thinking about the Taipei subway.
15. Two conceptions of proof

I once had a great opening line for an essay, "Leibniz knew
what a proof is. Descartes did not."72 But I could have said the
reverse, for there are two distinguishable concepts of proof. One
is that of perspicuous proof, which was the Cartesian idea, and I
think the Platonic ideal. The other is that of a proof as calculation,
fully explicated in 20th century symbolic logic. Mathematical
objects are born of perspicuous proof, not calculation.
From earliest times in the Mediterranean, there appears to
have been the feeling that in mathematics we investigate a special
kind of object, about which we discover a special type of truths.
There just are these objects, as we say "out there"—and this "out
there" conveys two things. First, they are not something mental,
72 "Leibniz and Descartes: Proof and Eternal Truths,"' a Dawes Hicks Lecture to the
British Academy, 1973, published in its Proceedings, and reprinted in my Historical
Ontology, Harvard, 2002, 200-226.
they have an existence independent of human minds, hence out

there, outside of human minds. On the other hand they are not in
space and time either. Hence the worry: they are not out there in
space, so where are they? Mathematical spheres are to some
extent like ball bearings, but very much not like ball bearings.
Hence the awe in ancient times, of certain theorems about, say,
regular solids. And when we move to numbers and finite groups,
there is not even the temptation, any more, to locate them in the
spatio-temporal world. Platonism is sometimes called realism,
because of quirks in the history of Western philosophy. I say
quirks, for platonism is, if anything, super-realism, which bears a
relation to ordinary reality not unlike that of the super-natural to
the natural.
16. A second crystallization within the mathematical

style?
Scholars emphasize that there was mathematics in West Asia
and North Africa long before Thales. My theme has been that the
mathematical style crystallized, with a new way of telling truth
about geometrical objects. Are there other events that one might
call a similar crystallization? I am happy to introduce another
legend, in this case, one for the discovery of algorithms and
algebra, the legend of Al-Khwarizmi (about 780-850) working in
the House of Wisdom at Baghdad, after whom the very word
"algorithm" is named. The name is not that well known, but, to
give a hint of his significance, the Iranian Research Organization
for Science and Technology has for some twenty years been
awarding the International Khwarizmi Prizes, its top awards in
Science and Technology.
It is clear that many extraordinary things happened in the
sciences in that particular flowering of civilization in Persia in the
ninth century. Medicine, chemistry, and much else bloomed.

Those who like grand tales of world history, in terms of empires
and battles, should include the Battle of the River Zab (750) as
one of the four or five major battles in the history of the world.
Until that time the political and cultural centre of Islam had been
Damascus, and Islamic culture was essentially Hellenistic. The
Greater Zab is a tributary of the Tigris; it begins in the mountains
of Turkey. Before 750 the centre of Islamic civilization had been
Damascus, and the culture was Hellenistic. After 750, when the
Abbassids defeated the Umayyads, Islam became oriental, rooted
not in Greece but in Persia. A capital of the new Persian Empire
was built from scratch in 763; this was Baghdad. During the
following century it became the most interesting place on the
planet to exercise the scientific imagination.
Just as Kant fantasized Thales (or some other), I fantasize
AI-Khwarizmi as the origin of a combinatorial or algorithmic or
calculating style of reasoning in the "western" tradition.
In line with Williams' schema, we would then obtain:
(*) A shift in conceptions of what it is to tell the truth about
quantities.
(**) The significant change took place in the 9th century, and its
emblem is Al-Khwarizmi.
Why talk about legendary beginnings? Because folklore may tell

us more about a deep common wisdom than systematic history
of the sciences does. You find the tale of a legendary trailblazer
for every style of scientific thinking. With probability, legend has
Pascal. With theoretical modelling we have Galileo—Husserl
spoke of the Galilean style long before Crombie ever came on the
scene. In experimentation I could mention the great
chemist-alchemist Geber. That is the Latin name for Abu Musa
Jabir ibn Hayyan c.721 - c.815, a great chemical pioneer who
worked mostly in Kufa south of modem Baghdad. There are of

course Chinese counterparts. For the laboratory style my
favourite legend, as you shall see in my third lecture, is that of
Robert Boyle and the air-pump—exactly the topic so admired by
Bruno Latour, but for reasons radically different from his. In each
case there is the cultural discovery of how to do something, of
how to find out.
In each case, what is discovered is a human potential. One
branch of modem cognitive science urges that we have innate
domain-specific capacities. Cynics say this is just a resurrection of
early-modem faculty psychology, fashionable in Spain about
1600. Certainly details are a bit sketchy, for all the present
enthusiasm for interpreting what are still very crude instruments
such as MRI and PET scans to study blood flow to the brain. But I
will go along with that, partly because it seems to me, on a lot of
old anecdotal evidence, that the ability to grasp proofs, especially
geometrical style proofs, employs different cognitive skills from
combinatorial reasoning. But then we had to discover how to
reason combinatorially and algorithmically too.
Here I tend towards Jean-Pierre Changeux, a neurobiologist
about whom I shall say more in a moment. When he was
interviewing me for a job he used a marvellous expression that so
far as I know he has never used in print—our "genetic
envelope"—that is the inherited cognitive capacities, which can
be developed in various ways, or not. I conceive of the ability to
make proofs as having long lain fallow in the genetic envelope of
the human brain, until that group of people around Thales,
whatever be his name. He grasped the possibility of proving
something about the triangle.
There is a true history here to be told. For my money, the
most exciting work is that I have already mentioned, by Reviel
Netz. I insist on his subtitle, A study in cognitive history. Netz tells
me everyone reads his history as about the missing diagrams in

ancient texts. But what he was really writing about was the
discovery of a cognitive capacity for using diagrams and words.
You see now, why Netz's book so fascinates me. I want someone
to write a book called, The Algorithms of Baghdad: A Study in
Cognitive History.
17. Mathematical objects

One of the longest running debates in the history of Western
philosophy is about mathematical objects and mathematical
truths. People called platonists hold that the objects have an
existence independent of any mind and any matter.73 Other
thinkers have long protested that they are products of the human
mind. In a more recent version, they are either the products of
collective sociology or of the structure of the human brain. For a
fairly recent shootout, see, Conversations about Mind, Matter, and
Mathematics, by my Paris colleagues Alain Connes (Fields Medal
winner, 1982) and Jean-Pierre Changeux (cellular neuroscientist).74
As you might expect, Connes argues that mathematical objects
are "just there" to be discovered, while Changeux argues that they
are products of the human cognitive system as physiologically
organized in the brain.
My take on this debate is part of the general approach
outlined in the first lecture. My "styles of scientific thinking"
73 The idea is ancient, but the label, in connection with mathematics appears to be
modem. It derives from the Paul Bemays, after whom von Neumann-Bernays set
theory is named. "Sur le platonisme dans les mathématiques," L 'Enseignement
mathématique 34 (1935), from a lecture given in 1934, and translated by Charles
Parsons for Paul Benacerraf and Hilary Putnam. Philosophy of Mathematics:
Selected Readings, 1964, 2nd ed., Cambridge: Cambridge University Press, 1983,
258-271.
74 Matiere a penser, Paris: Odile Jacob, 1989, Conversations on Mind, Matter, and
Mathematics, Princeton: Princeton University Press, 1995.
include a Galilean style, a laboratory style, a taxonomic style, and,

needless to say, a "geometrical style" and a "combinatorial style".
(The names are unimportant; they are mere pointers.) One claim
is that every style of reasoning arises from human capacities,
mostly but not entirely cognitive. These capacities are discovered
and developed at specific moments in history. Another claim is
that every style of reasoning "introduces" new types of objects,
and new ways of telling the truth. This is not a relativistic thesis
but a theory about the origins of objectivity.
It is obvious that, like some of the great philosophers whom I
have mentioned, I was absolutely bowled over by a phenomenon,
which Connes says
is peculiar to mathematics, and which is very hard to explain. One
can often, by means of tremendous effort, arrive at a complete list
of mathematical objects defined by very simple conditions. One
believes intuitively that the list is complete, and one tries to prove
that it is exhaustive. In trying to prove that one has exhausted the
list, one often discovers [unexpected] objects. Take for example the
theory of finite groups.75
Mathematicians first thought that there are just 6 types of finite

groups, a list complete by the end of the 19th century. During a
period quite late in the 20th century, exactly 20 more types were
discovered—the sporadic groups. And that is all there are: end of
the story. The last finite group to be found is called the monster,
for such it is, with more than (8)x(10!) elements. Assuming that
there is no deep underlying mistake in the proof, it feels as if this
last idiotic group was just there all the time, laying in wait for us,
with a monstrous grin on its face. This is one of the phenomena
that create a demand for a philosophy of mathematics, and then
for a platonistic ontology.
75 Matiere a penser p. 37.

It is important to distinguish two different things that

Changeux and his fellow brain scientists may be trying to do.
One is to give an account of what the brain does, when one of us
does some mathematics. How is it possible for this complex
organization of neurons to do this kind of work? No one will
deny the enormous interest of that type of research, which is in
its infancy. There is another type of research, cognitive
psychology as opposed to the science of the brain. It proposes
that there are specific cognitive abilities that are grounded in the
structure of the flesh-and-blood brain, and that these abilities are,
in the sense used by Chomsky, modular. But neither of these gets
at the experience, which is had in proving, of seeing that the
result must follow, or to sense of objects just lying in wait for us.
They pass by what I call the experience of doing mathematics.
Latour, who taught us all about science in action, looks at some
mathematics in action in great detail, but he has in my opinion
too small a diet of examples. He fails to attend to the motley of
mathematical life, some of which includes the experience of
proof.
18. Broaden your horizons!

I shall discuss kinds of realism, in general, in my fourth
lecture. In connection with mathematics, "realism" an all
purpose label for Plato & his friends. The debate between Connes
& Changeux is a "realist" debate. But consider: there are realist
debates about other sciences too. Throughout a lot of the
twentieth century there was a debate as to whether the
unobservable theoretical entities of physics are real, or are only
instruments to aid our thinking. And even nowadays, when we
can count atoms in an ultracold trap and have the sense that we
are observing macroscopic quantum phenomena, a macroscopic
wave function, for example, there is still the possibility of putting

up an anti-realist, positivist thesis. One of the smartest well
established philosophers of science continues to do just that,
namely Bas van Fraassen. He insists that our theories may be
"empirically adequate" but that we should not assert or believe
that there really are 16 atoms in a trap.
Likewise consider systematic taxonomy in biology. Ever since
Linnaeus, people have debated whether species, genera, families,
classes, orders, are "real", or are only ways of organizing the
incredible complexity of the natural world; a problem traditionally
vexing for plants, and a total mess for mushrooms and their ilk, let
alone bacteria. That is why, in my first lecture, I spoke of the
ontological debates that are a by—product of styles of scientific
thinking. In every case the objects are introduced by the style.
One has always said that these debates are unrelated. I urge you
to broaden your horizons, and say they are much of a piece.
19. Comparative anthropology of reason

Crombie, in the first instance a historian of the sciences of the
European high Middle Ages, Renaissance, and Early Modem
Period, deliberately titled his magnum opus, Styles of Scientific
Thinking in the European Tradition. I lay emphasis on the
"European". But Crombie was not wholly Eurocentric. He also
remarked that his travels between his native country, Australia,
and his adopted one, England, and the stopovers he made in
Asia en route, led him to contemplate a "comparative historical
anthropology of reason." I think that is a useful idea, albeit a
sketchy one.
I have been in Asia before, but this is the first time I have
been here to discuss my approach to scientific reasoning. In
particular I am on the Eastern edge of Asia. It is my tum to reflect
on the very distinct historical evolutions of sciences in, on the one

hand, Western Asia, (which includes ancient Mesopotamia),
North Africa, the Mediterranean, Northern Europe, and then the
Americas— and, on the other hand, Eastern Asia.
I must begin by saying that I have absolutely no interest in the
question which some westerners like to ask, "How come the
modem sciences are a product of western Europe, beginning in
the 17th century, while Chinese sciences never took off in the
same way, and even ancient Chinese mathematics was not a
stepping stone to modem mathematics?"
That looks like an interesting question. But I suspect that it is
no more and no less interesting than the question, "Why did
ancient Mediterranean mathematics reach its apogee in
Archimedes, and then come to a full stop, with (to exaggerate)
not a single act of creative mathematics taking place until a rather
different combinatorial tradition, emerging primarily in Baghdad,
was brought together with Greek geometrical thinking to
produce a new amazing take-off?" These questions, about why or
how come grand events happened or did not happen in the past,
are utterly beyond my competence. But also, perhaps just in
ignorance, I tend to dismiss them as not making much sense.
They were once a good place to start. But by now it must seem
that the grand events in question are the results of so many
contingencies that no plausible "Why?" question is there to be
answered. I think that is one of many reasons why Geoffrey
Lloyd, who now directs the Needham Research Institute in
Cambridge, England, asks different-sounding questions from
those of the great pioneer in Western studies of Chinese science,
Joseph Needham. Lloyd has recently adopted a language
superficially similar to but different from Crombie's and mine: he
writes of "styles of inquiry" in the ancient world.76 In particular

he addresses important questions about the kinds of institution
within which some kinds of inquiry thrive and others do not.
The title of this lecture is a philosopher's question: "Where do
mathematical objects come from?" That is a standard question of
Western philosophy from Plato to now. It was certainly a
question of ancient Greek philosophy. But it does not appear to
be a question of ancient Chinese philosophy, which also had a
rich mathematical tradition. We can ask, why not? More
generally, we can ask, "Why has mathematics so deeply coloured
Western philosophical thought, at least from the time of Plato?"
Perhaps this is not so frivolous a question. Perhaps there are
significant things to say about ancient Greek and Chinese
mathematics, each embedded in its own local historical
circumstances, which bear on this question.
Perhaps the very motley of mathematics is part of the answer:
that the kind of mathematics developed in early Mediterranean
civilizations invites a conception of mathematical objects, while
the kinds of mathematics developed at more or less the same
time in China do not. This may be consistent with yet another
type of answer, that the grammar of the languages in the West is
more inviting to questions about mathematical objects than are
the grammars of languages in the East. Both are consistent with
another type of answer, that the institutions within which
mathematics was practiced in China, and the public purposes
served, were rather different from those around the Mediterranean
Sea. When I say consistent I mean something a little more than
that: all three types of answer, and more, might and perhaps
should be used together, and all three brought to bear if one
thinks the question is worth answering. The question, to repeat,
76 G. E. R. Lloyd, Ancient Worlds, Modern Reflections, e.g. p.77.

is why have Western philosophers been obsessed with mathematics,

while Eastern ones have not?
20. Ancient Chinese mathematics

I cannot read Greek and I cannot read Chinese. All my
knowledge is secondary. I take as my canonical work The Nine
Chapters,77 of which scholarly editions have recently appeared in
French and in English. I know the French version because I talk a
good deal to one of its editors, the historian of Chinese
mathematics, Karine Chemla in Paris. There has been a
substantial project in Europe and America for re-discovering
ancient science beyond the traditional 19th century vision of, as
Kant put it, "that wonderful people, the Greeks." That is the
"Greek miracle" of which Latour speaks with such crushing
irony. For a while I was stubborn and conservatively European,
pro-Greek. After working through some of the work in The Nine
Chapters, I would say to Chemla, the light-bulb of proof never
went on over the head of your authors. They never realized and
conceptualized the possibility of making a deductive proof! They
never had that "A-ha!" experience which is at the core of the
phenomenon we call mathematics. Look, they solved problems
by successive approximations, they never proved anything!
Brilliant systems of approximation, a glimpse of a purely
computational mathematics, but no proofs. Or so I said.
Only slowly did I put my reaction together with the insights
of Lloyd and Netz. Demonstrative proofs were socially valuable
in the Athenian milieu because they provided a paradigm of
77 Karine Chemla et Guo Shuchun, Les neuf chapitres: Le c1assique mathématique

de la Chine ancienne et ses commentaires, Paris: Dunod, 2004. Tome I, p. 116: I
have not properly studied it ! Cf. Shen Kangsheng, John N. Crossley, Anthony W.-c.
Lun, The Nine Chapters on the Mathematical Art: Companion and Commentary,
Oxford University Press, 1999.
what it was to settle an argument. In a more authoritarian world,

arguments in the end were settled by recourse to superiors; what
the journeyman wanted was a system that worked. Chinese
mathematicians of roughly the same epoch as Archimedes had
an interest in similar problems, but they did not have to convince
"anyone" in the audience. Their world was hierarchical, the
society was authoritarian, often benevolently so. A process which
worked was good enough for an attentive master.
One might say that the Chinese were right and the proof-giving
Greeks were wrong. Approximation is the watchword of
pragmatism, in the sense of William James, or so we might say.
We want a set of procedures which in the end will lead careful
workers to more or less the same conclusions, so there can be no
undermining of authority.
This is very sketchy and simplistic, but it is a point at which
to stop. For a more serious contribution with which to continue, I
would follow some points outlined in a recent paper of Chemla's,
a paper with a very long title, "Geometrical Figures and
Generality in Ancient China and Beyond: Liu Hui and Zhao
Shang, Plato and Thabit Qurra."78 I regard this as a remarkable
step to making sense of the sentence I quoted from Kant,
The true method, so he found, was not to inspect what he
discerned in the figure, or in the bare concept of it, and from this,
as it were, to read off its properties; but to bring out what was
necessarily implied in the concepts that he had himself formed a
priori, and had put into the figure in the construction of by which
he presented it to himself.
A fundamental difference between Chinese and Greek geometry
was that the former investigated the properties of shapes by
approximations and sequences of procedures, while the latter
78 Science in Context, 18 (2005): 125-166.

aimed at perspicuous proofs. You might argue that the long

history of Western mathematics will tum out in the long run to
have been only a tedious detour to the future of mathematics. For
all we ever need in our relations with the material world is what
we demean by calling "approximations". We never need exact
solutions to differential equations. We only need answers to an
adequate approximation. In the future fast computation will
replace proof, and Eastern, rather than Western, mathematics
will be seen as having been on the right track from the beginning.
Was Crombie's geometrical style of thinking just a bad start to the
scientific enterprise, a historical accident?
Chemla's paper just mentioned provides a definitive instance
of this. She deliberately takes as her example what she calls, in
scare quotes, "the Pythagorean Theorem". Well, I know what
that theorem is—or so I thought. But there is a great difficulty of
even recognizing that theorem in the Chinese texts that Chemla
discusses. There is not exactly an "object" there to discern even in
the diagrams we can reconstruct. It is so different with Greek
mathematics as it has corne down to us, a body of work which
makes Kant's reading so inevitable.
21. The origin of the objects

The new classes of objects introduced by, or with the corning
into being of, new styles of scientific thinking make possible the
interminable Western ontological debates in those domains, for
example between platonism and nominalism in mathematics, or
scientific realism versus the various types of instrumentalism and
empiricism. I think that mathematical objects are not so different
in the nature of their origin from theoretical entities or new taxa
in systematic biology.
Ontological debates within the sciences result, then, from the
introduction of objects by styles of thinking. But there is more to

it than that, something more Eurocentric. In European language
we talk about these objects using sentences in which names for
the objects serve as grammatical subjects. This leads on to a third
point, emphasized by Nietzsche long ago: European languages
demand an existential presupposition for terms in the subject
position. European grammars generate the ontological obsessions.
These remarks do not go close to addressing what really
fascinated Kant and Russell, namely why or how we can have a
priori knowledge, of why it seems we can find out about things
just by thinking. I believe an answer to that question lies close to
what I have been saying, and that we have started to re-think
that group of questions in a way that will prove fruitful.
Unfortunately I shall not have time to do that rethinking in this
short course of lectures.79
79 For a sketch of how to go on, see my course of lectures at the College de France on
line at http://www.college-de-france.fr/default/EN/all/ins__pro/ pI157460409944.htm;
especially "Demonstration" and especially "La stablilité des styles de penseé
scientifique. "
Lecture 3
............
THE LABORATORY STYLE

OF THINKING AND DOING
1. Recapitulation
I shall begin by recalling the central novelties of the first
lecture. There were two starting points, both of them adaptations
of ideas developed by other scholars for other purposes. One was
associated with the historian of science, A. C. Crombie, from
whom I took the notion of a short list of enduring styles of
scientific thinking. The second was taken from the philosopher
Bernard Williams, the concept of truthfulness. I agreed with
Williams that truth does not have a history, but that truthfulness
does; indeed, that there are genealogies of truthfulness.
Crombie was an Australian who pursued his career in
Oxford, England. He was primarily a scholar of European
sciences in the high Middle Ages and the Renaissance; his hero,
like that of many other scholars, was Galileo. But contrary to
popular wisdom, he thought that the so-called "scientific
revolution" of the seventeenth century was not an autonomous
event of that period, but rather an evolution of developments
that began to flower in the 12th century. At a time when the
influence of Bachelard (in France) and Kuhn (in America) was

dominant, when philosophers and historians saw the past as a
sequence of mutations and revolutions, he saw plenty of
continuities and resumptions. His life work was summed up in 3
enormous volumes, Styles of Scientific Thinking in the European
Tradition, published near the end of his life, in 1994. In each case
he traced the trajectory of some six styles of scientific thinking
from the ancient Mediterranean world to modem Europe.
We can distinguish in the classic scientific movement a taxonomy
of six styles of scientific thinking [... ] distinguished by their objects
and their methods of reasoning.
We have three words here, styles, which are distinguished by

their objects, and by their methods of reasoning. In the case of
mathematics, we are familiar on the one hand with a distinction
between mathematical and other methods of reasoning, and on
the other, with a distinction between the abstract objects of
mathematics, and the objects of everyday life.
Do not get the idea that first we have a style of thinking,
which then introduces a new type of object and a new method of
reasoning. The styles are constituted by their methods and the
type of object with which they deal.
I made the claim that each style of thinking introduces a new
class of scientific objects. I proposed, moreover, that the
seemingly unrelated ontological debates about abstract objects,
about theoretical nonobservable objects in physics, or about the
taxa of systematic biology, are all by-products of the introduction
of new types of objects in the course of the emergence,
acceptance, and use of a new style of reasoning within specific
communities.
I urged one extension of Crombie's notion of styles of
scientific thinking, namely the idea that within his long tale of
THE LABORATORY STYLE OF THINKING AND DOING / 91
continuities, there are distinct breaks, which I called crystallizations.

A crystallization may occur deep in the past, as when I claimed
that the idea of demonstrative proof crystallized ways of
mathematical thinking about geometrical objects, perhaps in the
sixth century BCE. There might, I said in the second lecture, be
further crystallizations, such as the one that took place in
Baghdad early in the ninth century, when algorithmic and
algebraic thinking were clearly understood for the first time.
These notions are then combined with Williams's idea that
truthfulness about a subject matter may have a genealogy.
Williams gave two examples of sharp discontinuities, one in
telling the truth about the past, and the other in telling the truth
about the self. I proposed that we generalize his idea and apply it
to crystallizations that take place in distinct styles of scientific
thinking. With each crystallization there are two schemas:
(**) This significant change took place in the Y century, and its
emblem is Z.
For Williams's example of history, X = the past, Y = the fifth

century BCE, and Z = Thucydides. This emblematic figure was a
real person about whom we know quite a lot. He was what I
called an iconic trailblazer. But such icons need not be historical
personages at all, they are rather figures, half way between truth
and fiction, between history and legend, about whom a certain
radical change has been recognized in common belief and lore. In
the case of the second lecture, X = geometrical objects, Y = the
sixth century BCE, and Z = Thales. Thales may be purely
legendary, or if a historical personage is referred to, he may not
have done most of the things that lore attributes to him. He is an
emblem, and what I called an iconic trailblazer. Thucydides is
more truth than fiction, but Thales may be more fiction than
truth.
I noted in the second lecture that we should not be too
scrupulous in mimicking Crombie's sentences. With mathematics
all goes smoothly; there are pre-Ionian statements on geometrical
objects about which "Thales" can change the conception of what
it is to talk truly, namely by proving. But when we tum to what
happened in the time of Galileo, when there arose the idea that
God wrote the world in the language of mathematics, he urged a
new way of thinking about Nature. Yes, at the time of Galileo
there was a shift in conceptions of what it was to tell the truth
about X = motion, but the shift was far wider than that.
2. Experimental exploration and hypothetical modelling

Crombie's first style was mathematical, the topic of my
second lecture. I noted some fundamental differences in the
developments of mathematics in Ancient China and Ancient
Greece, and proposed that the latter gave rise to a Western
obsession with mathematical knowledge, truth, and objects, that
is unknown in other philosophical traditions. I suggested this
was partly due to the focus on perspicuous proof in canonical
Greek mathematics, while Ancient Chinese mathematics was
more concerned with process, and developed a remarkable series
of techniques of approximations. The "shift in conceptions"
about telling the truth was that after the time of the man whom
we call Thales, geometrical propositions could be proven.
Crombie catalogued exactly six ongoing styles of scientific
thinking. His fourth, fifth and sixth styles have to do with
populations and classes. I call them taxonomic, probabilistic, and
historico-genetic. I shall say almost nothing about them in these
lectures. But there is lots to say!
I first heard about the "six styles" from Crombie at a
conference in 1978, and that started me thinking anew about the

sciences; the first sign of my uptake was published in 1982, and I
have been thinking about these matters from time to time ever
since.80 Today I shall be concerned with his second and third
styles of thinking, but above all with a crystallization of the
second style, which leads me to speak of the laboratory style.
Here are the brief descriptions of the second and third styles that
I heard almost 30 years ago:
2. Experimental measurement, and exploration of more
complex observable relations.
3. Hypothetical construction of analogical models.
The laboratory style of thinking and doing should not be
thought of as an additional style but as a crystallization of the
second style, experimentation. It resonated with a slightly earlier
Galilean crystallization of the third style.
3. Recapitulation of 3 cautions
Before proceeding, I shall repeat three observations made in
previous lectures.
(a) A caution. Styles of scientific thinking are not sciences or
scientific disciplines, and they are not mutually exclusive.
Evolutionary biology uses lots of (1) mathematics, (2)
measurement and experimental exploration, (3) hypothetical
modelling and analogy, (4) taxonomy, (5) probability and
statistics, and yet it is our most viable example of a (6)
historico-genetic science. Most modem sciences use most of
Crombie's styles of scientific thinking.
(b) A cognitive conjecture. I postulate that each of Crombie's
80 Ian Hacking, "Language, Truth and Reason," Rationality and Relativism, ed. M.
Hollis and S. Lukes, Oxford: Blackwell, 1982, 48-66. " 'Style' for Historians and
Philosophers," Studies in History and Philosophy of Science 23 (1992), 1-20. Both
reprinted in Historical Ontology, Harvard, 2002.
styles is grounded in innate human capacities, which are

discovered, exploited, and developed in specific historical
situations. Thus the styles result from both cognition and culture;
they are the products of interactions between on the one hand
unique human endowments which are in tum the results of our
evolutionary heritage, and on the other hand, specific historical
events and developments. In virtue of the innate cognitive
element, it follows that even if a style of thinking evolved in the
first instance in a unique historical culture, it can later be
acquired by people of any other culture who choose to do so.
(c) Styles of scientific thinking change, evolve, divide, and
unite, in complex historical ways. We do not do mathematics
today in the manner of the so called "Nine Chapters" of classical
Chinese mathematics, with their commentaries, worked out over
centuries between 2000 and 1000 years ago. Yet at the same time
we recognize that work as "mathematics". As I have said, within
the evolving styles there may be breaks or crystallizations; two of
those are my topics today.
4. Exploration and analogy: styles (2) and (3)

Neither of Crombie's styles (2) and (3) is peculiarly
"European". One way to think about the history of Chinese
astronomy, for example, is to regard it as a deployment of these
two styles of scientific thinking from earliest times, namely
measurement and observation of complex relations on the one
hand, and on the other hypothetical modelling. The modelling
was typically of the movements in the heavens. I also count
Ionian speculations about the atomic nature of the world as
hypothetical modelling. The difference is that astronomical
modelling was controlled by observation, while atomism in
earliest times was uncontrolled speculation by imaginative
philosophers of nature.
Crombie characterized his second style as "experimental
measurement, and exploration of more complex observable
relations." Whereas only a few civilizations have developed
much that I would dignify by the name "mathematics" I believe
that all peoples have engaged in various activities worth
dignifying by such words as experiment and exploration. That is
true not only of people but also of animals and birds. Until
recently, birds have been thought to have "bird-brains," but now
we realize that crows, for example, are pretty smart, and perhaps
even in the league of the higher apes. I doubt, however, that I
would dignify any avian activities as measurement; perhaps that is
a species-specific, peculiar, that is, to human beings. But a great
many human civilizations have engaged in measurement. In the
beginning, many peoples seem to have used parts of the human
body as the standards for transportable measuring devices, for
example the royal foot, the priest's arm from elbow to finger tips,
or the length of an exemplary human thumb. I am disinclined to
see any distinct crystallization, any discovery of a new human
potential, in the long development of human curiosity,
experiment, exploration, or measurement.
5. Measurement
I could well be wrong about this. Perhaps there was a distinct
crystallization around measurement, the second clause in
Crombie's second styles of scientific thinking. That crystallization
would be the realization of transportable units. But since the
original units seem to have been eminently transportable, human
feet and thumbs, there may have been only a gradual passage
rather than a sharp break to standard transportable units.
Nevertheless, a primary use of measurement is in planning
buildings for habitation, worship, or protection. As soon as you

start making rigid dwelling places for family, for priests, to keep
out your enemies, or for the afterlife, you have got to do some
measuring. So we may guess that the first sustained need for
measurement was from builders.81
Transport is seldom emphasized in philosophical talk about
measurement. But it is the essence of measurement, and
continues into the work of modem national bureaux of standards.
The standard metre in Paris is a snare and a delusion for
philosophers. What you want is a standard metre to which
various tentative metres can be transported for comparison. We
think of the volt as a standard unit of electric potential, but
people who do standards need to construct apparatus that
provides the transportable volt, that is, a standard unit of electric
potential that can be taken to any laboratory or electrical plant in
the land.82
One of the greatest of scientific revolutions was Einstein's
realization that time is not transportable. We all carry watches
around with us, and think we are carrying the transportable
second. But accelerate clocks and away we go to the special
theory of relativity. The interest of the atomic clock is not its
accuracy but the fact that it relies on phenomena which, we hope,
pervade the universe, and so do not need to be transported.
81 The "primitive language" of the builders with which Wittgenstein opens the
PhilosophicalInvestigations may be primitive for them, but someone has to have cut
the slabs, for which they are calling, to a measure. I suppose that could be done
without speaking, but it does remind us of the immense social and cultural
complexity that even the simplest intelligible tale of human talk presupposes.
82 I first realized that transportation is the essence of measurement when visiting
Churchill Eisenhart, at the old U.S. Bureau of Standards in Gaithersburg, Virginia.
(The Bureau is now the U.S. National Institute of Standards and Technology.)
6. The hypothetico-deductive m,ethod

Crombie's third style is "the hypothetical construction of
analogical models." This too is a readily recognizable style, about
which I am inclined to say things similar to what I said about
experiment and exploration. Analogy is the very guide in life.83
Guesswork is our chief resource. Hypothesis is chiefly a fancy
word for reflective guessing, and it is part of human nature.
Popper went further, and assured us that amoeba make their way
about by much the same devices. Certainly all human societies
encourage guesswork and analogy, and from there it seems a
short step to making more intellectual models of the
environment using analogy and guesswork. In short, I do not see
any evident early breaks, distinct crystallizations, in this aspect of
our animal nature.
It will sound as if Crombie, in delineating his third style, is
describing something very familiar to logicians of science, namely
hypothetico-deductive reasoning. Twentieth century textbooks on
the philosophy of science, especially those of a positivist or
empiricist inclination, spoke of the hypothetico-deductive method
as the core of scientific reasoning. One makes a hypothesis or
conjecture H, which may contain references to theoretical
non-observable entities. One deduces observable consequences,
often in the form of a conditional proposition: If circumstances C
obtain, then result R is produced. One arranges an experiment so
that C obtains, and sees whether R comes about. If yes, then the
hypothesis is confirmed (Camap) or corroborated (Popper).
In my opinion this account completely ignores most
interesting aspects of experimental work, to which I shall tum in
83 Bishop Butler's phrase, that probability "is the very guide in life" is well known;
less well remembered is the fact that the phrase comes early in the Introduction to
his Analogy of Religion, Natural and Revealed (1736).
a moment. It does, however, have the great merit of being

logically transparent. The logical form applies well in the most
banal circumstances, when no theoretical or unobservable
entities are mentioned. It is the common sense of humankind, not
science. For example, I hear a noise every evening behind the
walls of my bedroom. I conjecture that there are mice living there.
I leave some cheese out, and reason that if there are mice in the
walls, the cheese will be gone by morning. The cheese disappears,
all but a few crumbs. I conclude I have mice indeed. The
conclusive verification comes when I set a trap and catch the
observable dead mouse. The inference, like all sound
non-deductive inference, may be mistaken. Exactly the same
noise continues after I have killed some mice. Oh dear, I have
birds nesting in the walls ...
Hypothetico-deductive reasoning, as thus far explained, is
certainly not a discovery of a community of thinkers and agents
in Greece or China or anywhere else. I imagine that human
beings have been using that method of reasoning for as long as
they have been able to talk, and arguably it is used by animals. It
is what Charles Sanders Peirce, the great American pragmatist,
called "abduction", and more recent writers have called
"inference to the best explanation".
As was mentioned in the first lecture, Peirce thought that
logic has three parts, deductive, inductive and abductive. This is
a statement of fundamental importance, not original with Peirce,
but more clearly understood, and more succinctly stated by him
than by any predecessor. As I stated, I do not count any of these
three as a style of scientific thinking; I hold them to be
universally practiced by human beings, even if their codification
in systems of logic is relatively recent. I file them with our
evolutionary heritage. And not as cultural discoveries of early
modem Europe, or of ancient China or ancient Egypt. Logic, in
the sense of Peirce's triad, is a human universal, and what is

universal has three aspects, deductive, inductive, and abductive.
7. Architectonic reasoning
I mentioned in the first lecture that there is a certain tension
between the early (1978) and final (1994) exposition of Crombie's
doctrine of styles. In the beginning the focus was on his own field
of specialization, the sciences of early modem Europe, that is, the
high middle ages (12th century) and on in to the Renaissance.
Later, he dwells on the continuity, and he tends to trace
everything back to the argumentative world of the Greeks. He
does, however, point to an event which he treats as relatively
gradual, but which, in rather different terms and organization, I
shall promote to a crystallization.
Crombie makes this move on page 1087 (!), in the second
volume of his book. There he states his view of what happened in
a single sentence:
The particular intellectual and artistic ambience of early modem

Europe came to make (3) the method of hypothetical modelling a
characteristically effective scientific combination of theoretical and
experimental exploration.
Whether he was fully conscious of it or not, Crombie here speaks

of the combination of the methods of two different styles. I shall
try to keep him to his original conception of distinct styles. I
regret that I shall say little about the artistic aspect, on which
Crombie rightly lays much weight. But we must not forget that
an important part of this local, contingent, historical sequence of
events lies in the developments in architecture and in the
development of specific techniques of perspective representation
developed in Italy and Flanders. It is also connected with a
Christian theological vision of the world. God is the divine
architect, and Man, in trying to understand the world, can do no

better than try to figure out how God did it.
Leibniz called theoretical modelling architectonic reasoning: to
figure out how the world works, you should reflect on how the
creator, the architect of the world, achieved his effects. The
Frontispiece to Crombie's three volumes is indicative. It shows a
drawing of "God the measurer." But not just God measuring, it is
God as the architect of the universe. The illustration is taken from
a late 13th century French Bible, illustrating the first line of the
Old Testament, "In the beginning God created Heaven and
Earth." God is standing over an ill formed spherical universe
holding a standard architect's tool, what is called a pair of
compasses, a device for drawing circles and measuring or
transferring distances.
It is a very good reminder, to connect hypothetical modelling
with architectural models and hence with measurement. We see
that Crombie's distinction between his second and third styles
curiously melts away, as is indicated in the single sentence I have
just quoted. Does this mean that his catalogue of six distinct styles
is just a sham? I think not. I believe that two crystallizations restore
the sacred six to their right relationships. The most important
crystallization of the second style-the one that changed the
world-is what I call the laboratory style of thinking and doing,
whose emergence I present in this lecture. The crystallization of
the third style is what I shall call Galilean.
8. The Galilean style

It is something of a commonplace in simplified histories of
early modem science, that around the time of Galileo, the
method of analogy and hypothetical modelling underwent a
remarkable sophistication that amounts to a mutation. I did not
THE LABORATORY STYLE OF THINKING AND DOING /101
make up the name, the Galilean style. Steven Weinberg, the

Nobel prize-winning cosmologist, recalled Husserl speaking of a
Galilean style for "making abstract models of the universe to
which at least physicists give a higher degree of reality than
would the ordinary world of sensation." Weinberg found this
astonishing, "for the universe does not seem to have been
prepared with human beings in mind."
I shall have to pass by that observation, about the universe
and the human mind. It does point to a nagging question. How
come we are so good at modelling the complex processes of
nature? I once made too strong an argument, that we have
evolved processes of reasoning that are "self-vindicating."84 That
is to say, the human mind grapples with the complex universe in
such a way that the answers obtained are what characterize what
is true about the universe. So the universe does not have to be
designed with us in mind, for us to have the experience of
cracking it. I shall touch on those issues in my fourth lecture, on
"Realism and anti-realism." As I say, I believe my proposal was
instructive but far too strong.
A far more radical proposal was made by Charles Sanders
Pierce. It is one that few philosophers take seriously. Sometimes I
feel that only I do! He gave it the rather scary name of
"evolutionary love". He suggested that the mind and the
universe evolved in harmony, so that by processes analogous to
those by which the universe settled down to stable laws, the
human mind settled down to certain structures that already
existed in the stable universe. Hence the mind resonates to the
universe. I take that idea seriously, not that I believe it.
To tum from metaphysics back to cosmology, the grammarian
84 'The Self-Vindication of the Laboratory Sciences," in A. Pickering, Science as

Practice and Culture (Chicago: University of Chicago Press, 1991),29-64. I am told
this paper was recently translated, in Taiwan, into Chinese.
Noam Chomsky picked up Steven Weinberg's remark, urging that

"we have no present alternative to pursuing the 'Galilean style' in
the natural sciences at least" in particular, that is necessary for
developing the theory of universal grammar, or so he urged. The
historian 1.B. Cohen then went on to compare the Galilean style,
with reference to Husserl, to what he called Newton's style.85
In the Crisis completed about 1936, Husserl did have a great
deal to say about Galileo, and much to say about style. (As noted
in the first lecture, Stil was a vogue word in the
German-speaking world during the 1930s.) But I cannot find
Husserl using exactly that label, "Galilean style." When the
phrase came into use, Cohen, Chomsky and Weinberg all worked
in the same area and crossed paths often enough, even if not
noticing the fact. Despite the references to Husserl, this very
specific sense of the words seems to have been devised in
Harvard Yard in the late 1970s, rather than by Husserl in 1936.
Although Husserl did not use the label in quite this way, his
long discussion of Galileo in the Crisis gets to the heart of the
matter better than any of the Harvard authors. He emphasized
the use of mathematical models to comprehend the universe,
both of the heavens and more importantly on the earth: and then
the revolutionary one-world guess, that the heavens and the
earth work by the same mechanics, as described mathematically.
Husserl took this to be a fundamental moment in the history of
European civilization: the mathematization of the (one) world.
He had the idea that only if we could recover that moment in
history could Europeans break free of the disaster that was about
to befall them. I cannot share that idea, but it reminds us how
seriously Husserl took it.
85 References to Weinberg, Chomsky and Cohen's discussions are found in my two

papers on styles reprinted in Historical Ontology, Harvard, 2002; specifically on
pages 162 and 179.
To hark back to Crombie's emphasis on theology, creation,

and architecture, never forget that Galileo insisted that God
wrote the Book of Nature in the language of mathematics.
Galileo saw himself as engaged in what Leibniz was later to call
architectonic reasoning.
The fact that Weinberg and Chomsky invoked the Galilean
style is instructive. For they are not simply great scientists. They
are scientists in wholly different fields, who share one thing in
common. To exaggerate, neither can tinker with the phenomena
that they investigate, they can only observe. In that respect they
are like the astronomers of ancient times. One man is a
cosmologist, the other a grammarian. Traditionally, all you can
do in cosmology is make models: you can not make experiments
on the cosmos. Weinberg is the author of a wonderful book, The
First Three Minutes-that's the first three minutes of the universe,
and we cannot experiment on that, although our cosmological
models are informed by innumerable experimental results. But to
exaggerate and even to parody, in cosmology all you can do is
make models and compare the consequences of your models
with observations. Likewise, in grammar, you cannot experiment
(again I exaggerate) and you can only compare the predictions of
your grammatical models by observing what people say or are
willing to say. In short, cosmology and grammar are paradigms
of non-laboratory hypothetical science. They are perfect
paradigms of Crombie's style (3) at its most pure.
I. B. Cohen, the historian, made a more detailed observation,
one that we can tie in with Crombie's idea of a seventeenth
century combination of methods of reasoning. Cohen spoke of
"two levels of ontology", one of mathematics, the other of
mensuration. Recall that Crombie's style (2) is the method (in
part) of measuring. His style (3) is the method not only of
modelling but also, with Galileo, as Husserl insisted, the method
of mathematical modelling. I do not much care, for a moment,

whether one says, with Cohen, that this is a combination of two
levels of ontology, or with Crombie, a combination of the
methods of two styles of scientific thinking. The point is
combination. But it is not simply a question of combining
measurement with modelling. What we now casually call "the
scientific method" is a combination of the Galilean style with
what I call the laboratory style. That is, a combination of the
crystallization of the third style with the crystallization of the
second. And what is that? Modelling plus the creation of
phenomena in the laboratory. Before we tum to that, we should
attempt a formulation of our (*) schemas for the Galilean style.
There was, I propose,
(*) A shift in conceptions of what it is to tell the truth about nature.
(**) This significant change took place early in the seventeenth

century, and its emblem is Galileo.
9. A new "Form of Life"

In the second lecture, on mathematics, I quoted an as yet
unpublished essay by Bruno Latour, in which he discusses a
book about ancient Greek mathematics by Reviel Netz. Latour
said that,
This is, without contest, the most important book of science studies
to appear since Shapin and Schaffer's Leviathan and the Air Pump.
That is the book well known in science studies, published more

than twenty years ago, and whose subtitle is Hobbes, Boyle, and the
Experimental Life.86 I said in Lecture II that I completely agree
with Latour about the two most important books published in
86 Steven Shapin and Simon Schaffer, Leviathan and the Air Pump: Hobbes, Boyle
and the Experimental Life, Princeton University Press, Princeton, N.J., 1985.
the field during the past 21 years-yet that I do so for reasons

completely opposite to his. I explained that in connection with
mathematics. Today it will become clear why I also so admire
this work about experiment, but use it for ends entirely different
from Latour's.
You can get a glimpse of Latour's interest in the book from
the subtitle of the translation of the book into French, which
Latour enabled through his own publisher. No longer is it
subtitled Hobbes, Boyle, and the Experimental Life. That is turned
into, Hobbes and Boyle between Science and Politics.87 There is
nothing incorrect about that, but it does redirect attention. The
English title, by the way, gives a hint of a central theme in the
book. The authors make heavy use of Wittgenstein's phrase
about forms of life. They see themselves as telling how and when
an experimental form of life came into being. They teach that
Boyle's experimental programme was, in Wittgenstein's phrases,
a new "language game" and a new "form of life." They say that,
we will make liberal, but informal, use of Wittgenstein's notions of
a "language-game" and a "form of life." We mean to approach
scientific method as integrated into patterns of activity. Just as for
Wittgenstein "the term 'language game' is meant to bring into
prominence the fact that the speaking of language is part of an
activity or a form of life," so we shall treat controversies over
scientific method as disputes over different patterns of doing
things and of organizing men to practical ends." (p. 15.)
The experimental program was, in Wittgenstein's phrase, a

"language-game" and a "form of life." (p. 22.)
I am too cautious a reader of Wittgenstein to follow our two

authors in using his words, but it is a valuable direction to
contemplate. By the "experimental programme" they mean more
87 Éditions la découverte, Paris, 1993.

106 SCIENTIFIC REASON
than Boyle's lab in Oxford; they imply the programme of

experimentation developed in the European seventeenth century.
Experimental exploration, in a more loose sense of the words,
evolved in many societies at many times around the globe. If I
were to choose to use Wittgenstein's words for my own purposes,
I would suggest that the crystallization of the laboratory style of
scientific thinking deployed a new language game, within a
novel form of life.
As the style evolved, other "language games" came into use.
If we think only of the published texts in scientific journals and
the like, there are excellent studies of that evolving genre.88 The
Internet has changed everything again. Pre-publication online is
the norm, and even journals of record pre-publish articles online,
sometimes months before they are in print. I doubt that
Wittgenstein would have called these language games, but his
phrase is now out there, for anyone to use. The trouble is that it is
so powerful a phrase, that it conveys a false impression of deep
understanding, and of knowing what you are talking about.
Some, like Schaffer and Shapin, use the phrase with precision.
Most do not. I prefer to leave it to the historical Wittgenstein.
Likewise one may say that the "forms of life," in which the
laboratory style is practiced, have changed. We may tend to
emphasize research laboratories too much. The laboratory has,
since the eighteenth century, been an arm of commerce and
industry. It has effected radical changes in geopolitics: witness
the development of German chemistry through the mastery of
synthetic dyes in the nineteenth century, which, with parallel
developments in steel and munitions processing, would, were it
not for political evil, have made Germany a far more dominant
player in world history than it became.
88 Charles Bazerman, Shaping Written Knowledge: The Genre and Activity of the
Experimental Article in Science, Madison, WI: University of Wisconsin Press, 1988.
Even the research laboratory has undergone immense

changes. Much has been written about "Big Science" epitomized
by the Manhattan Project and its consequences to the
administration of American science. More recently biotechnology
has totally changed the scientific landscape and of the forms of
life experienced therein. I shall return all too briefly to that point
at the end of this lecture. Now I take up the narrative of the Air
Pump where I left off.
10. A new actor: not a person but a piece of apparatus

Bruno Latour wrote an important essay about Leviathan and
the Air Pump, which was a prolegomenon to his own book soon
to appear, We Have Never been Modern.89 I too reviewed the book
the next year, drawing a very different lesson from his.90 I want
to reemphasize that the lessons I draw from the two most
important contributions to science studies in the past 21 years,
are not incompatible with Latour's. Not at all. They are
profoundly different emphases. It is the mark of greatness in an
author that different readers can learn different things from the
same words.
In my review, written nearly two decades ago, I was
absolutely astonished by what I learned from Shapin and
Schaffer's book. I shall so often speak of their book in this lecture
that I shall refer to them as S&S. I admired them because their
book was really new; as I put it then, they emphasized "a new
kind of character, a new kind of place, a new kind of writing, a
89 B. Latour, "Postmodem? No, Simply Amodem! Steps Towards an Anthropology

of Science,"~Studies in the History and Philosophy of Science, 21 (1990), 145-171.
90 "Artificial Phenomena,"' British Journal for the History of Science, 24 (1991),
235-241. This was late in the day for a review. The journal was rightly regretful
about the ignorant and complacent review of the hardbound copy that it published,
and used the occasion of a paperback edition to afford a second look at the book.
new kind of fact and a confrontation whose denouement was

revolutionary".
"The new protagonist is not a person but a piece of apparatus.
There had been plenty of instruments for measuring or
examining phenomena more minutely. The hero of this book is
something new, a device that creates effects that didn't
previously exist in isolation. It inaugurated laboratory science.
Before the air pump, one aimed at solving the phenomena given
in nature, usually in the heavens. Afterwards a new kind of
science answered to a new master, the phenomena that fleetingly
exist by artifice.
"The new place discussed in this book developed into the
laboratory itself, the site for the manufacture of phenomena -or,
if that sounds too paradoxical, for their purification. The
laboratory was to be a space at once open and shut. It had to be
public because according to the doctrine that evolved, any work
done in a laboratory can be done by anyone with adequate skills
and checked by anyone who is a good observer. It had to be
private because only a self-selecting few could know what was
going on, make anything work, or even tell when apparatus was
working. The nascent laboratory also led to a way of writing
scientific prose, writing that had to stand in for witnesses. At first
it had to make you think you were really there. The effect was
achieved not by adding local colour but by deleting it. The new
writing was persuasive because it was presented in a plain and
unadorned way, as if it were an exact description of what anyone
could have seen, without interpretation. It dealt in matters of fact,
and did so by making us think that matters of fact just come with
the territory, beyond controversy. This book is in part about how
it became settled what counts as 'matters of fact'. They really do
come with the territory - viz., the laboratory floor."
I was primed to read S&S in this way because of my book
Representing and Intervening, published in 1983. The second half

of the book, Intervening, was a plea for philosophers to take
experiments seriously. Philosophy of the sciences, especially the
physical sciences, had for decades been totally dominated by
theory. Just think of Carnap, or Popper, or Kuhn, or van Fraassen.
Experiment was a mere adjunct to theory. Popper asserted
openly what every contributor tended to take for granted, that
the experimenter cannot even begin until the theoretician has
done his work. Experiments were for verifying or corroborating
or articulating theories, or for showing that they were empirically
acceptable. I saw myself as starting a "Back to Francis Bacon
movement," little knowing that it was well under way. Latour
and Woolgar had already published their ethnography of the
laboratory, and Shapin and Schaffer were completing Leviathan
and the Air Pump. Next year Peter Galison was to publish How
Experiments End.
11. The new place: The laboratory

Words are often useful signals of events. In European history,
the word "laboratory" entered major European languages soon
after 1600, perhaps 1605 in English, and 1620 in French. It is first
of all a place, a piece of architecture, to quote the first definition
in the Oxford English Dictionary, which arranges its definitions
chronologically, it is "a building set apart for conducting practical
investigations in natural science, originally and especially in
chemistry, and for the elaboration or manufacture of chemical,
medicinal, and like products." The equivalent French dictionary,
Le Trésor de la langue française says, "Premises equipped with
installations and apparatus necessary for manipulation and
experiment, performed in the framework of scientific research, or
for the analysis of medicines or materials, for technical tests or
for scientific and technical education." S&S note, among other

things, the rapid increase in the number of places identified as
laboratories in London soon after 1600. This indicates that a new
word was in use, but also that a new kind of place was coming
into being. It is an outgrowth of alchemical research. S&S contrast
laboratories with alchemical cabinets. The new laboratory is
supposed to be a public space (p. 61). It is not a cabinet within
which men explore the secrets of nature and keep their findings
secret. Yet at the same time, as S&S emphasize, these new places
were not open to any public, but only to an elite whose
membership became defined as the members of the new
scientific societies, their associates, and their employees, and
aspirants hoping to join their ranks.
I like certain words in the French definition of the laboratory
just quoted. A laboratory is a place equipped with installations
and apparatus necessary for manipulation ... The part of my book
dedicated to experimental science was called Intervening, and
intervening is close kin to manipulating. As I went on to
emphasize in chapter 13, it is also a place for the "creation of
phenomena" using apparatus especially constructed for that
purpose. That is exactly what Robert Boyle was doing with his
Air Pump. He was creating a phenomenon almost never in
existence in a reliable reproducible form before, namely a
vacuum in a container. His was not the first contained vacuum.
Torricelli had reasoned that if you take a column of liquid
contained in glass up a mountain, and observe the level of the
liquid fall, you are observing a void, an empty space above the
liquid. There were the famous spheres of Magdeburg.
It is quite astonishing how much effort and money early
Europe was prepared to invest in creating a vacuum, which had
no conceivable practical value. Boyle put his own fortune into it,
and the British Government treated it as a major research
enterprise and funded it generously. Other instruments also were

richly funded, for example the chronometer, but that was of
straightforward commercial importance. It would enable British
ships to navigate, and hence to exploit, the globe more reliably.
They would reach distant parts and return with loot from what
they saw as the West, the South, and the Far East. The prosperity
and industrial development of modem Europe depended on
trade and then a colonial empire, and hence on the chronometer
for determining latitudes. But it is hard to see the practical point
of spending a large amount of national treasure in making a
better vacuum.
12. The creation of phenomena

When I wrote about the creation of phenomena I realized this
would seem like a gross exaggeration. I said that we bring into
being phenomena that did not exist anywhere in the universe
before. But I had to temporize, allowing it might be better to say
that we purify phenomena, or we realize phenomena. I said that
the laser was a new kind of phenomenon that simply did not
exist before 1950 anywhere in the universe. Numerous physicists
protested, but I think there is a growing awareness of how much
this way of looking at things makes sense. In my own present
hobby, very cold atoms, and what is called Bose-Einstein
Condensate, we can now read statements like this: "This state of
matter could never have existed naturally in the universe. So the
sample in our lab is the only chunk of this stuff in the universe,
unless it is a lab in some other solar system." Well, that was in a
proud press release from the first laboratory that produced this
new state of matter, Bose-Einstein Condensate, in July of 1995.91

To call a phenomenon that exists for such a short length of time
in extraordinarily artificial conditions a "new chunk of matter"
sounds like an exaggeration, even to me. But it was assuredly a
new phenomenon. Within the year another laboratory had
produced the same state of matter, and by now, twelve years
after the event, many labs are doing so, for example Professor
Yu's laboratory at National Tsing Hua University, or Professor
Han at National Chung Cheng University.
Cold atoms may seem a long way from Robert Boyle and the
air pump for creating a vacuum. I see them as its confirmation,
one of innumerable present culminations. The old alchemists
dreamed of transmuting base metals into gold. They made many
empirical discoveries in the course of their vain experimental
explorations. Boyle transmuted the dream of transmutation into
the creation of new phenomena. And now we have achieved the
alchemists' fantasy. In 1995 we were able to transmute a
substance, in this case the element rubidium, into a new state of
matter, Bose-Einstein Condensate, not a gas, not a liquid, not a
solid, but something new in the history of the universe.
13. What Thomas Hobbes saw clearly

Sometimes it is the first step that counts. Boyle is my iconic
representative for the first step. Many, although not very many,
workers around the same time, in different parts of Europe, were
doing the same thing. I choose Boyle for two reasons. One is
wholly fortuitous, namely that S&S have written that book. The
other is not. One man, and one man only, saw just what Boyle
91 "Physicists create a new state of matter at record low temperature. Joint release
by the National Institute of Standards and Technology and The University of
Colorado, 7/13/95, still to be found on-line.
was doing, and protested vehemently. That man was that old
curmudgeon, the philosopher Thomas Hobbes. Given what I
said about Stillman Drake and Galileo as a builder of apparatus
in Chapter II, you can, if you prefer, use Galileo as the icon not
only for the Galilean but also for the laboratory style. Galileo had
to contest with Cardinal Bellarmine, who said nothing to make
us reflect on, now. Boyle had Hobbes.
S&S show that Boyle's work was contested at the very outset,
on both scientific and philosophical grounds. The figurehead for
the confrontation was the other man in the subtitle of their book,
Hobbes, Boyle and the Experimental Life. Hobbes was the aging
philosopher, the veritable creator of the European theory of the
state. Hobbes's book that inaugurated modem political science,
Leviathan, furnished part of the title of the book by S&S about the
air pump. Now Latour, like all sensible people, sees Hobbes as
the author of Leviathan, and as such, the spokesman for a new
age, a new kind of political society. In Canada, and most of the
English speaking world, every freshman studying political
science has to try to read Hobbes, which is by no means easy, for
the language now seems archaic or unintelligible to children
weaned on television. Latour brilliantly read S&S as displaying
the way that so-called modernity began, with a division between
the social and the natural, with one body of expertise and
intervention pertinent to one, and a quite different one germane
to the other. Hence his subtitle for the book, "Between science
(Boyle) and politics (Hobbes)". Latour rejects the distinction in
favour of what he calls Cosmopolitics, and he campaigns for a
Parliament of Things. Here I have a much more modest project in
hand, and make much more modest use of Hobbes.
In 1660, Boyle was 33 and Hobbes was 72, which happens to
be my age. If I disagreed, on fundamental principles, with a
brilliant, well-endowed, and well-supported young man of 33,
how could I possibly win? I would be dismissed as an old fogey,

and probably rightly so. Boyle was the wave of the future. One
important contribution of the S&S book is that in an appendix
Schaffer has translated what had previously been an obscure
Latin pamphlet written by Hobbes, denouncing Boyle.
Hobbes saw exactly what Boyle was doing, and he hated it.
He foresaw that laboratory apparatus for generating phenomena
was radically new. He was dead against it. This was not a quarrel
about the relative weight of empirical evidence as against
deductive proof. The question was more profound and more
consequential. What shall be evidence? Is it to be what we find
among us, bring home from abroad, chart in the skies-or is to
be what we make with laboratory apparatus?92
Are there not enough, phenomena Hobbes asks, already
"shown by the high heavens and the seas and the broad Earth?"
His interlocutor replies that "There are some critical works of
nature, not known to us without method and diligence, in which
one part of nature, as I will say, by artifice, that is, produces its
way of working more manifestly than in one hundred thousand
of these everyday phenomena" (p. 351). "Artifice" is exactly the
right word. Boyle was inventing clever apparatus, artificial
devices, to achieve a vacuum. The importance of mechanical
devices in the scientific imagination of early modem Europe has
long been emphasized. There was a famous clock in the city of
Strasbourg in which artificial men paraded around telling the
hours. In my telling of the story, the air pump put artifice to new
ends, namely the creation of new phenomena. Hobbes saw that,
92 Hobbes posed this question in Dialogus physicus, published in 1661, a retort to

Boyle's 1660 New Experiments Physico-Mechanical (and Boyle in tum confuted
Hobbes in 1662). This dialogue was not translated in the French edition of S&S,
hence adding to the shift away from the material side, which I emphasize, and in the
direction of the social side of the book.
feared it, and hated it.

The interlocutor states one fundamental, and almost never
stated, rationale of laboratory science. It was as valid in the 1660s
as it is in our present decade, 450 years later. "Such are our
experiments, in which one discovered cause can be fitted to an
infinite number of common phenomena." Hobbes cynically
asks, "and what are they," these common phenomena infinite in
number? Today we might add, as sardonically as Hobbes, "and
tell us more about how you perform this wonderful act of fitting
a common cause to your wonderful new phenomena!"
In the same dialogue, Hobbes was equally prescient about the
authority of the laboratory itself. Gresham College, which
became the Royal Society of London, the mother of all modem
scientific academies, prided itself upon the open demonstration
of phenomena to witnesses. Thanks to that, everything was
public in order that there could be no doubt about what
happened. Hobbes nastily asks, "Cannot anyone who wishes
come, since as I suppose, they meet in a public place, and give his
opinion on the experiments which are seen ...?" -"Not at all"
Boyle is made to reply (p. 350). Only the elite few could enter
these laboratories, these colleges and societies. Closed clubs are
hardly new. Hobbes's point is the tension between the rhetoric of
public verification and the fact of private and heavily controlled
membership.
This aspect is developed at length by S&S, and has become
one of the best known contributions of their book to social
studies of knowledge. It is further developed by Steven Shapin in
his important book mentioned in my first lecture, A Social History
of Truth. 93 Shapin contends that it was the nature of the
presupposition of mutual trust and reliability, among a tiny elite
93 A Social History ofTruth: Civility and Science in Seventeenth Century England,

of the population of London, namely the gentry, that made

modem science possible. Sincerity was a presupposition of the
enterprise, and it required a social setting to be in place. It is
striking how the present practice of peer review for scientific
journals has absorbed much of the assumption of sincerity right
across the board, in all the sciences, including mathematics. Now
we ask peers to review, whereas once we relied on peers, that is,
English aristocrats and landed gentlemen.
14. The decisive conceptual shift

How shall we complete Williams's schemas (*) and (**) for the
crystallization of the style of experimental exploration and
measurement into the laboratory style? Given my take, in terms
of the creation of phenomena, it is not first of all a matter of a
shift in ways of telling the truth, but of ways of finding out,
which in tum lead to a shift in the conception of what it is to tell
the truth about the world. The new way to find out is to build
apparatus to create new phenomenon and to expose purified
ones. There was, then,
(*) A shift in conceptions of what it is to find out about nature, and
hence how to tell the truth about it.
(**) This significant change took place in the seventeenth century,

and its best emblem is Boyle.
Boyle was the trailblazer for the laboratory, and we are still doing
what he did, building apparatus to purify or produce new
phenomena. Sometimes the phenomena are created in order to
test standing theories, and sometimes the phenomena precede
any theoretical understanding. Even testing theories is not the
usual point, but rather the aim is to achieve more understanding
of the predicted phenomenon. Sometimes there are what in
Representing and Intervening I called "happy families" when both

work hand in hand-as is really most commonly the case
nowadays.
I shall give an example from my hobby. Superconducting
substances were first produced at the laboratory of Kaemerlingh
Onnes in Holland in 1908. This led on to superfluidity, produced
by Kapitsa at the Cavendish lab in Cambridge in 1937. No one
had much of a theoretical understanding of these phenomena
until 1957, with the theory devised by Bardeen, Schieffer and
Cooper. Conversely, Einstein proposed what is now called
Bose-Einstein Condensation of an ideal gas in 1925. But no one
could create the phenomenon until 1995. Moreover Bose
condensation was not created in order to test Einstein's theory,
but rather as part of a long research programme to see how it
could be done.
The fulfilment of the research dream was a technical
achievement of the laboratory, using apparatus so brilliant that it
deserved to win its Nobel Prize. Now any group of about six
people can do it, using the same apparatus-ideas. In fact as I
understand it, at Professor Yu's laboratory, almost all the creation
of a Bose-Einstein condensate at National Tsing Hua University
was done by a single graduate student, Hung-Wen Cho, whose
nice English nickname is "Motor". This is a nice analogy to a
point made by S&S: Boyle's air-pump cost a fortune in Research
and Development costs, and in the deployment of the most
powerful scientific minds of the day. Yet in a couple of decades
you could buy a knock-off imitation in a Paris shop for about the
price of a cheap laptop today. Making condensate in 1995 won a
Nobel shared between two laboratories. Now a first-class
graduate student anywhere in the world, with quite modest
support, can do the same thing.
15. Ontology: theoretical entities

Twenty or thirty years ago, entities that crop up in theories
but which are unobservable were the hot topic of general
philosophy of science, and they still stir up a lot of controversy.
They were debated under the heading of scientific realism, which
will be the topic of the fourth and final lecture (though in the
plural, realisms and anti-realisms). I claimed that ontological
debates about the sciences are a by-product of the introduction or
crystallization of scientific styles of reasoning.
The Greek atomists had unobservable theoretical entities
aplenty, and they had non-trivial properties, hooks and eyes, for
example. Popper said that was metaphysics, not science. I tend to
agree with Popper, although not because I accept his strong
criterion of demarcation, that to be scientific something must be
testable. The Greek atomists had a wonderful series of
speculations, which have informed thought ever since, but there
was simply no good reason to believe their stories. Indeed many,
and perhaps most, physicists were instrumentalists about atoms
and molecules until the beginning of the twentieth century. Most
geneticists may have been instrumentalists about genes and
chromosomes until mid-twentieth century. These are debates
within the sciences; they are not ontological positions such as
realism and anti-realism. The anti-realist about theoretical
entities says that there are none, or that there is never any ground
for asserting the existence of any of them.
The ontological debate arises from that combination of
Crombie's second and third styles. The entities occur in theories,
yes, and so we might focus on his third style, namely, theoretical
and analogical modelling, with special emphasis on its
crystallization, the Galilean style. But the vast proportion, of
ontologically debated theoretical entities, arises from the theories
evoked to help us understand the phenomena elicited or created

by the laboratory style. I shall say more about this in the fourth
lecture. We may notice here, however, that Boyle's production of
a vacuum in a container was in part intended to establish the
reality of "atoms and the void." Once you had established a
vacuum, then, at least within his corpuscularian philosophy of
nature, you had established there must be atoms as well.
16. The vacuum

Boyle did two things. He made a device that produced a
partial vacuum in a container, thereby defeating nature, if we
believe in the saying, "Natures abhors a vacuum". He also
convinced everyone, that that was what he had done. Science
studies tends to emphasize the social, and so it attends to the
second, so brilliantly analysed by S&S. I am very much a
materialist, so I am interested in the material object, the actual
Air Pump, and its clones, which soon became cheaply available
all over Europe. And I am interested in the phenomenon it
created, the (partial) vacuum in a container.
There is a remarkably instructive irony here, so powerful that
I sometimes think of nature playing tricks upon us still, always
hiding what she has up her sleeve. The idea of a secret of nature
is a very deep and powerful one, and is arguably at the very core
of the emergence of the scientific tradition in the ancient
Mediterranean world, Egypt, Mesopotamia, Greece. I have been
much impressed by a book by a Paris colleague of mine, Pierre
Hadot: The Veil of Isis: An Essay on the History of the Idea ofNature.94
I believe that it is a version of the story of what I call "finding
out", which began long before the human race embarked on
anything recognizably scientific. It begins with an obscure saying
94 Harvard University Press, 2006, translated from the French of2002.

of Heraclitus, "Nature likes to hide."95

The irony is that Boyle, trying to produce a vacuum in a
container, seems to have been making an idle gesture to produce
absolutely nothing. It was important in his day, for sure. He
strongly advocated the so-called corpuscularian philosophy,
which conceived of air being composed of little balls, the atoms,
bouncing around in the void. It was important to show that the
idea of a void was coherent. But after that, the void is nothing, is
it not?
Classically speaking the void is nothing. That is, within
classical nineteenth century physics, before the advent of
quantum mechanics. Classically speaking, the void at absolute
zero is where absolutely nothing is going on, in-nothing. But
the classical picture of the universe is false. Quantum
mechanically, the vacuum is full of quantum fluctuations. The
classic textbook on the vacuum says in its preface:
According to present ideas there is no vacuum in the ordinary
sense of tranquil nothingness. There is instead a fluctuating
quantum vacuum.96
Perhaps there is a certain truth in the statement of a physicist

who writes about science with enthusiasm and who is not
disinclined to invoke a Creator of the Universe in some of his
popular writing. In a book with the characteristically hyped title
of Superforce, P. C. W. Davies writes,
The vacuum holds the key to a full understanding of nature.97
95 See my review in the London Review of Books, 10 May, 2007, which explains
some of the reasons why I think this book is so important. Online at
http://www.lrb.co.uk/v29/n09/hackOl_.html.
96 Peter W. Milonni, The Quantum Vacuum: An Introduction to Quantum
Electrodynamics, Boston: Academic Press, 1997, p. 104.
97 New York: Simon and Schuster, 1985, p. 104.
Maybe he is right! Even more extraordinary than what I have just

reported, is that the quantum vacuum is a hive of quantum
activity at absolute zero. That is where Bose-Einstein Condensate
lives, trapped by laboratory apparatus in a thermal cloud of
other atoms within nanokelvins of zero. That is, within 10-9
degrees of zero, what is rightly called the ultracold. This may
take us back to Heraclitus. What a wonderful place for Nature to
hide her secrets, in a vacuum at absolute zero!
That is another reason to welcome Boyle and his air pump
into the pantheon. The vacuum, which seems to be nothing at all,
matters beyond his wildest dreams.
17. The biotechnicallaboratory

I am able to make fun of myself: I have now performed the
feat of connecting, into a seamless whole, the mystical ancient
Heraclitus and the most recent work in the laboratory. I should
conclude by getting real. It will be good to conclude thinking
about modern laboratories that are very different from the labs
that descend from Robert Boyle. I shall turn from physics to
biology, and from long ago to now.98 And it is not biology to
which we should direct our attention, but biotechnology.
Unfortunately, I know nothing about biotechnology. I think
like a physicist. So when I was writing Representing and
Intervening, I talked to physicists. Physics was still, in 1983, the
heartland of the natural sciences. I had known, since 1962, where
the future lay, but it was not my future. My first job after
98 This lecture was prepared for a day-long workshop on Science, Technology and
Society. Professor Ruey-Lin Chen, who was my host at this workshop, wrote me
that "Some STS friends suggest that I should ask you to talk a little about your
'imagination' of East Asian techno-science at the end of your lecture to the STS
community." What I say here about state biotechnology in China is adapted from
what I said in response to this request.
graduating was in Cambridge, England, and the first year I was a

fellow of a Cambridge college, two older fellows of that college
shared the crop of Nobel Prizes awarded that year in biology and
medicine, for DNA and RNA. I remember being thought
off-the-wall, when, a very few years later, in a crowd of
humanists, I said, that is where the world's action is. So by 1970 I
was saying, "Forget the past: These new ideas and techniques
will change the world!"
Later in the 1970s everybody of good sense, humanist or not,
knew that. But I did not retool. So today I know less about
biotechnology than an able high school student. In this,
unfortunately, I am not so different from many of my younger
philosophical colleagues.
The world is being changed by biotechnology. When, decades
ago, Bruno Latour and Steve Woolgar did their ethnography of a
biochemical lab that shared a Nobel prize, they were working in
a very traditional academic sort of place. Latour said the chief
product of the lab was inscriptions, which in the end generated
the stability of the science. I protested that the chief product of
the lab was a new substance, a synthesized peptide, Thyrotropin
Releasing Factor.99 My blunt Materialism and his visionary
Idealism were apparent even in those days! But those were the
old days.
By 1974 there was a conference of all the recombinant DNA
workers in the world, held at a nice resort on the California coast,
Asilomar. They wanted ethical guidelines to determine what
types of research were permissible to avoid creating monsters,
bacteria, say, that would destroy the world's rice crops. In those
days they still worked with experimental samples that might
occupy a litre. In a few years, as one of the participants (who was
99 Ian Hacking, "The Participant Irrealist at Large in the Laboratory," British

Journal for the Philosophy of Science, 39 (1988): 277-294.
one of the Cambridge Nobel prize-winners whom I mentioned,

and who by that time was running the European Research
Council) recalled, they were shipping such material around in
tanker trucks. New techniques interposed, and the speed of
research speeded up ten-thousand fold. There are places all over
the world where ten story buildings are occupied by nothing but
gene-sequencing machines. What is being made there will create
wealth in health, agricultural and other industries. The results
are not inscriptions but substances and techniques.
The initial research was done in academic laboratories.
Venture capital companies began forming in California in the
early 1970s. Now they can be said to dominate the field; certainly
profoundly important work continues to be done in academic
labs, but often it is venture capital that leads the way. This is
where science studies researchers of today should be learning
and analysing. In fact, it is an anthropologist who has led the way.
He is a classical anthropologist, who trained with Pierre
Bourdieu, and who later played a large part in introducing the
work of Michel Foucault to an American audience.
I am referring to Paul Rabinow, a Professor of Anthropology
at the University of California, Berkeley.100 I most strongly
recommend the work of Rabinow to anyone in science studies
who is interested in recent biotechnology. He is a model to us all.
He collaborates often with Nikolas Rose, a sociologist at the
London School of Economics, who has created a major research
centre there called BIOS, which describes itself as a
multidisciplinary centre for research into contemporary
developments in the life sciences, biomedicine and biotechnology.
100 A Machine to Make a Future: Biotech Chronicles, Princeton, 2005. Anthropos

Today: Reflections on Modern Equipment, Princeton, 2003. French DNA: Trouble
in Purgatory, Chicago, 1999. And my favourite, Making PCR : A Story of
Biotechnology, Chicago, 1996.
He has recently been visiting biotech labs in Shanghai, and I shall

use some things he said to me in conversation in what follows.
For venture capital in California is not the end of the game; next
is coming (or has come) a wave of work from enormous state
laboratories in China. It is tempting to call it state capitalism,
because of the similarities in practice to Californian venture
capitalism. It is not worth debating whether we should say
California does the capitalist model while Shanghai does the
socialist model, or the state capitalist model. The point is, from
the point of view of science studies, that we have a new way of
doing science before us.
Are we talking about a new kind of laboratory? What amazes
many visitors to Shanghai biotechnology establishments is the
immense person power at the PhD or mid-career level. The
scientists do still often go to America or Europe for training and
to hone skills, but increasingly they return. I recall speaking a
few years ago to my Paris colleague Philippe Kourilsky, who was
then head of the Pasteur Institute in Paris, France's premier
establishment for fundamental biomedical research. He had just
come back from Shanghai where he officiated at the opening of a
sort of clone or franchise of the Pasteur Institute, erected on the
site of the old French quarter in Shanghai, preserving some of the
elegant nineteenth century facades. That is mere appearance;
within that complex they were, when he was there, creating a
massive research enterprise, in remarkable new buildings, and in
no time at all by French standards. "They will completely
surpass everything we do within less than ten years," said
Kourilsky.
In America, biotechnological research continues to be done in
a mix of traditional academic laboratories and venture capital
companies. But increasingly, especially in the US, it is the venture
capital companies that are doing the cutting-edge work. Some of
them have grown to become major corporations in their own

right: Genentech, Celera, Symbio. There is a big difference
between the capitalist laboratory and the academic laboratory. It
goes right back to what I mentioned in connection with trust and
seventeenth century English gentlemen-the system of peer
review. Everything that happens in the academic setting is
governed by peer review. Not only everything that is accepted as
knowledge and published, but also all research that is funded.
Writing and re-writing grant proposals is a prime occupation of
academic researchers. Moreover, what is called peer review, at
the funding level, is typically not peer review at all. It is not
review by equals, which is the meaning of "peer". It is more
often the review of young researchers by successful old former
researchers. They do indeed furnish all sorts of advice. They are
wise old men determining who in the next generations will be
supported. I exaggerate. They are not all wise or old or men, but
certainly most of them are men, and many are old. The star of
venture biotechnology is quoted as saying "The greatest obstacle
to scientific and technological innovation is peer review of grant
proposals."
Private companies are different. There the measure of success
is a patentable product. I spoke about a new kind of activity that
began 400 years ago, the deliberate creation of new phenomena.
If only Boyle had been able to patent his air pump in those days,
he would have recovered his investments one hundred fold.
Compared to the academic laboratory, the world of
venture-capital funded research may seem cutthroat. Many
traditionalists deplore it. A "generation" is sometimes counted as
a span of 33 years. Within one such generation private capital has
transformed biotechnical research. The Universities frantically
try to keep up, establishing an office of patents. This department
is usually on what is (in fact if not law) at least equal footing with
ordinary academic departments, within the University pecking

order.
So what about technoscience in East Asia? It will affect the
present generation of bioscience as much as venture capital
changed the last one. And here, to connect with my earlier theme
of the origin of peer review, we should be well aware of the
potential of the transformation of the capitalist world into
state-managed research with immense resources. It is most
noticeable under what is officially the "communist" or at any rate
socialist regime across the Straits, the Peoples' Republic. The
difference from Western capital is not capitalism but the almost
unlimited body of actually or potentially trained scientific
labourers, and a method of labour management that has already
taken over the lion's share of Western basic manufacturing. The
transfer of manufacturing, which began in Taiwan, Korea, and
Japan, has been multiplied beyond measure in the People's
Republic. Virtually all the children's toys sold in America are
"Made in China". The same methods of manufacture, that is,
very large enterprises organized in small production units, with
a local team spirit, are now being applied to biotechnological
research.
There are just two criteria of success in this research, as
opposed to the development or production side of the new
technoscience. Peer review is wanted, yes, and this is generously
provided by chiefly American premium scientific journals. If a
paper is accepted there, then a researcher and the team have
established their credentials. The second is the patentable
product.
In a state system, money can be thrown at research groups.
Most of them will fail to meet the two criteria I have just
mentioned. Perhaps they will be jettisoned, the labourers
assigned to minor tasks for the rest of their lives, little differently
THE LABORA TORY STYLE OF THINKING AND DOING / 127
from what happens in a ruthlessly efficient capitalist company.

The differences lie (1) in the potential workforce of researchers,
on which there is no upper bound, and (2) in the organization
and morale of teams within larger structures.
I have mentioned the phrase "form of life", and have
declined to use it much. Yet here we have a new form of science
performed within what it seems natural to call a new form of life.
It is a question of immense resources, combined with, in effect, a
rather different kind of accountability from what we have
previously known. That can at present be marshalled more easily
by a hegemonic state than by any other kind of regime. We do
have a standard of comparison, the Manhattan project, when the
U. S. government threw untold sums of national treasure in a
pretty indiscriminate way, in order to achieve results that a
decade earlier might have been predicted to require half a
century of research. The immediate result was the bomb. The
long term effect on the organization of science is well known to
all sociologists of the sciences, namely "Big Science." Perhaps
that is a new scientific life form. But "Big" may prove to be
miniscule in comparison with what may take place in South-East
Asia within a generation.
18. Two kinds of laboratory?

I have pre-empted the word "laboratory" for what I call the
laboratory style of reasoning. Even though I have emphasized
what we do in the laboratory more than most philosophers. I
have still clung to a propositional approach to the sciences.
Science generates knowledge, which is expressed by propositions.
I have characterized styles of scientific reasoning in terms of their
ability to shift concepts about what it is be truthful about
something. Boyle produced a vacuum, but also, new ways of
telling the truth about such theoretical items.

S&S teach that the word "laboratory" in English, came into
being with a way of doing science and a form of life that began in
the seventeenth century. To speak of previous alchemical
laboratories is, on their view, an anachronism. Yes, chemistry
laboratories came into being, they teach, but they were not a
smooth continuation of the cabinet of the alchemist. But perhaps
the new biotechnology laboratories will lead us to see things
differently. The alchemical cabinets were for producing new and
often miraculous substances. That is exactly what biotechnology
laboratories are for. They are not primarily for finding out what
is true, but for finding out how to do things. They are for making,
and telling the truth is a mere instrument in the complicated
process of turning old substances into new ones. And yes, this is
scientific thinking-to repeat Crombie's key phrase. But it is
thinking in a new key.
Lecture 4
.........
REALISMS
AND ANTIREALISMS
1. Plurals
You might expect this lecture to be titled "Realism and
antirealism," but I want to emphasize that many very different
philosophical doctrines bear these names. In the end, I shall focus
on realisms and antirealisms that are connected with the sciences,
but even there we have "scientific realism" about theoretical
non-observable entities, and types of realism called "Platonism"
in mathematics that would be better called mathematical realism.
Although my interests end up being narrow, it is important to
glance at a larger canvas, so I shall begin by mentioning debates
about which I have, here, nothing constructive to say.
I do not mean to imply that nothing can be said. This lecture
was presented at Soochow University, at a time when it had a
three year research group at work, under exactly this heading,
"Realism and Antirealism." As I understand it, quite a few types
of realism and antirealism were under discussion, so that my
emphasis on realisms in the plural was unusually germane.
2. Not realism versus nominalism

There is a series of debates in the history of Western
philosophy which begins with Plato and Aristotle, and continues
to the present. These were most intense, I think, among the great
Arab philosophers in Mesopotamia and North Africa, and then
among the scholastic Christian philosophers of the European
high middle ages. What is at issue is described in many ways,
and many things were up for debate. They have two faces, one
ontological and one grammatical. Traditionally, the emphasis has
been on the ontological aspect, on questions about what there is.
But both the Islamic and the Christian scholastic philosophers
often turned the discussion in the direction of grammar, just as
twentieth century analytic philosophers turned to semantics. In
ontology, it is always tempting to engage in what Quine called
semantic ascent, to tum from a discussion of things to a
discussion of names for things.
One advantage of semantic ascent, at least for twentieth
century thinkers, is that it allows one to think that "it is all a
matter of words." Sometimes the effect is positively noble, as in
Rudolf Carnap's "principle of tolerance." Carnap proposed that
all ontological questions were external to knowledge, and had to
do with the choice of a language. Meaningful questions arose
only within the framework of a language which carried
presuppositions of existence (or not) with it.l01 Quine's critique
has of course superficially disabled this easy tolerance, but it is a
persistent theme in much analytic philosophy. I try to avoid
semantic ascent, which evades philosophical issues but does not
make them go away. I do not make them go away, but I think it
101 Rudolf Camap, " Empiricism, Semantics, and Ontology,"~ Revue internationale
de philosophie, 4 (1950): 20-40.
REALISMS AND ANTIREALISMS /131
helps to understand how they come into being.

One way to pose an ontological question is to take the
example of justice. Plato, as you know wrote the Republic, his
classic work of political philosophy, around the concept of justice.
Socrates keeps on asking, what justice is. His companions give
various examples of justice, but he keeps on protesting that they
are just examples, not justice itself. They propose definitions of
justice, and he interjects with counterexamples of two kinds.
There are cases that satisfy the definition, but which are examples
of unjust acts or arrangements. He also cites examples of just acts
or just arrangements that do not satisfy the definitions.
Sometimes one wants to say that it all boils down to the
question, "Is there anything in common among just acts and
arrangements, other than that we recognize them as just, other
than that they are called just?" The extreme nominalist replies,
"No!" There is nothing in common between just acts, except that
a Greek equivalent of the adjective "just" is applied to each of
them. There is no property of being just over and above the word
"just" and its use (in whatever language you may be speaking).
Of course the necessary parenthesis about a particular language
already makes one wonder if these questions are
language-relative, for the use of more or less inter-translatable
words in different languages is at best only approximately the
same. Leave such quibbles aside for the moment. Such extreme
nominalism could well be called "name-ism". Not very many
philosophers have been willing to state and defend extreme
name-ism, but some come close. I think of the English founder of
political science, Thomas Hobbes, in the seventeenth century, and
the American pragmatist Nelson Goodman in the twentieth.
At the opposite end of a spectrum is extreme realism, which
holds that justice itself is a real entity, over and above
arrangements, judgements, or acts that are just. Justice is not
merely the class of all just entities, but something that itself exists,
independently of any acts, arrangements, decisions or whatever
individual items you may think of. It is not merely an ideal for
which we may strive, but something that actually exists, even if
we never in fact attain it. This is one kind of realism. It is
ontological in character; it is about what exists, or to pound the
table, what really exists. This is not a body of dispute that yields
readily to semantic ascent, although doubtless Carnap's tolerance
is called for.
Times have changed. I spoke to one well-known British
philosopher who writes about realism, and asked if anyone still
spoke of realism as the opposite of nominalism. He tartly replied
"no" in the rather abrupt manner of a scholar accustomed to
dismissing the tedious questions of inept undergraduates. What
then did he think that realism/antirealism debates are about? See
the section about Michael Dummett below.
I do not believe that this philosopher's brusque answer was
justified. It may be regrettable, but philosophers of various types
will go on posing ontological problems, and say that they are
debating realism versus nominalism. Nominalism is very
properly taken to be one kind of antirealism. But since I am
concerned with the sciences, I shall, so far as is possible, avoid
this kind of realism and antirealism, one that can be expressed in
terms of universals. That is why I headed this section, "Not
realism versus nominalism."
3. Universals
Sometimes the abstract, real, entities that a realist claims to
exist are called universals. Hence what is at issue in such
realist-nominalist debates is sometimes called "the problem of
universals." That is a scholastic way of putting things, which
REALISMS AND ANTIREALISMS / 133
directs attention to grammar and semantics and away from

ontology. In fact the Latin word translated as "nominalism"
appears to have been invented by Spanish scholars in 1492, the
year that other Spaniards financed Columbus, captain of the first
European ship to reach the "West Indies" and Central America.
Bertrand Russell has a lovely seemingly simple discussion in
"The World of Universals", chapter IX of his 1912 primer, The
Problems of Philosophy. "When we examine common words," he
writes, "we find that, broadly speaking, proper names stand for
particulars, while other substantives, adjectives, prepositions,
and verbs, stands for universals." A little later on the same page,
he cheerfully said that, "Seeing that nearly all the words to be
found in the dictionary stand for universals, it is strange that
hardly anybody except students of philosophy ever realizes that
there are such entities as universals."
Someone reading Bertrand Russell might ask, well, what if
our language did not break up this way into substantives (that is,
common nouns), verbs, and so on? Would we be so inclined to
say that nearly all the words in the dictionary "stand for"
universals? Is it at all natural to even pose this problem, if one
speaks Chinese rather than a European language? In other
contexts Bertrand Russell himself argued that a great deal of
Western philosophy was simply wrong-headed because it was so
locked into the subject-predicate grammar and the corresponding
substance-attribute metaphysics. There is now a substantial
literature on this subject in English-language philosophy journals.
One such contribution is by Professor Wenzel of National Chi
Nan University.102 His philosophical work makes extensive
102 Christian Helmut Wenzel, "Chinese Language, Chinese Mind?" in C. Kanzian

and E. Runggaldier (eds.), Cultures: Conflict-Analysis-Dialogue, (Proceedings of
the 29th International Ludwig Wittgenstein Symposium, 2006), Frankfurt am Main:
Ontos, 2007, 296-314.
references to a recent anthropological study by Nisbett.l03 But

his real debt is proudly stated; he thinks that the clearest thinking
about language and thought, with special reference to Chinese
and German, is to be found in a long essay penned in 1827 by the
philosopher, philologist, William von Humboldt (1767-1835).104
Wenzel is planning to complete a book on this topic.
I do not here take a position on these issues, but I think it is
worth reflecting on this question: Might the roots of realism
debates lie not in universal aspects of nature and the human
mind, but in particular linguistic structures? More specifically,
might the issues of realism versus nominalism arise from
structures of European languages? Nietzsche probably thought
so.
4. "Ontology recapitulates philology."

Quine used this aphorism as an epigraph for his 1960 book,
Word and Object. He attributes it to James Grier Miller, one of the
founders of Systems Theory. I actually heard the phrase earlier
103 R. E. Nisbett, The Geography of Thought: How Asians and Westerners Think
Differently ... and Why, New York: Free Press, 2003. For a summary of criticisms of
Nisbett's type of results, see Geoffrey Lloyd, Cognitive Variations: Reflections on
the Unity and Diversity of the Human Mind, Oxford: Clarendon Press, 2007.
(Reviewed by me in the London Review of Books, I November 2007.)
104 William von Humboldt's systematic ideas about the heterogeneity of languages
is to be found in his famous monograph, On Language: On the Diversity of Human
Language Construction and Its Influence on the Mental Development of the Human
Species, Cambridge, Cambridge University Press, 1999. (Translated from Über die
Verschiedenheit des menschlichen Sprachbaus und seinen Einfluss auf die geistige
Entwicklung des Menschengeschlechts, 1836). Wenzel, however, attends to a text
that is less well known, a letter written in 1827 to Jean-Pierre Abel Rémusat, who
had been professor of Chinese at the College de France in Paris since 1814, the first
chair of Chinese in Europe. Lettre á M. Abel-Remusat: Sur la nature des formes
grammaticales en general et sur Ie genie de la langue chinoise en particulier, Paris:
Dondey-Dupré, 1827. (Reprinted by "Elibron Classics" 2005.) Wenzel provides
references to German and other scholarship on this letter.
than Quine's book, in lectures at Cambridge University by John

Wisdom, in the fall of 1956, but of course Grier may have coined
the phrase. (But systems theory was not exactly Wisdom's cup of
tea; I doubt he had ever heard of Grier.) Whoever first coined the
phrase, it was an obvious, nay inevitable, joke. It comes from a
catchphrase proposed by Ernst Haeckel in 1868: "Ontogeny
recapitulates Phylogeny." Haeckel was a great advocate of
Darwin's natural selection, and its prophet and spokesman in
Germany. "Ontogeny" is the growth of an organism, for example
of a human from fertilized egg to foetus to infant to adult.
"Phylogeny" is the evolutionary history of a species. So Haeckel
was urging that the growth of individual organisms from
conception followed the pattern by which the species, to which
that organism belongs, evolved. (This was once a popular idea,
but is no longer so.) At any rate, our philosophical aphorism is a
play on the biological one.
The maxim sounds cute but what does it mean? Quine
probably meant that by a suitable choice of grammar and
language, you do not need to commit yourself to more than a
very sparse ontology. Prune your syntax and your semantics, and
philosophy will mutely follow orders. Thus the maxim served
Quine's variety of nominalism well.
One could, however, use the aphorism to make a very
different suggestion: Ontological problems are by-products of the
grammar of the language spoken by the philosopher. That is not to say
that I subscribe to Quine's "ontological relativity." Quine held
that ontology is relative to a language. I do not make any claim
about ontology. I suggest something about the origin or source of
ontological problems. Thus I step back one pace from the
philosophies of Quine and his opponents. Thus I am proposing
something like, "the linguistic relativity of philosophical
problems."105
5. Against too much linguistic relativity: Nietzsche and

Chuang Tzu
The idea of linguistic relativity is an all too tempting a notion
to play with. People on their second or third exposure to
philosophy love to dabble in relativism of this and other sorts.
Although I am sympathetic, I do not want to overdo it. Some
philosophical concerns or instincts seem rather universal. One of
them is a fascination with names. It seems to transcend difference
in grammar, and to be recognizable in cultures whose linguistic
expression of names is very different from anything I well
understand. Without going too far afield, we find, in many
civilizations, philosophers puzzled by names and tending to
make quite remarkable comments about names. I shall just
mention two statements, one from 19th century German, and one
from Chinese written more than 2100 years earlier. Both appear
to be concerned with names and reality-and hence in some way
with realism and antirealism. First, Nietzsche in The Gay Science:
§ 58. Only as creators!- There is something that causes me the
greatest difficulty, and continues to do so without relief:
unspeakably more depends on what things are called than on what
they are. The reputation, name, and appearance, the usual
measure and weight of a thing, what it counts for-originally
almost always wrong and arbitrary, thrown over things like a
dress and altogether foreign to their nature and even to their
skin-all this grows from generation unto generation, merely
because people believe in it, until it gradually grows to be part of
105 For other senses of "linguistic relativity", see encyclopedia articles such as the
one by C. Swoyer in the Stanford online encyclopedia, http://plato.stanford.edu/
entries/relativism/supplement2.html, or L. Boroditsky, "linguistic relativity" in L.
Nadel (ed.), Encyclopedia af Cognitive Science, New York: Wiley.
the thing and turns into its very body: what at first was
appearance becomes in the end, almost invariably, the essence and
is effective as such! Only a fool would think it was enough to point
to this misty mantle of illusion in order to destroy the world that
counts as essential, so-called "reality"! We can destroy only as
creators! - But let us not forget this either: it is enough to create
new names and estimations and probabilities in order to create in
the long run new "things".106
One main point of this aphorism is, as its heading states: Only as
creators. A sub-theme then must be that we can undo a named
idea only by creating some positive concept in its place.
Deconstruction for its own sake is self-indulgent play. It is,
however, for the other thought that I single out this passage:
Unspeakably more depends on what things are called than on what
they are. And: It is enough to create new names and estimations
and probabilities in order to create new "things".
I happen cautiously to agree with Nietzsche when it comes to
names of kinds of people. I have used this quotation in a paper
stating my current position on an entirely different interest of
mine, namely the classification of people and the interaction of
classifications with the people themselves.107 But, subject to
qualifications again, I am not inclined to this way of thinking
when it comes to names of things. Needless to say, people are
"things", but it is the way in which people conceptualize
themselves, and other people conceptualize them, that makes the
difference. I am, if you want a label for me, very much a
materialist about non-sentient things, as will become more clear
towards the end of this talk. So Nietzsche's aphorism resonates
106 Friedrich Nietzsche, The Gay Science: with a Prelude in Rhymes and an
Appendix of Songs, translated by Walter Kaufmann from the 2nd ed. (1887), (New
York, Vintage Books, 1964), § 58.
107 "Kinds of People: Moving Targets," Proceedings of the British Academy 151
(2007): 285-318.
for me only when it comes to names for kinds of people; but it is

not an appealing doctrine about things.
I wish to put beside Nietzsche a small passage, equally out of
context, from Chuang Tzu, in what are called The Inner Chapters.
As I understand it, it is presumed that the Taoist philosopher
wrote these chapters himself, and they are not the work of later
commentators.
I shall take two remarkable translated sentences out of
context:
(1) A name is only the guest of reality.108
Pause to reflect on this. It is a beautiful saying, quite
regardless of what Chuang Tzu meant, or of whether it is a
correct translation of a Chinese sentence written 2300 years ago.
Pierre Hadot has said that 'To write the history of thought is
sometimes to write the history of series of misinterpretations'.109
He was telling the history of an even older adage attributed to
Heraclitus, "Nature likes to hide." It is important to the history of
Chinese philosophy to know exactly what Chuang Tzu meant,
but whatever he meant, that English sentence is remarkably
powerful, all by itself, out of context.
We usually think that misinterpretation is a terrible thing. We
ought to find out what the sage really meant! So we should, but
we should also welcome innovative misinterpretations that endure.
Nature herself is said to evolve by fruitful mistranscriptions of
108 Chuang Tzu, The Inner Chapters, trans. David Hinton, Washington, D. c.:
Counterpoint, p. 7. For the record, Par Lagerkvist (1891-1974), Nobel Prize for
literature 1951, published his autobiographical novel Guest of Reality (Gäst hos
Verkligheten, Stocklholm: Aldus/Bonnier 1967) in 1925. For a discussion of Chuang
Tzu even more out of context than mine, see, Mark Berkson, "Language: The Guest
of Reality-Zhuangzi and Derrida on Language, Reality, and Skillfulness," in P.
Kjellberg and P. J. Ivanhoe, Essays on Skepticism, Relativism, and Ethics in the
Zhuangzi, Albany: State University of New York Press, 1996. 97-126.
109 Pierre Hadot, The Veil of Isis: An Essay on the History of the Idea of Nature,
Harvard University Press 2006, p. 14.
genetic code. Misinterpretation can, on occasion, be more

creative than merely sound interpretation. Whatever I say about
a name being the guest of reality will probably be a
misinterpretation of what is possibly a mistranslation of what is
possibly an incorrectly transcribed ancient sentence. That does
not bother me. As it stands, (1) is exquisite; it also makes you
think. I hear it, first, as a strong commitment to a reality, wholly
independent of, and prior to, naming, classification, and any
human intellectual activity. Reality is just there, and occasionally
it welcomes this or that name as a good fit-but only as a guest.
That idea is not exactly "realism" in any of the philosophical
senses at which I gestured earlier. We might say that it is truly
realist, or even mystically realist, expressing a deep respect for
what is. A few pages on, however, we read another sentence that
seems to be a radical expression of nominalism:
(2) Naming things makes them real.110
That sounds just like Nietzsche! What follows next,
immediately changes the tone.
Why real? Real because real. Why nonreal? Nonreal because
nonreal. So the real is originally there in things, and the sufficient is
originally there in things. There's nothing that is not real, and
nothing that is not sufficient.
I shall take all this to express what I might call "really-real

realism," or, better, "really-realism." Unlike scholastic realism,
which expresses a commitment to the "reality" of universals,
concepts, and classes, these aphorisms appear to express a
profound respect for a reality that stands complete, no matter
what humans do or think.
Perhaps an English-speaking 21st century analytic
philosopher can see (1) and (2) as compatible in this way: A
110 The Inner Chapters, p. 23.

complete reality is prior to any conceptualization. When a name

is a welcome guest of that reality, it picks out a thing, or a kind of
thing, which is thereby real because it is a guest of that complete
reality. The same philosopher may suggest that the ancient sage
thinks of the complete reality as the ground of what we say and
know about things, but at the same time rejects the European
Enlightenment bid to provide foundations for knowledge. This
brings Chuang Tzu and Nietzsche surprisingly close together.
This little aside may indicate that I do not take too seriously
the idea that philosophical problems are relative to a language
group. To avoid misunderstanding, I should also say that I do
not take too seriously the idea that the German philosopher and
the Chinese one, separated by two millennia, had exactly the
same concerns.
10. General ontology and the special sciences

To return to my main theme, I said I do not want to talk
much today about realism versus antirealism in general. This is
because I am concerned with questions about existence that arise
in the sciences. For example, around 1970 analytic philosophers
of science began to talk about "scientific realism." They meant
the question whether the unobservable theoretical entities, such
as those postulated in physics, exist. Scientific realists said yes, an
entity exists if a theory about it is true, even if the entity is in
principle incapable of being observed. Scientific
antirealists-such as positivists and instrumentalists-said the
entities do not exist. They are only instruments of thought; the
terms that we use to express them do not denote entities that
exist. Bertrand Russell cast this in a precise form by claiming that
whenever possible we should substitute logical constructions for
inferred entities.
I interject here that positivism can be deadly. Historical

positivism, introduced in France by Auguste Comte in the 1830s,
was once a live philosophy, complete with churches which would
replace those of organized religion. I have been at one such
church, in the capital of the most southern province of Brazil,
Porto Allegre. The city square of that capital is like many, with an
imposing government building, with a cathedral, etc., but in the
centre is the magnificent gilded monument to positivism and
progress, many metres high. In the 1890s, after the restoration of
elected government, the first two governors were from the
Positivist Party. The second one died of small-pox because he did
not believe in germs and was not inoculated. He did not believe
that theoretical entities really exist.
A more modest and reflective positivism escapes that danger.
Bas van Fraassen's constructive empiricism maintains only that
we cannot assert that this or that theoretical entity exists. We
should not believe they exist, for we have no grounds for
believing or asserting that they do. Worse, in most cases, we
could not have any such grounds, no matter how empirically
adequate our theories are. But we are still allowed, by this
narrow philosophy, to act as if the theories are true-that is, when
theories are empirically adequate. When theories about
small-pox are empirically adequate, we should get inoculated
against the small pox, and so avoid the death that awaits the
heroic positivist.
I would like to maintain that realism debates in the special
sciences have nothing much to do with more general ontological
debates, or with any problem about universals. Certainly van
Fraassen does not sound much like a classic mediaeval
nominalist such as William of Ockham. But I have a difficulty in
pushing this thought very far. This is because I take the
traditional view that mathematics is a science. "Platonism" in
mathematics is a version of the realist doctrine about mathematical

objects, and anti-platonism is an antirealist doctrine. Both, I
claimed in lecture two, are by-products of the introduction of the
mathematical style of reasoning. We are right back to ontology,
for the abstract objects of mathematics were one of the pillars of
Plato's entire realistic philosophy.
For a nice confirmation of this, let us go back to Russell's
handy Problems of Philosophy. In my second lecture, "Where do
mathematical objects come from?", I wanted to show how
European philosophy had been obsessed with mathematics from
the word go. I started the discussion by quoting from Russell's
Problems: "The question which Kant put at the beginning of his
philosophy, namely 'How is pure mathematics possible?' is an
interesting and difficult one, to which every philosophy which is
not purely sceptical must find an answer." 111 That is from
chapter seven, on a priori knowledge, the chapter just before the
chapter on universals with which I began this talk. That chapter
ends with a lead on to "the following chapter, where we shall
find that it solves the problem of a priori knowledge, from which
we were first led to consider universals."112
Thus Russell saw realism versus antirealism, insofar as that
concerns the ontological reality of universals, as intimately
connected with realism and antirealism about mathematical
objects. Historically speaking he was beyond question correct. So
I cannot keep the general metaphysical issues so far away from
the sciences as I would have liked.
111 Bertrand Russell, The Problems of Philosophy, London: The Home University
Library, 1946, 84. Kant's question is stated in The Critique of Pure Reason,
translated by Norman Kemp Smith, London: Macmillan, 1929, 56 (B 20)
112 Ibid., p. 100.
7. Dummett's type of antirealism

I am not yet finished with types of realism that I shall not talk
about. (But which I spend too much time talking about!) Michael
Dummett began his career as a logician and philosopher of
mathematics. He was deeply attracted by intuitionism and
constructivism towards mathematics, yet at the same time
resisted Wittgenstein's Remarks on the Foundations of Mathematics,
at least in the way that he read them. He sharply distinguished
the denial of the law of the excluded middle, from the denial of
bivalence, the doctrine that propositions must have one of two
truth values. He turned realism into a thesis about bivalence.
Dummett applied this notion across the board. Consider a
man, now dead, who never in his life had the opportunity to
display or even to intimate whether he is courageous or not.
Then according to Dummett-and I tend to agree-the
proposition, "He was a courageous man," is not true. Likewise,
"He was not a courageous man," is not true. Thus on Dummett's
construal of realism as a commitment to bivalence, he is
antirealist about the courage of this man.
Dummett flirted with various antirealist doctrines about
history, and in general encouraged an anti realist discourse of
anti-bivalence that flourished particularly among British
philosophers a few years ago. Once again, that is an antirealism
that I shall pass by.
8. An allusion to Richard Rorty

The late Richard Rorty, who made a lecture tour in Taiwan
some years ago, thought that the entire family of
realist-anti realist debates was misguided. In particular he
thought that the ones that were going on in America during the
1980s were pointless. I recall him saying, in conversation, in what
was for him a pretty bitter tone of voice, "realism is

mickey-mouse." In the slang of those days, that meant too
easy-that the entire discussion was not worthy of a children's
comic.
Perhaps I should interject here a word about my attitude to
Rorty's philosophy. This is because the book based on his lectures
here in Taiwan has, I am told, been widely received and
acclaimed.l13
From the very beginning I have always been ambivalent
about his way of doing philosophy. I do not agree with his
running many kinds of discourse together, as part of an
undifferentiated conversation of mankind. Yes, it is an important
value to keep the conversation going, in as untrammelled a way
as possible. But I betray a very different instinct from Rorty's. I
am a splitter, a divider, an analyst; I believe that the making of
distinctions is not an end in itself, but it is one without which no
ends can be served. I always put plurals into titles! Example: I am
not talking about "realism" but about "realisms"
I was asked to write a puff for Rorty's last book, the fourth
volume of his collected essays published by Cambridge
University Press, which came out just before he died this year. I
spent a long time writing 60 words in order to express my
admiration but not to conceal my ambivalence. In the end, the
Press decided to put only one puff on the back of the book, mine.
I shall quote myself, for these words really demanded a lot of
work:
Wise and immensely readable, these essays hammer home John

Dewey's theme: Philosophy matters when it changes what we
want to talk about, and how we do it. In detail, they usually seem
113 Richard Rorty, Hope in Place of Knowledge: The Pragmatics Tradition in

Philosophy, Taipei: Institute of European and American Studies, Academia Sinica,
1999.
to me to be blissfully right or infuriatingly wrong: the fact that

they can itch so makes me deeply suspicious of a lot of the
philosophy that I hold dear.114
That is how I feel, and it is even how I feel about the proposal
that realism/antirealism debates are "mickey-mouse". Rorty said
that to me 20 or 30 years ago. It annoyed me then, but even so I
was inclined to agree with it, and hid from myself the fact that I
agreed with him. I am even more thus inclined today. I am also
inclined to make more use of Nelson Goodman's epithet, and to
speak of irrealism, using it to express a certain indifference to
traditional realist and anti realist debates.
9. Arthur Fine's Natural Ontological Attitude

Rorty referred to Arthur Fine as his "favourite philosopher of
science." 115 Fine's paper, "The natural ontological attitude,"
begins with the sentence, "Realism is dead."116 Rorty correctly
summarizes: Fine asserts "that we should be neither realists nor
antirealists, that the entire realism-antirealism issue should be set
aside." Well, I agree, it should be set aside. But it won't let itself
be shunted into retirement. Fine doubtless intended "Realism is
dead" to recall Nietzsche. Well, whatever else is the case about
God, existence-of-God debates have not gone away in Western
thought, since Nietzsche wrote his famous aphorism. I have
found it worth while to come to some understanding of the
appeal of both realisms and antirealisms. That has been one
theme of my first three lectures. Call it the by-product thesis. In
114 From the back cover of the paperback, Richard Rorty, Philosophy as Cultural
Politics. Philosophical Papers IV, Cambridge University Press 2007.
115 Ibid., p. 133.
116 Arthur Fine, "The Natural Ontological Attitude," in his The Shaky Game:
Einstein, Realism and the Quantum Theory, Chicago: University of Chicago Press,
1986.
the ontological debates that beset the sciences, the various types
of objects rejected by antirealists, the pound-on-the-table objects
asserted to exist by realists, are all by-products of the styles of
scientific thinking by which they are introduced.
I do not think that realism is dead. I am not interested either
in killing it off or in giving it life support. I am curious why so
many kinds of realism keep on thriving. But here I restrict my
query: why do realism debates continue to thrive among
philosophers who reflect on the sciences, although they do not
matter at all to scientists at work on their fields of expertise?
10. So what about my experimental argument for entity

realism?
My book of 1983, Representing and Intervening, came in two
parts divided by what I called a "break". The second part
included an argument for realism about unobservable theoretical
entities. Did I then think that realism was important? No. Have I
changed my mind? Hardly at all. First, let me highlight a couple
of lines from page 2 of the book.
Disputes about both reason and reality have long polarized
philosophers of science. [... J
Is either kind of question important? I doubt it. We do want to

know what is really real and what is truly rational. Yet you will
find that I dismiss most questions about rationality, and am a
realist on only the most pragmatic of grounds.l17
Is either kind of question important? To repeat myself, I doubt it.

Yes, I still doubt it, 25 years after these words were published.
Incidentally, since I have just mentioned Rorty, the great
neo-pragmatist of our times, I did not say I myself was a
117. R&I, p. 2.
pragmatist, that is, a card-carrying member of the

Peirce-James-Dewey philosophy. I said that if I was a realist at all,
it was on pragmatic grounds, not on grounds of pragmatism.
Peirce remains a hero of mine, just as Dewey was always Dick
Rorty's role-model. I have explained all this in a recent paper
explaining why I am not a pragmatist.118
My book was a plea for experiment. The key sentence was,
Experiment has a life of its own, independent of theory. The first
part of the book dutifully explained in a slightly light-hearted but
academically sound way, what one needed to know about recent
debates in philosophizing about theoretical science. When the
book was published, philosophers were simply not interested in
experiment. I wanted to open a door. Scientific realism was all
the vogue in 1980. I used the raging controversy about scientific
realism as a peg on which to hang my pleafor experiments.
I was lucky. Unknown to me, lots of young people were
beginning to open the door in their own ways. By the end of the
decade, experiment was positively fashionable among
philosophers, historians and above all sociologists of science.
Thus there wasno incompatibility in my own mind, between
my experimental argument for scientific realism, and what I said
on page 2, that I doubted that realism (or rationality) debates
were important. Sometimes I wish I had asked Rorty for
permission to quote his remark, "realism is mickey mouse."
11. What was the experimental argument?

(a) If, (NOT "only if")
Many people remember the sentence early in the book, "If

you can spray them, then they are real." This was made in
118 "On not being a pragmatist: eight reasons and a cause," in Cheryl Misak (ed.),
The New Pragmatists, Oxford, 2007, 32-49.
connection with an electron gun that sprayed polarized electrons

in order to attain certain well understood effects on a
super-cooled super-conducting super-fluid ball of niobium.
An astounding number of readers first took this to mean not
only what it says, but also, "Entities are real only if you can spray
them." I never thought that. It simply never occurred to me so I
did not guard against it. Ernan McMullen was one honest
philosopher of science who confessed to me, with apologies, that
he had really made that elementary error of reading.
More generally, my thought was that the standard debates
about scientific realism were always indecisive because they
were conducted at the level of theory and often of semantics.
They always were conducted as if the debaters fully subscribed
to what John Dewey called "the spectator theory of knowledge".
As if we just look and talk, and never do anything in the sciences.
My general theme was, only if you get away from the spectator
theory and start realizing that science is doing as well as
reasoning, will you lose interest in the debates that so flourished
in the 1980s.
12. What was the experimental argument?

(b) The strongest argument (NOT a conclusive one)
My experimental argument was given in chapter 12. The first

sentence of that chapter was:
Experimental work provides the strongest evidence for
scientific realism.
I continued:
This is not because we test hypotheses about entities. It is because
entities that in principle cannot be 'observed' are regularly
manipulated to produce a new phenomenon and to investigate
other aspects of nature. They are tools, instruments not for
thinking but for doing. (p. 262.)
And so on. I went on to make the consideration, to my mind,

rather compelling. My intention was always to escape the view of
the spectator. I go into the lab in order to look at a team of
experimenters. I go in to try to participate, to be to the best of my
ability (which is not much) a participant observer. Hence I would
have the experience of trying to think about how to make a piece
of apparatus use an entity in order to do something else. Note
that in the preface to the book I thank a colleague who welcomed
me into his laboratory in order to learn how to use a microscope,
and how I broke a lot of glass trying to make things. Today I have
a research assistant, a PhD student in Toronto, Eran Tal, doing
just the same thing in a cold atom (quantum optics) laboratory,
aided by two physics grad students.
I have never thought of the experimental argument as more
than, as the opening sentences states, the strongest argument for
scientific realism. The argument can in a particular case carry
conviction, it can be compelling, but it is not thereby conclusive.
I also thought that if an entity has not yet got to the stage of
being manipulated in order to be a tool for finding out about
something else, then we do not yet have a compelling argument
for its existence. I did not intend to say, that in those
circumstances that we have no argument, or that we cannot
reasonably think that the entity exists. I certainly did not intend
to say, "failing manipulability, the entity does not exist."
13. A strong argument can always have a false

conclusion.
Teachers of elementary logic have to remind their pupils that
excellent arguments can from time to time lead us astray. The
strongest type of argument for entity realism need not be
conclusive. Usually critics mentioned use old chestnuts for their

counterexamples: phlogiston, or the all pervasive aether of early
electromagnetic theorizing. It is not clear to me that one
manipulated these conjectured "entities" to interfere with and
learn about other objects, or that one designed instruments to
capitalize on one's ability to use them. So let me take a (counter-)
example I understand in greater detail, and which fits the form of
my argument better.
My quoted statement about the "strongest evidence" may be
ambiguous. Scientific realism is not a carte blanche doctrine that
every entity propounded in any theory is "real". It is the doctrine
that when one is using such an entity to investigate other entities,
antirealism about it does not make sense. It is not that one is
reasonably convinced that it is real, and hence that the general
doctrine of the reality of (some) theoretical entities is true. The
stance of the experimenter could be called a performative
argument.
My original experimental realism comes to this:
(a) Experimental work provides the strongest evidence for the
reality of an unobservable theoretical entity.
I always had in mind realism about this or that entity-using my

trite example of the electron, or more interestingly polarized
electrons. I now think I should not have spoken of "evidence" at
all. For that makes it seem as if we are "inferring" the existence of
an entity that we use. In short, it invites a return to what John
Dewey called "the spectator theory of knowledge." Science is not
a spectator sport. It is a game to be played, and those who play
hockey do not infer the existence of the puck: they hit it, they
move it, they aim, they usually miss, but sometimes they score
goals with it.
Because I wrote in terms of best evidence, most commentators
on my discussion have refused to leave the stands. They argue

that the argument must be a case of inference to the best
explanation. Nothing could have been further from my mind. It
is spectators who explain.l19
I choose arbitrarily one recent clear paper that sums up two
misreadings of my idea in a single sentence. Mauricio Suarez
state my "metaphysical claim" as,
Metaphysical Experimental Realism: Manipulation is a sufficient
condition on reality: x is real if x can be manipulated.120
First, I have to repeat that I said "strongest evidence," not that

manipulability is a sufficient condition, a proof of the reality of x.
Second, and much more important, to my way of thinking, my
statement did not say manipulated, full stop. What I wrote was
quoted above. I spoke of entities that are, regularly manipulated
to produce a new phenomenon and to investigate other aspects
of nature. They are tools, instruments not for thinking but for
doing.
Manipulation for the joy of manipulation is not much
different from spectator sports. I said manipulated with a purpose.
Or rather several purposes. To produce new phenomena. To
investigate other aspects of nature. To find something out.
Moreover, do not forget that modifier, regularly. A one-off
manipulation to try to produce new phenomena does not cut it.
Suarez is at pains to distinguish a metaphysical from an
epistemological version of experimental realism. The latter states
119 Some of the critiques are, D. Resnik, "Hacking's Experimental Realism,"

Canadian Journal of Philosophy. 24 (1994): 395-412. R. Reiner and R. Pierson,
"Hacking's Experimental Realism: An Untenable Ground," Philosophy of Science
62 (1995): 60-69.
120 "Experimental Realism Defended: How Inference to the Most Likely Cause
Might be Sound," in Stephan Hartmann, Carl Hoefer, Luc Bovens (eds.), Nancy
Cartwright's Philosophy of Science, Routledge, 2008.
that "Manipulation is a necessary and sufficient condition on

causal warrant: our belief that x exists acquires this special kind of
warrant if and only if we believe that we manipulate x." If I have
to choose between the two, my intentions were and are
metaphysical and not epistemological. My regrettable talk of
evidence might suggest I had some epistemological claim in
mind. In fact I was talking about evidence for the metaphysical
claim. Please remember, however, my doubt that the
metaphysical question matters much!
14. Best but not conclusive

I would prefer not to have put (a) in terms of evidence, but I
did so and have to live with that. (a) does not say even of any
particular case that the evidence is conclusive, only that it is
compelling. Thus I am unmoved by counter-examples that
mention the use of phlogiston to produce various effects-and by
all similar examples. I do not know enough about the phlogiston
story to know what Priestley or whomever thought he was doing.
But I am prepared to accept on faith that his experimental work
gave him compelling evidence for the reality of phlogiston. It was,
alas, not conclusive.
My own example is from the early days of photography.
Edmond Becquerel (1820-1891) was part of a great dynasty; his
son shared the Nobel Prize for the discovery of radioactivity, and
his father was a great pioneer of electrochemistry. In 1839, when
he was 19 years old, Edmond communicated a paper which with
hindsight can be read as the first demonstration of the
photo-electric effect.l21 In those days, one explained basic effects
121 Edmond Becquerel, 'Recherches sur les effets de la radiation chimique de la

lumière solaire, au moyen des courants èlectriques', Comptes rendus hebdomedaire
des seances de I'Academie des Sciences, 9 (1839): 145-149. Cf. Edmond Becquerel,
La Lumière: ses causes et ses effets, 2 vols., Paris: Firmin Didot Frères, II, p. 122.
of photochemistry by the conjecture that the sun emitted two

kinds of radiation, light rays and chemical rays. That was taught
in lectures and textbooks at the Polytechnique, the French
cathedral of mathematics and physics. So Edmond was using
chemical rays to produce differences in electric potential, and he
went on to more and more imaginative experiments. I am sure he
was convinced of the reality of chemical rays precisely because
he could use them to investigate other phenomena, such as
electric potential and above all change the properties of various
chloride emulsions in what we now call photography. In the
same year, 1839, Daguerre explained the secret of his marvellous
process for producing images, the daguerrotype; by 1840, young
Edmond Becquerel himself used chemical rays to produce
superb images-photographs-of Bridges over the Seine, the
Jardin de Tuilleries, etc. He had a compelling argument for the
existence of these theoretical entities, the chemical rays. Alas, like
many compelling arguments, it was not conclusive.122 But this is
not a counterexample to (a), about the strongest evidence for
scientific realism about this or that entity.
My argument was not about looking at some experiment, it
was about doing. It was above all an explication of the conviction
of experimenters, that sometimes the entities that they use are as
real as their left hands. It is replied, we all know that scientists
feel that way, so what? I say, they do not just feel that way; it is a
reasonable conviction based on what they can do with various
entities. In particular I pointed to engineering: what I had in
mind was designing and building a piece of apparatus in order
The experiment has been replicated by Jerôme Fatet; see his online seminar dated 26
January 2005, « Edmond Becquerel: La naissance de l'actinomètre electrochimique ».
122 Daniela Monaldi pointed this out to me, after reading what I wrote about
Becquerel in a preprint, "Another new world is being constructed right now: the
ultracold," Berlin: Max Planck Institute for the History of Science, 2006.
to do something with an entity, such as polarized electrons-or

chemical rays.
15. When we can not interfere

1 went on to write a paper about gravitational lenses, just at
the time that these were big news, that is, 1986, after the first four
had been detected.123 Gravitational lensing is a consequence of
the general theory of relativity. Let there be a big heavy star a
long way off. Let there be a light source yet further off, behind
the first star. Then light from that source will be bent by the star,
perhaps on each side of the star, just as light is bent by reflection
when passing through a suitable glass lens. If we are lucky, and
conditions are just right, gravitational lenses will be an
extraordinary way to bring distant objects "closer", in the
metaphorical sense in which a telescope is sometimes said to
bring a distant object closer.
1 took gravitational lenses as a perfect example of something
with which we could not interfere, something we just could not
manipulate. Einstein had thought, on the basis of a calculation
made on the back of an envelope, that we would never detect a
gravitational lens. That was, in his opinion, wonderful. We
would know that this phenomenon is going on, all over the
universe, but never see it! He was wrong. In the 1980s
gravitational lenses began to be detected; now they are
commonplace. They are in fact used to gain knowledge of the
distant universe (I confess that some of the claimed uses are a bit
exaggerated, more for getting grants than for studying the
distant universe.) But we still cannot interfere with them; we
cannot focus them, although we can do cunning things with the
123 "Extragalactic Reality: The Case of Gravitational Lensing," Philosophy of

Science. 56 (1989). 555-581.
bent light they deliver us for free. Compare the cosmic rays,
amazing sources of cheap high-energy beams, used even in the
1930s for discovering sub-atomic particles. There we can
manipulate the beams, but not the sources.
I ended the paper perhaps a little too flippantly. I suggested
that one can still be a sort of phenomenalist about gravitational
lenses, regarding them as theoretical entities that admirably fit
and explain the phenomena. I even made such a suggestion
about black holes-that we are always going to be in a Duhemian
position with respect to those, that all we can do is "save (solve)
the phenomena." That is because we are only spectators: in fact
in the case of black holes, we are not even spectators, we are
spectators of the havoc we think they wreak.
I was getting carried away, as you will see from the fact that
this paper about astrophysics ends by quoting poetry. I should
have stopped a little earlier in the paper, saying simply the
gravitational lensing was, at the time, breaking news in
observational astronomy, of immense potential, but that the
experimental argument for realism would never apply. Thus, the
strongest argument for realism could not be invoked. I did not
mean positively to assert anti-realism about gravitational lenses. I
never intended to claim that they are definitely not real! To
anyone who thought I implied we ought not to believe they are
real, I apologize. I meant only that we have more compelling
reasons for asserting the existence of polarized electrons than of
gravitational lenses.
16. The self-vindication of the laboratory sciences 124

I published a paper with this title after I began to develop my
124 "The Self-Vindication of the Laboratory Sciences," in A. Pickering, Science as

Practice and Culture, Chicago: University of Chicago Press, 1991,29-64.
philosophical adaptation of Crombie's idea of styles of scientific

thinking. I understand that it has recently been translated into
Chinese.
The paper was not about realism or antirealism. Indeed it
was written with the firm conviction that debates about realism
are by-products of the laboratory sciences. If anyone wishes to
apply Nelson Goodman's epithet, "irrealist," to that paper, I will
grudgingly accept it. Why grudging? Because irrealism does
have a tincture of antirealism, and by now I do not want to be
realist or antirealist. Perhaps I am pleased by the illusion that
Representing and Intervening was a realist work, while this paper is
an antirealist one. Both, I want to say, are neither.
The argument was that experiment, theory and apparatus
constitute plastic resources, a term that I took from Andrew
Pickering. Each can be moulded and adapted to fit the other.
Duhem had already shown how theory can be moulded to take
account of recalcitrant experimental results, but also, how one
can mould an account of experiment to preserve theory. I added
that there are many layers of theory and of experiment, including
theories of the apparatus, and also the material stuff is a layer,
the physical apparatus. All of these are changed and modified in
the course of "getting" an experimental result. The result is a
kind of self-vindication which is not viciously circular, indeed it
is incredibly painful and often cannot be done. When it cannot be
done, it is research that never gets published and is erased from
history. Most research is erased.
17. The stability of the sciences

One aim of the paper was to address the stability of the
sciences. Philosophers had lived through an era of literal
revolution-of space, time, and causality. But-let us say after
1950-revolutions stopped occurring in physics. It was as if we

had got it right. I have even argued that from now on there will
be no more revolutions, only surprises. I argue this as one small
point in a long discussion of ultracold experiments. I can only
point at these thoughts now. I happen to believe that physics
itself is changing to make my thesis about self-vindication more
and more apparent. Here is a recent sentence that I noticed in my
current hobby, cold atoms.
"[... ] our results point to the fact that the [Bose-Hubbard model] is
sufficient to explain all the features discovered in the experiment
and that the experiment was a clean realization of the model as
expected."125
The model is right because it explains how the experiment went,

and the experiment was a good one because it fit the model. How
is that for self-vindication?
125 S. R. Clark and D. Jaksch, "Signatures of the Superfluid to Mott-insulator

Transition in the Excitation Spectrum of Ultracold Atoms," New Journal of Physics,
8, [8] (2006), pp 160-178 on p. 177.
............
BIBLIOGRAPHY
Ariew, Roger. "Review of A. C. Crombie's Style of Scientific Thinking in

the European Tradition: The History of Argument and Explanation
Especially in the Mathematical and Biomedical Science and Arts." Isis.
86 (1997): 82-83.
Bazerman, Charles. Shaping Written Knowledge: The Genre and Activity of
the Experimental Article in Science. Madison, WI: University of
Wisconsin Press, 1988.
Beck, Ulrih. Risk Society: Towards A New Modernity. London: Sage, 1992.
Becquerel, Edmond. La Lumière: ses causes et ses effets. 2 vols. Paris:
Firmin Didot Frères, 1867-68.
____ . "Recherches sur les effets de la radiation chimique de la
lumière solaire, au moyen des courants èlectriques." Comptes rendus
hebdomedaire des seances de I'Academie des Sciences 9 (1839): 145-149.
Berkson, Mark. "Language: The Guest of Reality-Zhuangzi and
Derrida on Language, Realty, and Skillfulness." In Kjellberg, P. and
Ivanhoe, P. J. eds., Essay on Skepticism, Relativism, and Ethics in the
Zhuangzi, pp. 97-126. Albany: State University of New York Press,
1996.
Bernays, Paul. "On Platonism in Mathematics." In Benacerraf, Paul and
Putnam, Hilary eds., Philosophy of Mathematics: Selected Readings, pp.
BIBLIOGRAPHY /159
258-271. 2nd ed. Cambridge: Cambridge University Press, 1983.

(Translated by Charles Parsons from "Sur Ie platonisme dans les
mathématiques," L'Enseignement mathématique 34 (1935).
Bernstein, Richard. "The Rage against Reason." In McMullen, E. ed.,
Construction and Constraint, pp. 189-221. Notre Dame: Notre Dame
Boroditsky, Lera. "Linguistic Relativity." In Nadel, L. ed., Encyclopedia of
Cognitive Science, London: MacMillan Press, 2003: 917-921.
Bourdieu, Pierre. Meditations pascaliennes. Paris: Seuil, 1997.
____ . Pascalian Meditations. Stanford: Stanford University Press,
2000.
Butterfield, Herbert. The Origins of Modern Science, 1300-1800. London
and Cambridge: Deighton Bell, 1949.
Butterworth, Brian. What Counts: How Every Brain is Hardwired for Math.
New York: Free Press, 1999.
Carnap, Rudolf. "Empiricism, Semantics, and Ontology." Revue
internationale de philosophie 4 (1950): 20-40.
____ . The Logical Structure of the World: Pseudoproblems in Philosophy.
Trans. Rolf A. George. Berkeley: University of California Press, 1969.
Charles Bazerman. Shaping Written Knowledge: The Genre and Activity of
the Experimental Article in Science. Madison, WI: University of
Wisconsin Press, 1988.
Chemla, Karine. "Geometrical Figures and Generality in Ancient China
and Beyond: Liu Hui and Zhao Shang, Plato and Thabit Qurra."
Science in Context 18 (2005): 125-166.
Chemla, Karine et Guo Shuchun. Les neuf chapitres: Ie classique
mathématique de la Chine ancienne et ses commentaires. Paris: Dunod,
2004.
Chomsky, Noam. Rules and Representations. New York: Columbia
Chuang, Tzu. The Inner Chapters. Trans. David Hinton. Washington, D.
C.: Counterpoint, 1998.
Clark, Stephen R. and Dieter, Jaksch. "Signatures of the Superfluid to
Mott-Insulator Transition in the Excitation Spectrum of Ultracold

Atoms." New Journal of Physics 8 (2006): 160-178.
Coetzee, John M. The Lives of Animals. Princeton: Princeton University
Press, 1999.
Cohen, I. Bernard. "The Principia, Universal Gravitation, and the
"Newtonian Style," in Relation to the Newtonian Revolution in
Science: Notes on the Occasion of the 250th Anniversary of Newton's
Death." In Bechle, Zev ed., Contemporary Newtonian Research, pp.
21-108. Dordrecht: D. Reidel, 1982.
Connes, Alain and Changeux, Jean-Pierre. Matiere it penser. Paris: Odile
Jacob, 1989.
_____ . Conversations on Mind, Matter, and Mathematics. Translated
M. B. DeBevoise. Princeton: Princeton University Press, 1995.
Crombie, Alistair C. "Philosophical Presuppositions and Shifting
Interpretations of Galileo." In Hintikka, J., Gruender, D. and Agazzi, F.
eds., Theory Change, Ancient Axiomatics, and Galileo's Methodology:
Proceedings of The 1978 Pisa Conference on The History and Philosophy of
Science, vol. I, pp. 271-286. Dordrecht: D. Reidel, 1981.
_____ . Styles of Scientific Thinking in the European Tradition: The
History of Argument and Explanation Especially in the Mathematical and
Biomedical Sciences and Arts. 3 vols. London: Duckworth, 1994.
Dahaene, Stephen. The Number Sense: How the Mind Creates Mathematics.
Oxford: Oxford University Press, 1987.
Daston, Lorraine and Galison, Peter. Objectivity. New York: Zone Books,
2007.
Davidson, Arnold. The Emergence of Sexuality: Historical Epistemology and
the Formation of Concepts. Cambridge, Mass.: Harvard University Press,
200l.
Davidson, Donald. Essays on Actions and Events. 2nd ed. Oxford: Oxford
University Press, 200l.
____ . Inquiries into Truth and Interpretation. 2nd ed. Oxford: Oxford
University Press, 200l.
____ . Subjective, Intersubjective, Objective. Oxford: Oxford
BIBLIOGRAPHY/161

____ . Problems of Rationality. Oxford: Oxford University Press,
2004.
____ . Truth, Language, and History. Oxford: Oxford University Press,
2005.
____ . Truth and Predication. Cambridge, Mass: Belknap, 2005.
Davies, Paul C. W. Superforce. New York: Simon and Schuster, 1985.
Drake, Stillman. "Galileo's Discovery of the Law of Free Fall." Scientific
American Journal 228(5) (1973): 84-92.
____ . Galileo at Work. Chicago: The University of Chicago Press,
1978.
Elwick, James. Styles of Reasoning in the British Life Sciences: Shared
Assumptoions, 1820-1858. London: Pickering & Chatto, 2007.
Fine, Arthur. "The Natural Ontological Attitude." In The Shaky Game:
Einstein, Realism and the Quantum Theory, pp. 112-135. Chicago: The
Fleck, Ludwik. Genesis and Development of a Scientific Fact. Chicago: The
University of Chicago Press, 1979. (Translated by T. J. Trenn and R. K.
Merton from the German: Entstehung und Entwicklung Einer
Wissenschaftlichen Tatsache. Einführung in Die Lehre von Denkstil und
Denkkollektiv. Basel: Benno Schwabe, 1935).
Fodor, Jerry. The Modularity of Mind: An Essay on Faculty Psychology.
Cambridge, Mass.: MIT Press, 1983.
____ . "Modules, Frames, Fridgeons, Sleeping Dogs, and the Music
of the Spheres." In Garfield, J. ed., Modularity in Knowledge
Representation and Natural-Language Understanding, pp. 25-36.
Cambridge Mass.: MIT Press, 1987.
Foucault, Michel. Les mots et les choses: Une archéologiedes sciences
humaines. Paris: Gallimard, 1966.
____ .The Order of Things: An Archaeology of Human Sciences. New
York: Vintage Books, 1973.
Galison, Peter. How Experiments End. Chicago: The University of Chicago
Press, 1987.
Galton, Francis. English Men of Science: Their Nature and Nurture. London:
Macmillan, 1874
Gould, Stephen J. and Lewontin, Richard C. "The Spandrels of San
Marco and the Panglossian Paradigm: A Critique of the Panglossian
Paradigm." Proceedings of the Royal Society of London, B, Biological
Sciences 205 (1979): 581-598.
Gould, Stephen J. "The Exaptive Excellence of Spandrels as a Term and
Prototype." Proceedings of the National Academy of Sciences 94 (1997):
10750-10755.
Hacking, Ian. "Leibniz and Descartes: Proof and Eternal Truths." In
Dawes Hicks Lecture on Philosophy British Academy, pp. 3-16. London:
Oxford University Press, 1973.
____ . The Emergence of Probability. Cambridge: Cambridge
University Press, 1975; second edition, with a new "Introduction
2006." Cambridge: Cambridge University Press, 2006.
___ . "What Is Logic?" Journal of Philosophy 86 (1979): 285-319.
____ . "Language, Truth and Reason." In Hollis, M. and Lukes, S.
eds., Rationality and Relativism, pp. 48-66. Oxford: Blackwell, 1982.
____ . "Was There a Probabilistic Revolution 1800-19307" In Krüger,
L., Daston, L. J. and Heidelberger, M. eds., The Probabilistic Revolution,
vol. I: Ideas in History, pp. 45-58. Cambridge, Mass.: M.LT. Press,
1987.
____ . "The Participant Irrealist at Large in the Laboratory." British
Journal for the Philosophy of Science 39 (1988): 277-294.
____ . "Extragalactic Reality: The Case of Gravitational Lensing."
Philosophy of Science 56 (1989): 555-581.
____ . "Artificial Phenomena." British Journal for the History of
Science 24 (1991): 235-241.
____ . "The Self-Vindication of the Laboratory Sciences." In
Pickering, A. ed., Science as Practice and Culture, pp. 29-64. Chicago:
The University of Chicago Press, 1991,.
____ . "'Style' for Historians and Philosophers." Studies in History
and Philosophy of Science 23 (1992): 1-20.
BIBLIOGRAPHY/ 163
____ . "The Disunities of the Sciences." In Galison, P. and Stump, D.

J. eds., The Disunity of Science. Boundaries, Contexts and Power, pp. :
37-74. Stanford: Stanford University Press, 1996.
____ . "What Mathematics Has Done to Some and only Some
Philosophers." In Smiley, T. J. ed., Mathematics and Necessity, pp.
83-138. London: British Academy, 2000.
____ . Historical Ontology. Cambridge, Mass.: Harvard University
Press, 2002.
____ . "Critical Notice of Bernard Williams Truth and Truthfulness:
An Essay in Genealogy." Canadian Journal of Philosophy 34 (2004):
137-148.
____ . "La science de la science chez Pierre Bourdieu." in
Bouveresse, J. et Roche, Daniel eds., La Liberté par la connaissance.
Pierre Bourdieu (1930-2002), pp. 147-162. Paris: Odile Jacob, 2004.
____ . Another New World is Being Constructed Right Now: the
Ultracold. Berlin: Max Planck Institute for the History of Science, 2006.
(Preprint 316).
____ . "Kinds of People: Moving Targets." Proceedings of the British
Academy 151 (2007): 285-318.
____ . "Almost Zero." London Review of Books, 10 May, 2007.
____ . "On not Being a Pragmatist: Eight Reasons and a Cause." In
Misak, Cheryl ed., New Pragmatists, pp. 32-49. Oxford: Oxford
____ . "Husserl on the Origins of Geometry." In Hyder, David ed.,
Husserl's Epistemology. Standford: Stanford University Press.
(forthcoming)
Hadot, Pierre. Le voile d'Isis. Essai sur l'histoire de I 'idee de nature. Paris:
Gallimard, 2004.
____ . The Veil of Isis: An Essay on the History of the Idea of Nature.
Trans. Michael Chase. Cambridge, Mass.: Harvard University Press,
2006.
Hesse, Herman. The Glass Bead Game: A Tentative Sketch of the Life of
Magister Ludi Joseph Knecht, together with Knecht's Posthumous Writings.
London: Jonathan Cape, 1970.

Hintikka, J., Gruender, D. and Agazzi, F. eds., Theory Change, Ancient
Axiomatics, and Galileo's Methodology: Proceedings of The 1978 Pisa
Conference on The History and Philosophy of Science, volume. 1.
Dordrecht: D. Reidel, 1981.
Husserl, Edmund. The Crisis of European Science and Transcendental
Phenomenology: An Introduction to Phenomenological Philosophy. Trans.
D. Carr from the German of 1936. Evanston, Ill.: Northwestern
Iliffe, Robert. "Rational Artistry." History of Science 36 (1998): 329-357.
Kant, Immanuel. The Critique of Pure Reason. Trans. Norman Kemp Smith.
London: Macmillan, 1929.
____ . Anthropology From a Pragmatic Point of View. Trans. and ed. R.
Louden. Cambridge: Cambridge University Press, 2006.
Kitcher, Philip. The Nature of Mathematical Knowledge. Oxford: Oxford
Lagerkvist, Par. Gäst hos Verkligheten. Stockholm: Aldus/Bonnier, 1967.
Lambek, Michael. The Weight of Past: Living with History in Matanga,
Madagascar. London: Palgrave, 2002.
Latour, Bruno. "Postmodern? No, Simply Amodern! Steps Towards an
Anthropology of Science." Studies in the History and Philosophy of
Science 21 (1990): 145-171.
____ . "The Netz-Works of Greek Deductions." Social Studies of
Science. 38 (2008): 441-459.
Latour, Bruno and Woolgar, Steve. Laboratory Life: The Social
Construction of Scientific Facts. Beverly Hills: Sage Publications, 1979.
____ . Laboratory Life. Princeton: Princeton University Press, 1986.
Lloyd, Geoffrey E. R. Ancient Worlds, Modern Reflections: Philosophical
Perspectives on Greek and Chinese Science and Culture. Oxford: Oxford
____ . Cognitive Variations: Reflections on the Unity and Diversity of
the Human Mind. Oxford: Clarendon Press, 2007.
Magruder, Kerry V. "Review: Style of Scientific Thinking in the
BIBLIOGRAPHY /165
European Tradition." Sixteenth Century Journal 26 (1995): 406-410.

Meadows, Arthur J. "Paradigms are Not the Only Answer-Style of
Scientific Thinking in the European Tradition (Volumes I, II and III)
by Alistair Crombie." New Scientist 143 (1994): 38.
Milonni, Peter W. The Quantum Vacuum: An Introduction to Quantum
Electrodynamics. Boston: Academic Press, 1997.
Netz, Reviel. The Shaping of Deduction in Greek Mathematics: A Study in
Cognitive History. Cambridge: Cambridge University Press, 1999.
____ and Noel, William. The Archimedes Codex: Revealing the Secrets
of the World's Greatest Palimpsest. London: Wiedenfield and Nicholson,
2007.
Nietzsche, Friedrich. The Gay Science: With a Prelude in Rhymes and an
Appendix of Songs. Trans. Walter Kaufmann. New York: Vintage Books,
1964.
Nisbett, Richard E. The Geography of Thought: How Asians and Westerners
Think Differently and Why. New York: Free Press, 2003.
Quine, Willard V. 0., Word and Object. Cambridge, Mass: MIT Press,
1960.
Rabinow, Paul. Making PCR: A Story of Biotechnology. Chicago: The
____ . French DNA: Trouble in Purgatory. Chicago: The University of
Chicago Press, 1999.
____ . Anthropos Today: Reflections on Modern Equipment. Princeton:
Princeton University Press, 2003.
____ . A Machine to Make a Future: Biotech Chronicles. Princeton:
Reiner, R. and Pierson, R. "Hacking's Experimental Realism: An
Untenable Ground." Philosophy of Science 62 (1995): 60-69.
Resnik, David. "Hacking's Experimental Realism." Canadian Journal of
Philosophy 24 (1994): 395-412.
Rogers, Timothy T. and McClelland, James L. Semantic Cognition: A
Parallel Distributed Processing Approach. Cambridge, Mass.: MIT Press,
2004.
Rorty, Richard. Hope in Place of Knowledge: The Pragmatics Tradition in

Philosophy. Taipei: Institute of European and American Studies,
Academia Sinica, 1999.
____ . Philosophy as Cultural Politics. Philosophical Papers IV.
Cambridge: Cambridge University Press, 2007.
Russell, Bertrand. The Problems of Philosophy. London: The Home
University Library, 1946.
Russell, Bertrand and Whitehead, Alfred N. Principia Mathematica.
Cambridge: Cambridge University Press, 1910, 1912, and 1913.
Shapin, Steven and Schaffer, Simon. Leviathan and the Air Pump: Hobbes,
Boyle and the Experimental Life. Princeton: Princeton University Press,
1986.
____ . Leviathan et la pompe à air: Hobbes et Boyle entre science et
politique. Traduction Thierry Pielat. Paris: Editions la decouverte,
1993.
Shapin, Steven. A Social History of Truth: Civility and Science in Seventeenth
Century England. Chicago: The University of Chicago Press, 1994.
Shen, K., Crossley, John N. and Lun, Anthony W.-C. The Nine Chapters on
the Mathematical Art: Companion and Commentary. Oxford: Oxford
Sperber, Dan. "In Defense of Massive Modularity." In Dupoux, E. ed.,
Language, Brain and Cognitive Development: Essays in Honor of Jacques
Mehler, pp. 47-57. Cambridge, Mass: MIT Press, 2002.
Suarez, Mauricio. "Experimental Realism Defended: How Inference to
the Most Likely Cause Might be Sound." In Hartmann, S., Hoefer, C.
and Bovens, Luc eds., Nancy Cartwright's Philosophy of Science, London:
Routledge, 2008.
Uzan, Jean-Philippe. "The Fundamental Constants and Their Variation:
Observational and Theoretical Status." Reviews of Modern Physics 75
(2003): 403-455.
Von Humboldt, William. Lettre a M. Abel-Remusat: Sur la nature desformes
grammaticales en general et sur Ie genie de la langue chinoise en particulier.
Paris: Dondey-Dupre, 1827.
BIBLIOGRAPHY / 167
____ . On Language: on the Diversity of Human Language Construction

and Its Influence on the Mental Development of the Human Species. Ed.
Michael Losonsky, trans. Peter Heath. Cambridge: Cambridge
Watts, Isaac. Logic, or The Right Use of Reason in the Enquiry After Truth
With a Variety of Rules to Guard Against Error in the Affairs of Religion
and Human Life, as well as in the Sciences. London, 1724.
Weinberg, Stephen. "The Force of Nature." The Bulletin of American
Academy of Arts and Sciences 29 (1956).
Wenzel, Christian Helmut. "Chinese Language, Chinese Mind?" In
Kanzian, C. and Runggaldier, E. eds., Cultures:
Conflict-AnaLysis-Dialogue, pp. 296-314. Frankfurt am Main: Ontos,
2007.
Williams, Bernard. Ethics and the Limits of Philosophy. Cambridge, Mass.:
Harvard University Press, 1985.
____ . Truth and Truthfulness: An Essay in Genealogy. Princeton:
Wittgenstein, Ludwig. Philosophical Investigations. Trans. G. E. M.
Anscombe. Oxford: Blackwell, 1953.
____ . Remarks on the Foundations of Mathematics. Eds. G. H. von
Wright, R. Rhees, and G. E. M. Anscombe, trans. G. E. M. Anscombe.
London: Basil Blackwell, 1956.
Wölfflin, Heinrich. Principles of Art History: The Problem of the
Development of Style in Later Art. Trans. M. D. Hottinger. New York:
Dover, 1929.
............
INDEX OF SUBJECTS
A priori 55, 57, 59, 63, 74, 85, 87, 103

142 Athens 30, 33, 37, 59
Abbassids 76 Authenticity 30, 34
Abduction 45-46,98 Axioms 13,72
Accuracy 33, 39, 41, 43, 96 Baghdad 37, 75-78, 82, 91
Ah-ha! 73 Bivalence 143
Air pump 43, 67-68,104,107-110, Bose-Einstein Condensate 111-112,
112-114,121,125,167 121
Alchemist 76,112,128 Calculation 13,72-74,154
Analogy 93-94, 97, 100, 117 like the Taipei subway 74
Analytic philosophy 21, 62, 130 Calligraphy 6
Anthropology 6,8,21,48-50,52, Chemical rays 153-154
68-69,81, 107, 123, 165; China 6,8, 14, 26-28, 33, 61
comparative 81; historical 6, 8, 83,85,92,98, 121, 124, 126,
21,81; philosophical viii, 12,24, 160; ancient science in 26;
39,66-68,70,86-87 biotechnology in 121;
Antirealism 23, 129, 132, 136, 140, philosophy 101, 156
142-143,145,150,156 Chronometer 41, 111
Approximation 84-86 Cognition and culture 4, 94
Archaeology ofknowledge 14 Cognitive science 3, 48,77;
Archaeology of Knowledge capacities 11-12, 23, 44, 54-55,
(Foucault) 16 63, 71, 95; history 55, 77, 84, 87,
Architectonic reasoning 99-100, 89, 95-96, 184
INDEX OF SUBJECTS / 169
Cold atoms 52,111-112,157 First Three Minutes (Weinberg)

Combinatorial reasoning 55, 77 103
Conversations about Mind, Matter, Formof life 105-106, 127-128
and Mathematics (Connes & Genealogy 2, 29-31, 33-34, 91,
Changeux) 78 164, 168
Corpuscularian philosophy Genetic envelope 77
119-120 Geometry 5, 13, 22, 26, 36, 38,40,
Crisis of European Civilization 44,55-56,58,60,62,65,69,85,
(Husser!) 65 164
Crystallization 9, 13-14, 16,20,26, Glass Bead Game (Hesse) 27
36,38,41-45,58,75,91,93,95, Grammar 22,38,41,45,54,83,
99-100, 104, 106, 116, 118 102-103, 130, 133, 135-136
Deconstruction 137 Gravitational lenses 154-155
Denkstil18, 162 Greece 6,8, II, 16,26,33,37,48,
Dinosaurs 22 76,92,98, 119; atomists 118;
Disciplinary matrix 19 mathematics 37, 59, 66, 69-70,
Disunity 51,164 86,92, 104; miracle 70, 84;
Emblem 36, 38, 45, 58-59, 76, 91, science 26
104, 116 Gresham College 115
Emergence of Probability "Hardness of the logical must" 37,
(Hacking) 14 63
Empiricism 23,86, 130, 141, 160; Historicism 25
constructive 141 History 1-9, 11-13, 16-18,25,
Evolution: 30-37,39,42,44-48,51,55-56,
evolutionary psychology 53 59,66,69-72,75-79,86,89,91,
Evolutionary love 101 93-94,102,106-107,109,112,
Experiment 4, 41, 44, 57, 67, 95, 115,119,130,135,138,143,
97, 103, 105, 109, 147, 153, 153,156,159,161-168;
156-157 historical time 34-35
Experimental argument for entity History of the present 4
realism 146 Hypothesis, method of 45
"Experiment has a life of its own" Hypothetico-deductive 41,97-98
147 "If you can spray them, then they
Explanation 7-8, 23,34,53,98, are real" 147
151,159,161; explanatory Inclined plane 41
power 34, 38; inference to the Induction 45
best 98, 151 Innate capacities 47
Finite groups 7, 75, 79 Inscriptions 122-123
Instrumentalism 23,86 Modernity 14, 113, 159

Irrealism 145-156 Module 38-39, 54
Laboratory 4,6,9,39,41-44,46, "Motley of mathematics" 51,
65-67,77,79,93,96, 100-101, 72-73,83
103-104,106-111,113-117,119, MRI 77
121-122,124-125,127,149, Music 5, 27, 38, 162
155-156, 163, 165; Mutation 16, 100
biotechnology 8, 9, 107, 121, Name-ism 131
122, 123, 124, 125, 128 National Chung Cheng University
Laboratory Life (Latour) 66 xvi, 112
Language game 105-106 National Tsing Hua University xv,
Leviathan and the Air Pump 112, 117
(Shapin and Schaffer) 43,67-68, "Natural history of human beings"
104, 107, 109 48
Logic 23-24, 45-46, 56, 59, 65, 74, Nature and nurture 50
98, 149, 163, 168 Network 57, 68-71
Madagascar 35, 165 Neuroscience 48
Magdeburg, spheres of 110 Nominalism 86, 130-135, 139
Manhattan Project 107 Objectivity 23-24,30,47,79, 161
Manipulation 109-11 0, 151-152 Objects 9-11, 21-23, 40, 47-48,
Marxist historiography 5 58-59,61-63,65,71-72,74-75,
Materialism 122 78-81,83,86-87,90-92, 142,
Mathematics 7-8, 13,23,26-27, 146, 150, 154; geometrica140,
36-37,40-41,44,47,51-54,56, 58-59,71,75,91-92;
59,61-64,66,69-75,78-80, mathematical 22, 59-63, 65, 72,
82-84,86,90,92-95,103-105, 74,78-79,83,86,142
116,129,141-143,153,159, Ontological debates 21, 23, 47-48,
161, 164, 166, 168; and 81,86,90,118,141,146
philosophy 135; Chinese 8, 56, Ontology 23, 74, 79, 93, 102-103,
82-84; 92, 94; Greek 26,37,59, 118,130,133-135,140,142,
66,69-70,86,92,104 160, 164; Natural Ontological
Measurement 10, 42, 93-96, 100, Attitude 145, 162; recapitulates
104, 116 philology 134
Meno (Plato) 51, 60 Paradigms 12, 103, 166
Methods of reasoning 9, 11, 15, 21, Past 2-4, 10, 16-17, 22, 24, 30,
38,46, 90, 103 33-36,40,58,68,71,82,90-91,
Mice 98 105, 107, 122, 165
Misinterpretation 138-139 Pasteur Institute 124
INDEX OF SUBJECTS/171
Peer review 44, 116, 125-126 degrees of 121; guest of 138,

PET 77 139, 159; mathematica162
Phenomena, creation of 104, Reason 1-4,7,19,25-26,28-29,36,
110-111,116 46,53-57,62,65,68-69,72,77,
Philology 134 81,93,98, 118, 121, 142, 146,
Photography 152 160, 163, 165, 168; rage against
Phylogeny 135 28
Plastic resources 156 Remarks on the Foundations of
Platonism 23, 75,86,129,141-142, Mathematics (Wittgenstein) 51,
159 64,l43
Porto Allegre 141 Representing and Intervening
Positivism 21,141; logical21, 34, (Hacking) 109, 116, 121, 146
37,63,65,74,98,140; Republic (Plato) 131
Positivist Party 141 Research programmes 12,61
Postulation 10 Revolution 11, 16,28-29,36,
Pragmatism 85, 147 56-58,72,89, 156, 161, 163;
Principia Mathematica Cultural28; scientific 11,72,89,
(Whitehead and Russell) 65 96
Probability 14-17,45,76,93,97, Risk society 14
163 Royal Society of London 44,53,
Problems of Philosophy (Russell) 69, 115, 163
133 Science studies 24, 66-68, 104, 107,
Proof 13, 36-37, 39,40,47, 51, 119,123
57-58,60-64,69,71-74,79-80, Secrets of nature 110
84-86,91-92, 114, 151, 163; Self-authentication 47
perspicuous 52, 64, 72-74, 86, Self-vindication 155-157
92 Semantics 31-32,130,133,135,
Pythagorean theorem 22, 86 148, 160; semantic ascent 130,
Realism 22-23, 47, 75, 80, 86,101, 132
118,129-132,134,136,139-151, Shanghai 9,124
153, 155-156, 162, 166-167; Shaping of Deduction in Greek
entity 132-133; Metaphysical Mathematics (Netz) 37, 59, 66
Experimental Realism 151; Sincerity 33, 39, 43-44, 116
really-realism 139; scientific Social History of Truth (Shapin) 44,
22-23,86,117,129,140, 115
147-150, 153 Soochow University xv, 129
Reality 62, 75,101,119,136-140, Spandrel53-54
142, 146, 150-154, 159, 163; Specific styles:
Algorithmic 37, 46, 76; Themata 12

Experimental exploration 42-43, Theoretical entities 22, 43, 47, 80,
46; Galilean 41-42, 46, 65, 76, 86,118,140-141,146,150,153,
78,100-104,118; Geometrical 155
77-78,86; Historico-genetic 92; Theory and experiment 52
Hypothetical modelling 6, 17, Third World (Popper) 26
42,46,92-94,99-100; Thought-collective 18
Laboratory 9,39,42-44,46,65, Thyrotropin Releasing Factor 122
67, 76, 78, 93, 100, 104, 106, Tolerance, principle of 130
112,116,118,127; Trailblazer 36,42,45,58,76,91,
Mathematical9, 46,75,78, 142; 116
Statistical 7; Taxonomic 9,39, Transcendental aesthetic 38
46,78 Trust 44, 115, 125
Stability 122, 156 Truth, has no history 31; social
"Style", the word 18-20,40 history of 4
Style of inquiry 20; of reasoning 20, Truthfulness 2, 15,21,29,32-34,
76,79,90,127,142 39,43,47,58,89,91,164,168
Styles ofScientific Thinking in the Universals 23, 46,132-133,139,
European Tradition (Crombie) 7, 141-142
9, 11,81,90 "Unspeakably more depends on
Styles of scientific thinking and what things are called" 136-137
doing 93; as constitutedby Vacuum 110, 112, 114, 119-121,
methods and objects 9; 127, 166
Crombie's list of six 43; not Venture capital 9, 123-124, 126
sciences 7, 93 Verification, method of21
Superconducting 117 Whig history 4
Superfluidity 117 Zab, Battle of the River 75
Synthetic a priori 55
Template 7, 10, 12
INDEX OF NAMES
A Bourdieu, Pierre 24, 29, 46, 123,

160,164
Agazzi, F. 6, 161, 165 Bouveresse, Jacques 24, 164
AI-Khwarizmi 37, 75, 76 Boyle, Robert 42-45,67,77,
Archimedes 26-28, 67, 69-70, 82, 104-106,110,112-117,119-121,
85, 166 125, 127, 167
Ariew, Roger 12, 159 Butler, Bishop 97
Aristotle 1,31,59,130 Butterfield, Herbert 4, 160
B c
Bachelard, Gaston 15, 16,90 Carnap, Rudolf 18, 97, 109, 130,
Bardeen, John 117 132,160
Barnes, Barry 68 Changeux, Jean-Pierre 77-78, 80,
Bazerman, Charles 106, 159-160 161
Beck, Ulrich 14, 159 Chemla, Karine 20,84-86, 160
Becquerel, Edmond 152-153, 159 Cho, Hung-Wen 117
Bellarmine, Robert 113 Chomsky, Noam 38, 41-42, 54, 80,
Bernstein, Richard 28, 160 102-103, 160
Bloor, David 68 Chuang, Tzu 136, 138, 140, 160
Bose, Satyendra 111-112, 117, 121, Coetzee, John Maxwell28, 161
157 Cohen, Bernard 1. 102-103, 161
Comte, Auguste 48, 141
Connes, Alain 78-80, 161
Cooper, Leon N. 117 49, 123, 162

Costello, Elizabeth 28 Frege, Gottlob F. L. 26
Crombie, Alistaire. C. ix, xi, 6-13,
15-20,30,37,40-43,46,76, G
81-82,86,89-90,92-95,97,
Galileo, Galilei 11, 17, 26, 40-42,
99-100, 103, 118, 128, 156, 159,
45,70,89,92,100,102-104,
161, 166
113,161-162
D Galison, Peter 23-24, 51, 109,
161-162,164
Daguerre, Louis-Jacques-Mande Gardner, Martin 73
153 Garfield, Jay L. 38, 162
Darwin, Charles R. 5, 135 Geber (Abu Musa Jabir ibn
Daston, Lorraine 23, 24, 161, 163 Hayyan) 76
Davidson, Arnold 19 Gödel, Kurt 65
Davidson, Donald 31-32 Goodman, Nelson 22,131,145,
Davies, Paul C. W. 120, 162 156
Dennett, Daniel 15 Gould, Stephen Jay 53, 163
Descartes, Rene 62-65, 73-74, 163 Gruender, David 6, 161, 165
Dewey, John 144, 147-148, 150 Guo, Shuchun 20, 84, 160
Diderot, Denis 30, 45
Drake, Stillman 41, 113, 162 H
Dummett, Michael 132, 143
Hacking, Ian ix-xiii, 14,24,93,
E 122,151,163,166
Hadot, Pierre 119, 138, 164
Einstein, Albert 96, 111-112, 117, Haeckel, Ernst 135
121, 145, 154, 162 Han, Dian-Jiun 112
Eisenhart, Churchill 96 Heraclitus 120-121, 138
Elwick, James 20, 162 Herodotus 35
Euclid 13,37,67 Hesse, Herman 27-28, 164
Hintikka, Jaakko 6, II, 161, 165
F Hobbes, Thomas 43-44, 67,
104-105,112-115,131,167
Fermat, Pierre de 51,72
Hollis, Martin 19, 93, 163
Fine, Arthur 145,162
Holton, Gerald 12
Fleck, Ludwik 18, 26, 162
Hume, David 4, 34, 50
Fodor, Jerry 38, 162
Husserl, Edmund G. A. 2, 25-26,

Foucault, Michel xi, 4, 14-16,29,
INDEX OF NAMES / 175
41-42,62,65,76, 101-103, M
164-165
Hyder, David 26,62, 164 Magruder, Kerry 12, 165
Mannheim, Karl 18
I McClelland, James L. 39, 166
Meadows, Arthur J. 12, 166
Iliffe, Robert 12, 165
Miller, James G. 134
J Milonni, Peter W. 120, 166
James, William 20,39, 85, 134, N

147, 162, 166
Needham, Joseph 82
K Netz, Revie137, 59, 66-71, 77, 84,
104,165-166
Kafka, Franz 28 Nietzsche, Friedrich W. 23, 29, 87,
Kant, Immanue16, 12,24,36,38, 134, 136-140, 145, 166
42,49-50, 55-59, 62, 65, 76, Nisbett, Richard E. 134, 166
84-87, 142, 165
Kapitsa, Pyotr 117 o
Kirwan, Christopher 31
of Ockham, William 141
Kourilsky, Philippe 124
Koyré, Alexandre 4, 41 P
Krüger, Lorenz 16, 163
Kuhn, Thomas 12, 15, 19,90, 109 Pascal, Blaise 17,45, 76
Peirce, Charles S. 45-46, 98, 147
L Pickering, Andrew 20, 101,
155-156, 162-163
Lakatos, Imre 12
Planck, Max 1, 153, 164
Lambek, Michael 35, 165
Plato 1,7,26,51-52,61,63-65,
Latour, Bruno 66-71, 77, 80, 84,
69-72,80,83,85, 130-131, 142,
104-105, 107, 109, 113, 122,
160
165
Popper, Karl 26, 46, 97, 109, 118
Leibniz, Gottfried Wilhelm 1,
Pythagoras 51
23-24,62-65,74,100,103,163
Lewontin, Richard 53, 163
Q
Linnaeus, Carl 17, 45, 81
Lloyd, Geoffrey E. R. 8,20,26,37, Quine, Willard V.O. 130, 134-135
59,60,71,82-84,134,165
Lukes, Steven 19,93, 163
R U
Rabinow, Paul 123, 166 Uzan, Jean-Philippe 1, 167

Roche, Daniel24, 164
Rogers, Timothy T. 39, 166 v
Rorty, Richard 143-147, 167 van Fraassen, Bas 48,81,109,141
Rose, Nikolas 123
Vernant, Jean-Pierre 6
Ross, William D. 31
Von Humboldt, William 167
Rousseau, Jean-Jacques 30, 45
Russell, Bertrand 62-65,87, 133, w
140, 142, 167
Watts, Issac 24, 168
s Weinberg, Steven 41-42,101-103,
Schaffer, Simon 43-44, 67-68, 71, 168
104,106-107,109,114,167 Wenzel, Christian H. 133-134, 168
Schlick, Moritz 2, 21 Whewell, William 51
Shapin, Steven 43-44,67-68, 71, Whitehead, Alfred North 70, 167
104,106-107,109,115,167 Williams, Bernard A. O. 1-2,21,
Sima Tan 33 29-30,33-36,39-40,43,47,58,
Spengler, Oswald 18 71,76,89,91,116,164,168
Sperber, Dan 38, 167 Wisdom, John 37, 75, 135
Suarez, Mauricio 151, 167 Wittgenstein, Ludwig 37, 48, 51,
62-65, 73, 96, 105-106, 133, 143,
T
168
W ölffiin, Heinrich 17-18, 168
Tal, Eran 149
Tarski, Alfred 31-32 Woolgar, Steve 66,109,122,165
Thales 36, 38,42,45,57-59,69, Y
75-77,91-92
Thucydides, 33, 58 Yu, He Albert 112, 117
Torricelli, Evangelista 110

Hacking - Scientific Reason

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

Hacking - Scientific Reason

Uploaded by

Copyright:

Available Formats

SCIENTIFIC REASON

Edited by Der-Lan YEH and Jeu-Jenq YUANN

Published by Si-Chen LEE

Distributed by National Taiwan University Press

Copy editing by Bookman Books Ltd.

Retail Price NT$ 350

Lecture 1 ON THE HISTORICAL ROOTS OF

18. The laboratory style 42

Lecture 2 WHERE DO MATHEMATICAL OBJECTS

17. Mathematical objects 78

Lecture 3 THE LABORATORY STYLE OF THINKING

Lecture 4 REALISMS AND ANTIREALISMS 129

In November 2007, the philosophical community of Taiwan

The focus of Hacking's thought is revealed in the title,

Far from conceiving scientific reason in accordance with its

Describing science from a historical dimension is by no

The two aspects of innate human ability and the development

orientation through which scientific reason can be understood.

While taking Crombie's historical survey of science as a

discovered, and even phenomena created? Would what is within

The act of identifying philosophical objects within styles of

Styles of scientific reasoning exist naturally and culturally

convincing way of conveying the "truthfulness" of the past, the

Professor Hacking's message gives us the opportunity to

National Taiwan University and the Department of Philosophy,

These chapters are rewritten versions of a series of four

The audiences were wonderful, and significant changes have

Chung Cheng University. I also was able to meet clinicians and

ON THE HISTORICAL ROOTS

These lectures address very traditional subjects of philosophical

1 A standard recent survey paper is, Jean-Philippe Uzan, "The Fundamental

he calls truthfulness. Ways to tell the truth have a history, or what

2 Bernard Williams, Truth and Truthfulness: An Essay in Genealogy, Princeton:

that we are ill-advised to forget, even when it violates the rules of

1. Learning how to learn

I should emphasize at the start that I am a philosopher who

• There may also be a genetic way of understanding, most

China. There, if we wanted to shake up our preconceptions, we

2. Styles of scientific thinking

4 The proceedings were published in J. Hintikka, D. Gruender, and F. Agazzi (eds.)

In 1994 Crombie finally produced his life work, the

6 London: Duckworth, 1994.

science uses all of them. That goes for relatively "lowbrow"

Two aspects of Crombie's title should be noticed. There is the

7 Geoffrey Lloyd, Ancient Worlds, Modern Reflections: Philosophical Perspectives

laboratory science, the laboratory style of thinking and doing.

4. Styles as constituted by methods and objects

We can establish in the classic scientific movement a taxonomy of

8 A. C. Crombie, Styles of Scientific Thinking in the European Tradition, vol. I, p.

things, which are found in all languages, but objects bearing a

5. A crude template for organizing the past

The active promotion and diversification of the scientific methods

(a) the simple postulation established in the mathematical sciences,

The first three of these methods concern essentially the science of

individual regularities, and the second three the science of the

9 A.C. Crombie, "Philosophical Perspectives and Shifting Interpretations of

10 R. Iliffe, "Rational Artistry," History of Science, 36 (1998): 329-357, an extended

current practices. It helps immeasurably to have a template to

tell the story of each style of scientific thinking as having at least

11 Ian Hacking, The Emergence of Probability, Cambridge University Press, 1975;

found rich elements of other systems of thought, in which we can

13 I say "inherited", For Anglophone philosophers of my generation, it was of

notice that I do not now call the emergence of probability in 1650