Latour Pandora's Hope

Reality of Science Studies
Pandora ’s H ope
E ssa ys o n the R e a l it y
of S c ie n c e St u d i e s
Bruno Latour
HARVARD UNIVERSITY PRESS

Cambridge, Massachusetts
London, England
Copyright © 1999 by the President and Fellows of Harvard College
All rights reserved
Printed in the United States of America
Library of C ongress Cataloging - in -P ublication D ata
Latour, Bruno.
Pandora’s hope : essays on the reality of science studies / Bruno Latour
p. cm.
Includes bibliographical references and index.
ISBN 0-674-65335-1 (alk. paper). — ISBN 0-674-65336-X (pbk. : alk. paper)
1. Realism. 2. Science—Philosophy. I. Title
Q175.32.R42L38 1999
501—dc2i 98-50061
Second printing, 2000

To Shirley Strum, Donna Haraway, Steve Glickman,
and their baboons, cyborgs, and hyenas
Acknowledgments
Several chapters of this book are based on papers first published else
where. I have made no attempt to preserve the form of the originals
and have gutted them whenever it was necessary for the main argu
ment. For the sake of readers without a prior knowledge of science
studies, I have kept the references to a minimum; more annotation can
be found in the original publications.
I thank the editors and publishers of the following journals and
books, first for having accepted my strange papers, and then for allow
ing them to be conjoined here: “Do Scientific Objects Have a History?
Pasteur and Whitehead in a Bath of Lactic Acid,” Common Knowledge 5,
no. 1 (1993): 76-91 (translated by Lydia Davis). “Pasteur on Lactic Acid
Yeast—A Partial Semiotic Analysis,” Configurations 1, no. 1 (1993): 127-
142. “On Technical Mediation,” Common Knowledge 3, no. 2 (1994): 29-
64. “Joliot: History and Physics Mixed Together,” in Michel Serres,
ed.. History of Scientific Thought (London: Blackwell, 1995), 611-635.
“The ‘Pedofil’ of Boa Vista: A Photo-Philosophical Montage,” Common
knowledge 4, no. 1 (1995): 145-187. “Socrates’ and Callicles’ Settlement,
or Ihe Invention of the Impossible Body Politic,” Configurations, 5, no. 2
(Spring i997):i89-240. “A Few Steps toward the Anthropology of the
Uonoclastic Gesture,” Science in Context 10, no. 1 (1998): 62-83.
So many people have read preliminary drafts of parts of this work
lhal I have lost track of what is theirs and what is mine. As usual,
Michel t alion and Isabelle Stengers have provided essential guidance.
Behind Ihe mask of an anonymous referee, Mario Biagioli was instru-
menlal in giving (lu* book ils final shape. For more than ten years I
ha\e henehled liom I indsav Waleis’s generosity as an editor. Once
\ 11
ACKNOWLEDGMENTS
viii
again he offered a shelter for my work. My main gratitude goes, how
ever, to John Tresch, who streamlined the language and logic of the
manuscript. If readers are not satisfied with the result, they should try
to imagine the jungle through which John managed to blaze this tan
gled trail!
I should warn the reader that this is not a book about new facts, nor
it is exactly a book of philosophy. In it, using only very rudimentary
tools, I simply try to present, in the space left empty by the dichotomy
between subject and object, a conceptual scenography for the pair hu
man and nonhuman. I agree that powerful arguments and detailed
empirical case studies would be better, but, as sometimes happens in
detective stories, a somewhat weaker, more solitary, and more adven
turous strategy may succeed against the kidnapping of scientific disci
plines by science warriors where others have failed.
One last caveat. Throughout this book I use the expression “science
studies” as if this discipline exists and is a homogeneous body of work
with a single coherent metaphysics. It would be an understatement to
say that this is far from the case. Most of my colleagues disagree with
my portrayal. As I do not enjoy being isolated and instead thrive on
the conversations involved in a collective undertaking, I present sci
ence studies as if it is a unified field to which I belong.
Contents
1. “Do You Believe in Reality?” 1

Newsfrom the Trenches of the Science Wars
2. Circulating Reference 24
Sampling the Soil in the Amazon Forest
Science’s Blood Flow 80
An Examplefrom Joliot’s Scientific Intelligence
4. From Fabrication to Reality 113
Pasteur and His Lactic Acid Ferment
5. The Historicity of Things 145
Where Were Microbes before Pasteur?
a. A Collective of Humans and Nonhumans 174
Following Daedalus’s Labyrinth
/ The Invention of the Science Wars 216
The Settlement of Socrates and Callicles
s A Politics Freed from Science 236
'The body Cosmopolitic
'» I'lie Slight Surprise of Action 266
Fails, Fetishes, Factishes
( 'niu lusion 293

What ( 011I1 main e ] \ ill I ice Pa in lain s Hope?
CONTENTS
Glossary 303
Bibliography 312
Index 317
Words and phrases that I use in a technical sense

a u t h o r ’s n o t e :
are marked with an asterisk; for their definitions see the Glossary.
Lucifer is the chap who bringsfalse light . . .
Iam shrouding them in the darkness of truth.
—Lakatos to Feyerabend
C H A P T E R O N E
“Do You Believe in Reality?”

News from the Trenches of the Science Wars
“I have a question for you,” he said, taking out of his pocket a crum
pled piece of paper on which he had scribbled a few key words. He
look a breath: “Do you believe in reality?”
“But of course!” I laughed. “What a question! Is reality something
we have to believe in?”
I le had asked me to meet him for a private discussion in a place I
lovind as bizarre as the question: by the lake near the chalet, in this
\t range imitation of a Swiss resort located in the tropical mountains of
Iei esopolis in Brazil. Has reality truly become something people have
lo believe in, I wondered, the answer to a serious question asked in a
hushed and embarrassed tone? Is reality something like God, the topic
<>( a confession reached after a long and intimate discussion? Are there
people on earth who don’t believe in reality?
When I noticed that he was relieved by my quick and laughing an-
.wrr, I was even more baffled, since his relief proved clearly enough
111.11 he had anticipated a negative reply, something like “Of course not!
I >o v<>u think I am that naive?” This was not a joke, then: he really was
«oik erned, and his query had been in earnest.
' I have two more questions,” he added, sounding more relaxed. “Do
« know more than we used to?”
But of course! A thousand times more!”
Bui is science cumulative?” he continued with some anxiety, as if
In did not want to be won over too fast.
I guess so,” I replied, “although I am less positive on this one, since
i In si ien< es also loi gel so much, so much of their past and so much of
i
PANDORA'S HOPE
their bygone research programs—but, on the whole, let’s say yes. Why
are you asking me these questions? Who do you think I am?”
I had to switch interpretations fast enough to comprehend both the
monster he was seeing me as when he raised these questions and his
touching openness of mind in daring to address such a monster pri
vately. It must have taken courage for him to meet with one of these
creatures that threatened, in his view, the whole establishment of sci
ence, one of these people from a mysterious field called “science stud
ies,” of which he had never before met a flesh-and-blood representa
tive but which—at least so he had been told—was another threat to
science in a country, America, where scientific inquiry had never had a
completely secure foothold.
He was a highly respected psychologist, and we had both been in
vited by the Wenner-Grenn Foundation to a gathering made up of
two-thirds scientists and one-third “science students.” This division
itself, announced by the organizers, baffled me. How could we be pit
ted against the scientists? That we are studying a subject matter does
not mean that we are attacking it. Are biologists anti-life, astronomers
anti-stars, immunologists anti-antibodies? Besides, I had taught for
twenty years in scientific schools, I wrote regularly in scientific jour
nals, I and my colleagues lived on contract research carried out on be
half of many groups of scientists in industry and in the academy. Was I
not part of the French scientific establishment? I was a bit vexed to be
excluded so casually. Of course I am just a philosopher, but what
would my friends in science studies say? Most of them have been
trained in the sciences, and several of them, at least, pride themselves
on extending the scientific outlook to science itself. They could be la
beled as members of another discipline or another subfield, but cer
tainly not as “anti-scientists” meeting halfway with scientists, as if the
two groups were opposing armies conferring under a flag of truce be
fore returning to the battlefield!
I could not get over the strangeness of the question posed by this
man I considered a colleague, yes, a colleague (and who has since be
come a good friend). If science studies has achieved anything, I
thought, surely it has added reality to science, not withdrawn any
Irom it. Instead of the stuffed scientists hanging on Ilie walls ol the
armchair philosophers ol science ol the past, we have poiliaycd lively
ch.uacleis, immersed in Ilicit laboialoi ies, lull ol passion, loaded with
DO YOU BELIEVE IN REALITY?
instruments, steeped in know-how, closely connected to a larger and

more vibrant milieu. Instead of the pale and bloodless objectivity of
science, we have all shown, it seemed to me, that the many
nonhumans mixed into our collective life through laboratory practice
have a history, flexibility, culture, blood—in short, all the characteris-
Iics that were denied to them by the humanists on the other side of the
campus. Indeed, I naively thought, if scientists have a faithful ally, it is
we, the “science students” who have managed over the years to inter
est scores of literary folk in science and technology, readers who were
convinced, until science studies came along, that “science does not
think” as Heidegger, one of their masters, had said.
The psychologist’s suspicion struck me as deeply unfair, since he did
nol seem to understand that in this guerrilla warfare being conducted
in (he no-man’s-land between the “two cultures,” we were the ones be
ing attacked by militants, activists, sociologists, philosophers, and
technophobes of all hues, precisely because of our interest in the inner
workings of scientific facts. Who loves the sciences, I asked myself,
more than this tiny scientific tribe that has learned to open up facts,
machines, and theories with all their roots, blood vessels, networks,
ilu/omes, and tendrils? Who believes more in the objectivity of sci
ence (han those who claim that it can be turned into an object of in-
11111ry7
Then I realized that I was wrong. What I would call “adding realism
D> science” was actually seen, by the scientists at this gathering, as a
lineal to the calling of science, as a way of decreasing its stake in truth
im I Iheir claims to certainty. How has this misunderstanding come
ibe ml ? I low could I have lived long enough to be asked in all serious
ness lhis incredible question: “Do you believe in reality?” The dis-
i nu e between what I thought we had achieved in science studies and
\ hal was implied by this question was so vast that I needed to retrace
m\ sieps a bit. And so this book was born.
TIk*Strange Invention of an “Outside” World

I In ie is no nalural situation on earth in which someone could be
i I eil (Ins sliangest of all questions: “Do you believe in reality?” To
i I suc h a question one has to become so distant from reality that the
!• n ol An///i; it cutiicly becomes plausible and (his lear itsell has an
P A N D O R A ’S HOPE
intellectual history that should at least be sketched. Without this de

tour we would never be able to fathom the extent of the misunder
standing between my colleague and me, or to measure the extraordi
nary form of radical realism that science studies has been uncovering.
I remembered that my colleague’s question was not so new. My
compatriot Descartes had raised it against himself when asking how
an isolated mind could be absolutely as opposed to relatively sure of
anything about the outside world. Of course, he framed his question
in a way that made it impossible to give the only reasonable answer,
which we in science studies have slowly rediscovered three centuries
later: that we are relatively sure of many of the things with which we
are daily engaged through the practice of our laboratories. By Des
cartes’s time this sturdy relativism*, based on the number of relations
established with the world, was already in the past, a once-passable
path now lost in a thicket of brambles. Descartes was asking for abso
lute certainty from a brain-in-a-vat, a certainty that was not needed
when the brain (or the mind) was firmly attached to its body and the
body thoroughly involved n its normal ecology. As in Curt Siodmak’s
novel Donovan’s Brain, absolute certainty is the sort of neurotic fantasy
that only a surgically removed mind would look for after it had lost ev
erything else. Like a heart taken out of a young woman who has just
died in an accident and soon to be transplanted into someone else’s
thorax thousands of miles away, Descartes’s mind requires artificial
life-support to keep it viable. Only a mind put in the strangest posi
tion, looking at a world from the inside out and linked to the outside by
nothing but the tenuous connection of the gaze, will throb in the con
stant fear of losing reality; only such a bodiless observer will desper
ately look for some absolute life-supporting survival kit.
For Descartes the only route by which his mind-in-a-vat could re
establish some reasonably sure connection with the outside world was
through God. My friend the psychologist was thus right to phrase his
query using the same formula I had learned in Sunday school: “Do
you believe in reality?”—“Credo in unum Deum,” or rather, “Credo
in unam realitam,” as my friend Donna Haraway kept chanting in
Teresopolis! After Descartes, however, many people thought that go
ing through God to reach the world was a bit expensive and far
fetched. They looked for a shortcut. They wondcied whelliei the
world could directly send us enough information to produce a stable

image of itself in our minds.
But in asking this question the empiricists kept going along the
same path. They did not retrace their steps. They never plugged the
wriggling and squiggling brain back into its withering body. They
were still dealing with a mind looking through the gaze at a lost out
side world. They simply tried to train it to recognize patterns. God
was out, to be sure, but the tabula rasa of the empiricists was as dis
connected as the mind in Descartes’s times. The brain-in-a-vat simply
exchanged one survival kit for another. Bombarded by a world re
duced to meaningless stimuli, it was supposed to extract from these
slimuli everything it needed to recompose the world’s shapes and sto-
i ics. The result was like a badly connected TV set, and no amount of
liming made this precursor of neural nets produce more than a fuzzy
set of blurry lines, with white points falling like snow. No shape was
i (.'cognizable. Absolute certainty was lost, so precarious were the con
nections of the senses to a world that was pushed ever further outside.
I here was too much static to get any clear picture.
I he solution came, but in the form of a catastrophe from which we
.ue only now beginning to extricate ourselves. Instead of retracing
l heir steps and taking the other path at the forgotten fork in the road,
philosophers abandoned even the claim to absolute certainty, and set
tled instead on a makeshift solution that preserved at least some ac-
( ess (o an outside reality. Since the empiricists’ associative neural net
was unable to offer clear pictures of the lost world, this must prove,
Ihev said, that the mind (still in a vat) extracts from itself everything
II needs to form shapes and stories. Everything, that is, except the real-
11\ 11self. Instead of the fuzzy lines on the poorly tuned TV set, we got
ilie fixed tuning grid, molding the confused static, dots, and lines of
Ilie empiricist channel into a steady picture held in place by the mind-
.( I s predesigned categories. Kant’s a priori started this extravagant
lt»nn of constructivism, which neither Descartes, with his detour
ilnough God, nor Hume, with his shortcut to associated stimuli,
\ mild ever have dreamed of.
Now, with the Konigsberg broadcast, everything was ruled by the
mind iisell and reality came in simply to say that it was there, indeed,
iikI not imaginai y! I <n the banquet of reality, the mind provided the
P A N D O R A ’ S HOPE
food, and the inaccessible things-in-themselves to which the world

had been reduced simply dropped by to say “We are here, what you
eat is not dust,” but otherwise remained mute and stoic guests. If we
abandon absolute certainty, Kant said, we can at least retrieve univer
sality as long as we remain inside the restricted sphere of science, to
which the world outside contributes decisively but minimally. The
rest of the quest for the absolute is to be found in morality, another a
priori certainty that the mind-in-the-vat extracts from its own wiring.
Under the name of a “Copernican Revolution”* Kant invented this
science-fiction nightmare: the outside world now turns around the
mind-in-the-vat, which dictates most of that world’s laws, laws it has
extracted from itself without help from anyone else. A crippled despot
now ruled the world of reality. This philosophy was thought, strangely
enough, to be the deepest of all, because it had at once managed to
abandon the quest for absolute certainty and to retain it under the
banner of “universal aprioris,” a clever sleight of hand that hid the lost
path even deeper in the thickets.
Do we really have to swallow these unsavory pellets of textbook
philosophy to understand the psychologist’s question? I am afraid
so, because otherwise the innovations of science studies will remain
invisible. The worst is yet to come. Kant had invented a form of
constructivism in which the mind-in-the-vat built everything by itself
but not entirely without constraints: what it learned from itself had to
be universal and could be elicited only by some experiential contact
with a reality out there, a reality reduced to its barest minimum, but
there nonetheless. For Kant there was still something that revolved
around the crippled despot, a green planet around this pathetic sun. It
would not be long before people realized that this “transcendental
Ego,” as Kant named it, was a fiction, a line in the sand, a negotiating
position in a complicated settlement to avoid the complete loss of the
world or the complete abandonment of the quest for absolute cer
tainty. It was soon replaced by a more reasonable candidate, society*.
Instead of a mythical Mind giving shape to reality, carving it, cutting
it, ordering it, it was now the prejudices, categories, and paradigms of
a group of people living together that determined the représentai ions
of every one of those people. This new definition, however, in spile
of the use of the word “social,” had only a snpei filial lesemblame to
the realism to which we science students have become attached, and

which I will outline over the course of this book.
First, this replacement of the despotic Ego with the sacred “society”
did not retrace the philosophers’ steps but went even further in dis
tancing the individual’s vision, now a “view of the world,” from the
definitely lost outside world. Between the two, society interposed its
filters; its paraphernalia of biases, theories, cultures, traditions, and
standpoints became an opaque window. Nothing of the world could
pass through so many intermediaries and reach the individual mind.
People were now locked not only into the prison of their own catego
ries but into that of their social groups as well. Second, this “society”
it self was just a series of minds-in-a-vat, many minds and many vats to
he sure, but each of them still composed of that strangest of beasts: a
detached mind gazing at an outside world. Some improvement! If
prisoners were no longer in isolated cells, they were now confined to
l he same dormitory, the same collective mentality. Third, the next
shift, from one Ego to multiple cultures, jeopardized the only good
Ilung about Kant, that is, the universality of the a priori categories, the
only bit of ersatz absolute certainty he had been able to retain. Every-
<me was not locked in the same prison any more; now there were many
pi isons, incommensurable, unconnected. Not only was the mind dis-
( onnected from the world, but each collective mind, each culture was
disconnected from the others. More and more progress in a philoso
phy dreamed up, it seems, by prison wardens.
But there was a fourth reason, even more dramatic, even sadder,
ih.il made this shift to “society” a catastrophe following fast on the
Iice Is of the Kantian revolution. The claims to knowledge of all these
Ih>01 minds, prisoners in their long rows of vats, were now made part
«il an even more bizarre history, were now associated with an even
more ancient threat, the fear of mob rule. If my friend’s voice quivered
is lie asked me “Do you believe in reality?” it was not only because he
Ic.ued llial all connection with the outside world might be lost, but
dim e all because he worried that I might answer, “Reality depends on
whatever the mob thinks is right at any given time.” It is the réso
n a n t e ol these two fears, the loss of any certain access to reality and
l lie in v a sio n by the mob, Ihat makes his question at once so unfair and
.1) SCI IOIIN
But before we disentangle this second threat, let me finish with the
first one. The sad story, unfortunately, does not end here. However in
credible it seems, it is possible to go even further along the wrong
path, always thinking that a more radical solution will solve the prob
lems accumulated from the past decision. One solution, or more ex
actly another clever sleight of hand, is to become so very pleased with
the loss of absolute certainty and universal a prions that one rejoices in
abandoning them. Every defect of the former position is now taken to
be its best quality. Yes, we have lost the world. Yes, we are forever pris
oners of language. No, we will never regain certainty. No, we will
never get beyond our biases. Yes, we will forever be stuck within our
own selfish standpoint. Bravo! Encore! The prisoners are now gagging
even those who ask them to look out their cell windows; they will “de
construct,” as they say—which means destroy in slow motion—any
one who reminds them that there was a time when they were free and
when their language bore a connection with the world.
Who can avoid hearing the cry of despair that echoes deep down,
carefully repressed, meticulously denied, in these paradoxical claims
for a joyous, jubilant, free construction of narratives and stories by
people forever in chains? But even if there were people who could say
such things with a blissful and light heart (their existence is as uncer
tain to me as that of the Loch Ness monster, or, for that matter, as un
certain as that of the real world would be to these mythical creatures),
how could we avoid noticing that we have not moved an inch since
Descartes? That the mind is still in its vat, excised from the rest, dis
connected, and contemplating (now with a blind gaze) the world (now
lost in darkness) from the very same bubbling glassware? Such people
may be able to smile smugly instead of trembling with fear, but they
are still descending further and further along the spiraling curves of
the same hell. At the end of this chapter we will meet these gloating
prisoners again.
In our century, though, a second solution has been proposed, one
that has occupied many bright minds. This solution consists of taking
only a part of the mind out of the vat and then doing the obvious
thing, that is, offering it a body again and putting the reassembled ag
gregate back into relation with a world that is no longer a spectacle at
which we gaze but a lived, sell evident, and unreflexive extension of
ourselves. In appearance1, the piog,less is immense, and the descent
into damnation suspended, since we no longer have a mind dealing

with an outside world, but a lived world to which a semi-conscious
and intentional body is now attached.
Unfortunately, however, in order to succeed, this emergency opera-
Iion must chop the mind into even smaller pieces. The real world, the
one known by science, is left entirely to itself. Phenomenology deals
only with the world-for-a-human-consciousness. It will teach us a lot
about how we never distance ourselves from what we see, how we
never gaze at a distant spectacle, how we are always immersed in the
world’s rich and lived texture, but, alas, this knowledge will be of no
use in accounting for how things really are, since we will never be able
lo escape from the narrow focus of human intentionality. Instead of
exploring the ways we can shift from standpoint to standpoint, we will
always be fixed in the human one. We will hear much talk about the
leal, fleshy, pre-reflexive lived world, but this will not be enough to
«over the noise of the second ring of prison doors slamming even
more tightly shut behind us. For all its claims to overcoming the dis-
i.mee between subject and object—as if this distinction were some-
lb mg that could be overcome! as if it had not been devised so as not to
Iie overcome!—phenomenology leaves us with the most dramatic split
m ibis whole sad story: a world of science left entirely to itself, en-
11u ly cold, absolutely inhuman; and a rich lived world of intentional
lances entirely limited to humans, absolutely divorced from what
lb mgs are in and for themselves. A slight pause on the way down be
lote sliding even further in the same direction.
Why not choose the opposite solution and forget the mind-in-a-vat
iliogelher? Why not let the “outside world” invade the scene, break
i lie glassware, spill the bubbling liquid, and turn the mind into a brain,
inio a neuronal machine sitting inside a Darwinian animal strug-
•Img lor ils life? Would that not solve all the problems and reverse the
I iial downward spiral? Instead of the complex “life-world” of the
pbmomenologisls, why not study the adaptation of humans, as natu-
i ilr.ls have studied all other aspects of “life”? If science can invade ev-
i i\lbmg, it surely can put an end to Descartes’s long-lasting fallacy
uni make the mind a wriggling and squiggling part of nature. This
(mid leilamly please my Iriend the psychologist—or would it? No,
bii.mxe the mgicdients that make up this “nature,” this hegemonic
md all eiuompassmg naluie*, whiili would now include the human
lO
species, are the very same ones that have constituted the spectacle of
a world viewed from inside by a brain-in-a-vat. Inhuman, reduction
ist, causal, law-like, certain, objective, cold, unanimous, absolute—
all these expressions do not pertain to nature as such, but to nature
viewed through the deforming prism of the glass vessel!
If there is something unattainable, it is the dream of treating nature
as a homogeneous unity in order to unify the different views the sci
ences have of it! This would require us to ignore too many controver
sies, too much history, too much unfinished business, too many loose
ends. If phenomenology abandoned science to its destiny by limiting it
to human intention, the opposite move, studying humans as “natural
phenomena,” would be even worse: it would abandon the rich and
controversial human history of science—and for what? The averaged-
out orthodoxy of a few neurophilosophers? A blind Darwinian pro
cess that would limit the mind’s activity to a struggle for survival to
“fit” with a reality whose true nature would escape us forever? No, no,
we can surely do better, we can surely stop the downward slide and re
trace our steps, retaining both the history of humans’ involvement in
the making of scientific facts and the sciences’ involvement in the
making of human history.
Unfortunately, we can’t do this, not yet. We are prevented from
returning to the lost crossroads and taking the other path by the dan
gerous bogeyman I mentioned earlier. It is the threat of mob rule
that stops us, the same threat that made my friend’s voice quake and
quiver.
The Fear of Mob Rule

As I said, two fears lay behind my friend’s strange question. The first
one, the fear of a mind-in-a-vat losing its connection to a world out
side, has a shorter history than the second, which stems from this tru
ism: if reason does not rule, then mere force will take over. So great is
this threat that any and every political expedient is used with impu
nity against those who are deemed to advocate force against reason.
But where does this striking opposition between the camp of reason
and the camp of force come from? It comes from an old and venerable
debate, one that probably occurs in many places but tlial is staged
most clearly and influentially in Plato’s (ioiyjns In Ibis dialog,, wlntli I
11
will examine in more detail in Chapters 7 and 8, Socrates, the true sci
entist, confronts Callicles, another of those monsters who must be in
terviewed in order to expose their nonsense, this time not on the
shores of a Brazilian lake but in the agora in Athens. He tells Callicles:
"You’ve failed to notice how much power geometrical equality has among
1uxIs and men, and this neglect of geometry has led you to believe that
one should try to gain a disproportionate share of things” (508a).1
Callicles is an expert at disproportion, no doubt about that. “I
1hink,” he boasts in a preview of Social Darwinism, “we only have
10 look at nature to find evidence that it is right for better to have
.1 greater share than worse. . . The superior person shall dominate
1he inferior person and have more than him” (483c-d). Might makes
Tight, Callicles frankly admits. But, as we shall see at the end of this
hook, there is a little snag. As both of the two protagonists are quick to
point out, there are at least two sorts of Mights to consider: that of
( .(Hides and that of the Athenian mob. “What else do you think I’ve
hern saying?” Callicles asks. “Law consists of the statements made by
111 assembly of slaves and assorted other forms of human debris who
11mid be completely discounted if it weren ’tfor thefact they do hauephysi-
, ,il strength at their disposal” (489c). So the question is not simply the
opposition of force and reason, Might and Right, but the Might of the
.oh 1ary patrician against the superior force of the crowd. How can the
1(nnbined forces of the people of Athens be nullified? “Here’s your po
il ion, then,” Socrates ironizes: “a single clever person is almost bound
io be superior to ten thousand fools; political power should be his and
1hey should be his subjects; and it is appropriate for someone with po-
IiIh al power to have more than his subjects” (490a). When Callicles
peaks of brute force, what he means is an inherited moral force supe-
1101 io that of ten thousand brutes.
Tail is it fair for Socrates to practice irony on Callicles? What sort of
ile.pioportion is Socrates himself setting in motion? What sort of
pi >uri is he trying to wield? The Might that Socrates sides with is the
>,i,u 1 <>/ irasou, “(he power of geometrical equality,” the force which
iules over gods and men,” which he knows, which Callicles and the
mol) ignore. As we shall see, (here is a second little snag here, because
1 I 11 .c I lie 1ei enl 1 1.111' . 1 .11 11mi In l\< >1 >■11 VV.ilei h eld (( ) \ l o n l : ( K l o n l I Iniversily Tress,
1 1 111
12
there are two forces of reason, one directed against Callicles, the ideal
foil, and the other directed sideways, aimed at reversing the balance of
power between Socrates and all the other Athenians. Socrates is also
looking for a force able to nullify that of “ten thousand fools.” He too
tries to get the biggest share. His success at reversing the balance of
forces is so extraordinary that he boasts, at the end of the Gorgias, of
being “the only real statesman of Athens,” the only winner of the big
gest share of all, an eternity of glory that will be awarded to him by
Rhadamantes, Aeacus, and Minos, who preside over the tribunal of
hell! He ridicules all the famous Athenian politicians, Pericles in
cluded, and he alone, equipped with “the power of geometrical equal
ity,” will rule over the citizens of the city even beyond death. One of
the first of many in the long literary history of mad scientists.
“As if your slapdash history of modern philosophy is not enough,”
the reader may complain, “do you also have to drag us all the way back
to the Greeks just to account for the question asked by your psycholo
gist in Brazil?” I am afraid both of these detours were necessary, be
cause only now can the two threads, the two threats, be tied together
to explain my friend’s worries. Only after these digressions can my po
sition, I hope, be clarified at last.
Why, in the first place, did we even need the idea of an outside world
looked at through a gaze from the very uncomfortable observation
post of a mind-in-a-vat? This has puzzled me ever since I started in the
field of science studies almost twenty-five years ago. How could it be
so important to maintain this awkward position, in spite of all the
cramps it gave philosophers, instead of doing the obvious: retracing
our steps, pruning back the brambles hiding the lost fork in the road,
and firmly walking on the other, forgotten path? And why burden this
solitary mind with the impossible task of finding absolute certainty in
stead of plugging it into the connections that would provide it with all
the relative certainties it needed to know and to act? Why shout out of
both sides of our mouths these two contradictory orders: “Be abso
lutely disconnected!” “Find absolute proof that you are connected!”
Who could untangle such an impossible double bind? No wonder so
many philosophers wound up in asylums. In order to justify such a
self-inflicted, maniacal torture, we would have to be pursuing a loltiei
goal, and such indeed has been the case. This is the plaie while the
Iwo Ihreads connect : it is in oi dei to a\ oui Ihe inhuman 11owd Ihat we
13
need to rely on another inhuman resource, the objective object un-

11niched by human hands.
lb avoid the threat of a mob rule that would make everything lowly,
monstrous, and inhuman, we have to depend on something that has
no human origin, no trace of humanity, something that is purely,
blindly, and coldly outside of the City. The idea of a completely outside
oild dreamed up by epistemologists is the only way, in the eyes of
moialists, to avoid falling prey to mob rule. Only inhumanity will quash
inhumanity. But how is it possible to imagine an outside world? Has
m\one ever seen such a bizarre oddity? No problem. We will make
i lu- world into a spectacle seen from the inside.
Io obtain such a contrast, we will imagine that there is a mind-in-a-
ii that is totally disconnected from the world and accesses it only
ilnongh one narrow, artificial conduit. This minimal link, psycholo-
r.is are confident, will be enough to keep the world outside, to keep
ihe mind informed, provided we later manage to rig up some absolute
mt .ms of getting certainty back—no mean feat, as it turns out. But this
i\ we will achieve our overarching agenda: to keep the crowds at bay. It
i Iici ause we want to fend off the irascible mob that we need a world
11m i is totally outside—while remaining accessible!—and it is in order
i*i i each this impossible goal that we came up with the extraordinary
m ml ion of a mind-in-a-vat disconnected from everything else, striv-
iii)' lor absolute truth, and, alas, failing to get it. As we can see in Fig-
1111 i i, epistemology, morality, politics, and psychology go hand in hand and
n< 11lining at the same settlement*.
I Ins is the argument of this book. It is also the reason the reality of
• h nc e studies is so difficult to locate. Behind the cold epistemological
|in si ion—can our representations capture with some certainty stable
I. limes of the world out there?—the second, more burning anxiety is
•I i\ s lurking: can we find a way to fend off the people? Conversely,
hi 11111(1any définition of the “social” is the same worry: will we still be
11■!< lo use objective reality to shut the mob’s too many mouths?
I\ 11lend s question, on the shore of the lake, shaded by the cha-
I i . iooI liom the tropical noontime sun in this austral winter, be-
ima s clear at last: “Do you believe in reality?” means “Are you will-
iii ■In at t epl this settlement of epistemology, morality, politics, and
I i lio lo g y ? ” to whit h I he quit, k and laughing answer is, obviously:
ill ( >! t o u i se not ! Who do you think 1am? I low could I believe real
PANDORA'S HOPE
14
God
Theology
Society
Figure 1.1 The modernist settlement. For science studies there is no sense in talk
ing independently of epistemology, ontology, psychology, and politics—not to
mention theology. In short: “out there,” “nature”; “in there,” the mind; “down
there,” the social; “up there,” God. We do not claim that these spheres are cut off
from one another, but rather that they all pertain to the same settlement, a settle
ment that can be replaced by several alternative ones.
ity to be the answer to a question of belief asked by a brain-in-a-vat

terrified of losing contact with an outside world because it is even
more terrified of being invaded by a social world stigmatized as inhu
man?” Reality is an object of belief only for those who have started
down this impossible cascade of settlements, always tumbling into a
worse and more radical solution. Let them clean up their own mess
and accept the responsibility for their own sins. My trajectory has al
ways been different. “Let the dead bury the dead,” and, please, listen
for one minute to what we have to say on our own account, instead of
trying to shut us up by putting in our mouths the words that Plato, all
those centuries ago, placed in the mouths of Socrates and Callicles to
keep the people silent.
Science studies, as I see it, lias made two relaied discoveries that
were very slow in coming because of Ilie power ol Ilie* sel Ilemeiil Iliai I
have now exposed as well as lor a lew olliei leasons I will explain
15
I. iler. This joint discovery is that neither the object nor the social has the
inhuman character that Socrates’ and Callicles’ melodramatic show re-
tHiired. When we say there is no outside world, this does not mean
Ihat we deny its existence, but, on the contrary, that we refuse to grant
il Ihe ahistorical, isolated, inhuman, cold, objective existence that it
was given only to combat the crowd. When we say that science is so-
<ul, the word social for us does not bear the stigma of the “human de
là is,” of the “unruly mob” that Socrates and Callicles were so quick to
invoke in order to justify the search for a force strong enough to re-
\ ci se the power of “ten thousand fools.”
Neither of these two monstrous forms of inhumanity—the mob
down there,” the objective world “out there”—interests us very
nint h. And thus we have no need for a mind- or brain-in-a-vat, that
. nppled despot constantly fearful of losing either “access” to the
\ mid or its “superior force” against the people. We long neither for
i he absolute certainty of a contact with the world nor for the absolute
, ( i lainty of a transcendent force against the unruly mob. We do not
/m k certainty, because we never dreamed of dominating the people. For
II, i here is no inhumanity to be quashed with another inhumanity. Hu-
in ms and nonhumans are enough for us. We do not need a social
mid to break the back of objective reality, nor an objective reality to
ilei ice the mob. It is quite simple, even though it may sound incredi
ble m these times of the science wars: we are not at war.
\s soon as we refuse to engage the scientific disciplines in this dis
pute about who should hold sway over the people, the lost crossroads
i i ( discovered, and there is no major difficulty in treading along the
m jdcc ted path. Realism now returns in force, as will be made obvious,
I In>|>c, in later chapters, which should look like milestones along the
i. mic to a more “realistic realism.” My argument in this book recapitu-
I iii •. die halting “two steps forward, one step back” advance of science
i mlics along this long-forgotten pathway.
\\c started when we first began to talk about scientific practice*
uid thus offered a more realistic account of science-in-the-making,
i. mndiiig it lirmly in laboratory sites, experiments, and groups of col
li (j*iics, as I do in Chapters 2 and 3. Facts, we found, were clearly fabri-
111<I I'hen realism gushed forth again when, instead of talking about
>b|i ( Is and objet livily, we began to speak ol nonhumans* that were so
il di cd thiouj’h the lahoialmy and with which scientists and engi
16
neers began to swap properties. In Chapter 4 we see how Pasteur

makes his microbes while the microbes “make their Pasteur”; Chapter
6 offers a more general treatment of humans and nonhumans folding
into each other, forming constantly changing collectives. Whereas ob
jects had been made cold, asocial, and distant for political reasons, we
found that nonhumans were close, hot, and easier to enroll and to en
list, adding more and more reality to the many struggles in which sci
entists and engineers had engaged.
But realism became even more abundant when nonhumans began
to have a history, too, and were allowed the multiplicity of interpreta
tions, the flexibility, the complexity that had been reserved, until then,
for humans (see Chapter 5). Through a series of counter-Copernican
revolutions*, Kant’s nightmarish fantasy slowly lost its pervasive
dominance over the philosophy of science. There was again a clear
sense in which we could say that words have reference to the world and
that science grasps the things themselves (see Chapters 2 and 4).
Naïveté was back at last, a naïveté appropriate for those who had
never understood how the world could be “outside” in the first place.
We have yet to provide a real alternative to that fateful distinction be
tween construction and reality; I attempt to provide one here with the
notion of “factish.” As we see in Chapter 9, “factish” is a combination
of the words “fact” and “fetish,” in which the work of fabrication has
been twice added, canceling the twin effects of belief and knowledge.
Instead of the three poles—a reality “out there,” a mind “in there,”
and a mob “down there”—we have finally arrived at a sense of what I
call a collective*. As the explication of the Gorgias in Chapters 7 and 8
demonstrates, Socrates has defined this collective very well before
switching to his bellicose collusion with Callicles: “The expert’s opin
ion is that co-operation, love, order, discipline, and justice bind heaven
and earth, gods and men. That’s why they call the universe an ordered
whole, my friend, rather than a disorderly mess or an unruly shambles"
(507e- 508a).
Yes, we live in a hybrid world made up at once of gods, people, stars,
electrons, nuclear plants, and markets, and it is our duty to turn it
into either an “unruly shambles” or an “ordered whole,” a cosmos as
the Greek text puts it, undertaking what Isabelle Slengers gives the
beautiful name of eosmopolilies* (Slengers I99t>) Ome theie is no
longer a mind in a vat looking, lluough the ga/e at an outside woild,
17
Ilie search for absolute certainty becomes less urgent, and thus there is
no great difficulty in reconnecting with the relativism, the relations,
ilie relativity on which the sciences have always thrived. Once the so-
ual realm no longer bears these stigmata branded upon it by those
wlio want to silence the mob, there is no great difficulty in recognizing
llie human character of scientific practice, its lively history, its many
( onnections with the rest of the collective. Realism comes back like
I>lood through the many vessels now reattached by the clever hands of
ilie surgeons—there is no longer any need for a survival kit. After fol
lowing this route, no one would even think of asking the bizarre ques-
i ion “Do you believe in reality?”—at least not of asking us!
The Originality of Science Studies

cvertheless, my friend the psychologist would still be entitled to
Im»se another, more serious query: “Why is it that, in spite of what you
i l.iitn your field has achieved, I was tempted to ask you my silly ques
tion as if it were a worthwhile one? Why is it that in spite of all these
philosophies you zigzagged me through, I still doubt the radical real-
i in you advocate? I can’t avoid the nasty feeling that there is a science
n going on. In the end, are you a friend of science or its enemy?”
I hi ce different phenomena explain, to me at least, why the novelty
. 'I 'science studies” cannot be registered so easily. The first is that we
no situated, as I said, in the no-man’s-land between the two cultures,
iniK 11 like the fields between the Siegfried and Maginot lines in which
I it in h and German soldiers grew cabbages and turnips during the
plumy war” in 1940. Scientists always stomp around meetings talking
IImM11 “bridging the two-culture gap,” but when scores of people from
•>ni side the sciences begin to build just that bridge, they recoil in hor-
0-1 and want to impose the strangest of all gags on free speech since
<m1.ili s: only scientists should speak about science!
lust imagine if that slogan were generalized: only politicians should
p< al about politics, businessmen about business; or even worse: only
mi , will speak about rats, frogs about frogs, electrons about electrons!
pi ci li implies by definition the risk of misunderstanding across the
hit *1*gaps between dillcrent species. If scientists want to bridge the
1 h i nil 111e divide lor good, Ihey will have to get used to a lot of noise
uni \rs, mote lhan .1 hide Ini ol nonsense Aller all, (he humanists
18
and the literati do not make such a fuss about the many absurdities ut
tered by the team of scientists building the bridge from the other end.
More seriously, bridging the gap cannot mean extending the unques
tionable results of science in order to stop the “human debris” from be
having irrationally. Such an attempt can at best be called pedagogy, at
worst propaganda. This cannot pass for the cosmopolitics that would
require the collective to socialize into its midst the humans, the non
humans, and the gods together. Bridging the two-culture gap cannot
mean lending a helping hand to Socrates’ and Plato’s dreams of utter
control.
But where does the two-culture debate itself originate? In a division
of labor between the two sides of the campus. One camp deems the
sciences accurate only when they have been purged of any contamina
tion by subjectivity, politics, or passion; the other camp, spread out
much more widely, deems humanity, morality, subjectivity, or rights
worthwhile only when they have been protected from any contact
with science, technology, and objectivity. We in science studies fight
against these two purges, against both purifications at once, and this is
what makes us traitors to both camps. We tell the scientists that the
more connected a science is to the rest of the collective, the better it is, the
more accurate, the more verifiable, the more solid (see Chapter 3)—
and this runs against all the conditioned reflexes of epistemologists.
When we tell them that the social world is good for science’s health,
they hear us as saying that Callicles’ mobs are coming to ransack their
laboratories.
But, against the other camp, we tell the humanists that the more
nonhumans share existence with humans, the more humane a collective
is—and this too runs against what they have been trained for years to
believe. When we try to focus their attention on solid facts and hard
mechanisms, when we say that objects are good for the subjects’
health because objects have none of the inhuman characteristics they
fear so much, they scream that the iron hand of objectivity is turning
frail and pliable souls into reified machines. But we keep defecting and
counter-defecting from both sides, and we insist and insist again that
there is a social history of things and a “thingy” history of humans,
but that neither “the social” nor “the objective world” plays the role
assigned to it by Socrates and Callicles in their grotesque melodrama
If anything, and heat* we tan be* lightly att used ol a slight I,it k ol
19
\ minetry, “science students” fight the humanists who are trying to in-
\ mi a human world purged of nonhumans much more than we combat
i!>«• epistemologists who are trying to purify the sciences of any con-
i imination by the social. Why? Because scientists spend only a frac-
11**n of their time purifying their sciences and, frankly, do not give a
d min about the philosophers of science coming to their rescue, while
i lie humanists spend all their time on and take very seriously the task
■•I liceing the human subjects from the dangers of objectification and
i< iMention. Good scientists enlist in the science wars only in their
Imie time or when they are retired or have run out of grant money,
IMil ilie others are up in arms day and night and even get granting
i|>«,ncies to join in their battle. This is what makes us so angry about
i Mr suspicion of our scientist colleagues. They don’t seem to be able to
dillerentiate friends from foes anymore. Some are pursuing the vain
■h r.mi of an autonomous and isolated science, Socrates’ way, while we
h«• pointing out the very means they need to reconnect the facts to the
h ililies without which the existence of the sciences cannot be sus-
i lined. Who first offered us this treasure trove of knowledge? The sci-
>niisis ihemselves!
I Imd this blindness all the more bizarre because, in the last twenty
i us. many scientific disciplines have joined us, crowding into the
urn no man’s-land between the two lines. This is the second reason
• n iice studies” is so contentious. By mistake, it is caught in the mid-
ll< oI another dispute, this one within the sciences themselves. On one
nIf (here are what could be called the “cold war disciplines,” which
i ill look superficially like the Science of the past, autonomous and de-
i i«lied from the collective; on the other side there are strange imbro-
lm\ of politics, science, technology, markets, values, ethics, facts,
Mu h cannot easily be captured by the word Science with a capital S.
II l Mere is some plausibility in the assertion that cosmology does not
lit e ihe slightest connection with society—although even that is
mug. as Plato reminds us so tellingly—it is hard to say the same of
....... ipsychology, sociobiology, primatology, computer sciences, mar-
I i lni)1, soil science, cryptology, genome mapping, or fuzzy logic, to
ii une )iisl a lew of these active zones, a few of the “disorderly messes”
i hii I.ill's would call them. On the one hand we have a model that
nil applies the e.u lier slogan the less connected a science the
In III i while on llie olliri we have m a n v disciplines, uncert ai n of
PANDORA'S HOPE
ZO
their exact status, striving to apply the old model, unable to reinstate
it, and not yet prepared to mutter something like what we have been
saying all along: “Relax, calm down, the more connected a science is
the better. Being part of a collective will not deprive you of the
nonhumans you socialize so well. It will only deprive you of the po
lemical kind of objectivity that has no other use than as a weapon for
waging a political war against politics.”
To put it even more bluntly, science studies has become a hostage in
a huge shift from Science to what we could call Research (or Science
No. 2, as I will call it in Chapter 8). While Science had certainty, cold
ness, aloofness, objectivity, distance, and necessity, Research appears
to have all the opposite characteristics: it is uncertain; open-ended;
immersed in many lowly problems of money, instruments, and know
how; unable to differentiate as yet between hot and cold, subjective
and objective, human and nonhuman. If Science thrived by behaving
as if it were totally disconnected from the collective, Research is best
seen as a collective experimentation about what humans and nonhumans
together are able to swallow or to withstand. It seems to me that the
second model is wiser than the former. No longer do we have to
choose between Right and Might, because there is now a third party
in the dispute, that is, the collective*; no longer do we have to decide
between Science and Anti-Science, because here too there is a third
party—the same third party, the collective.
Research is this zone into which humans and nonhumans are
thrown, in which has been practiced, over the ages, the most extraor
dinary collective experiment to distinguish, in real time, between
“cosmos” and “unruly shambles” with no one, neither the scientists
nor the “science students,” knowing in advance what the provisional
answer will be. Maybe science studies is anti-Science, after all, but in
that case it is wholeheartedly/or Research, and, in the future, when the
spirit of the times will have taken a firmer grip on public opinion, it
will be in the same camp as all of the active scientists, leaving on the
other side only a few disgruntled cold-war physicists still wishing to
help Socrates shut the mouths of the “ten thousand fools” with an un
questionable and indisputable absolute truth coming from nowhere.
The opposite of relativism, we should never lorgel, is calk'd absolul
ism (Bloor f 1976 I 1991 ).
1 a m being a hi I di si ngenuous, I know bee a list' I liei e is a I Ini d 1ea
21
son that makes it hard to believe that science studies could have
so many goodies to offer. By an unfortunate coincidence, or maybe
Ihrough a strange case of Darwinian mimicry in the ecology of the so-
i ial sciences, or—who knows?—through some case of mutual contam
ination, science studies bears a superficial resemblance to those pris
oners locked in their cells whom we left, a few pages ago, in their slow
descent from Kant to hell and smiling smugly all the way down, since
111c'y claim no longer to care about the ability of language to refer to re
ality. When we talk about hybrids and imbroglios, mediations, prac
tice, networks, relativism, relations, provisional answers, partial con
nections, humans and nonhumans, “disorderly messes,” it may sound
is if we, too, are marching along the same path, in a hurried flight
liom truth and reason, fragmenting into ever smaller pieces the caté
gories that keep the human mind forever removed from the presence
«*1 reality. And yet—there is no need to paper it over—just as there is a
lij’lit inside the scientific disciplines between the model of Science and
ilie model of Research, there is a fight in the social sciences and the
humanities between two opposite models, one that can loosely be
*.illed postmodern* and the other that I have called nonmodern*. Ev-
• i \ ihing the first takes to be a justification for more absence, more de-
l*imking, more negation, more deconstruction, the second takes as a
I'iooI of presence, deployment, affirmation, and construction.
Hie cause of the radical differences as well as of the passing resem-
M.mces is not difficult to ferret out. Postmodernism, as the name indi-
• tics, is descended from the series of settlements that have defined
modernity. It has inherited from these the disconnected mind-in-the-
»i s quest for absolute truth, the debate between Might and Right, the
i idieal distinction between science and politics, Kant’s con
iine tivism, and the critical urge that goes with it, but it has stopped
U lieving it is possible to carry out this implausible program success
fully In this disappointment it shows good common sense, and that is
■'incl liing to say in its favor. But it has not retraced the path of moder-
imi\ all (he way back to the various bifurcations that started this im-
I'ovuble project in the first place. It feels the same nostalgia as mod-
*1111sin, except that it tries to take on, as positive features, the
" <i whelming failures <>l the rationalist project. I lence its apology on
l»i hall ol fallu les and the Sophists, its rejoicing in virtual reality, its
I*Iui ills in g o( “mast ei nan aIives,” il s <.la ini that it is good to be stuck
22
inside one’s own standpoint, its overemphasis on reflexivity, its mad

dening efforts to write texts that do not carry any risk of presence.
Science studies, as I see it, has been engaged in a very different
nonmodern task. For us, modernity has never been the order of the
day. Reality and morality have never been lacking. The fight for or
against absolute truth, for or against multiple standpoints, for or
against social construction, for or against presence, has never been the
important one. The program of debunking, exposing, avoiding being
taken in, steals energy from the task that has always seemed much
more important to the collective of people, things, and gods, namely,
the task of sorting out the “cosmos” from an “unruly shambles.” We
are aiming at a politics of things, not at the bygone dispute about
whether or not words refer to the world. Of course they do! You might
as well ask me if I believe in Mom and apple pie or, for that matter, if I
believe in reality!
Are you still unconvinced, my friend? Still uncertain if we are fish
or fowl, friends or foes? I must confess that it takes more than a small
act of faith to accept this portrayal of our work in such a light, but
since you asked your question with such an open mind, I thought you
deserved to be answered with the same frankness. It is true that it is a
bit difficult to locate us in the middle of the two-culture divide, in the
midst of the epochal shift from Science to Research, torn between the
postmodern and the nonmodern predicament. I hope you are con
vinced, at least, that there is no deliberate obfuscation in our position,
but that being faithful to your own scientific work in these troubled
times is just damned difficult. In my view, your work and that of your
many colleagues, your effort to establish facts, has been taken hostage
in a tired old dispute about how best to control the people. We believe
the sciences deserve better than this kidnapping by Science.
Contrary to what you may have thought when you asked me for this
private conversation, far from being the ones who have limited science
to “mere social construction” by the frantic disorderly mob invented
to satisfy Callicles’ and Socrates’ urge for power, we in science studies
may be the first to have found a way to free the sciences from politics—
the politics of reason, that old settlement among epistemology, moral
ity, psychology, and theology. We may be the first to have lived non
humans from the politics ol objet tivily and humans limn the politic s
of subjei lilicalion. The disc iphnes themselves, the lut Is and the ai li
23
I.icts with their beautiful roots, their delicate articulations, their many
tendrils, and their fragile networks remain, for the most part, to be in-
\ cstigated and described. I try my best, in the pages that follow, to un-
1.ingle a few of them. Far from the rumblings of the science wars in
wInch neither you nor I want to fight (well, maybe I won’t mind firing
.1 lew shots!), facts and artifacts can be part of many other conversa-
nons, much less bellicose, much more productive, and, yes, much
li icndlier.
I liave to admit I am being disingenuous again. In opening the black
l»o\ of scientific facts, we knew we would be opening Pandora’s box.
I lu re was no way to avoid it. It was tightly sealed as long as it re-
iiMined in the two-culture no-man’s-land, buried among the cabbages
ind (he turnips, blissfully ignored by the humanists trying to avoid all
1lie dangers of objectification and by the epistemologists trying to fend
<«Il all the ills carried by the unruly mob. Now that it has been opened,
\ 1(h plagues and curses, sins and ills whirling around, there is only
«'iic lliing to do, and that is to go even deeper, all the way down into
Ilie almost-empty box, in order to retrieve what, according to the ven-
II iMe legend, has been left at the bottom—yes, hope. It is much too
•li ( p lor me on my own; are you willing to help me reach it? May I
i\ c you a hand?
C H A P T E R T W O
Circulating Reference
Sampling the Soil in the Amazon Forest
The only way to understand the reality of science studies is to follow

what science studies does best, that is, paying close attention to the de
tails of scientific practice. Once we have described this practice from
up close as other anthropologists do when they go off to live among
foreign tribes, we will be able to raise again the classic question that
the philosophy of science attempted to solve without the help of an
empirical grounding: how do we pack the world into words? To begin
with I have chosen a discipline, soil science, and a situation, a field trip
in the Amazon, that will not require too much previous knowledge. As
we examine in detail the practices that produce information about a
state of affairs, it should become clear how very unrealistic most of the
philosophical discussions about realism have been.
The old settlement started from a gap between words and the
world, and then tried to construct a tiny footbridge over this chasm
through a risky correspondence between what were understood as to
tally different ontological domains—language and nature. I want to
show that there is neither correspondence, nor gaps, nor even two dis
tinct ontological domains, but an entirely different phenomenon: cir
culating reference*. To capture it, we need to slow our pace a bit and
set aside all our time-saving abstractions. With the help of my camera,
I will attempt to bring some sort of order to the jungle of scientific
practice. Let us turn now to the first freeze-lrame of this photo-
philosophical montage. If a picture is worth a thousand words, a map,
as we shall see, can be worth a whole lorcsl.
On the left in h'igme z.i is a laige savanna. ( >n the i ight ahi upllv be
-I
CIRCULATING REFERENCE
25
Figure 2.1
C.iiis the outskirts of a dense forest. One side is dry and empty, the
oilier wet and teem ing w ith life, and though it may look as if local in-
11.1hi 1ants have created this edge, no one has ever cultivated these
finds and no line has traced the border, w hich extends for hundreds of
I ilometers. Although the savanna serves as a pasture for some land-
owners’ cattle, its lim it is the natural edge of the forest, not a man-
ni.ide boundary.
I il I le figures lost in the landscape, pushed off to the side as in a
|um (ing by Poussin, point at interesting phenom ena w ith their fingers
nul pens. The first character, pointing at some trees and plants, is
I dileusa Setta-Silva. She is Brazilian. She lives in this region, teaching
holany at the small university in the little tow n of Boa Vista, the capi
tal of the Amazonian province of Roraima. Just to her right another
Imi son looks on attentively, sm iling at w hat Edileusa is showing him.
\11nand C’hauvel is from France. He has been sent on this trip by
1 tKSTOM, the research institute of the French form er colonial em
pire, the “agency lor the developm ent of cooperative scientific re-
• .11eh."
Armand is not a bolanisl bill a pedologist (pedology is one ol Ihe
26
soil sciences, not to be confused with either geology, the science of
subsoil, or podiatry, the medical art of treating feet) ; he resides about
a thousand kilometers away in Manaus, where ORSTOM finances his
laboratory in a Brazilian research center known as INPA.
The third person, taking notes in a small notebook, is Héloïsa
Filizola. She is a geographer, or rather, as she insists, a geo
morphologist, studying the natural and social history of the shape of
the land. She is Brazilian like Edileusa, but from the south, from Sâo
Paulo, which is thousands of kilometers away, almost another country.
She is also a professor at a university, though one far larger than the
one in Boa Vista.
As for me, I’m the one taking this picture and describing this
scene. My job as a French anthropologist is to follow these three at
work. Familiar with laboratories, I decided for a change to observe a
field expedition. I also decided, being something of a philosopher, to
use my report on the expedition as a chance to study empirically the
epistemological question of scientific reference. Through this photo-
philosophical account I will bring before your eyes, dear reader, a
small part of the forest of Boa Vista; I will show you some traits of my
scientists’ intelligence; and I will strive to make you aware of the labor
required for this transport and that reference.
What are they talking about on this early morning in October 1991,
after driving the jeep over terrible roads to reach this field site, which
for many years now Edileusa has been carefully dividing into sections,
where she has been noting the growth patterns of the trees and the so
ciology and demography of the plants? They are talking about the soil
and the forest. Yet because they belong to two very different disci
plines, they speak of them in different ways.
Edileusa is pointing to a species of fire-resistant trees that usually
grow only in the savanna and that are surrounded by many small seed
lings. Yet she has also found trees of this same species along the edge
of the forest, where they are more vigorous but do not shade any
smaller plants. To her surprise she has even managed to find a few of
these trees ten meters into the forest, where they tend to die from
insufficient light. Might the forest be advancing? Fdileusa hesitates.
For her, the large tree that you see in the background of this picture
may be a scout sent by the forest as an advance guaid, or perhaps
a rear guard, sat 1diced by the 1cheating finest to the meuiless en
27
. machinent of the savanna. Is the forest advancing like Birnam Wood

low ard Dunsinane, or is it retreating?
I liis is the question that interests Armand; this is why he has come
li <»iu so far away. Edileusa believes the forest is advancing, but she can
in •( be certain because the botanical evidence is confused: the same
urc may be playing either of two contradictory roles, scout or rear
•li.ii d. For Armand, the pedologist, at first glance it is the savanna that
must be eating up the forest little by little, degrading the clay soil nec-
■..iry for healthy trees into a sandy soil in which only grass and small
luubs can survive. If all her knowledge as a botanist makes Edileusa
i<lu with the forest, all his knowledge of pedology makes Armand lean
inward the savanna. Soil goes from clay to sand, not from sand to
11\ everyone knows that. Soil cannot avoid degradation; if the laws
-I pedology do not make this clear, then the laws of thermodynamics
In mid.
I In is our friends are faced with an interesting cognitive and disci-
l'lui.h y conflict. A field expedition to resolve it was easy to justify. The
ni h I* world is interested in the Amazon forest. The news that the Boa
i 1.1 forest, on the outskirts of dense tropical zones, is advancing or
i i mmling should indeed be o f interest to businessmen. It was equally
i lo justify mixing the know-how o f botany with that o f pedology
h i single expedition, even though such a combination is unusual. The
h mi o f translation* that allows them to obtain funding is not very
mi > I will not deal at length with the politics surrounding this expedi-
11 >n since in this chapter I wish to concentrate on scientific reference
• i philosopher, not on its “context” as a sociologist. (I apologize in
i I u n e to the reader, because I am going to om it many aspects o f this
i Id 11 ip that pertain to the colonial situation. What I want to do here
t>> mimic as m uch as possible the problems and vocabulary o f the
l.d.isophers in order to rework the question o f reference. Later I w ill
• ink (he notion o f context, and in Chapter 3 I will correct the dis-
11Mlion between content and context.)
In Ibe morning before leaving we meet on the terrace o f the little
f< I icstauranl called Eusebio (Figure 2.2). We are in the center of
I • 1 \ isla, a lather rough frontier town where th e gorimperos sell the
Id that they have exliai led by shovel, by mercury, by gun, from the
11 I .md Itom (he Vanomnmi.
I 01 1his expedition, Aimaud (on the 1ighl) has asked lot the help
28
Figure 2.2
of his colleague René Boulet (the m an w ith the pipe). French like
Armand, René is also a pedologist from ORSTOM but based in Sâo
Paulo. Here are two men and two women. Two Frenchm en and two
Brazilians. Two pedologists, one geographer, and a botanist. Three
visitors and one “native.” All four are leaning over two kinds of maps
and pointing at the precise location of the site m arked out by Edileusa.
Also on the table is an orange box, the indispensable topofil, w hich I
will discuss later.
The first map, printed on paper, corresponds to the section of the
atlas, com piled by Radambrasil on a scale of one to one million, that
covers all of Amazonia. I will soon learn to put quotation marks
around the w ord “covers,” since, according to my inform ants, the
beautiful yellow, orange, and green colors on the map do not always
correspond to the pedological data. This is the reason they wish to
zoom in, using black-and-white aerial photographs on a scale of one to
fifty thousand. A single inscription* would not inspire trust, but the
superposition of the two allows at least a quick indication of the exact
location of the site.
This is a situation so trivial that we lend to forget its novelty: here
29
ai e four scientists whose gaze is able to dominate two maps of the

very landscape that surrounds them. (Both of Armand’s hands and
Tdileusa’s right hand must continually smooth out the corners of the
map, otherwise the comparison would be lost and the feature they are
trying to find would not appear.) Remove both maps, confuse carto
graphic conventions, erase the tens of thousands of hours invested in
Kadambrasil’s atlas, interfere with the radar of planes, and our four
dentists would be lost in the landscape and obliged once more to be
gin all the work of exploration, reference marking, triangulation, and
quaring performed by their hundreds of predecessors. Yes, scientists
master the world, but only if the world comes to them in the form of
Iwo dimensional, superposable, combinable inscriptions*. It has al-
ays been the same story, ever since Thales stood at the foot of the
Pyramids.
Note, dear reader, that the owner of the restaurant seems to have
Ilie same problem as our researchers and Thales. If the owner had not
11lien the number 29 in big black letters on the table on the ter
race, he would be unable to navigate his own restaurant; without such
in u kings he would not be able to keep track of the orders or distribute
(he bills. He looks like a mafioso as he lowers his enormous belly into a
Chair when he arrives in the morning, but he, even he, needs inscrip
tions to oversee the economy of his small world. Erase the numbers
In cribed on the table, and he would be as lost in his restaurant as our
tit mists would be in the forest without maps.
In the previous picture our friends were immersed in a world in
lilch distinct features could be discerned only if pointed out with a
Anger. Our friends fumbled. They hesitated. But in this picture they
ill 0 sure of themselves. Why? Because they can point with their fingers
(0 phenomena taken in by the eye and susceptible to the know-how
f their age-old disciplines: trigonometry, cartography, geography. In
DCCounting for knowledge thus acquired, we should not forget to men
tion the rocket ship Ariane, orbiting satellites, data banks,
IIaftspeople, engravers, printers, and all those whose work here mani-
f (S itself as paper. There remains that gesture of the finger, the “in-
U par excellence. “Here, there, I, Edileusa, I leave words behind and
I designate, on the map, on the restaurant table, the location of the site
here we will go later, when Sandoval the technician comes to get us
lniheJeepM
30
How does one pass from the first image to the second—from igno
rance to certainty, from weakness to strength, from inferiority in the
face of the world to the domination of the world by the human eye?
These are the questions that interest me, and for which I have traveled
so far. Not to resolve, as my friends intend, the dynamic of the forest-
savanna transition, but to describe the tiny gesture of a finger pointed
toward the referent of discourse. Do the sciences speak of the world?
This is what they claim, and yet Edileusa’s finger designates a single
coded point on a photograph that bears a mere resemblance, in certain
traits, to figures printed on the map. At the restaurant table we are
quite distant from the forest, yet she talks about it with assurance, as if
she had it under her hand. The sciences do not speak of the world but,
rather, construct representations that seem always to push it away, but
also to bring it closer. My friends want to discover whether the forest
advances or recedes, and I want to know how the sciences can be at
the same time realist and constructivist, immediate and intermediary,
reliable and fragile, near and far. Does the discourse of science have a
referent? When I speak of Boa Vista, to what does the spoken word re
fer? Do science and fiction differ? And one additional query: how
does my way of talking about this photomontage differ from the man
ner in which my informants speak of their soil?
Laboratories are excellent sites in which to understand the produc
tion of certainty, and that is why I enjoy studying them so much, but
like these maps, they have the major disadvantage of relying on the
indefinite sedimentation of other disciplines, instruments, languages,
and practices. One no longer sees science stammer, making its debut,
creating itself from nothing in direct confrontation with the world.
In the laboratory there is always a preconstructed universe that is mi
raculously similar to that of the sciences. In consequence, since the
known world and the knowing world are always performing in con
cert with each other, reference always resembles a tautology (Hacking
1992). But not in Boa Vista, or so it seems. Here science does not blend
well with the garimperos and the white waters of the Rio Branco. Whal
luck! In accompanying this expedition 1 will be able to follow the trail
of a relatively poor and weak discipline that will, before my eyes, take*
its first steps, just as I would have been able to obsei ve the teeterings
and loiterings of geography had I. in past cenlmies, inn through I>1a/11
alter (ussieu or I Imnboldl
Iliac in the girat loiesl (ligmc » {). a hmi/nnlal hi.null is Imr
31
Figure 2.3
t.•milled against an otherw ise uniform ly green background. On this

hi mu h, attached to a rusty nail, is a little tin tag on w hich is w ritten
1In n u m b e r 234.
In the thousands of years in which hum ans have traveled through
'hi loi est, slashing and burning in order to cultivate it, no one had
. 1 belore had the peculiar idea o f attaching num bers to it. It took a
. j. n ir,1, <»i perhaps a forester designating trees to be felled. In either
1 1 1In*, num bering of trees is, we must assume, the w ork of a metic-
111n 1. bookkeeper (M iller 1994)*
Ml. 1 an hour in the jeep, we have arrived at the plot of land that
1 III. le.a has been charting lor many years. I.ike the owner of the res-
1,1111 ml m the previous picture, she would not be able to remember
PANDORA'S HOPE
32
the differences between patches of the forest for very long without
marking them in some way. She has therefore placed tags at regular
intervals so as to cover the few hectares of her field site in a grid of
Cartesian coordinates. These numbers will allow her to register the
variations of growth and the emergence of species in her notebook.
Each plant possesses what is called a reference, both in geometry
(through the attribution of coordinates) and in the management of
stock (through the affixing of specific numbers).
Despite the pioneering quality of this expedition, it turns out, I am
not assisting at the birth of a science ex nihilo. My pedological col
leagues cannot fruitfully begin their work unless the site has already
been marked out by another science, botany. I thought I was deep in
the forest, but the implication of this sign, “234,” is that we are in a lab
oratory, albeit a minimalist one, traced by the grid of coordinates. The
forest, divided into squares, has already lent itself to the collection of
information on paper that likewise takes a quadrilateral form. I redis
cover the tautology that I believed I was escaping by coming into the
field. One science always hides another. If I were to tear down these
tree tags, or if I were to mix them up, Edileusa would panic like those
giant ants whose paths I disturb by slowly passing my finger across
their chemical freeways.
Edileusa cuts off her specimens (Figure 2.4). We always forget that
the word “reference” comes from the Latin referre, “to bring back.” Is
the referent what I point to with my finger outside of discourse, or is it
what I bring back inside discourse? The whole object of this montage
is to answer that question. If I appear to be taking a roundabout route
to the response, it is because there is no fast-forward button for un
reeling the practice of science if I want to follow the many steps be
tween our arrival at the site and the eventual publication.
In this frame Edileusa extracts, from the broad diversity of plants,
specimens that correspond to those recognized taxonomically as
Guatteria schomburgkiana, Curatella americana, and Connnarus favosus.
She says she recognizes them as well as she does the members of her
own family. Each plant that she removes represents thousands of the
same species present in the forest, in the savanna, and on the border of
the two. It is not a bouquet of flowers she is assembling but evidence
that she wants to keep as a reference (using here another sense of the
word). She must be able to retrieve what she writes in her notebooks
33
Mgmv z.4
PAND OR A’S HOPE
34
and refer to it in the future. In order to be able to say that Afulamata

diasporis, a common forest plant, is found in the savanna but only in
the shadow of a few forest plants that manage to survive there, she
must preserve, not the whole population, but a sample that will serve
as a silent witness for this claim.
In the bouquet she has just picked we can recognize two features of
reference: on the one hand an economy, an induction, a shortcut, a
funnel in which she picks one blade of grass as the sole representative
of thousands of blades of grass; and on the other hand the preserva
tion of a specimen that will later act as guarantor when she is in doubt
herself or when, for various reasons, colleagues may doubt her claims.
Like the footnotes used in scholarly works to which the inquisitive
or the skeptical “make reference” (yet another use of the word), this
armful of specimens will guarantee the text that results from her field
expedition. The forest cannot directly give its credit to Edileusa’s text,
but she can be credited indirectly through the extraction of a represen
tative guarantor, neatly preserved and tagged, that can be transported,
along with her notes, to her collection at the university in Boa Vista.
We will be able to go from her written report to the names of the
plants, from these names to the dried and classified specimens. And if
there is ever a dispute, we will, with the help of her notebook, be able
to go back from these specimens to the marked-out site from which
she started.
A text speaks of plants. A text has plants for footnotes. A leaflet rests
on a bed of leaves.
What will happen to these plants? They will be transported further,
placed in a collection, a library, a museum. Let us see what will happen
to them in one of these institutions, because this step is much better
known and has been more often described (Law and Fyfe 1988; Lynch
and Woolgar 1990; Star and Griesemer 1989; Jones and Galison 1998).
Then we will focus again on the intermediary steps. In Figure 2.5 we
are in a botanical institute, quite far from the forest, in Manaus. A cab
inet with three ranks of shelves constitutes a work space crisscrossed
in columns and rows, x- and y-axes. Each compartment shown in this
photograph is used as much for classification as for tagging and pres
ervation. This piece of furniture is a theory, only slightly heavier than
the tag in Figure 2.3 but much more capable of organizing this office, a
35
Figure 2.<;
36
perfect intermediary between hardware (since it shelters) and soft

ware (since it classifies), between a box and the tree of knowledge.
The tags designate the names of the collected plants. The dossiers,
files, and folders shelter not text—forms or mail—but plants, the very
plants that the botanist removed from the forest, that she dried in an
oven at 40 degrees Celsius to kill the fungi, and that she has since
pressed between newspapers.
Are we far from or near to the forest? Near, since one finds it here in
the collection. The entire forest? No. Neither ants, nor trapdoor spi
ders, nor trees, nor soil, nor worms, nor the howler monkeys whose
cry can be heard for miles are in attendance. Only those few speci
mens and representatives that are of interest to the botanist have
made it into the collection. So are we, therefore, far from the forest?
Let us say we are in between, possessing all of it through these dele
gates, as if Congress held the entire United States; a very economical
metonymy in science as in politics, by which a tiny part allows the
grasping of the immense whole.
And what would be the point of transporting the whole forest here?
One would get lost in it. It would be hot. The botanist would in any
case be unable to see beyond her small plot. Here, however, the air
conditioner is humming. Here, even the walls become part of the mul
tiple crisscrossed lines of the chart where the plants find a place that
belongs to them within the taxonomy that has been standardized for
many centuries. Space becomes a table chart, the table chart becomes
a cabinet, the cabinet becomes a concept, and the concept becomes an
institution.
Therefore we are neither very far from nor very close to the field
site. We are at a good distance, and we have transported a small num
ber of pertinent features. During the transportation something has
been preserved. If I can manage to grasp this invariant, this je ne sais
quoi, I believe, I will have understood scientific reference.
In this little room where the botanist shelters her collection (Figure
2.6) is a table, similar to that in the restaurant, on which the speci
mens brought back from distinct locations at different times are now
displayed. Philosophy, the art of wonderment, should consider this ta
ble carefully, since it is where we see why the botanist gains so much
more from her collection than she loses by distancing herself from the
I'iguiv 2.(i
38
forest. Let us first review what we know of that superiority before

again attempting to follow the intermediary steps.
The first advantage: comfort. In leafing through the pages of news
print, the researcher makes the dried stems and flowers visible so she
can examine them at leisure, writing just beside them as if the stems
and flowers could imprint themselves directly onto the paper or at
least become compatible with the paper world. The supposedly vast
distance between writing and things is now only a few centimeters.
A second advantage, just as important, is that once classified, speci
mens from different locations and times become contemporaries of
one another on the flat table, all visible under the same unifying gaze.
This plant, classified three years ago, and this other, obtained more
than a thousand kilometers away, conspire on the table to form a syn
optic tableau.
A third advantage, again equally decisive, is that the researcher can
shift the position of specimens and substitute one for another as if
shuffling cards. Plants are not exactly signs, yet they have become as
mobile and recombinable as the lead monotype characters of a print
ing press.
Hardly surprising, then, that in the calm and cool office the botanist
who patiently arranges the leaves is able to discern emerging patterns
that no predecessor could see. The contrary would be much more sur
prising. Innovations in knowledge naturally emerge from the collec
tion deployed on the table (Eisenstein 1979). In the forest, in the same
world but with all of its trees, plants, roots, soil, and worms, the bota
nist could not calmly arrange the pieces of her jigsaw puzzle on her
card table. Scattered through time and space, these leaves would never
have met without her redistributing their traits into new combina
tions.
At the card table, with so many trumps in hand, every scientist be
comes a structuralist. No need to look any further for the martingale
that wins every time against those who sweat in the forest, those
crushed beneath the complex phenomena that are maddeningly pres
ent, indiscernible, impossible to identify, reshuffle, and control. In los
ing the forest, we win knowledge of it. In a beautiful contradiction, the
English word “oversight” exactly captures the two meanings of this
domination by sight, since it means at once looking at something from
above and ignoring it.
39
In the naturalist’s collection things happen to plants that have never

occurred since the dawn of the world (see Chapter 5). The plants
find themselves detached, separated, preserved, classified, and tagged.
They are then reassembled, reunited, redistributed according to en
tirely new principles that depend on the researcher, on the discipline
of botany, which has been standardized for centuries, and on the insti
tution that shelters them, but they no longer grow as they did in the
great forest. The botanist learns new things, and she is transformed ac
cordingly, but the plants are transformed also. From this point of view
there is no difference between observation and experience: both are
constructions. ^Through its displacement onto this table, the inter
lace between forest and savanna becomes a hybrid/nixture of scien
tist, botany, and forest, the proportions of whlcfil will have to calcu-
l He later.
Still, the naturalist does not always succeed. In the upper-right-hand
coiner of the photograph something scary is brewing: an enormous
pile of newspaper stuffed with plants brought back from the site and
awaiting classification. The botanist has fallen behind. It is the same
loi y in every laboratory. As soon as we go into the field or turn on an
Instrument, we find ourselves drowning in a sea of data. (I too have
llus problem, being incapable of saying all that can be said about a
In Id trip that took only fifteen days.) Darwin moved out of his house
oon after his voyage, pursued by treasure chests of data that cease-
li ssly arrived from the Beagle. Within the botanist’s collection, the for-
c t, reduced to its simplest expression, can quickly become as thick as
(lie* tangle of branches from which we started. The world can return to
1nnlusion at any point along this displacement: in the pile of leaves to
la indexed, in the botanist’s notes which threaten to submerge her, in
llu reprints sent from colleagues, in the library where the issues of
Joui nais are piling up. We have barely arrived when we must leave; the
In si instrument is hardly operational when we must think of a second
li vice to absorb what its predecessor has already inscribed. The pace
must be accelerated if we are to avoid being overwhelmed by worlds
ol dees, plants, leaves, paper, texts. Knowledge derives from such
01 t merits, not from simple contemplation of the forest.
We now know the advantages of being in an air-conditioned mu-
i inn, but we have gone too quickly over the transformations that
I ilileusa made the forest undergo I have opposed too abruptly the im
40
age of the botanist pointing to the trees and that of the naturalist in
control of specimens on the worktable. In passing directly from the
field to the collection, I must have missed the decisive go-between. If
I say that “the cat is on the mat,” I may seem to be designating a
cat whose actual presence on said mat would validate my statement.
In actual practice, however, one never travels directly from objects to
words, from the referent to the sign, but always through a risky inter
mediary pathway. What is no longer visible with cats and mats, be
cause they are too familiar, becomes visible again as soon as I take a
more unusual and complicated statement. If I say “the forest of Boa
Vista advances on the savanna” how can I point to that whose pres
ence would accord a truth-value to my sentence? How can one engage
those sorts of objects into discourse; to use an old word, how can one
“educe” them into discourse? One needs to go back to the field and
carefully follow, not only what happens inside collections, but how
our friends are collecting data in the forest itself.
In the photograph in Figure 2.7, everything is a blur. We have left
the laboratory and are now in the midst of the virgin forest. The re
searchers can only be distinguished as khaki and blue spots on a green
background, and at any moment they could disappear into the Green
Hell of the forest if they move away from one another.
René, Armand, and Héloïsa are having a discussion around a hole in
the ground. Holes and pits are to pedology what a specimen collection
is to botany : the basic craft and the focus of obsessive attention. Since
the structure of soil is always hidden beneath our feet, pedologists can
display its profile only by digging holes. A profile is the assemblage of
the successive layers of soil, designated by the beautiful word “hori
zon.” Rainwater, plants, roots, worms, moles, and billions of bacteria
transform the parent material of the bedrock (studied by geologists)
into many different “horizons,” which the pedologists learn to distin
guish, classify, and envelop in a history that they call “pedogenesis”
(Ruellan and Dosso 1993).
In accordance with the habits of their profession, the pedologists
wanted to know whether the bedrock was, at a certain depth, different
beneath the forest than beneath the savanna. Here was a simple hy
pothesis that would have put an end to the controversy between bot
any and pedology: neither the forest nor the savanna is receding, the
border that separates them reflects a difference in soil The superstruc
41
Figure 2.7
•mic would be explained by the infrastructure, to use an old M arxist

mi laphor. Yet, as they soon discover, at depths below fifty centim e-
1.1 . 1he soil under the savanna and the soil under the forest appear ex-
tit 11v 1lie same. The hypothesis from infrastructure does not hold.
;<ailing in the bedrock seems to explain the difference in the
n| mi licial horizons—clayey beneath the forest and sandy beneath the
h mua. The profile is “bizarre,” and that makes my friends all the
millr exciled.
In die pielure in Figure 2.8, René is standing and aim ing at me w ith
.11 ne.11ument com bining com pass and clisim eter in order to establish
. In .1 topographic bearing. W hile taking advantage of the situation to
h ip a pic ture, I play the m inor role, well suited to my height, of an
111j■1111u*i)1 pole so that René can m ark precisely w here the pedologists
li<»111«I dig their holes. Lost in the forest, the researchers rely on one of
du oldest and most prim itive techniques for organizing space, claim-
ill}* 1 place with stakes driven into the ground to delineate geom etric
liapes against I he background noise, o ra l least to perm it the possibil-
ii\ ut then 1ecognit ion.
\11b1 nergcd in Ilie forest again, they are forced to count on the oldest
42
Figure 2.8
of the sciences, the m easure of angles, a geom etry w hose m ythical o ri
gin has been recounted by Michel Serres (Serres 1993). Once m ore a
science, pedology, m ust follow the tracks of an older discipline, sur
veying, w ithout w hich we would dig our holes haphazardly, trusting
to luck, incapable of creating on graph paper the precise map that
René would like to draw. The succession of triangles will be used as a
reference and will be added to the num bering of square sections of the
field site already done by Edileusa (see Figure 2.3). In order for the bo
tanical and pedological data to be superposed on the same diagram
later, these two bodies of reference m ust be com patible. O ne should
never speak of “d ata”—w hat is given—but rather of sublata, that is, of
“achievem ents.”
René’s standard practice is to reconstitute the surface soil along
transects, the extrem e lim its of w hich contain soils that are as differ
ent as possible. Here, for example, it is very sandy beneath the savanna
and very clayey beneath the forest. He proceeds by approxim ate grada
tions, first choosing two extrem e soils, then taking a sample in the
middle. Starting again, he continues in this way until he obtains ho
mogeneous horizons. Ilis m ethod recalls both artillery (it approxi-
43
mates by finding medians), and anatomy (it traces the geometry of ho
rizons, true “organs” of the soil). If I were playing the historian, not
the philosopher in pursuit of reference, I would discuss at length the
fascinating paradigm of what René calls “structural pedology,” how
It distinguishes itself from others and the controversies that arise
from it.
To get from one point to another the pedologists cannot use a sur-
eyor’s chain of measurement; no agriculturist has ever leveled this
oil Instead they use a wonderful instrument, the Topofil Chaix™
(Figure 2.9), a device that their Brazilian colleagues have perversely
named a “pedofil,” and of which Sandoval, in this photograph, reveals
the mechanism by opening its orange box. So much depends upon an
Orange pedofil. . .
Aspool of cotton thread unrolls evenly and spins a pulley that acti
nies the cogwheel of a counter. Setting the counter to zero, then un-
Inding the thread of Ariadne behind him, the pedologist can get
fl om one point to the next. Upon arrival at his destination, he simply
(Uls the thread with a blade set near the spool and ties off the end to
prevent any untimely unrolling. A glance at the window on the coun
ter tells the distance he has traveled to within a meter. His path be-
Omes a single number easily transcribed into a notebook and—a dou
ble advantage—takes on material form in the thread that remains in
place Losing an expensive and distracted pedologist in the Green Hell
I Impossible: the cotton thread will always bring him back to camp. If
Hansel and Gretel had had access to a “Topofil Chaix à fil perdu n° de
référence I-8237,” their tale would have unwound very differently.
After a few days’ work the field site is littered with threads that en
tangle our feet. Still, as a result of the compass’s measurements of an-
I and the pedofil’s measurements of lines, the land has become a
I1 o laboratory—a Euclidean world where all phenomena can be reg-
>1
Itied by a collection of coordinates. Had Kant used this instrument,

ould have recognized in it the practical form of his philosophy.
1 Ihe world to become knowable, it must become a laboratory. If
t in loi est is to be transformed into a laboratory, the forest must be
|> 11ed to be rendered as a diagram (Hirshauer 1991). In the extrac-
<>l a diagram from a confusion of plants, scattered locations be
lt mai ked and measured points linked by cotton threads that ma*
1 i (01 spiritualize) lines in a netwoik composed of a succession
PANDORA'S HOPE
44
Figure 2.9
of triangles. Equipped only w ith the a priori form s of intuition, to use

K ant’s expression again, it w ould be im possible to draw these sites to
gether, short of teaching, somehow, a limbless m ind-in-a-vat how to
use such equipm ent as compasses, clisimeters, and topofils.
Sandoval the technician, the only person on the expedition w ho is
native to the region, has dug the largest part of the hole shown in Fig
ure 2.10. (O f course had I not artificially severed the philosophy from
the sociology, I w ould have to account for this division of labor be
tween French and Brazilians, mestizos and Indians, and I would have
to explain the male and female distributions of roles.) Armand, here
leaning on the drill, is rem oving core samples by collecting earth in
the small cham ber at its tip. Unlike Sandoval’s tool, the m attock that
is lying on the ground now th at its task is complete, the drill is a piece
of laboratory equipm ent. Two rubber stoppers placed at 90 centim e
ters and at one m eter allow it to be used both as an instrum ent for
m easuring depth and, by pushing and twisting, as a sam pling tool. The
pedologists examine the soil sample, then Héloïsa collects it in a plas
tic bag on which she w rites the num ber ol the hole and the depth at
which it was taken.
46
As with Edileusa’s specimens, most of the analyses cannot be per

formed in the field but must be done in the laboratory. The plastic
bags here begin a long voyage that will take some of them to Paris, via
Manaus and Sâo Paulo. Even if René and Armand are able to judge on
the spot the quality of the earth, its texture, its color, and the activity
of earthworms, they cannot analyze the soil’s chemical composition,
its grain size, or the radioactivity of the carbon it contains without
costly instruments and skill that one does not easily find among the
poor garimperos or the wealthy landowners. On this expedition, the
pedologists are the vanguard for the distant laboratories to which they
will take their samples. The samples will remain attached to their orig
inal context solely by the fragile link of the numbers inscribed in black
felt-tip pen on the little transparent bags. If, like me, you should ever
run into a gang of pedologists, one word of advice: never offer to carry
their suitcases, which are enormous and stuffed with the bags of earth
they tote from one part of the world to another and with which they
will quickly fill your refrigerator. The circulation of their samples
traces a network on the Earth as dense as the cotton webs spun by
their topofils.
What industrialists call the “traceability” of references depends, in
this case, on the reliability of Héloïsa. Sitting in front of the hole, the
group members rely on her for the careful maintenance of the field
notebook. For each sample she must record the coordinates of the lo
cation, the number of the hole, the time and depths at which it was
collected. In addition, she must note down all the qualitative data her
two male colleagues can extract from the lumps of earth before they
slide them into the bags.
The success of the entire expedition depends on this little logbook,
equivalent to the protocol book that regulates the life of any labora
tory. It is this book that will allow us to return to each data point in or
der to reconstitute its history. The list of questions that was decided
on at the restaurant is imposed on each sequence of action by Héloïsa.
It is a grid that we must systematically fill with information. Héloïsa
acts as guarantor of the standardization of experimental protocols, so
that we take the same kinds of samples from each location and in the
same way. The protocols ensure the compatibility and therefore the
comparability of the holes, and the notebook then allows for continu
ity in time as well as in space. Ilcloisa does not only handle tags and
47
pioiocols. A geomorphologist, she adds her two cents to all the con-
( i sations, allowing her expatriate colleagues to “triangulate” their
Judgments through hers.
I istening to Héloïsa call us to order—having repeated the informa-
lion dictated to us by René and twice verified the inscriptions on the
I»igs—it seems to me that never before has the forest of Boa Vista
known such discipline. The indigenous people who once traveled
llnough this place probably imposed rites on themselves as well, per
il ips as fastidious as those of Héloïsa, but surely not so strange. Sent
l»y institutions that are thousands of kilometers away, obliged at all
ctsl s to maintain the traceability of the data we produce with minimal
l« loi mation (while transforming them totally by ridding them of
IIn n local context), we would have seemed extremely exotic to the in-
ligenous people. Why take such care in sampling specimens whose
I 11in es are visible only at such a distance that the context from which
Ilu y were taken will have disappeared? Why not remain in the forest?
Iiy not “go native”? And what about me, standing here, useless,
Hms dangling, incapable of distinguishing a profile from a horizon—
mi I not even more exotic, exacting from the hard labor of my infor-
II) mis the bare minimum for a philosophy of reference that will be of
nl( icst only to a very few colleagues in Paris, California, or Texas?
Iiy not become a pedologist? Why not become an indigenous soil
tilt i lor, an autochthonous botanist?
Io understand these small anthropological mysteries we must draw
In c i to the beautiful object in Figure 2.11, the “pedocomparator.” On
Ik savanna grass, we see a series of empty little cardboard cubes
lly 1led to form a square. More Cartesian coordinates, more columns,
nh11e 1ows. These little cubes rest in a wooden frame that allows them
lu stowed away in a drawer. With the cleverness of our pedologists,
1 id wilh the addition of a handle, clasps, and a padded flap that serves
i llexible cover for all the cardboard cubes (not visible in the photo-
1ipli), this drawer can also be transformed into a suitcase. The suit-
I « pel mils the simultaneous transportation of all the clods of earth
li il li.ive since become Cartesian coordinates, and their collection in
li il thus becomes a pedolibrary.
I lie the cabinet m Figuie 2.5, the pedocomparator will help us
II p Ilie pKidual dilleience between abstract and concrete, sign and
1nlluie Willi Us handle, its wooden liame, its padding, and its card
48
Figure 2.11
board, the pedocomparator belongs to “things.” But in the regularity

of its cubes, their disposition in columns and rows, their discrete char
acter, and the possibility of freely substituting one column for an
other, the pedocomparator belongs to “signs.” Or rather, it is through
the cunning invention of this hybrid that the world of things may be
come a sign. With the next three photographs we will try to under
stand more concretely the practical task of abstraction and what it
means to load a state of affairs into a statement.
I will be obliged to employ vague terms—we do not have as discrim
inating a vocabulary for speaking of the engagement of things into dis
course as we do for speaking of discourse itself. Analytic philosophers
keep themselves busy trying to discover how we can speak of the
world in a language capable of truth (Moore 1993 )- Curiously, even
though they attach importance to the structure, coherence, and valid
ity of language, in all their demonstrations the world simply awaits
designation by words whose truth or falsehood is guaranteed solely by
its presence. The “real” cat waits quietly on its proverbial mat to con
fer a truth-value on the sentence “the cat is on the mat.” Yet to achieve
certainty the world needs to stir and transform itself much more than
49
words (see Chapters 4 and 5). It is this, the other neglected half of ana
lytic philosophy, that analysts must now acknowledge.
For the time being, the pedocomparator is empty. This instrument
i .111 be added to the list of empty forms that has been getting longer
during the expedition: Edileusa’s plot of land, divided into squares by
numbers inscribed on tags that are nailed to trees; the marking of the
holes with Rene’s compass and topofil; the numbering of the samples
md the disciplined sequence of the protocol controlled by Héloïsa. All
Ihose empty forms are set up behind the phenomena, before the phe
nomena manifest themselves, in order for them to be manifested. Ob-
1ured in the forest by their sheer number, phenomena will be able at
I isl to appear, that is, to stand out against the new backgrounds we
I I ive astutely placed behind them. In my eyes and in those of my
11lends, pertinent traits will be bathed in a spotlight as white as the
i mpty pedocomparator or the graph paper, very different in any case
horn the deep greens and grays of the vast and noisy forest, where
ome birds whistle so obscenely that the locals call them “flirting
Imds.”
I11 Figure 2.12, René abstracts. After cutting the earth with a knife,
lu 1emoves a clod, from a depth dictated by the protocol, and deposits
II m one of the cardboard cubes. With a felt-tip pen Héloïsa will code
Ilie edge of the cube with a number that she will also record in her
notebook.
Consider this lump of earth. Grasped by René’s right hand, it re-
I mis all the materiality of soil—“ashes to ashes, dust to dust.” Yet as it
I placed inside the cardboard cube in René’s left hand, the earth be-
111lies a sign, takes on a geometrical form, becomes the carrier of a
numbered code, and will soon be defined by a color. In the philosophy
I sc icnce, which studies only the resulting abstraction, the left hand
loi s not know what the right hand is doing! In science studies, we are
imludextrous: we focus the reader’s attention on this hybrid, this mo-
niiiit of substitution, the very instant when the future sign is ab-
li it led from the soil. We should never take our eyes off the material
1iglit of this action. The earthly dimension of Platonism is revealed
lu IIns image. We are not jumping from soil to the Idea of soil, but
Imm continuous and multiple clumps of earth to a discrete color in a
i omeinc cube coded in x and y coordinates. And yet René does not
pose piedeleimined categoiies on a shapeless hoiizon; he loads his
50
igure 2.12
51
I»( docomparator with the meaning of the piece of earth—he educes it,
Im articulates* it (see Chapter 4). Only the movement of substitution
» which the real soil becomes the soil known to pedology counts. The
immense abyss separating things and words can be found everywhere,
list 1ibuted to many smaller gaps between the clods of earth and the
iibes cases-codes of the pedocomparator.
What a transformation, what a movement, what a deformation,
li it an invention, what a discovery! In jumping from the soil to the
li iwer, the piece of earth benefits from a means of transportation that
111 longer transforms it. In the previous photograph we could see how
In soil changed states; in Figure 2.13 we see how it changes location.
II iv mg made the passage from a clump of earth to a sign, the soil is
I 1 able to travel through space without further alterations and to re
in 1m intact through time. At night, in the restaurant, René opens the
ibmet suitcases of the two pedocomparators and contemplates the
ms of cardboard cubes regrouped in rows corresponding to holes
m 1olumns corresponding to depths. The restaurant becomes the an-
1
ol a pedolibrary. All the transects have become compatible and

mpaiable.
t >m.e filled, the cubes gather clods of earth on the way to becoming
n but we know that the empty compartments, either humble ones
lu Ihose or famous ones like those of Mendeleev, are always the most
ipoilant part of any classification scheme (Bensaude-Vincent 1986;
inly 1977). When we compare them, the compartments define what
li 11 loi us to find, and we are able to plan the next day’s labor in ad
mit since we know what we must gather. Thanks to the empty com-
1111111ills, we see the blanks in our protocol. According to René, “It is
111docomparator that tells us if we have finished a transect.”
I lit In si great advantage of the pedocomparator, as “profitable” as
Iml,mist’s classification in Figure 2.6, is that in it all the different
mI ilis liom all the different depths become visible simultaneously,
iiyli Ihoy were extracted over the course of a week. Thanks to the
I 11ompaiator, the differences in color become manifest and form a
I 01 t liai I; all of the disparate samples are embraced synoptically.
II lou si savanna transition has now been translated, through the ar
il 1nu ill ol nuanced shades of brown and beige, into columns and
i Iniiisilion now giaspable because ihe instrument has given us
It nit Hi on (lie eat III
PANDORA'S HOPE
52
Figure 2.1 {
53
Look at René in the photograph: he is master of the phenomenon

ili.il a few days earlier was tucked away in the soil, invisible, and dis-
|u i sed in an undifferentiated continuum. I have never followed a sci-
lue, rich or poor, hard or soft, hot or cold, whose moment of truth
is not found on a one- or two-meter-square flat surface that a re-
i ii cher with pen in hand could carefully inspect (see Figures 2.2 and
f>) The pedocomparator has made the forest-savanna transition into
I 111juratory phenomenon almost as two-dimensional as a diagram, as
II ulily observed as a map, as easily reshuffled as a pack of cards, as
Imply transported as a suitcase, about which René jots down notes
lule peacefully smoking his pipe, having taken a shower to wash off
lin dust and earth that are no longer useful.
And I, of course, ill-equipped and thus short on rigor, I bring back to
In leader, by superposing pictures and text, a phenomenon, that of
11 ( irculating reference*, that was until now invisible, purposely mud-
llnl by epistemologists, dispersed in the practice of scientists, and
ilecl up in the knowledges that I now calmly display with a cup of tea
1 1md at my house in Paris, while reporting what I observed at the
1
1tier of Boa Vista.

Another advantage of the pedocomparator, once it is saturated with
il 1 ,i pattern emerges. And here again, as with Edileusa’s discover-
11 would be astounding were this not the case. Invention almost
lys follows the new handle offered by a new translation or trans-
il ilion. The most incomprehensible thing in the world would be
I Ilie pattern to remain incomprehensible after such rearrange-
111s
I Ids expedition, it too, via the intermediary of the pedocomparator,
overs or constructs (we will choose between those two verbs in
I iplei 4, before realizing in Chapter 9 why we do not have to
111 e) an extraordinary phenomenon. Between the sandy savanna
I (lie clayey forest, it seems that a twenty-meter-wide strip of land
I uls out at the border, on the savanna side. This strip of land is am-
iious, more clayey than the savanna but less so than the forest. It
ldi I appear that the forest casts its own soil before it to create condi-
II I ivoi able lo its expansion. Unless, on the contrary, the savanna is
1 tiling the woodland humus as it prepares to invade the forest. The
til Inns sceiunos that my Iriends discuss, at night in the restaurant,
now gauged by the weight <>( evidence They become possible in
PANDORA'S HOPE
54
terpretations of the matters of fact that are solidly in place in the grid
of the pedocomparator.
One scenario will eventually become text, and the pedocomparator
will become a table in an article. There now needs to be only one last,
tiny transformation.
On the table, in the table/chart, in Figure 2.14, we see the forest on
the left and the savanna on the right, the reverse of Figure 2.1, give
or take a few transformations. (Since there are not enough compart
ments in the pedocomparator, the series of samples must be altered,
breaking the beautiful order of the table and requiring us to devise an
ad hoc reading convention.) Beside the open drawers there is a dia
gram drawn on millimeter-ruled graph paper and a table drawn on
straight-ruled paper. The coordinates of the samples, taken by the
team along a given transect, are recaptured in a vertical cross-section,
while the chart sums up color variations as a function of depth at a
given set of coordinates. A transparent ruler negligently placed on the
drawer further ensures the transition from furniture to paper.
In Figure 2.12 René moved from concrete to abstract in one quick
gesture. He was moving from thing to sign and from the three
dimensional earth to the two and a half dimensions of the table/chart
In Figure 2.13 he had slipped from the field site to the restaurant: the
drawers convert into a suitcase, permitting Rene’s movement from an
uncomfortable and underequipped location to the relative comfort ol
a café, and in principle nothing (except Customs officers) can stop the
transportation of this drawer/suitcase/chart anywhere in the world,
or its comparison with all other profiles in all other pedolibraries.
In Figure 2.14 another transformation as important as the others be
comes evident, but one that, under the name of inscription*, has re
ceived more attention than the others. We move now from the instru
ment to the diagram, from the hybrid earth/sign/drawer to paper.
People are often surprised that mathematics can be applied to the
world. In this case, for once, the surprise is misplaced. For here we
must ask how much the world needs to change in order for one kind ol
paper to be superposed on a geometry of another kind without suffer
ing too much distortion. Mathematics has never crossed the great
abyss between ideas and things, but it is able to cross the tiny gap hi
tween the already geometrical pedocomparator and the piece of milli
meter ruled paper on which Rene has recouled the data fiom the sum
55
Figure 2.14
I'l' ll is easy to cross this gap—I can even m easure the distance w ith
• plastic ruler: ten centim eters!
V abstract as the pedocom parator is, it rem ains an object. It is
lii< i than the forest, yet heavier than the paper; it is less corrupt-
1 b 1ban the vibrant earth, but m ore corruptible than geom etry; it is
.... . mobile than the savanna, but less mobile than the diagram that I
" 1.1 send by phone if Boa Vista had a fax machine. As coded as the
i I*»t <nnparator is, René cannot insert it into the text of his report. He
•tt "iily hold it in reserve, keeping it for future com parisons if he ever
in', lo have doubts about his article. W ith the diagram, in contrast,
1 it hesl -savanna transition becomes paper, assimilable by every arti-
1 in the world, and transportable to every text. The geom etric form
1 1In diagram renders it com patible w ith all the geom etric transfor-
......... 1hal have ever been recorded since centers o f calculation* have
1 in I Whal we lose in m atter through successive reductions of the
il wr regain a hundredfold in the branching off to other forms that
b h d m lions—w ritten, calculated, and archival—make possible.
In 1hr report that we are preparing to write, only one rupture will
in m i , a gap as tiny and as immense as all t he steps we have just fol-
56
lowed: I mean th e gap that divides our prose from the annex of dia
grams it will refer to. We will write about the forest-savanna transi
tion, which w e w ill show within the text through the medium of a
graph. The scientific text is different from all other forms of narrative.
It speaks of a referent, present in the text, in a form other than prose: a
chart, diagram, equation, map, or sketch. Mobilizing its own internal
referent*, the scientific text carries within itself its own verification.
In Figure 2.15 is the diagram that combines all the data obtained
during the expedition. It appears as “figure 3” in the written report oi
which I am one o f the proud authors and of which the title page reads
Relations between Vegetation Dynamics and the Differ

entiation of Soils in the Forest-Savanna Transition Zone
in the Region of Boa Vista, Roraima, Amazonia (Brazil)
Report on Expedition in Roraima Province, October 2-
14,1991
E. L. Setta Silva (1), R. Boulet (2), H. Filizola (3),
S. do N. Morais (4), A. Chauvel (5) and B. Latour (6)
(1) MIRR, Boa Vista RR, (2.3) USP, Sao Paulo, (3-5)
INPA,
Manaus, (6) CSI, ENSMP, (2.5) ORSTOM Brazil
Let us quickly retrace our steps back down the road we have trav
eled while following our friends. The prose of the final report spewI
of a diagram, which summarizes the form displayed by the layout
of the pedocomparator, which extracts, classifies, and codes the soil
which, in the end, is marked, ruled, and designated through the cns
crossing of coordinates. Notice that, at every stage, each element Ik
longs to matter by its origin and to form by its destination; it is nl)
stracted from a too-concrete domain before it becomes, at the 11c I
stage, too concrete again. We never detect the rupture between tiling
and signs, and we never face the imposition of arbitrary and disc it It*
signs on shapeless and continuous matter. We see only an unbiol ti
series of well-nested elements, each of which plays the role of sign loi
the previous one and of thing for the succeeding one.
At every stage we find elemental yjo rn is ol mathematics, which lit
used to collect m a tte r thiough the mediation of a piadice embodltt
57
limiti i'honzon © .Homonte de fe ie s deminhocas (#.<U dejections d t vers de lerre)

limite de couleur (chromil © .H om onle com estr. mjmelcnar (as f e w lip.cai diminuent p. biixo)
limite de texture(saM A«ft) ® .Horieonte «* pouco mais daro Ao que os horieontes 9 vperiorrs(m.ex.na sarana
limite de texture (sabl.)” Q -H oritonte de t r a «51940,p*r jo x ta p o siç io ou cancinva. J
Figure's. Coupe du tran se ct •)
Figure 2.15
I f.*10up of researchers. On each occasion a new phenomenon is

in i from this hybrid of form, matter, skilled bodies, and groups.
In 1emember René, in Figure 2.12, placing the brown earth into the
II i urdboard cube that was then immediately marked with a num-
I Ile did not divide the soil according to intellectual categories, as
(lit Kantian mythology; rather, he conveyed the meaning of each
immenon by making matter cross the gap that separated it from
in
II I id, if we flip quickly through these photographs, we become
in Ilut, even if my inquiry had been more meticulous, each stage
1I11 1c*veal a rift as complete as those which follow and precede it.
i I might, like a new Zeno, to multiply the intermediaries, there is
1 1 1csemblance between stages so that we can merely superpose
in I ompare the two extremes in Figures 2.1 and 2.15. The differ-
ImIween them is no wider than that between the lumps of earth
I li d by René (Figure 2.12) and the data-points that they become in
I 1doiomparator. Whether I choose the two extremes or multiply
nli 1mediaries, I find this same discontinuity.
I 11ici e is also a continuity, since all the photographs say the same
mil lepresent the same forest-savanna transition, made ever
1111.un and precise at each stage. Our field report indeed refers
II iiic winch indeed refers to the Boa Vista forest. Our report
1 lo Ilie si 1«mge dynamics o( vegetal ion that appear to allow the
I lo deleal ihe savanna, as il ihe liees had turned sandy soil into
58
clay to prepare for growth in the twenty-meter-wide strip of land. But

these acts of reference are all the more assured since they rely not so
much on resemblance as on a regulated series of transformations,
transmutations, and translations. A thing can remain more durable
and be transported farther and more quickly if it continues to undergo
transformations at each stage of this long cascade.
It seems that reference is not simply the act of pointing or a way of
keeping, on the outside, some material guarantee for the truth of a
statement; rather it is our way of keeping something constant through
a series of transformations. Knowledge does not reflect a real external
world that it resembles via mimesis, but rather a real interior world,
the coherence and continuity of which it helps to ensure. What a beau
tiful move, apparently sacrificing resemblance at each stage only to
settle again on the same meaning, which remains intact through sets
of rapid transformations. The discovery of this strange and contradic
tory behavior is worthy of the discovery of a forest able to create its
own soil. If I could find the solution to that puzzle, my own expedition
would be no less productive than that of my happy colleagues.
In order to understand the constant that is maintained throughout
these transformations, let us consider a small apparatus as ingenious
as the topofil or the pedocomparator (Figure 2.16). Since our friends
cannot easily bring the soil of Amazonia back to France, they must be
able to transform the color of each cube using a label, and if possible a
number, that will make the samples of soil compatible with the uni
verse of calculation and allow the scientists to benefit from the advan
tage that all calculators lend to every manipulator of signs.
But won’t relativism rear its monstrous head as we attempt to qual
ify the nuances of brown? How can we dispute tastes and colors? An
the French saying goes, “So many heads, so many opinions.” In Figim
2.16 we see René’s solution for repairing the ravages of relativism.
For thirty years he has toiled in the tropical soils of the world cai ry
ing a small notebook with rigid pages: the Munsell code. Each page ol
this little volume groups together colors of very similar shades. Thoi p
is a page for the purplish reds, another for the yellowish reds, anotlu 1
for the browns. The Munsell code is a relatively universalized noi m, ll
is used as a common standard for painters, paint manufactuieis, cai
tographers, and pedologists, since page by page it ai ranges all the nil
ances of all the tolois ol the speed um by assigning each a numlui
59
Figure 2.16
I hr num ber is a reference that is quickly understandable and repro-

■lut ible by all the colorists in the w orld on the condition that they use
1hr same com pilation, the same code. By telephone, you and a sales-
Ih■1son cannot m atch samples of wallpaper, but you can, based on a
■•h>i chart the salesperson has given you, select a reference num ber.
I hr Munsell code is a decisive advantage for René. Lost in Roraima,
1ii le so tragically local, he is able to become, through the interm edi-
ii \ 111 his code, as global as it is possible for a hum an being to be. The
•11it 11it* color of this particular soil sample becomes a (relatively) uni-
■1 il number.
\i 1his m om ent, the pow er of standardization (Schaffer 1991) is less
"ii testing to me than a stupefying technical trick—the little holes
•hill have been pierced above the shades of color. Though seemingly
iv. out of reach, the threshold between local and global can now
• lossed instantaneously. Still, it takes some skill to insert the soil
'iiiplc into the Munsell code. In order for the soil sample to quality as
.... nber, René must in fact be able to m atch, superpose, and align the
il 1lod ol earth, which he holds in his hand, with the standardized
I 1 t hosen as a reference, lb accomplish this, he passes the soil sain-
P A N D O RA ’S HOPE
60
pie beneath the openings made in the notebook and, by successive ap
proximations, selects the color closest to that of the sample.
There is, as I have said, a complete rupture at each stage between
the “thing” part of each object and its “sign” part, between the tail end
of the soil sample and its head. That abyss is all the wider because our
brains are incapable of memorizing color with precision. Even if the
soil sample and the standard were no farther apart than ten or fifteen
centimeters, the width of the notebook, this would be enough for
Rene’s brain to forget the precise correspondence between the two.
The only way the resemblance between a standardized color and a soil
sample can be established is by piercing holes in the pages that allow
us to align the rough surface of the lump of soil with the bright and
uniform surface of the standard. With less than a millimeter of dis
tance separating them, then and only then can they be read synopti
cally. Without the holes, there can be no alignment, no precision, no
reading, and therefore no transmutation of local earth into universal
code. Across the abyss of matter and form, René throws a bridge. It is a
footbridge, a line, a grappling hook.
“The Japanese have made one without holes,” René says; “I cannol
use it.” We are always amazed by the minds of scientists, and justly so,
but we should also admire their utter lack of trust in their own cogni
tive abilities (Hutchins 1995). They doubt their brains so much that
they need to invent little tricks like this to ensure their understanding
of the simple color of a soil sample. (And how could I make the readei
understand this work of reference without the photographs that I have
taken, images that must be viewed at exactly the same time as the
story I am relating is being read? I am so afraid of making a mistake in
my account that I myself do not dare lose sight of the photographs
even for an instant.)
The rupture between the handful of dust and the printed number is
always there, though it has become infinitesimal because of the holes
Through the intermediary of the Munsell code, a soil sample can hi
read as a text: “10YR3/2”—further evidence of the practical Platonism
that turns dust into an Idea via the two callused hands firmly holding 11
notebook/instrument/calibrator.
Let us follow in more detail the trail displayed in Figure 2.16, sketc h
ing the lost road of reference lor ourselves. Rene has extracted hi
lump of eailh, 1enouncing the too lich and loo complex soil The holt
6l
Ini urn, allows the framing of the lump and the selection o f its color by
I muring its volume and texture. The little flat rectangle o f color is
Ilien used as an intermediary between the earth, summarized as a
nlor, and the number inscribed under the corresponding shade. Just
i we are able to ignore the volume o f the sample in order to concen-
li lie on the color o f the rectangle, w e are soon able to ignore the color
ii older to conserve only the reference number. Later, in the report,
i will om it the number, w hich is too concrete, too detailed, too pre-
I c, and retain only the horizon, the tendency.
I (ere w e find the same cascade as before, o f w hich only a tiny por-
Inii (the passage from the sample color to that o f the standard) rests
ii lesem blance, on adequatio. All the others depend only on the
u>servation o f traces that establish a reversible route that makes it

ii sible to retrace o n e’s footsteps as needed. Across the variations o f
i il lcrs/form s, scientists forge a pathway. Reduction, com pression,
i ii I mg, continuity, reversibility, standardization, com patibility w ith
I .md num bers—all these count infinitely m ore than adequatio
me No step—except one—resem bles the one that precedes it, yet in
it t nd, w hen I read the field report, I am indeed holding in m y hands
I loi est o f Boa Vista. A text truly speaks o f the w orld. H ow can re
in hi.m ce result from this rarely described series o f exotic and m inus-
ili Ii ansform ations obsessively nested in to on e another so as to keep
mi Ilung constant?
In I igure 2.17 w e see Sandoval squatting, the shaft o f the m attock
ill leslin g under his arm, contem plating the n ew hole he has just
I Standing, H éloïsa is thinking about the few anim als in this green-
i loi est. She is w earing a g eo lo g ist’s pouch, an am m unition belt the
Ii ol w hich is studded w ith eyelets too narrow for cartridges but
II mted for carrying the colored p en cils indispensable to the pro-
ional cartographer. In her hand she holds the fam ous notebook,
j>i otocol book that m akes it true that w e are in a vast, green labora-
1 She is w aiting to op en it and to take n otes n o w that both
lologisls have finished their exam in ation and reached agreem ent,
im and (on the left) and René (o n the right) are engaged in the
(hi 1 si 1ange exercise o f “earth tastin g.” In one hand each o f them
i 1 1) en a bit o f soil sam pled from the h ole at a depth d ictated by
I »i i s p io to c o l. 1’hey have d elicately spat on the dust and now,
ill Ilu*o lh ei hand, they slo w ly knead it Is this for the pleasure o f
62
Mgure 2.17
63
molding figurines? No, it is to extract another judgment, one that no

longer involves color, but rather texture. Unfortunately, for this pur-
I»»se there is no equivalent of the Munsell code, and if there were one,
( wouldn’t know how to get it here. To define granularity in a stan-
111dized manner, one would need half of a well-equipped laboratory.
( onsequently, our friends must content themselves with a qualitative
I I Ihat rests on thirty years’ experience and that they will later com-
IIne with laboratory results. If the soil is easily molded, it is clay; if it
1umbles under one’s fingers, then one is dealing with sand. Here is an
i|»|».iiently very easy trial that amounts to a sort of laboratory experi-
11ii ni in the hollow of one’s hand. The two extremes are easily recog-
nl ible, even by a beginner like me. It is the intermediate compounds
I s.md and clay that make the differentiation difficult and crucial,
line we are interested in qualifying the subtle modifications of the
1insition soils which are more clayey toward the forest and more
indy toward the savanna.
I it king any kind of gauge, Armand and René rely on a back-and-
illi discussion of their judgments of taste, as my father would do
In n he tasted his Corton wines.
Simdy-clay or clayey-sand?”
No, I would say clayey, sandy, no sandy-clay.”
Wail, mold it a bit more, give it some time.”
( )l ay, yes, let’s say between sandy-clay and clayey-sand.”
Ilelo'isa, make a note: at P2, between five and seventeen centime-
1 areno-argiloso a argilo-arenoso. ” (I forgot to mention that we are al-
111 ilmg constantly between French and Portuguese, the politics of
nip1mge being added to the politics of race, gender, and disciplines.)
I lie combination of discussion, know-how, and physical manipula
it illows for the extraction of a calibrated qualification of texture
Ii il uni immediately replace, in the notebook, the soil that can now
tin own away. A word replaces a thing while conserving a trait that
Unis it. Is this a term-to-term correspondence? No, the judgment
< not resemble the soil. Is this metaphorical displacement? No
I mi so than a correspondence. Is it metonymy? Not that either, since
II i we lake a handful of soil for the whole horizon, we keep only
h il Is on the paper of the notebook and none of the earth that was
«I to qualify it. Is this compression of data? Yes, definitely, since
I hi woitls occupy the location of the soil sample, but it is a change of
P A ND O R A ’S HOPE
64
state so radical that now a sign appears in place of a thing. Here it is no

longer a question of reduction but of transubstantiation.
Are we crossing the sacred boundary that divides the world from
discourse? Obviously yes, but we have already crossed it a good ten
times. This new leap is no more distant than the preceding one, in
which the earth extracted by René, cleaned of blades of grass and
worm feces, became evidence in a test of its resistance to molding; or
the one before that, in which Sandoval dug the P2 hole with his mat
tock; or the following one, in which, on the diagram, the whole hori
zon from five to seventeen centimeters takes on a single texture, allow
ing, through induction, the coverage of the surface from a point; or
the n + i transformation that permits a diagram drawn on millimeter
ruled graph paper to play the role of internal referent for the written
report. There is nothing privileged about the passage to words, and all
stages can serve equally to allow us to grasp the nesting of reference
In none of the stages is it ever a question of copying the preceding
stage. Rather, it is a matter of aligning each stage with the ones thal
precede and follow it, so that, beginning with the last stage, one will be
able to return to the first.
How can we qualify this relation of representation, of delegation,
when it is not mimetic yet is so regulated, so exact, so packed with re
ality, and, in the end, so realistic? Philosophers fool themselves when
they look for a correspondence between words and things as the ulti
mate standard of truth. There is truth and there is reality, but then
is neither correspondence nor adequatio. To attest to and guarantee
what we say, there is a much more reliable movement—indirect, cross
wise, and crablike—through successive layers of transformation
(James [1907] 1975). At each step, most of the elements are lost but also
renewed, thus leaping across the straits that separate matter and foi m
without aid other than, occasionally, a resemblance that is more tenu
ous than the rails that help climbers over the most acrobatic passes
In Figure 2.18 we are on the site, toward the end of our expedition
and René is commenting on a diagram on graph paper of a veil it it
cross-section of the transect that we have just dug and exammtd
Torn, dirty, stained with sweat, incomplete, and sketched in pent II
this diagram is the direct predecessor of the one in Figuie 2.15 1out1
the one to the other theie are indeed transfoi mations, which mt ItuIm
processes of selection, cenlenng, lettenng, and cleaning, but (lu
65
Figure 2.18
minor in com parison w ith the transform ations through w hich we

' 1 |usl passed (Tufte 1984).
In Ilie middle of the photograph René is indicating a line w ith his
1 .1 gesture we have followed from the first (see Figures 2.1 and
. 1 1111less it is pointed in anger as a prelude to a fist, the extension of
111,|c\ linger always signals an access to reality even when it targets
..... I,icce of paper, an access w hich in this case nonetheless encom-
1he totality of the site, which, paradoxically, has entirely disap-
■11even as we are sweating at the center of it. This is the same re-
,1 ,,I space and tim e we have already seen many tim es: thanks to
,i|.lions, we are able to oversee and control a situation in w hich
tl, ..iibmerged, we becom e superior to that which is greater than
,M.I we are able to gather together synoptically all the actions that
......I over many days and that we have since forgotten.
j ,,, il,,, diagram not only redistributes the tem poral flux and inverts
I,,, 11, hit al order of space, il reveals to us features th at previously
invisible even though they were literally under the feet of our
, 1,,,. 1-.1*. Il is impossible lor us to see the forest-savanna transition
, 1,, ,11 , toss sections, to <|ualily il in homogeneous horizons, and
66
to mark it with data-points and lines. René points with his finger
made of flesh and attracts the gaze o f the living onto a profile whose
observer could never exist. The observer would have to not only reside
under the earth like a mole but be able to cut the soil as if with a blade
hundreds of meters long and replace the confusing variation of
forms with homogeneous hatchings! To say that a scientist “occupies
a standpoint” is never very useful, since she w ill immediately move to
another through the application o f an instrument. Scientists never
stand in their standpoint.
Despite the implausible vista it offers, the diagram adds to our in
formation. On one paper surface we combine very different sources
that are blended through the intermediary of a homogeneous graphi
cal language. The positions o f the samples along the transect, the
depths, the horizons, the textures, and the reference numbers of the
colors can be added to one another by superposition—and the reality
we had lost is replaced.
René, for instance, has just added to the diagrams the worm feces I
have mentioned. According to my friends, it seems that the worms
may carry the solution to the enigma within their particularly vora
cious digestive tracts. What produces the strip of clayey soil in the sa
vanna at the edge of the forest? Not the forest, since this strip extends
twenty meters beyond the protective shadow and nourishing humid
ity of the trees. Not the savanna either, since, let us remember, it al
ways reduces clay into sand. What is this mysterious action at a dis
tance that prepares the soil for the arrival of the forest, ascending the
thermodynamic slope that continues to degrade the clay? Why not the
earthworms? Might they be the catalyzing agents of the pedogenesis?
In modeling the situation, the diagram allows for the imagining ol
new scenarios, which our friends discuss passionately while consider
ing what is missing and where to dig the next hole to get back to the
“raw data” with their pick and drill (Ochs, Jacoby, et al. 1994).
Is the diagram that René holds in his hand more abstract or moi 1
concrete than our previous stages? More abstract, since here an
infinitesimal fraction of the original situation is preserved; more con
crete, since we can grasp in our hands, and see with our eyes, the es
sence of the forest-savanna transition, summarized in a few lines I
the diagram a construction, a discovery, an invention, or a conven
tion? All four, as always. The diagram is ion\tnutcd by the labois ol
67
li\e people and by passing through successive geometrical construc-

l ions. We are well aware that we have invented it and that, without us
iml the pedologists, it would never have appeared. Still, it discovers a
loi in that until now has been hidden but that we retrospectively feel
is already there beneath the visible features of the soil. At the same
lime, we know that without the conventional coding of judgments,
Ini ms, tags, and words, all we could see in this diagram drawn from
l lu1earth would be formless scribbles.
All of these contradictory qualities—contradictory, that is, for us
philosophers—ballast this diagram with reality. It is not realistic; it
I tes not resemble anything. It does more than resemble. It takes the
lit c of the original situation, which we can retrace, thanks to the proto-
il book, the tags, the pedocomparator, the record cards, the stakes,
iml finally, the delicate spiderweb woven by the “pedofil.” Yet we
mnot divorce this diagram from this series of transformations. In
•I ilion, it would have no further meaning. It replaces without re-
Ih mg anything. It summarizes without being able to substitute com-
I lely for what it has gathered. It is a strange transversal object, an
i mnent operator, truthful only on condition that it allow for passage
Iv cm what precedes and what follows it.
( hi Ihe last day of the expedition we find ourselves in the restaurant,
11ansformed into a meeting room for our mobile laboratory, in
It i lo write a draft of our report (Figure 2.19). René is holding the
completed diagram in his hand and commenting on it, point-
ith a pencil for the benefit of Edileusa and Héloïsa. Armand has
I Imished reading the only thesis that has been published on our
1m 1 of the forest, and he has opened it to pages of color photo-
iplis obtained by satellite. In the foreground rest the notebooks of
mlhropologist who is taking this picture—one more form of re-
I Img amid forms of inscription. We are again among maps and
II Iwo dimensional documents and published literature, already
li I.11 horn the site where we have labored for ten days. Have we,
11 leluined to our starting point (see Figure 2.2)? No, because we
have gained these diagrams, these new inscriptions we are at-
mplmg lo interpret and to insert as an appendix and as evidence
I 1 1 n.u 1alive we aie negotiating together, paragraph by paragraph,
I o languages, lench and Poi luguese. Let me quote a passage from
1
1 1 one
PANDORA'S HOPE
68
Figure 2.19
The interest of this expedition report stems from the fact that, in
the first phase of work, the conclusions of the approaches of botany
and pedology appear contradictory. Without the contribution of the bo
tanical data, the pedologists would have concluded that the savanna is ad
vancing on the forest. The collaboration of the two disciplines in this
case has forced us to ask new questions of pedology, (italics in the
original)
Here we are on much m ore fam iliar terrain—rhetoric, discourse,

epistemology, and the w riting of articles—busy w ith the weighing ol
argum ents for and against the advance of the forest. N either philoso
phers of language, nor sociologists of controversy, nor semioticians,
nor rhetoricians, nor scholars of literature will have much difficult)
here.
As thrilling as will be the transform ations that Boa Vista will tm
dergo from text to text, I do not, for the m om ent, wish to follow them
W hat interests me now is the transform ation undergone by the soil
now bound up in words. I low to sum m arize this? I need to draw, not ,1
diagram on graph paper like that ol my colleagues, but at least .1
sketch, a schema that will allow me to locale and point to what I, in
69
my own field of science studies, have discovered: a discovery brought
luck from the underworld, worthy of our lowly brethren, the earth
worms.
The philosophy of language makes it seem as if there exist two dis
jointed spheres separated by a unique and radical gap that must be re
duced through the search for correspondence, for reference, between
words and the world (Figure 2.20). While following the expedition to
hoa Vista, I arrived at a quite different solution (Figure 2.21). Knowl-
ulge, it seems, does not reside in the face-to-face confrontation of a
mind with an object, any more than reference designates a thing by
means of a sentence verified by that thing. On the contrary, at every
(age we have recognized a common operator, which belongs to mat
in at one end, to form at the other, and which is separated from the
(age that follows it by a gap that no resemblance could fill. The opéra
it us are linked in a series that passes across the difference between
Ilungs and words, and that redistributes these two obsolete fixtures of
Ilie philosophy of language: the earth becomes a cardboard cube,
01ds become paper, colors become numbers, and so forth.
An essential property of this chain is that it must remain reversible.
I lie succession of stages must be traceable, allowing for travel in both
duections. If the chain is interrupted at any point, it ceases to trans-
Im>11 truth—ceases, that is, to produce, to construct, to trace, and to
u induct it. The word “reference” designates the quality of the chain in its
ninety, and no longer adequatio rei et intellectus. Truth-value circulates
lu 1e like electricity through a wire, so long as this circuit is not inter-
l uplcd.1
Correspondence
11 iiu 1 10 I he “s.illulionisl’s” (James [1907] W75) conception of the feat of cor-

I poiideiKe implies that theie is a gap between woild and words that reference
aim In budge
70
Elements of representation
Form
I Matter Gap
Chain of elements
Upstream
Representation
00 I? 00
◄--------------
Downstream
Figure 2.21 The “deambulatory” conception of reference follows a series of trans
formations, each of them implying a small gap between “form” and “matter”; ref
erence, in this view, qualifies the movement back and forth as well as the quality of
the transformation; the key point is that reference, in this model, grows from the
center toward the two extremities.
Another property is revealed by the comparison of my two

sketches: the chain has no limit at either end. In the prior model (Fig
ure 2.20), the world and language existed as two finite spheres capable
of self-enclosure. Here, on the contrary, we can elongate the chain
indefinitely by extending it at both ends, by adding other stages—yet
we can neither cut the line nor skip a sequence, despite our capacity to
summarize them all in a single “black box.”
In order to understand the chain of transformation, and to grasp the
dialectic of gain and loss that, as we have seen, characterizes each
stage, we must look from above as well as at the cross-section (Figure
2.22). From forest to expedition report, we have consistently re
represented the forest-savanna transition as if drawing two isosceles
triangles covering each other in reverse. Stage by stage, we lost local
ity, particularity, materiality, multiplicity, and continuity, such that,
in the end, there was scarcely anything left but a few leaves of paper.
Let us give the name reduction to the first triangle, whose tip is all
that finally counts. But at each stage we have not only reduced, we
have also gained or regained, since, with the same work of ic
representation, we have been able to obtain much greater compatibil
ity, standardization, text, calculation, circulation, and 1dative univei
sality, such that by the end, inside the field lepoit, we hold not only
71
I I 'me 2.22 The transformation at each step of the reference (see Figure 2.21)
II iv be pictured as a trade-off between what is gained (amplification) and what
I In l (reduction) at each information-producing step.
dl ol Boa Vista (to which we can return), but also the explanation of
II dynamic. We have been able, at every stage, to extend our link with
ilu.tdy established practical knowledge, starting with the old trigo-
11 iineliy placed “behind” phenomena and ending up with all of the
ecology, the new findings of “botanical pedology.” Let us call this
1oiul triangle, by which the tiny transect of Boa Vista has been en-
1d with a vast and powerful basis, amplification.
( )iii philosophical tradition has been mistaken in wanting to make
In nomena* the meeting point between things-in-themselves and cat-
01 les of human understanding (Figure 2.23; also see Chapter 4). Re-
I 1 1 empiiicists, idealists, and assorted rationalists have fought
i ( lessly among themselves around this bipolar model. Phenomena,
ivei, aie not found at the meeting point between things and the
mi ol llie human mind; phenomena are what circulates all along
In level sihle chain ol tiansloi mations, at each step losing some prop-
II li lo gam otheis that lendei them compatible with already
72
Figure 2.23 In the Kantian scenography, phenomena reside at the meeting point
between the inaccessible things in themselves and the categorizing work made by
the active Ego; with circulating reference, phenomena are what routinely circu
lates through the cascade of transformations.
established centers of calculation. Instead of growing from two fixed

extremities toward a stable meeting point in the middle, the unstable
reference grows from the middle toward the ends, which are continually
pushed further away. To understand how Kantian philosophy has
muddled the triangles, a fifteen-day expedition is all that is required.
(All that is required, I hasten to add, on condition that I am not asked
to speak of my work in the same lavish detail in which the pedologists
report theirs: fifteen days would then become twenty-five years of
hard labor at controversies with scores of dear colleagues equipped
with decades worth of data, instruments, and concepts. I portray my
self here, without fear o f contradiction, as a simple spectator with
easy access to the knowledge of my informants. A reflexivity that
could follow every thread at once is, I would be the first to admit, be
yond me.)
Is it possible, with the help of my schema, to understand, visualize,
and detect why the original model of philosophers of language is so
widespread, when this slightest inquiry quickly reveals its impossibil
ity? Nothing could be simpler; all we need to do is obliterate, bit by
bit, each o f the stages we have witnessed in this photomontage (Figuie
2.24).
Let us block in the extremities o f the chain as il one w eie the lefei
73
Circulating reference
Mediations from matter to form
-----------------------------------------------------------------------------------------------►
oo LI
■<--------------------------------
Mediations from matter to form
l'he canonical view
o
Creation
Erasure of all mediations
Creation of a gap
o
o f one material to replace the lost of one formal
extremity: mediations and of extremity:
the world a longing for language
corresDondence
1 1>me 2.24 To obtain the canonical model of words and world separated by an
I• ss and related by the perilous bridge of correspondence, one has simply to con-
I It 1 l lie circulating reference and to eliminate all mediations as being unneces-
II 1111ermediaries that render the connection opaque. This is possible only at the
I Divisional) end of the process.
ni Ihe forest of Boa Vista, and the other were a phrase, “the forest of
I it 1 Vista.” Let us erase all the mediations that I have delighted in de-
IIhmg. In place of the forgotten mediations, let us create a radical
II> one capable of covering the huge abyss that separates the state-
1util I utter in Paris and its referent six thousand kilometers away. Et
a we have returned to the former model, searching for something
lill Ihe void we have created, looking for some adequatio, some re-
mblance between two ontological varieties that we have made as
iiiul.u as possible. It is hardly surprising that philosophers have
11 unable to reach an understanding on the question of realism and
I ill 1sin they have taken the two provisional extremities for the en-
1 i ham, as if they had tried to understand how a lamp and a switch
till I coiiespond" to each oilier after cutting the wire and making
limp ‘ga/e out” at the “exlei nal” switch. As William James said in
I povvului style
74
The intermediaries which in their concrete particularity form a

bridge, evaporate ideally into an empty interval to cross, and then,
the relation of the end-terms having become saltatory, the whole ho
cus-pocus of erkenntnistheorie begins, and goes on unrestrained by
further concrete considerations. The idea, in ‘meaning’ an object sep
arated by an ‘epistemological chasm’ from itself, now executes what
Professor Ladd calls a ‘salto mortale’ . . . The relation between idea
and object, thus made abstract and saltatory, is thenceforward op
posed, as being more essential and previous, to its own ambulatory
self, and the more concrete description is branded as either false or
insufficient. (James [1907] 1975, 247-248)
The next morning, after drafting the expedition report, we load the
precious cardboard boxes containing the earthworms preserved in
formaldehyde, and the neatly tagged little bags of earth, into the jeep
(Figure 2.25). And this is what philosophical arguments that wish to
link language to the world by a single regular transformation cannot
successfully explain. From text we return to things, displaced a little
further. From the restaurant-laboratory we set out for another labora
tory a thousand kilometers away, in Manaus, and from there to Jussieu
University in Paris, another six thousand kilometers away. Sandoval
will return to Manaus alone with the precious samples that he must
preserve intact despite the arduous trek that lies ahead. As I have said,
each stage is matter for what follows and form for what precedes it,
each separated from the other by a gap as wide as the distance be
tween that which counts as words and that which counts as things.
They are getting ready to leave, but they are also preparing to return.
Each sequence flows “upstream” and “downstream,” and in this way
the double direction o f the movement of reference is amplified. To
know is not simply to explore, but rather is to be able to make your
way back over your own footsteps, following the path you have just
marked out. The report that we drafted the night before makes this
much clear: another expedition is required, to study the activity of
those suspicious earthworms at the same field site:
From a pedological point of view, admitting that the forest is advanc

ing on the savanna implies;
1. that the forest and the biological activity particular to it trans
form a sandy soil into a clayey-sandy soil in the top 15 to 20 centime
ters;
76
2. that this transformation would begin in the savanna in a 15- to

30-meter band at the edge.
While these two notions are difficult to conceive when starting
from the assumptions of classical pedology, it is necessary, taking
into account the solidity of the arguments derived from biological
study, to test these hypotheses.
The clay enrichment of superior horizons cannot be accomplished
by neoformation (lacking a known source of aluminum [aluminum
is responsible for the creation of clay out of the silica contained
in quartz]). The only agents capable of accomplishing this are the
earthworms, whose activity on the studied site we have been able to
verify, and which dispose of large quantities of koalinite contained in
the horizon to a depth of 70 cm. The study of this worm population
and the measure of its activity will therefore supply essential data for
the continuation of this research.
Unfortunately, I will not be able to follow the next expedition.

While the other members of the team say au revoir to Edileusa, I must
say adieu. We are leaving by plane. Edileusa is staying in Boa Vista,
pleased by an intense and friendly collaboration that was new to her,
and she will continue to watch over her field site, which, because of
the superposition of pedology and botany, has just increased in impor
tance. And her plot will thicken more once we add the science of
earthworms. Constructing a phenomenon in successive layers renders
it more and more real within a network traced by the displacements
(in both senses) of researchers, samples, graphics, specimens, maps,
reports, and funding requests.
For this network to begin to lie—for it to cease to refer—it is
sufficient to interrupt its expansion at either end, to stop providing
for it, to suspend its funding, or to break it at any other point. II
Sandoval’s jeep swerves, breaking the jars of earthworms and scatter
ing the little packages of earth, the whole expedition will have to be re
peated. If my friends cannot find the funding to return to the field, we
will never know if the sentence in the report about the role of the
earthworms is a scientific truth, a gratuitous hypothesis, or a fiction
And if I lose all my negatives at the photo shop, how will anyone know
whether I have lied?
Air conditioning at last! Finally, a space that looks moie like a lahn
ratory (Figure 2.26). We are in Manaus, at INPA, in an old woil
77
I'ij'iuv i:i()
78
room transformed into an office. On the wall, Radambrasil’s map of

Amazonia and Mendeleev’s chart. Offprints, files, slides, canteens,
bags, cans o f gasoline, an outboard motor. Smoking a cigarette,
Armand writes the final version of the report on his laptop computer.
The forest-savanna transition o f Boa Vista continues its transforma
tions. Once typed in and saved on disk, the transition w ill circulate by
fax, electronic mail, diskette, preceding the suitcases heavy with the
earth and earthworms that will undergo various series of new trials
in various laboratories selected by our pedologists. The results w ill re
turn to thicken the piles of notes and files on Armand’s desk, in sup
port o f his request for funding to return to the field. The unending
round of scientific credibility: each turn absorbs more of Amazonia
into pedology, a motion that cannot stop lest significance and
signification be immediately lost.
Smoking a cigar, I too am writing my report on my laptop. Back in
Paris, I am sitting at a desk cluttered with books, files, and slides, in
front of an immense map of the Amazon basin. Like my colleagues, I
extend the network of the forest-savanna transition—all the way to
philosophers and sociologists, to the readers o f this book. The section
of the network that I am constructing, however, is made, not of the
sort of references enacted by the other scientists, but of allusions ami
illustrations. My schemas do not refer in the same way as their dm
grams and maps. Unlike Armand’s inscription o f the soil of Boa Visla
my photographs do not transport that o f which I speak. I am writing a
text of empirical philosophy that does not re-represent its evidence in
the manner of my pedologist friends, and hence the traceability of my
subject matter is not sufficiently immutable to permit the reader’s le
turn to the field. (I will leave it to the reader to measure the distance
that separates the natural and social sciences, for that mystery would
require another expedition, one that would study the role of the ban
tarn empiricist that I have been playing.)
You can now look at a map of Brazil in an atlas, at the area around
Boa Vista, but not for a resemblance between the map and the sit*
whose story I have been recounting. This whole tired question ol (lie
correspondence between words and the world stems from a simplt*
confusion between epistemology and the history of art. We have (al t II
science for realist painting, imagining that it made an exact copy of
the world. The sciences do something else cnluely paintings loo
79
loi Ihat matter. Through successive stages they link us to an aligned,

11 msformed, constructed world. We forfeit resemblance, in this
model, but there is compensation: by pointing with our index fingers
lo leatures of an entry printed in an atlas, we can, through a series of
nnlormly discontinuous transformations, link ourselves to Boa Vista,
h i us rejoice in this long chain of transformations, this potentially
mHess sequence of mediators, instead of begging for the poor plea-
mes of adequatio and for the rather dangerous salto mortale that James
i meely ridiculed. I can never verify the resemblance between my
iliul and the world, but I can, if I pay the price, extend the chain of
i mslormations wherever verified reference circulates through con-
I ml substitutions. Is this “deambulatory” philosophy of science not
11111' lealist, and certainly more realistic, than the old settlement?
C H A P T E R T H R E E
Science’s Blood Flow

An Example from Joliot’s Scientific Intelligence
Now that we have begun to understand that reference is something

that circulates, everything is going to change in our understanding
of the connections between a scientific discipline and the rest of its
world. In particular, we are going to be able to reconnect many of the
contextual elements that we had to abandon in the previous chapter.
With more than a little exaggeration, science studies can be said to
have made a discovery not totally dissimilar to that of the great Wil
liam Harvey himself. . . By following the ways in which facts circulate,
we will be able to reconstruct, blood vessel after blood vessel, the
whole circulatory system of science. The notion of a science isolated
from the rest of the society will become as meaningless as the idea of a
system of arteries disconnected from the system of veins. Even the no
tion of a conceptual “heart” of science will take on a completely differ
ent meaning once we begin to examine the rich vascularization that
makes the scientific disciplines alive.
To exemplify this second point I w ill take a canonical example, this
time not from a science as green and friendly as pedology but from
one as heavy and somber as atomic physics. My intention is not to add
to the history and anthropology o f physics as many o f my colleagues
have so excellently done (Schaffer 1994; Pickering 1995; Galison 1997),
but to recast the meaning of the little adjective “social.” If, in Chaptei
2, 1 had to abandon m ost o f the threads leading outward to the context
of the expedition, in this chapter I will leave out most of the technical
content to concentrate on the threading itself. This will allow me to in
troduce the little bit of classic sociology o f science that we need to
80
S C I E N C E ’S BLOOD FLOW
8l
i ontinue, and to help readers who believe science studies aims to pro
vide a “social” explanation of science abandon this prejudice. Once we
ne equipped with a different notion of reference and a renewed con-
i eption of the social, it will be possible to integrate the two with an al-
11*1native definition of the object. I wish I could go faster, but going
I ist, in these matters, is a sure recipe for simply repeating the old set-
llement without any hope of illuminating the new one that is still
II naked in darkness.
A Little Example from Joliot

In May 1939 Frédéric Joliot, advised by his friends in the Ministry of
Vai and by André Laugier, the director of the recently established
1 RS (Centre National de la Recherche Scientifique, France’s Na
in nul Center for Scientific Research), entered into a very subtle le-
il agreement with a Belgian company, the Union Minière du Haut-
I iianga. Thanks to the discovery of radium by Pierre and Marie Curie
ml the discovery of uranium deposits in the Congo, this company had
uome the most important supplier to all the laboratories in the
in kl that were feeling their way toward the production of the first
ulilidal nuclear chain reaction. Joliot, like his mother-in-law Marie
in ic before him, had found a way of getting the company involved. In
i I Ihe Union Minière used its radioactive ores only as a source of ra-
11111M, which it sold to doctors; immense heaps of uranium oxide were
11 lying about at its waste sites. Joliet planned to build an atomic re
lui, lor which he would need a huge quantity of uranium; this made
li 11 had been a mere waste product of the production of radium into
milhing valuable. The company promised Joliot five tons of ura-
1111oxide, technical assistance, and a million francs. In return, all the
in h scientists’ discoveries would be patented by a syndicate which
tild distribute the profits fifty-fifty between the Union Minière and
I NRS.
U inwhile, in his laboratory at the Collège de France, Joliot and his
1111.1m research colleagues, Hans Halban and Lew Kowarski, were
I lug loi an arrangement just as subtle as the one that had brought
111i c i Ihe mlei ests of the Ministry of War, the CNRS, and the Un
it 111 i c i c But this lime it was a matter of coordinating the appar-
1
ll 11leconulable behaviois ol atomic pai tides. The principle of

82
fission had just been discovered. When bombarded by neutrons, each
atom of uranium broke in two, liberating energy. This artificial radio
activity had a consequence that was immediately grasped by several
physicists: if under bombardment each atom of uranium gave off two
or three other neutrons which in turn bombarded other atoms of ura
nium, an extremely powerful chain reaction would be set in motion
Joliot’s team immediately set to work to prove that such a reaction
could be produced, and that it would open the way to new scientific
discoveries and to a new technique for producing energy in unlimited
quantities. The first team able to prove that each generation of neu
trons did indeed give birth to an even greater number would gain con
siderable prestige in the highly competitive scientific community, in
which the French occupied, at that time, a position of the first rank.
Determined to pursue this important scientific discovery, Joliot and
his colleagues continued to publish their findings, despite the urgent
telegrams Leo Szilard was sending them from America. In 19*4
Szilard, an émigré from Hungary and a visionary physicist, had taken
out a secret patent on the principles of construction of an atomic
bomb. Worried that the Germans too would develop an atomic bomb
as soon as they could be certain that the neutrons emitted were mou»
numerous than those present at the beginning, Szilard fought to en
courage self-censorship by all anti-Nazi researchers. He could nnl
however, prevent Joliot from publishing a final article in the English
journal Nature in April 1939, which showed that it might be possiblp
to generate 3.5 neutrons per fission. On reading this article, physicIsl
in Germany, England, and the Soviet Union all had the same thought
they immediately reoriented their research toward bringing about M
chain reaction and just as quickly wrote to their governments to ah it
them to the vast importance of this research, to inform them ol
its dangers, and to request immediate provision of the enormous u»
sources needed to test Joliot’s claim.
Around the world about ten different teams became passional! I
engaged in the attempt to produce the first artificial nuclear chain u»
action, but only Joliot and his team were already in a position to lull
this into an industrial or military reality. Joliot’s first problem wa I
slow down the neutrons emitted by the first fissions, for if these wi I
too fast they would not set off the reaction. The team looked loi it
moderator that could slow the ncutions without absoibing (lu m 1»
SCIENCE'S BLOOD FLOW
83
Minting them back; thus the ideal moderator would have a set of
I<>|)c‘i ties very difficult to reconcile. In their workshop at Ivry, they
1li il different moderators under different configurations, for example
II illin and graphite. It was Halban who drew their attention to the
1 ive advantages of deuterium, an isotope of hydrogen, twice as
1 y but with the same chemical behavior. It could take the place of
liogen in water molecules, which then became “heavy.” From ear-
1 v oik he had done on heavy water, Halban knew that it absorbed
lew neutrons. Unfortunately, this ideal moderator had one major
1 luck: there was only one atom of deuterium in every 6,000 atoms
li diogen. It cost a fortune to obtain heavy water, and it was pro-
1il on an industrial scale at only one plant in the world, which be-
i il lo the Norwegian company Norsk Hydro Elektrisk.
I ion I Dautry, a graduate of the Ecole Polytechnique and a senior
I ( 1vant who became the French Minister of Armaments only too
illy before the defeat of France in World War II, was also kept in-
IIti il ol Joliot’s work from the very beginning. He had been in favor
lioi’s agreement with the Union Minière and did everything he
llo support the team at the Collège de France and the early days
lu ( NRS, attempting to integrate, as much as French tradition al-
I military and advanced scientific research. Although he did not
|ollot’s leftist political opinions, he had the same confidence in
I 1op ess of knowledge and the same passion for national inde-
I me Joliot promised an experimental reactor for civilian use
li might eventually lead to the construction of a new type of ar-
111 Dautry and other technocrats offered Joliot generous sup
in le asking him to change his priorities: if the bomb was practi-
ll must be developed first and very quickly.
I ms calculations on the slowing of neutrons, Joliot’s hypothesis
li isibility of the chain reaction, and Dautry’s conviction about
1 it y ol developing new armaments became even more closely
hud when it came to obtaining the heavy water from Norway,
llit ' phony war” was taking place between the Siegfried and
I I Inol lines, spies, bankers, diplomats, and German, English,
Il nul Noiwegian physicists fought over twenty-six containers
1 tgians lud given the French lo prevent the Germans from
hold ol Ihem Allei an evenllul few weeks the containers
I |ohnl s possession Halban and Kowaiski, both foreigners and
84
therefore suspect, had been put out to pasture by the French secret
service for the duration of the operation. Once it was completed, they
were authorized to return to the laboratory at the Collège de France,
where under the protection of Dautry and the military, they set to
work to combine the uranium from the Union M inière and the heavy
water from the Norwegians with the calculations that Halban worked
out every day with the confusing data from their primitive Geigei
counter.
How to Link the History of

Science with That of France
How should we understand this story, so well told by the American
historian Spencer Weart (1979), of which I have given only a summai y
of a single episode? Two major misunderstandings have made the pro
ject of science studies of mapping the circulatory system o f science in
comprehensible. The first is the belief that science studies seeks a “so
cial explanation” o f scientific facts; the second the belief that it deal
only with discourse and rhetoric, or at best epistemological questions
but does not care about “the real world outside.” Let us clear up each
of these misunderstandings in turn.
Science studies, to be sure, rejects the idea of a science disconnected
from the rest of society, but this rejection does not mean that it cm
braces the opposite position, that of a “social construction” of reality
or that it ends up in some intermediary position, trying to sort out
“purely” scientific factors from “merely” social ones (see the end ol
Chapter 4). What science studies rejects is the entire research program
that would try to divide the story of Joliot into two parts: one foi (lit*
legal problems with the Union Minière, the “phony war,” Dautry’s nu
tionalism, the German spies; and the other for neutrons, deutcimm
the absorption coefficient of paraffin. A scholar of this period would
then have two lists of characters corresponding to two stories: m (lit*
first, the history of France from 1939 to 1940; in the second, the hi
tory of science in the same period. The first list would deal with poll
tics, law, economics, institutions, and passions; the other with idiit
principles, knowledge, and procedures.
We might even imagine two subprofessions, two dilleienl 1 ind «I
85
liislorians, one preferring explanations by pure politics, the other by

Imi e science. The first kind o f explanation is usually called externalist*
IIni the second internalist*. In this period of 1939-1940, these two his-
•1ics would have no points of intersection. The one would speak of
tlolf Hitler, Raoul Dautry, Edouard Daladier, and the CNRS, but not
I neutrons, deuterium, or paraffin; the other would talk about the
Imt iple of the chain reaction, but not about the Union Minière or the
inks that owned Norsk Hydro Elektrisk. Like two teams of civil engi-
II s working in two parallel valleys in the Alps, they both would do
II ( normous amount of work without ever knowing of each other’s
I lence.
( >f course, once this division between human and nonhuman actors
i d 1awn, everyone would admit that there remained a slightly mud-
(I area of hybrids, which might be found perhaps in one column,
III ips in the other, or perhaps in neither. To deal with this “twilight
in " externalists and internalists would have to borrow factors from
li olher’s lists. One might say, for example, that Joliot “mixed
political concerns with purely scientific interests. Or one might
lint the plan to slow neutrons with deuterium was, of course, a
1111lie project, but that it was also “influenced” by extrascientific
(ms Szilard’s project of self-censorship was not “strictly scientific,”
might say, because it introduced military and political consider-
11s into the free interchange of ideas of pure science. In this way,
I Ilung that appears mixed is explained by reference to one of two
II lly pure constituents; politics and science.
It nee studies could be defined as the project whose aim is to do
1 with this division altogether. The story of Joliot as told by
in 11 Weart is a “seamless web” which cannot be torn in two with
in 11 mg both the politics of the time and the atomic physics in-
i| 111hensible. Instead of following the parallel valleys, the purpose
h ut i* studies is to dig a tunnel between them by putting together
It mis, which attack the problem from opposite ends and hope to
I In (lie middle.
following Ilalban’s arguments on cross-sections (Weart 1979),
li t o i k lude that deuterium has decisive advantages, the analyst of
11 1 Is led, without piejudice and without postulating a great divide
111 science and politics, through an imperceptible tra n sitio n into
li ollice, and liom then? into the plane ol Jacques Allier, a
86
banker and flying officer who was the secret agent sent by France to
outwit the fighters of the Luftwaffe. Starting on the science side of the*
tunnel, the historian ultimately arrives on the other side, with war and
politics. But en route she might meet a colleague coming from the
other direction who started with the industrial strategy of the Union
Minière and, through another imperceptible transition, ended up very
interested in the method of extraction of uranium 235, and subse
quently in Halban’s calculations. Starting from the politics side, this
historian, willingly or not, becomes involved in mathematics. Instead
of two histories which do not intersect at any point, we now have peo
pie who tell two symmetrical stories which include the same elements
and the same actors, but in the opposite order. The first scholar expected
to follow Halban’s calculations without having to deal with the
Luftwaffe, and the second imagined that he could look at the Union
Minière without having to do any atomic physics.
They were both mistaken, but the paths they traced, thanks to the
opening of the tunnel, are much more interesting than they had ex
pected. In fact, by following without prejudice the interconnected
threads of their reasoning, science studies will reveal a posteriori the
work the scientists and the politicians had to do to become so inexti I
cably bound together. It wasn’t determined in advance that all Ilu
elements of Weart’s account should be mixed together. The Union
Minière could have carried on producing and selling copper without
bothering about radium or uranium. If Marie Curie and later Frédei k
Joliot had not worked at getting the company interested in the woi It
done in their laboratories, an analyst from the Union Minière would
never have had to do nuclear physics. When discussing Joliot, Weill I
would never have had to speak of the Upper Katanga. Conversely, oui p
he had envisioned the possibility of a chain reaction, Joliot could have*
directed his research at some other topic, without having to mobill p
in order to produce a reactor, nearly all of France’s industrialists ami
enlightened technocrats. Writing about prewar France, Weart would
not have had to mention Joliot.
In other words, the project of science studies, contrary to wh.il 8»I
ence warriors have tried to make everyone believe, is not to stale a f>t
ori that there exists “some connection” between science and sot It I
because the existence of this connection depends on what the ai ton ho
done or not done to establish it. Science studies meiely pi ovules th
87
mm.ms of tracing this connection when it exists. Instead of cutting the

<loidian knot—on the one hand pure science, on the other pure poli-
»«s it struggles to follow the gestures of those who tie it tighter. The
>«i.i I history of the sciences does not say: “Look for society hidden in,
i fund, or underneath the sciences.” It merely asks some simple ques-
I *11s: “In a given period, how long can you follow a policy before hav-
nj/ lo deal with the detailed content of a science? How long can you
imi ne the reasoning of a scientist before having to get involved with
In details of a policy? A minute? A century? An eternity? A second?
II we ask of you is not to cut away the thread when it leads you,
In(*ugh a series of imperceptible transitions, from one type of ele-
i Ml lo another.” Ail the answers are interesting and count as major
III Ior anyone who wishes to understand this imbroglio of things
nd people—including, of course, the data that might show that there
m>l Ihe slightest connection, at a given time, between a piece of sci-
ii i and the rest of the culture.
II is not enough to say that the connections between science and
IIIus form a very tangled web. To refuse any a priori division be
lli Ilie list of human or political actors and that of ideas and proce-
11 s is no more than a first step, and an entirely negative one at that.
11msl also be able to understand the series of operations by which
Industrialist who wanted only to develop his business found him-
I Ioited to do calculations o f the rate of absorption of neutrons by
i illin, or how someone who wanted nothing but a Nobel Prize set
ill oiganizing a commando operation in Norway. In both cases the
11 vocabulary is different from the final vocabulary. There is a
lotion * o f political terms into scientific terms and vice versa. For
m muging director of the Union Minière, “making money” now
tii (o some extent, “investing in Joliot’s physics”; while for Joliot,
mi Misdating the possibility o f a chain reaction” now means in part
I lug out for Nazi spies.” The analysis of these translation opéra
it in ikes up a large portion o f science studies. The idea of transla-
ii pi ovules the two teams o f scholars, one coming from the side of
llli mid going toward the sciences and the other coming from the
ol lhe sciences and following the circulating references, with the
I in o I guidance and alignment that gives them some chance of
ling in the middle lather than missing each other.
I I il follow an elemental y opei at ion o f tianslation, so as to under
88
stand how in practice one passes from one register to another. Dautry
wants to ensure France’s military strength and the self-sufficiency of
its energy production. Let us say that this is his “goal”—whatever psy
chology we wish to impute to him. Joliot wants to be the first in the
world to produce controlled artificial nuclear fission in the laboratory;
this is his goal. To call the first ambition “purely political” and the sec
ond “purely scientific” is completely pointless, because it is the “impu
rity” alone that will allow both goals to be attained.
Indeed, when Joliot met Dautry he did not particularly try to change
Dautry’s goal, but to position his own project in such a way that
Dautry would see the nuclear chain reaction as the fastest and most
certain way of achieving national independence. “If you use my labo
ratory,” Joliot may have said, “it will be possible to gain a significant
lead over other countries, and perhaps even to produce an explosive
that goes beyond anything we know.” This transaction is not of a com
mercial nature. For Joliot it is not a question of selling nuclear fission,
since it doesn’t even exist yet. On the contrary, the only way he can
make it exist is to receive from the Minister of Armaments the person
nel, the premises, and the connections that will enable him, in thi
middle of a war, to obtain the tons of graphite, the uranium, and the li
ters of heavy water that are needed. Both men believe that, since if
is impossible for either to achieve his goal directly, political and scicn
tific purity are in vain, and that it will thus be best to negotiate an ai
rangement that modifies the relation between their two original goals
The operation of translation consists of combining two hitherto dil
ferent interests (waging war, slowing down neutrons) to form a single
composite goal (see Figure 3.1). Of course there is no guarantee that
one or the other of the parties isn’t cheating. Dautry may be squandoi
ing precious resources by letting Joliot fool around with his neutron
while the Germans are massing their tanks in the Ardennes. Con
versely, Joliot may feel he is being forced to build the bomb before IIn
civilian reactor. Even if the balance is equal, neither of the parties, as I
shown in the diagram, will be able to arrive at exactly his original goal
There is a drift, a slippage, a displacement, which, depending on (In
case, may be tiny or infinitely large.
In the case we are using as an example, Joliot and Daulry did not
achieve their goals until fifteen years later, alter a ten ible defeat, win n
General de Gaulle created the CFA, the Commiss.11 lat a I’f'neigli
89
IWlore translation
Dautry Goal : national

independence
Joliot
Goal : first to master
chain reaction
Aller translation
Goal : national
independence 5
Goal : first to master
chain reaction
__ ^ New goal : a laboratory for
chain reaction
and future national
independence
mi 11 One should be careful not to fix interests a priori; interests are “trans-
I 11 That is, when their goals are frustrated, actors take detours through the
11 ol others, resulting in a general drift, the language of one actor being substi-
I d loi the language of another.
hunique (Atomic Energy Commission). What is important in such

1«»pei ation of translation is not only the fusion of interests that it al
lnit the creation of a new mixture, the laboratory. In fact, the
» 11 al Ivry became the crucial juncture that would allow the joint re-
ilion of both Joliot’s scientific project and the national independ-
» so close to Dautry’s heart. The laboratory’s walls, its equipment,
I ill, and its resources were brought into existence by both Dautry
I )ohot. It was no longer possible to tell, among the complex of
I 1 mobilized around the copper sphere filled with uranium and
II tl Im, what belonged to Joliot and what to Dautry.
10 (udy a single negotiation or translation in isolation would be
I s Joliol's labors could not of course be confined to ministerial
11 1 I laving gained his laboialoi y, he now had to go and negotiate
• llu tn u iro n s them selves W.is it one thing lo peisuade a minister to
90
provide a stock of graphite, and quite another to persuade a neutron to

slow down enough to hit a uranium atom so as to provide three more
neutrons? Yes and no. For Joliot it wasn’t very different. In the morn
ing he dealt with the neutrons and in the afternoon he dealt with the
minister. The more time passed, the more these two problems became
one: if too many neutrons escaped from the copper vessel and lowered
the output of the reaction, the minister might lose patience. For Joliot,
containing the minister and the neutrons in the same project, keeping
them acting and keeping them under discipline, were not really dis
tinct tasks. He needed them both.
Joliot crossed and recrossed Paris, moving from mathematics to
law and to politics, sending telegrams to Szilard so the flow of publica
tions needed to promote the project would continue, telephoning his
legal adviser so the Union Minière would keep sending uranium, and
recalculating for the nth time the absorption curve obtained with his
rudimentary Geiger counter. Such was his scientific work: holding
together all the threads and getting favors from everybody, neu
trons, Norwegians, deuterium, colleagues, anti-Nazis, Americans
paraffin . . . No one said being a scientist was a simple job! To be intelh
gent, as the word’s etymology indicates, is to be able to hold all these
connections at once. To understand science is, with Joliot’s help (and
Weart’s), to understand this complex web of connections without
imagining in advance that there exist a given state of society and u
given state of science.
It is now easier to see the difference between science studies and Ilit
two parallel histories that it replaces. In order to explain all the polill
cal and scientific imbroglios, the two teams of historians always had In
see them as regrettable intermixings of two equally pure registers All
their explanations therefore had to be couched in terms of “distoi
tion,” of “impurity,” or at best of “juxtaposition.” For them, purely pn
litical or economic factors were added to purely scientific ones. Wlu I
these historians saw only confusion, science studies sees a slow, con
tinuous, and entirely explicable substitution of a certain kind of coil
cern and a certain kind of practice for another. There are in fact mi
ments when, if one holds firmly the calculation of the cross section 11
deuterium, one also holds, through substitutions and translations (It
fate of France, the future of industry, the destiny of physics, a pale ill rt
good paper, a Nobel Prize, and so on
91
With the help of another diagram it is possible to extend the con-

11ast between these two types of inquiry into the connections of sei
ne e. The left side of Figure 3.2 portrays the separation between sei
nee and politics in its most common form: there is a nucleus of
1icntific content surrounded by a social, political, and cultural “envi-
11minent,” which can be called the “context” of science. On the basis
•I such a separation it is possible to offer either externalist explana
tions or internalist ones and to feed the contradictory research pro-
1uns of our two teams of scholars. The members of the first team
ill use the vocabulary of context* and will attempt (sometimes) to
netrate as far as they can into the scientific content; the members of
In second will use the vocabulary of content* and will remain within
» centrai conceptual core. For the first, what explains science is soci-
.ilthough usually only the surface of the discipline is in question:
oiganization, the relative status of different workers, or the errors
I later shown to have produced. In the second the sciences explain
nselves, without need of external assistance since they provide their
111ommentary about themselves and develop from their own inner
1 is To be sure, the social environment can either hinder or encour-
Ihoir development, but it never forms or constitutes the very con-
tit ni the sciences.
( )n the right side of Figure 3.2 is the science studies program that
I be called the translation model* (Callon 1981). It should now be
II l lut there is no relation between the two paradigms. Science
I lli s does not occupy a position inside the classical debate between
in 1 1st and externalist history. It entirely reconfigures the ques-
1
II I lie only thing one can say is that the successive chains of trans
it involve, at one end, exoteric resources (which are more like what
1 id «ibout in the daily papers), and at the other end, esoteric re-
1 1 (which are more like what we read in university textbooks).
IIn se two ends are no more important and no more real than the
11t'inities of reference in the previous chapter, and for the same
n I veiything important happens between the two, and the same
in liions serve to carry the translation in both directions. In this
ml model the same methods are used to understand science and
I Sc icnce studies has never had any interest whatsoever, at least
» ll in piovidmg a social explanation of any given piece of sci-
II ll hail, it would have laded immediately, since nothing in the
92
Goal i :
exoteric
A
Society or
Chain of translations
explanation Goal 4 :
esoteric
MODEL 1 MODEL 2
Figure 3.2 In Model 1 science is conceived as a core surrounded by a corona ol

social contexts that are irrelevant to the definition of science; thus internalist and
externalist explanations have little in common. In Model 2 the successive transla
tions have made esoteric and exoteric vocabularies have something in common
and the distinction between internalist and externalist explanations is exactly a*,
small (or as large) as the chain of translation itself.
ordinary definition of w hat society is could account for the connection

between a M inister of A rm am ents and neutrons. It is only because ol
Joliot’s w ork that this connection has been made. Science studies loi
lows those implausible translations w hich mobilize in completely un
expected ways fresh definitions of w hat it is to make w ar and fresh
definitions of w hat the w orld is made of.
The Progressive Packing of N onhum ans

into H um an D iscourse
Now that the first misunderstanding has been cleared up, it will In
easier to deal with the second, especially with (he help of what \vi
learned about circulating reference in Chapter 2. Scientists not <>111\
blur, in their daily practice, the boundary between their pure esoleiit
science and the impure exoteric realm of society, they also bliu tin
boundary between the domain ol discourse and what Ihe world is lil •
S C I E N C E ’S B L O O D FLOW
93
Philosophers of science like to remind us, as if this were the epitome

til good common sense, that we should never confuse epistemological
questions (what our representation of the world is) and ontological
questions (what the world is really like). Unfortunately, if we followed
Ilie philosophers’ advice we would not understand any scientific activ
ity. since confusing those two supposedly separate domains is pre-
Isely what scientists spend much of their time doing. Joliot not only
II ii)slates social and scientific considerations more and more inti-
ni ilely, he also mixes up epistemological and ontological questions
iume thoroughly every day. It is only because of this gradually accu
mulating confusion that what he says about chain reactions can be
III en more and more seriously by others.
I (insider this sentence: (1) “Every neutron liberates 2.5 neutrons.”
Ilus is what one reads today in encyclopedias. This is what is called a
i mitific fact.” Now let us take another sentence: (2) “Joliot claims
Ini each neutron liberates 3 to 4 neutrons, but that’s impossible; he
i no proof; he’s far too optimistic; that’s the French all over, count-
np llieir chickens before they’re hatched; and in any case, it’s incredi-
I dangerous; if the Germans read his claim, they’ll believe it’s all
1 silile and work on it seriously.” Unlike sentence (1), sentence (2)
M not conform to the stylistic rules governing the appearance of
h nl die facts: it cannot be read in any encyclopedia. Its dated charac-
1 Is easily discernible (somewhere between 1939 and 1940) and it
I lil lie ascribed to a fellow physicist (such as Szilard, who had found
1 11ven at that time in Enrico Fermi’s laboratory on the South Side of
lilt igo). We may note that these two sentences have a section in
Minion, the statement or dictum*: “each neutron liberates x neu-
II , and a very different part, made up of an ensemble of situa
it people, and judgments, called the modifier or modus*.
I have shown many times, a convenient marker of the appear-
1 id a scientific fact is that the modifier drops entirely and only the
111111 is maintained. The elimination of these modifiers is the result
I (miel unes the goal of scientific controversy (as we will see in
1ipU 1 4, when Pasteur steps back from his yeast cells to let them
il loi lliemselves). For example, if Joliot and his group have done
Ii woik sutcesslully, his colleagues will move imperceptibly from
k o i k I sentence lo a third, more respectful one: (3) “The Joliot
III unis (o have pioved that every neutron liberates three neu-
PANDO RA’S HOPE
94
trons; that’s very interesting.” A few years later we will read sentences
like this: (4) “Numerous experiments have proven that each neutron
liberates between 2 and 3 neutrons.” One more effort, and we arrive al
the phrase with which we started: (1) “Each neutron liberates 2.5 neu
trons.” A little later this sentence, without a trace of qualification,
without author, without judgment, without polemics or controversies,
without even any allusion to the experimental mechanism that made
it possible, will enter into a state of even greater certainty. Atomic
physicists will not even speak of it, will even stop writing it—except
in an introductory course or a popular article—so obvious will it have
become. From lively controversy to tacit knowledge, the transition
is progressive and continuous—at least when everything goes well,
which is, of course, very rarely.
How are we going to account for this progressive shift from (2) to
(1) through (3) and (4)? Are we going to say, to use the tired cliche
that they tend “asymptotically” toward the true state of affairs? Al e
we going to say that (2) is still a human statement marked by language
and history while (1) is not a statement at all and has escaped history
and humanity altogether? The traditional way to answer these ques
tions is to try to identify among these statements the ones that com*
spond to a state of affairs and the ones that have no reference. Bui
again, science studies is not the research program that would take u
position in this classical debate. As we saw in Chapter 2, it is interested
in a rather different question: how can the world be progressively
packed into discourse through successive transformations so that u
stable flow of reference in two directions may ensue? How can Joliol
get rid of the qualifications that hedge the scientific fact he wishes (u
establish? The answer to this question explains why there can be no
other history of science than science studies, as I am defining it here
Joliot may be convinced in his own mind that the nuclear chain ic
action is feasible and that it will lead in a few years to the consli uc
tion of an atomic reactor. If, however, each time he states this possibll
ity, his colleagues add qualifications—such as “It is ridiculous li
believe that [dictum],” “It is impossible to think that [dictum],” “Il I
dangerous to imagine that [dictum],” “It is contrary to theory to claim
that [dictum]”—Joliot will find himself utterly powerless. lie camm!
b y h im se lf transform the statement he is proposing into a scientific I u I
that the others accept; by definition, he needs the others to hi mg ubtm!
95
IIns transformation. It was Szilard who had to admit, “I am now con-

mced that Joliot can make his reactor work,” even if he immediately
uliled, “as long as the Germans don’t get ahold of it if they occupy
l'ai is.” In other words, to reuse a slogan I have often employed, the
l.ilement’s fate is in the hands of others, in the hands mostly of dear
nlleagues, who are for this reason both loved and hated (the fewer
lliey are, and the more esoteric or important the statement in ques-
l Ion, the more they will be loved or hated).
I am not trying to stress here the regrettable “social dimension” of
«icnce that would serve to prove that scientists are only human, all
lo o human. Controversy is not something that would disappear if re-
ii i hers would only be “really scientific.” There is no way to skip any
I Ihe steps toward conviction; one might as well imagine Joliot im-
uliately writing an encyclopedia article on the operation of a nu-
« n power plant! It is always necessary to convince the others first,
m by one. The others are always there, skeptical, undisciplined, inat-
nl ive, uninterested; they form the social group that Joliot cannot do
llliout.
loliot, like all researchers, needs the others, needs to discipline them
mllo convince them; he is not able to do without them and lock him-
II up in the Collège de France, alone with his firm conviction that he
11ght. He is not, however, completely without weapons of his own.
pile the slanderous claim of the science warriors, science studies
i never said that the “others” mixed up in the conviction process
ii .ill humans. On the contrary, the whole effort of science studies
i been to follow the extraordinary mixtures of humans and non-
iiii ms that scientists had to devise in order to convince. Into his dis
ions with colleagues Joliot can introduce other resources than the
i l.issically handed down to him by rhetoric.
I Ins is the very reason he was in such a hurry to slow down the neu-
i wilh deuterium. Alone, he could not force his colleagues to be
lli in. If he could get his reactor going for only a few seconds, and
I ( t>111cl get evidence of this event that was sufficiently clear that no
i (uilil accuse him of seeing only what he wanted to see, then Joliot
uliI no longer be alone. With him, behind him, disciplined and su-
i I id by lus collaborators, and properly lined up, the neutrons of
ii ii loi could be made visible in the form of a cross sectional dia-
uii I he expei iment m the shed at Iviy was very expensive, but it
96
was precisely this expense that would force his esteemed colleagues to
take his article in N a tu re seriously. Science studies, once again, does
not take a position in a classical debate—is it rhetoric or proof that
finally convinces scientists?—but reconfigures the whole question in
order to understand this strange hybrid: a copper sphere built to con
vince.
For six months Joliot was the only one in the world who had at his
disposal the material resources allowing him to mobilize both col
leagues and neutrons around and inside a real reactor. Joliot’s opinion
by itself could be swept aside with a wave of the hand; Joliot’s opinion
supported by Halban’s and Kowarski’s diagrams, diagrams obtained
from the copper sphere in the shed at Ivry, could not so easily be casl
aside—the proof being that three countries at war immediately set to
work at building their own reactors. Disciplining men and mobilizing
things, mobilizing things by disciplining men; this is a new way ol
convincing, sometimes called scientific research.
In no way is science studies an analysis of the rhetoric of science, ol
the discursive dimension of science. It has always been an analysis ol
how language slowly becomes capable of transporting things them
selves w ith o u t deformation th rou gh transformations. The notion of tlx
huge gap between words and world made it impossible to understand
this progressive loading—as did the very distinction between rhetoi k
and reality, the political origins of which I will examine in Chapter
But getting rid of a nonexistent gap and of an even less real correspon
dence between two nonexistent things—words and world—is not ill
all the same thing as saying that humans are forever stuck in the prison
of language. It implies exactly the opposite. Nonhumans can be loadt d
into discourse exactly as easily as ministers can be made to understand
neutrons. As we will see in Chapter 6 , this is the simplest of all thing
to do. Only the sway of the modernist settlement could make IIII
commonsensical evidence appear bizarre.
What seemed shocking at first in this new paradigm was that it did
not rely on the myth of a heroic break a w a y from society, convent Ion
and discourse, a mythical break that would let the solitary stienll I
discover the world as it is. To be sure, we no longer portray sc lent I I
as those who abandon the realm of signs, politics, passions, and lid
ings in order to discover the world of cold and inhuman things li
themselves, “out there.” But that does not mean that we poiliay tlx h
S C I E N C E ’S B LO O D FLOW
97
is talking to humans, to humans only, because those they address in

Iheir research are not exactly humans but strange hybrids with long
I ills, trails, tentacles, filaments tying words to things which are, so to
peak, behind them, accessible only through highly indirect and im
mensely complex mediations of different series of instruments. The
II ni h of what scientists say no longer comes from their breaking away
Imm society, convention, mediations, connections, but from the
iIci y provided by the circulating references that cascade through a
i«.il number of transformations and translations, modifying and con-
li lining the speech acts of many humans over which no one has
n durable control. Instead of abandoning the base world of rhetoric,
i 'iimentation, calculation—much like the religious hermits of the
i l scientists begin to speak in truth because they plunge even more
«ply into the secular world of words, signs, passions, materials, and
udialions, and extend themselves ever further in intimate connec-
iiis with the nonhumans they have learned to bring to bear on their
( ussions.
II the traditional picture had the motto “The more disconnected a
li me the better,” science studies says, “The more connected a sci-
t Ihe more accurate it may become.” The quality of a science’s ref
ill e does not come from some salto mortale out of discourse and so-
I in order to access things, but depends rather on the extent of its
ni loi mations, the safety of its connections, the progressive accu-
il ilion of its mediations, the number of interlocutors it engages, its
till y lo make nonhumans accessible to words, its capacity to inter-
iikI to convince others, and its routine institutionalization of these
(see Chapter 5). There do not exist true statements that corre-
iid lo a state of affairs and false statements that do not, but only
I Iiukhis or interrupted reference. It is not a question of truthful sci-
I who have broken away from society and liars who are influ-
I by Ihe vagaries of passion and politics, but one of highly con-
I d suentists, such as Joliot, and sparsely connected scientists
li d only lo words.
In Imbioglio with which this chapter began is not a regrettable as-
I il suenlilic production; it is the result of that very production. At
I>(mil one finds people and things mixed up, opening up a con-
1 01 pulling an end lo one. Il, aller Joliot had outlined his pro-
I I ) lull y had not leceived a (avoiable lesponse liom his advisers.
98
Joliot would not have obtained the resources to mobilize the tons ol
graphite his experiment demanded—and if he had not been able to
convince Dautry’s advisers, he would not have been able to convince
his own colleagues. It is the same scientific work that led him to go
down to the shed at Ivry, to go up to Dautry’s office, to approach his
colleagues, to go back over his calculations. It is this same disciplining
and disciplined labor that led him to concern himself with the devel
opment of the CNRS—without which he would not have had col
leagues sophisticated enough in the new physics (Pestre 1984) to
find his arguments interesting; to give lectures to the workers in the
Communist suburbs—without which there would not have been wide
spread support for scientific research as a whole; to get the directoi 1
of the Union Minière to visit his laboratory—without which he would
not have received the tons of radioactive waste needed for his reactoi
to write articles for Nature—without which the very goal of his ic
search would have been foiled; but, above all else, to struggle to get the
damned reactor working.
As we will see, the energy with which Joliot pushed Szil.ml
Kowarski, Dautry, and all the others is proportional to the number ol
resources and interests he had already mobilized. If the reactor fails It
each neutron liberates no more than one other neutron, then all Iht v
accumulated resources will scatter and disperse. It will no longei In»
worth going to all this trouble. This line of research will be seen ft
costly, useless, or premature, and Joliot’s words will begin to lie I
lack reference. What matters for science studies is that a heteioga
neous assembly of hitherto unrelated elements now shares a com 11km
fate within a common collective, and that Joliot’s words will bee (Mil
true or false according to what circulates throughout this entire m wl
assembled collective. It is too late to claim that ontological an
epistemological questions should be kept neatly distinct. Became
Joliot’s work these questions are now tied to one another, and the 11 1
vance of what he says to what the world is really like now hinge
what happens in the copper sphere at Ivry.
The Circulatory System of Scientific Facts

Translation operations transform political questions into question
technique, and vice versa; during a contioversy, operations oi eon
SCIENCE’S BLOOD FLOW
99
Imn mobilize a mixture of human and nonhuman agents. Instead of

defining a priori the distance between the nucleus of scientific con
tint and its context, an assumption that would render incomprehensi-
lili1the numerous short-circuits between ministers and neutrons, sci-
iue studies follows leads, nodes, and pathways no matter how
looked and unpredictable they may look to traditional philosophers
I science. If it is impossible, by definition, to give a general descrip-
lon once and for all of the unpredictable and heterogeneous links that
plain the circulatory system that keeps scientific facts alive, it must
ei theless be possible to outline the different preoccupations that
I i i searchers will hold simultaneously if they want to be good scien-
I I,
111 us try to enumerate the various flows that Joliot must take into
i mint simultaneously and that together guarantee the reference for
li il lie says. All at the same time, Joliot must get the reactor to work;
ii nice his colleagues; interest the military, politicians, and industri-
l give the public a positive image of his activities; and, last but
I Ii .ist, understand what is going on with these neutrons that have
•me so important to the parties he has interested in their fate,
i .u e five types of activities that science studies needs to describe
Mi l l seeks to begin to understand in any sort of realistic way what
i ii scientific discipline is up to: instruments, colleagues, allies,
lit and finally, what I will call links or knots so as to avoid the his-
il luggage that comes with the phrase “conceptual content.” Each
In e live activities is as important as the others, and each feeds
Ii Into itself and into the other four: without allies, no graphite, and
no reactor; without colleagues, no favorable opinion from
II and thus no expedition to Norway; without a way of calculat-
tIn neutrons’ rate of reproduction, no assessment of the reactor,
i piool, and thus no colleagues convinced. In Figure 3.3 I have
il Ihe five different loops that science studies needs to consider
I 1 lo 1econstruct the circulation of scientific facts.
Mobilization o f the World
III I loop one has lo follow can be called the mobilization of the
il we uiulei stand by this very genetal expression all the means
I I h nonluimans ate ptogtessively loaded into discourse, as we
lO O
3
Alliances
(allies)
2
Autonomization
(colleagues)
(instruments)
Figure 3.3 Once we abandon the core/context model, it is possible to deploy itl
alternative one. Five loops have to be taken into account simultaneously for mi
realistic rendering of science; in this model, the conceptual element (links tit
knots) is still in the middle, but it is situated more like a central knot tying the loti
other loops than like a stone surrounded by a context.
saw in Chapter 2. It is a matter of moving toward the world, mal In

it mobile, bringing it to the site of controversy, keeping it engage 11
and making it available for arguments. In certain disciplines, such it
Joliot’s nuclear physics, this expression primarily designates the /
struments and major equipment that, at least since World War II, It 1
made up the history of Big Science. For many other discipline I
will also designate the expeditions sent around the world over the |M
three or four centuries to bring back plants, animals, trophies, 11
cartographical observations. We saw an example of this in Chaple I
with the soil of the Amazon forest becoming more and moie meibl
and beginning a long voyage, through a series of transfoi mal Ithi
toward the University of Paris. For still other disciplines the wt I
"mobilization” will mean neither instruments, equipment, noi t pi I
tions, but surveys, the questionnaires that have galheiod mloimill
about the stale of a society or an economy.
101
Whatever the kinds of mediations put to work, this loop is con-

11lied with doing practically what Kant called a Copernican Revolu-
it m, though he hardly realized how very practical was the activity des-
ii.ited by this grandiose expression: instead of moving around the
lijects, scientists make the objects move around them. Our friends
In pedologists were lost in the middle of an indecipherable land-
ipe (see Figure 2.7); once safely back in Manaus they had all the
ilological horizons mapped out and could now master at a glance
i lorest that had previously dominated them. As can be seen in the
1 mlispiece of the book by Mercator, the sixteenth-century geogra-
1111 who first used the term atlas, the demiurgic task of Atlas—that of
ii mg the world on his shoulders—has now been transformed into
m .itlas” and requires no more heroic force than that of turning the
1 i s of a beautifully printed book that the cartographer holds in
li.mds.
lins first loop deals with expeditions and surveys, with instru
its and equipment, but also with the sites in which all the objects of
oild thus mobilized are assembled and contained. For instance,
1 m Paris alone, the galleries of the Museum d’Histoire Naturelle,
iiillections of the Musée de l’Homme, the maps of the Service
I M.iphique, the databases of the CNRS, the files of the police, and
i quipment of the physiology laboratories of the Collège de
1111* all these are so many crucial objects of study for those who
II lo understand the mediation through which humans, speaking to
mol her, increasingly speak truthfully about things. Thanks to a
mvey and new data, an economist formerly without resources
I 111 spitting out reliable statistics at the rate of thousands of col-
pci minute. An ecologist whom nobody used to take seriously
now intervene in a debate with beautiful satellite photographs
IIlow her, without budging from a Paris laboratory, to observe the
11111 ol the forest in Boa Vista. A doctor, accustomed to treating
I i ase by case at the operating table, now has access to tables of
loins based on hundreds of cases, provided by the hospital’s re-
11vice.
I want lo understand why these people begin to speak more au-
II illvely and with moie assurance, we have to follow this mobili-
11ol Ihe woild, thanks lo which things now present themselves in
nil 111it icndcis (hem immcdialcly useful in (lie arguments that
102
scientists have with their colleagues. Through this mobilization tin-

world is converted into arguments. To write the history of the firsi
loop is to write the history of the transformation of the world into ini
mutable and combinable mobiles*. In brief, it is the study of the wril
ing of the “great book of nature” in characters legible to scientists, 01
to put it another way, it is the study of the logistics that are so indis
pensable to the logics of science.
Autonomization
To convince someone, a scientist needs data (or more exactly sublata)

but also someone to convince! The aim of the historians of the second
part of the vascular system is to show how a researcher finds col
leagues. I call this second loop autonomization, because it concerns tin
way in which a discipline, a profession, a clique, or an “invisible col
lege*” becomes independent and forms its own criteria of evaluation
and relevance. We always forget that specialists are produced from
amateurs in the same way soldiers are made out of civilians. Thu y
have not always been scientists and researchers. It was necessary, wit It
great effort, to extract chemists from alchemists, economists from ju
rists, sociologists from philosophers; or to obtain the subtle mixtuu
that produce biochemists out of biologists and chemists, social ps
chologists out of psychologists and sociologists. The conflict of disi I
plines is not a brake on the development of science, but one of its mu
tors. The increase in the credibility of experiments, expeditions, uni
surveys presupposes a colleague capable of both criticizing and u
ing them. What would be the use of obtaining ten million colored im
ages from a satellite, if there were only two specialists in the woil
who could interpret them? An isolated specialist is a contradiction h
terms. No one can specialize without the concurrent autonomi/alli I
of a small group of peers. Even in the middle of the Amazon, mil
friends the soil scientists never stopped speaking in a virtual ami 11
colleagues with whom they were constantly arguing in absentia, n
the wooded landscape had been transformed into the wooden pun
ing of a conference room.
The analysis of scientific professions is certainly the easiest put
science studies and the one most easily undeislandable by suent I I
who are never short of gossip on this topic It deals with the liistoi
103
I sudations and learned societies, as well as those of small cliques,

1«nips, and clusters which form the seeds of all relationships among
M r archers. More generally, this analysis deals with the criteria by
Inch one can distinguish, in the course of history, between a scientist
nul .1virtuoso, an expert and an amateur, a central and a marginal re
lic her. How does one establish the values for a new profession, the
in liculous control over titles and over barriers to entry? How does
in impose a monopoly of competence, regulate the internal demog-
ipliy of a field, and find jobs for students and disciples? How does
in 1esolve the innumerable conflicts of competence between the pro
ion and its neighboring disciplines, between, say, botany and
ilology?
In addition to the history of professions and disciplines, the second
II includes the history of scientific institutions*. There must be orga-
11ions, resources, statutes, and regulations to keep the crowds of
III igues together. It isn’t possible, for instance, to think of French
11(1* without the Académie, the Institut, the grandes écoles, the
US lhe Bureau de Recherches Géologiques et Minières, and the
ni et Chaussées. The institutions are as necessary for the resolution
oui 1oversies as is the regular flow of data obtained in the first loop,
pi oblem for the practicing scientist is that the skills demanded for
i ciind activity are entirely different from those of the first. A
1I1tgist may be great at digging trenches and keeping worms in vats
In middle of the forest but utterly useless when it comes to writing
1 and talking to colleagues. And yet one has to do both. Circu-
1clerence does not stop with the data. It has to flow further and
luce* other colleagues as well. But things are even more compli-
I I(m scientists, because the circulation does not stop at this sec-
I nip either.
Alliances
n (imnents can be developed, no discipline can become autono-

no new institution can be founded without the third loop,
h I c ill alliantes. Groups that previously wouldn’t give each other
llini ol day may be cm oiled in the scientists’ controversies. The
n must be made intei ested in physics, industrialists in chemis-
111 in caitogiaphy, leatheis in educational theory, congiessmen
104
in political science. Without this labor of making people interested,

the other loops would be no more than armchair traveling; without
colleagues and without a world, the researcher won’t cost much but
won’t be worth much either. Immense groups, rich and well endowed,
must be mobilized for scientific work to develop on any scale, for ex
peditions to multiply and go farther afield, for institutions to grow, foi
professions to develop, for professorial chairs and other positions to
open up. The skills required for getting others interested are again dii
ferent from those necessary for setting up instruments and for pro
ducing colleagues. One may be very good at writing convincing tech
nical papers and terrible at persuading ministries that they cannot go
on without science. As in the case of Joliot, these tasks can even In
somewhat contradictory: his alliances bring in many strangers, liki
Dautry and his advisers, whereas the work of autonomization aims aI
limiting the discussion to his fellow physicists.
As we saw in the preceding section, it is not a question of histoi i
ans finding a contextual explanation for a scientific discipline, but ol
the scientists themselves placing the discipline in a context sufficiently
large and secure to enable it to exist and endure. It is not a matter ol
studying the impact of the economic base on the development of llit
scientific superstructure, but of finding out how, for example, an in
dustrialist could improve his business by investing in a solid-slat 1*
physics laboratory, or how a state geological service could expand
by attaching itself to a department of transportation. The alliant t
do not pervert the pure flow of scientific information but are wli tl
makes this blood flow much faster and with a much higher pul i»
rate. Depending on the circumstances, these alliances can take in mi
merable forms, but this enormous labor of persuasion and liaison I
never self-evident: there is no natural connection between a militai
man and a chemical molecule, between an industrialist and an fit <
tron; they do not encounter each other by following some natural In
clination. This inclination, this clinamen has to be created, the soi I (I
and material world has to be worked on to make these alliances ip
pear, in retrospect, inevitable. This presents an immense and passion
ately interesting history, probably the most important for umli 1
standing our own societies: the history of how new nonhumans li 1 e
become entangled in the existence of millions of new humans ( h*
Chapter 6).
105
Public Representation
I i n if the instruments were in place, if peers had been trained and

11.1iplined, if well-endowed institutions were ready to offer a home to
1Ins wonderful world of colleagues and collections, and if government,
ndnslry, army, social security, and education provided the sciences
llli wide support, there would still be a great deal of work to be done.
I Ins massive socialization of novel objects—atoms, fossils, bombs, ra-
111 statistics, theorems—into the collective, all this agitation, and all
In se controversies would present a terrible shock to people’s every-
II pi actice, would risk overturning the normal system of beliefs and
pinions. It would be astonishing if it were otherwise, for what is sci-
1111 lor if not to modify the associations of people and things? The
11111' scientists who had to travel the world to make it mobile, to con-
11111' colleagues and lay siege to ministers and boards of directors,
1 have to take care of their relations with another outside world of
I Ilians: reporters, pundits, and the man and woman in the street. I
ill IIns fourth loop public representation (if we can free this expression
in Ilie stigma associated with “PR”).
I nnlrary to what is often suggested by science warriors, this new
I hlc world is no more outside than the three previous ones: it sim-
lias other properties and brings people with other qualities and
mpi'lences into the fray. How have societies formed representations
1111science is; what is a people’s spontaneous epistemology? How
inli Ilust do they place in science? How can this confidence be
1 lin'd in different periods and for different disciplines? How, for
I met*, was Isaac Newton’s theory received in France? How was
inks Darwin’s theory greeted by English clerics? How was
loi Ism accepted by French trade-unionists during the Great War?
did economics little by little become one of the stock topics of
Ilk Ians? How was psychoanalysis gradually absorbed into daily
liological discussions? How are DNA fingerprinting specialists
In on llie witness stand?
I ll t all the others, this loop requires from scientists a completely
I 1ml sol of skills, unrelated to those of the other loops, and yet de-
min ml loi them all. One may be very good at convincing govern-
nt mliiislcis but completely unable to field questions on a talk
Mow could one pioduce a discipline that would modify every-
106
one’s opinion, and nonetheless expect passive acceptance by all? II

primatologists, ethologists, and geneticists produce entirely differed I
genealogies for sex roles, aggression, and maternal love, how can they
be surprised if large sections of the public take umbrage? Every as
tronomer recalculating the numbers of planets turning around stai n
knows that everything will change if throngs of other life forms aie
suddenly added to the definition of the human collective. This fouilh
loop is all the more important because the three others largely de
pend on it. A major part of advanced research in molecular biology in
France, for example, depends on a private charity’s annual telethon (o
combat muscular dystrophy. Every argument for or against genetic di
terminism will feed back into this funding. Our sensitivity to the puli
lie representation of science must be all the greater because informa
tion does not simply flow from the three other loops to the fourth, 11
also makes up a lot of the presuppositions of scientists themselvi
about their objects of study. Thus, far from being a marginal append
age of science, this loop too is part and parcel of the fabric of facts ami
cannot be left to educational theorists and students of media.
Links and Knots
To reach the fifth loop is not to reach the scientific content at last, a 11
the four others were simply conditions for its existence. From the In I
circle on, we have not departed for one moment from the corn si 1
scientific intelligence at work. As is clear from Figure 3.3, we It t
not been going endlessly around the mulberry bush and evading lit
“conceptual content,” as science warriors are wont to say. We It 1
simply followed the veins and arteries and arrive now, inevitably il
the pumping heart. Why does this fifth loop, which I call links a
knots so as to avoid for the moment the word “concept,” have the 11jm
tation of being much harder to study than the rest? Well, it /s mu
harder. I don’t pretend to crack it now, but simply to redefine its lo|u
ogy, so to speak, one of the reasons for its solidity.
This hardness is not that of a pit inside the soft flesh of a pe.u h II
that of a very tight knot at the center of a net. It is hard because II h
to hold so many heterogeneous resources together. Ol com si 11
heart is important for understanding the circulaloiy system ol Ilu li
107
mm body, but Harvey certainly did not make his famous discovery by
onsidering the heart on one side and the blood vessels on another.
I lu1same is true for science studies. If one takes the content on one
lili* and the context on the other, the flow of science becomes incom-
| n licnsible, and so does the source of its oxygen and nutriment, as
( II as their means of entering the bloodstream. What would happen
Ihere were no fifth loop? The other four would die off at once. The
•»i Id would stop being mobilizable; disgruntled colleagues would flee
ill directions; allies would lose interest; and so would the general
ihlic, after expressing either its shock or its indifference. But this
ilh would ensue just as quickly if any of the other four loops were
Hnil.
Ilns point is always one of the first casualties of the science wars. Of
in sc Joliot “has thoughts”; of course he “has concepts”; of course his
in i* has some content. But when science studies seeks to under-
iih 11he centrality of the conceptual content of science, it tries first to
Im what periphery this content plays the role of the center, of what
n md arteries it is the pumping heart, of what net it is the knot, of
pathways it is the intersection, of what commerce it is the clear-
house. If we imagine Joliot to be now circulating along the loop
I m ikes up the center of Figure 3.3, we understand why he tries so
IIy and so earnestly to find a way of keeping together, all at the
Iline, his instruments, his colleagues, the officials and industrial-
III has gotten involved, and the public.
S Joliot can succeed only by understanding the chain reaction—
In held better do it quick, before Szilard does it first, before the
111ms a1rive in Paris, before the two hundred liters of heavy water
II IIn. Norwegians run out, and before Halban and Kowarski are
d In lice, denounced by their neighbors as foreigners. Yes, there
tin 01 y, yes, the calculation of the cross-section made at night by
II 11 will make all the difference, yes, the knowledge they have
1 <d about the neutrons will put them on the brink of a decisive
1I lyi beloie the debacle of May 1940 brings an end to it all. But
id ill the 1est for this calculation to be the theory of something.
I I Indeed a conceptual core, but this is not defined by the pre-
j liions located at the furthest remove from the others; on the
111 ll is what keeps them all together, what strengthens their co-
PAN D O RA ’S HOPE
108
hesion, what accelerates their circulation. Science warriors defend the

conceptual content of science with the wrong sort of metaphor. They
want it to be like an Idea floating in Heaven freed from the pollution of
this base world; science studies wants to understand it as more like
the heart beating at the center of a rich system of blood vessels, or
better yet, like the thousands of alveoli in the lungs which allow the
blood to be reoxygenated.
The difference in metaphors is not trivial. What science studies
most wants to be able to explain is the relation between the size of this
fifth loop and the four others. A concept does not become scientific
because it is farther removed from the rest of what it holds, but be
cause it is more intensely connected to a much larger repertoire of re
sources. Goat trails do not need turnpikes. Elephant hearts are a loi
bigger than those of mice. The same goes for the conceptual contenl
of a science: hard disciplines need bigger and harder concepts than
soft ones, not because they are more remote from the rest of the world
of data, colleagues, allies, and spectators—the four other loops—bui
because the world that they churn, steer, move, and connect is vastly
bigger.
The content of a science is not something contained; it is itself a
container. Indeed, if etymology is of any help, its concepts, its Begnjje
(from greifen, to seize or to grasp), are what hold a collective tightly to
gether. Technical contents are not astounding mysteries put in the way
of those who study science by the gods to humble them by reminding
them of the existence of another world, a world that escapes from hli
tory; nor are they provided for the amusement of epistemologists lo
enable them to look down on all those who are ignorant of science
They are part of this world. They only grow here, in our world, he
cause they are what makes it up by linking together more and molt»
elements in bigger and bigger collectives (as we will see in Chapter ft )
For this point to be something more than an empty declaration ol In
tent, I should obviously get much closer to the technical content than I
have in my sketch of Joliot. But I cannot do this before substituting In
the next chapters, a new definition of what it is for a human to di il
with a nonhuman for the old subject-object dichotomy. In the mean
time, I can simply place concepts, links and knots, in a dilferent po I
tion so that when we learn about the esoteric content of a science wc»
immediately look for the four other loops that give il meaning
109
The Enucleation of Society out of the Collective

||ow can I convince my scientific friends that by studying the
iscularization of scientific facts we gain in realism and science gains
In hardness? Perhaps this is so commonsensical that it appears hereti-
il for a little while at least. The more connected a science the stur-
Iit 1 it is; how could that be any simpler? And yet, for political reasons
llnl will become clear in Chapter 7, epistemologists have trans-
I a med this very simple fact of life into a complete mystery. For episte-
inologists the scientific disciplines have to become solid and reli-
ibli without being connected through any sort of vessels to the rest
I Iheir world. The heart will be required to pump in and out, but
Imic will be no input and no output, no body, no lungs, and no vascu-
II system. Science warriors deal with nothing but an empty heart
i l ’hlly lit on an operating table; science studies treats a bloody,
ill lobbing, tangled mess, the entire vascularization of the collec-
I i And the first group makes fun of the second because its mem-
1 look messy and have blood on their white coats, and accuses them
I I 'imring the heart of science! Indeed, how can we talk to one an-
IIk 1?
Nil, as at the end of Chapter 2, we also have to account for how the
i|il msible, irrealist model can be extracted from the realist one pro-
1d by science studies. A new paradigm should always be able to un-
I I md the one it claims to replace. As we saw in Figure 2.24, the no-
II ol a yawning gap between words and world was obtained by
I Ing all the mediations and interrogating only the two extremes
1 lug each other like two distant bookends, thus artificially creating
pioblem” of reference. The mutilation of the circulatory system
1Unee is even more gruesome (see Figure 3.4). If one fails to pay
i illcnlion to the entirety of scientific endeavor (Figure 3.4a) one
in gi I Ihe impression that there exists on the one hand a series of
nllngencies (the corona) and on the other hand, at the center, a con-
III il 1unlent that counts most (Figure 3-4b). Here, it will take only
II 'blest lapse of attention, the slightest bit of carelessness, and
il III be it! The rich and fragile webs will be cut, distanced from
llilngs they connect and assemble. Another tiny slip, and the nu
ll ul surntihc content" will be separated from what will become,
mill isl a cunlmgent lusUmcal “context” (Figure 3.4c). We will
PANDO RA’S HOPE
HO
Figure 3.4 As in Figure 2.24, it is possible to extract the canonical model from the
new one by erasing key mediations. If the conceptual dimension—the center cirtli
in (a)—is excised from the other four, it will be transformed into a core (b), whilt
the four other loops, now disconnected, when reconnected will form a sort of con
text of no relevance for defining the inner core of science (c).
have shifted from one branch of geometry to another, from knots to

surfaces.
Only with inattention and the careless use of different analytical
scalpels can one get the model of content vs. context from the hetei o
geneous and multiple labor of scientists. The whole of this laboi
then becomes obscure, because one no longer sees the essential con
necting point, which is all the diverse elements that the theories ami
concepts theorize and bring together. Instead of the continuous ami
curved path of translations, one runs into an iron curtain separating
the sciences from “extrascientific” factors, just as a long gray wall ol
concrete used to cut off the circulation through Berlin’s delicate sys
tem of lanes, tramlines, and neighborhoods. Epistemologists, dl
couraged when faced with these objects so hard and so durable (It il
they seemed to come from another world, could only send them It
a Platonic Heaven and connect them to one another in an enlutl
phantasmagorical history, which is sometimes called the “concept ml
history of science” despite the fact that there is no longer anything III
torical about it and thus nothing scientific about it either (see Ch.iplt 1
5). The damage has been done: long trajectories of solid ideas ami
principles now appear to hover above a contingent history 111 c 1
many foreign bodies.
The worst is yet to come: historians, economists, sociologists «
customed to studying all the aspects I have listed, become disc0111apt d
111
11 .ill this strangeness bobbing over their heads and leave the concep-
lu 11 core of the sciences to the scientists and philosophers, modestly
mi tenting themselves with wading through “social factors” and “so-
I il dimensions.” This modesty would do them honor if, in abandon
ing (lie study of scientific and technical content, they did not also ren-
li i incomprehensible the very social existence that they claim to study
mil to which they claim to restrict themselves. Indeed, what is most
nous about this entirely artificial separation between the nucleus
imI Ilie cell, of theories from what they theorize, is not that it enables
nli llectual historians to postulate this ahistorical, endless unfolding
I purely” scientific ideas. The real danger lies in the corresponding
lu I among social scientists that by lining up previously “enucleated”
nlexls it is possible to account for the existence of societies without
m* to deal with science and technology.
In place of a collective of humans and nonhumans we now have two
it 11Id series of artifacts that never intersect: ideas on the one hand
I wniety* on the other. The first series, which results in the dreams
pistemology and the knee-jerk defensiveness of science warriors, is
iply annoying and puerile; the second, which results in the illusion of
i ml world, is far more damaging, at least for those like me who try
pi ul ice a realistic philosophy. The whole of the modern world is
i li impossible to understand by this invention of an enucleated so
il i onlcxt.
111 us suppose, for example, that a historian is studying the military
I Ions and programs of France during World War II. As we have
il opeialions of translation made Joliot’s laboratory indispensable
llu 11inducting of French military affairs. Now, Joliot himself could
I «l lus reactor to start except by discovering a new radioactive ele-
111 plu Ionium, which kicks off the chain reaction far more easily.
I hI ms of military affairs, following the series of translations, must
II ilily become interested in the history of plutonium; more pre-
I IIns inevitability is a function of Joliot’s work and his success.
li uenlisls’ activities over the last three or four hundred years,
long can one study a military man before finding oneself in a lab-
I il I At most a quarter of an hour if one studies postwar science,
I mi iybe an hour if one is dealing with the previous century
till 19Hz, Alder 1997) Consequently, to write military history
h ul lool mg at the laboiatoi ies that make up this history is an ab
112
surdity. It is not a matter of disciplinary principles, knowing whether

or not one has the right to approach history without paying attention
to science and technology; it is a question oifact —whether or not the
players studied by historians mixed their lives and their feelings with
nonhumans mobilized by laboratories and scientific professions. If the
answer is yes, as it most certainly is in this example, it is unthinkable
not to put back into the game the plutonium that Joliot and the mili
tary used, in their different ways, to make war and peace.
Now we can begin to measure the huge misunderstanding of those
who say that science studies provides “a social explanation of science.”
Yes indeed, it offers an explanation, but of the artifactual origin of a use
less concept o f society* that has been obtained by the enucleation of
scientific disciplines out of their collective existence. What remains af
ter this excision is, on the one hand, a society of humans among them
selves and, on the other, a conceptual core. It would be even more ab
surd to say that science studies seeks to reconcile a social explanation
with a conceptual explanation, if we understand these as the two dis
tinct kinds of explanation that keep the parallel series of artifacts from
ever intersecting. To retie two artifacts together makes for a third aril
fact, not for a solution! From Figure 3.4 it should be obvious that sim
ply grafting a large corona of social factors onto the inner core of
science, as in 3.4c, will not return us to the rich vascularization of
scientific facts circulating through the five loops of 3.4a. The mcla
phors, the paradigms, the methods are entirely different and wholly
incompatible. No matter how strange it may seem to science warrioi
and, yes, to most social scientists, in order to regain a sense of realism lit
the study of science, one has to abandon the notion of a society altogethi r
No wonder: as we will see in Chapters 7 and 8, this conception of soi I
ety was invented for reasons entirely unsuited to providing an expin
nation of anything.
C H A P T E R F O U R
From Fabrication to Reality

Pasteur and His Lactic Acid Ferment
i‘ have now made two moves that should begin to modify for good
lIn settlement* laid out in the first chapter. The notion of a world “out
Ihue” to which a mind-in-a-vat tries to get access by establishing
•me safe correspondence between words and states of affairs should
in be seen for what it is: a very unrealistic position for science, so
Hied, so cramped that it can only be explained by powerful political
i ilives (which we will examine later in this book). In Chapter 2 we
pm to understand that reference is not something that is added to
mis, but that it is a circulating phenomenon, whose deambulation—
hoiiow, once again, William James’s term—should not be inter-
ii| •11(I by any saltation if we want words to refer to the things progres-
I 1ly packed into them. Instead of the vertical abyss between words
ml vvoild, above which the perilous footbridge of correspondence
mill hang, we now have a sturdy and thick layering of transverse
lilts lluough which masses of transformations circulate.
I In n in Chapter 3 we realized what an impossible double bind the
I 1Itlcment imposed on the scientist: “Be entirely cut off from the
I III ol society, psychology, ideology, people”; and at the same time,
ihsolulely, not relatively, sure of the laws of the world outside.”
tin I this contradictory injunction, we realized that the only rea-
II il ill' Ihe only realistic way for a mind to speak truthfully about the
1hi Is lo rvumnect through as many relations and vessels as possible
II hill (lie 1ich vascularization that makes science flow—and of course
I mi ms that theie is no longer any “mind” (Hutchins 1995). The
it it I liions a scientific discipline has, the moie chance there is for
m
PANDO RA’S HOPE
114
accuracy to circulate through its many vessels. Instead of the impossi

ble task of freeing science from society, we now have a more manage
able one: that of tying the discipline as much as possible to the rest of
the collective.
And yet nothing is solved. We have simply begun to extract our
selves from the most blatant defects of the old settlement. We have not
yet found a better one. More reality has to be taken into account if we
want to continue. In Chapters 2 and 3 we left the world, so to speak,
intact. Our friends the soil scientists, Joliot and his colleagues, were
doing many things, but the soil itself, the neutrons themselves were
behaving as if they had been there all along, waiting to be metamor
phosed into so many stakes, diagrams, maps, arguments and broughl
to bear on the realm of human discourse. This is obviously not enough
to explain how we can talk truthfully about a state of affairs. It does
not matter how much we modify the notion of reference, if we are nol
also able to modify our understanding of what the entities of the
world do when they come into contact with the scientific community
and begin to be socialized into the collective*.
From the very beginning of science studies the solution has been to
use the words “construction” and “fabrication.” To take account ol
this transformation of the world effected by scientists, we have spoken
of “the construction of facts,” “the fabrication of neutrons,” and othei
similar expressions which throw science warriors into fits and which
they now fling back at us. I would be the first to admit that there ai t
many problems with this way of accounting for action. First, although
“construct” and “ fabricate” are terms for technical activities, it hap
pens that, under the pens of sociologists and philosophers working
within the narrow space that the modern settlement allowed them
technology has been rendered almost as obscure as science (as we will
see in Chapter 6). Second, this account implies that the initiative of ac
tion always comes from the human sphere, the world itself doing hi t U
more than offering a sort of playground for human ingenuity (in Ilit
discussion of the “factish” in Chapter 9 I will seek to counteract this)
Third, speaking of construction implies a zero-sum game, with a hxul
list of ingredients; fabrication merely combines them in other way
Finally, and this is much more worrying, the old settlement has kid
napped the notions of construction and fabrication, turning them Inin
FROM FABRICATION TO REALITY
115
weapons in a polarized battle against truth and reality. All too often
Ilie implication is that if something is fabricated it is false; likewise, if
il is constructed it must also be deconstructible.
These are the main reasons why the more we in science studies
liowed the constructivist character of science, the deeper was the
misunderstanding between us and our scientific friends. It is as if we
ci e undermining science’s claim to truth. Yes, we were undermining
mnething, but something else altogether. Although we were a bit slow
In icalize it, we were shaking the foundations of the very idiom of con-
Ina tion andfabrication we had earlier taken for granted—and also, as
i*will see in Chapter 9, the basic notions of action and creation. Con-
lmetion and fabrication, even more than reference and “conceptual
mitent,” have to be reconfigured totally, like all the other concepts
Ili it have been handed down to us, if we really wish to understand sci-
»me in action. This reconfiguration is what I hope to accomplish in
IIns chapter by visiting yet another empirical site, this time Louis Pas-
li in \ laboratory. Let us follow in some detail the “Mémoire sur la fer-
iiu illation appelée lactique,”1which historians of science consider to
In one of Pasteur’s most important papers.
I lie text is ideal for our purposes since it is structured around two
omlnned dramas. The first one modifies the status of a nonhuman
nid Ihat of a human. It converts a nonentity, the Cinderella of chemi-
il llieory, into a glorious and heroic character. In parallel, Pasteur’s
Iilinon, the Prince Charming, triumphs against all odds and reverses
I li lug’s theory: “The stone which the builders refused has become the
h iici stone.” And then there is a second drama, a reflexive one, a mys-
I 1 Ilut appears only at the end: who is constructing the facts, who is
III 11 1mg the story, who is pulling the strings? Is it the scientist’s prej-
1III i's, or is it the nonhumans? Thus to the ontological drama is
1Mi il an epistemological one. We will be able to see, using Pasteur’s
II words, how a scientist solves for himself and for us two of the ba-
pioblems of science studies. First let us turn to the uplifting story
I I imlciella the-yeast.
1 I 111inlly 11.instated into English by J. B. Conant in the Harvard Case Studies in Ex-
I1111 ni il Suence (Conant 1957). I have completed and modified the translation in
1 il pi m s I he 1 tenth text tan be found in volume 2 of Pasteur’s complete works.
IIn b 111giound see (icison (1974)
ll 6
The First Drama: From Attributes to Substance

In 1856, some time after the fermentation of brewer’s yeast became his
primary interest, Pasteur related the discovery of a yeast peculiar to
lactic acid. Today lactic fermentation is no longer an object of discus
sion, and dairies, creameries, and cheese manufacturers the world
over can order by mail as much yeast as they need. But one has only to
“place oneself in the conditions of the period” to appreciate the origi
nality of Pasteur’s report. In the middle of the nineteenth century, in
scientific circles where Liebig’s chemistry held sway, the claim that a
specific microorganism could explain fermentation amounted to a
step backward, since it was only by ridding itself of obscure vitalist ex
planations that chemistry had won its laurels. Fermentation had been
explained in a purely chemical way, without the intervention of any
living thing, by an appeal to the degradation of inert substances. In
any case, specialists in lactic fermentation had never seen any micro
organisms associated with the transformation of sugar.
At the beginning of Pasteur’s paper, lactic acid fermentation has no
clear-cut, isolable cause. If a yeast is involved it is nothing but an al
most invisible by-product of a purely chemical mechanism of fermen
tation, or even worse, it is an unwelcome impurity that would hindoi
and spoil the fermentation. By the end of the paper, however, the yeasl
has become a full-blown entity in its own right, integrated into a class
of similar phenomena; it has become the sole cause of fermentation
In the course of a single paragraph Pasteur takes the yeast through till
entire transformation:
Under the microscope, when one is not forewarned, it is hardly pos si
ble to distinguish it from casein, disaggregated gluten, etc.; in such a
way that nothing indicates that it is a separate material or that it
was produced during the fermentation. Its apparent weight always
remains very little as compared to that of the nitrogenous matei nil
originally necessary for the carrying out of the process. Finally, vei y
often it is so mixed with the mass of casein and chalk that there would
be no reason to suspect its existence. (§7)
And yet Pasteur concludes this paragraph with this brave and sin
prising sentence: “It is this [the yeast], nevertheless, that plays the pi In
cipal role.” The abrupt transformation is not only that ol the yeasl «
117
11 acted from nothingness to become everything, it is also that of the

h i nee Charming, Pasteur himself. At the beginning of the paper, his
opinion counts for nothing against Liebig’s and Berzelius’s powerful
ihoories. At the end of the paper, Pasteur triumphs over his enemies
md his view wins the day, defeating the chemical account of the fer
mentation. He begins:
The facts [that make the cause of lactic acid fermentation so obscure]
Ilien seem veryfavorable to the ideas of Liebig or to those of Berzelius
.. These opinions gain more credit daily... These works all agree in re
acting the idea of some sort of influence from organization and life as
a cause of the phenomena that we are considering. (§5)
And again he concludes the paragraph with a defiant sentence that

li Meets the weight of previous arguments: “I have been led to an en-
k I)i different point of view.” But to accompany this elevation of
( liulcrella and this triumph of Prince Charming, another, more wide-
1mging transformation is necessary. The capacities of the natural
Iu Id are modified between the beginning and the end of the story. At
Ilu start of the paper the reader lives in a world in which the relation
I • Iween organic matter and ferments is that of contact and decay:
In the eyes of [Liebig] a ferment is an excessively alterable substance

Ilui decomposes and thereby excites fermentation in consequence of
Ils alteration which communicates a disintegrating disturbance to
Ilie molecular group of the fermentable matter. According to Liebig,
ne h is the primary cause of all fermentations and the origin of most
iniilagious diseases. Berzelius believes that the chemical act of fer-
im nlalion is to be referred to the action of contact (§5)
Al Ihe end the reader lives in a world in which a ferment is as active

my other already identified life form, so much so that it now feeds
II the oiganic material, which instead of being its cause, has become
I loodV
V lioevei judges impartially the results of this work and that which I
lull slmilly publish will recognize with me that fermentation ap-
pi us to be coirelative to life and to the organization of globules, and
>/ lo (lieu death and putrefaction, no more than fermentation is a
I In nomenon due lo contact m which the transformation of sugar
PA N D O RA ’S HOPE
118
takes place in the presence of the ferment without giving up any
thing to it or taking anything from it.(§22)
Let us now follow the main nonhuman character of the story to
see through how many different ontological stages this entity is forced
to pass before becoming something like a well-recognized substance
How does a scientist explain in his own words this emergence of a new
actor out of other entities that he has to destroy, redistribute, and reas
semble? What happens to this actant x that will soon be named lactic
acid fermentation yeast? Like the forest-savannah limit in Chapter 2,
the new entity is first a circulating object undergoing trials and sub
mitted to an extraordinary series of transformations. At the beginning
its very existence is denied:
Until now minute researches have been unable to discover the develop
ment of organized beings. Observers who have recognized some of
those beings have at the same time established that they were acci
dental and spoiled the process. (§4)
Then Pasteur’s main experiment allows “a forewarned observer" to

detect such an organized being. But this object x is stripped of all
its essential qualities, which are redistributed among elementary sens»
data:
If one carefully examines an ordinary lactic fermentation, there ai e
cases where one can find, on top of the deposit of the chalk and ni
trogenous material, spots of a gray substance which sometimes form a
layer [formant quelquefois zone] on the surface of the deposit. At othei
times, this substance is found adhering to the upper sides of the ves
sel, where it has been carried by the movement of the gases. (§7)
When it solidifies [prise en masse] it looks exactly like ordinary pressed

or drained yeast. It is slightly viscous, and gray in color. Under the ml
croscope, it appears to be formed of little globules or very shorl seg
mented filaments, isolated or in clusters, which form irregular fl.il 1
resembling those of certain amorphous precipitates. (§10)
It would be hard for something to have less existence than that I II I

not an object but a cloud of transient perceptions, not yet the pud)
cates of a coherent substance. In Pasteur’s philosophy o( suenu lit
119
phenomena precede what they are the phenomena of. Something else
is necessary to grant* an essence*to make it into an actor: the series of
I iboratoiy trialsdhrougli which the object z proves its mettle. In the
next paragraph Pasteur turns it into what I have called elsewhere “a
n.une of action*”: we do not know what it is, but we know what it does
Iloin the trials conducted in the lab. A series of performances* pre-
(des the definition of the competence* that will later be made the sole
iuse of these very performances.
About fifty to one hundred grams of sugar are then dissolved in each
liter, some chalk is added, and a trace of the gray material I have just
mentioned from a good, ordinary lactic fermentation is sprinkled in
. On the very next day a lively and regularfermentation is manifest.
I lie liquid, originally very limpid, becomes turbid; little by little the
ili.ilk disappears, while at the same time a deposit is formed that
■Iows continuously and progressively with the solution of the chalk.
I lie gas that is evolved is pure carbonic acid, or a mixture in variable
Iimportions of carbonic acid and hydrogen. After the chalk has disap-
><at cd, if the liquid is evaporated, an abundant crystallization of lac-
I lie of limeforms overnight, and the mother liquor contains variable
i|ii.mlities of the butyrate of this base. If the proportions of chalk and
ug.ir are correct, the lactate crystallizes in a voluminous mass right in
Ilie liquid during the course of the operation. Sometimes the liquid
monies very viscous. In a word, we have before our eyes a clearly
uuacterized lactic fermentation, with all the accidents and the usual
oinplications of this phenomenon whose external manifestations
in well known to chemists. (§8)
\ i do not yet know what it is, but we do know that it can be sprin-
il that it triggers fermentation, that it renders a liquid turbid, that
m il es the chalk disappear, that it forms a deposit, that it generates
lli.il it forms crystals, that it becomes viscous (Hacking 1983). As
now it is a list of entries recorded in the laboratory notebook,
bia disjecta which do not yet pertain to one entity—properties
I lug lor the substance they belong to. At this point in the text, the
III is so fragile, its envelope* so indeterminate, that Pasteur notes
ill in pi ise its ability to travel:
II 1an be collected and transported for great distances without los-
11 Ils a'livily, which is weakened only when the material is dried
120
or when it is boiled in water. Very little of this yeast is necessary

to transform a considerable weight of sugar. These fermentations
should preferably be carried on so that the material is protected from
the air, so that they will not be hindered by vegetation or foreign
infusoria. (§10)
Maybe shaking the flask will make the phenomenon disappear,
maybe exposing it to the air will destroy it. Before the entity is safely
underwritten by a fixed ontological substance, Pasteur has to add pre
cautions that he will soon find useless. Not yet knowing what it is, he
has to fumble, investigating all sides of the vague boundaries he has
sketched around the entity in order to determine its precise contours
But how can he increase the ontological status of this entity, how
can he transform these fragile, uncertain boundaries into a sturdy en
velope, how can he move from this “name of action” to the “name of a
thing”? If it acts so much, must the entity be an actor? Not necessarily
Something more is needed to turn this fragile candidate into a full
blown actor which will be designated as the origin of those actions
another act is necessary to conjure up the substrate of these predi
cates, to define a competence that will then be “expressed” or “manI
fested” through so many performances in laboratory trials. In llu
main section of the paper, Pasteur does not hesitate. He uses evei y
thing at hand to stabilize the noumenal substrate of this en til y
granting it an activity similar to that of brewer’s yeast. Borrowing llu
metaphor of growing plants allows him to evoke the processes of do
mestication and cultivation, the firmly established ontological stain
of plants, as a way of giving shape to his aspiring actor:
Here we find all the general characteristics of brewer’s yeast, and these
substances probably have organic structures that, in a naluial
classification, place them in neighboring species or in two connected
families. (§11)
There is another characteristic that permits one to compare this new

ferment with brewer’s yeast: if brewer’s yeast instead of the lac Ik
ferment is sown in limpid, sugared, albuminous liquid, brewer’s yeast
will develop, and with it, alcoholic fermentation, even though Ihe*
other conditions of the operation remain unchanged. One should
not conclude from this that the chemical composition of the two
yeasts is identical any more than that the chemical composition oi
two plant \ is the same because they giew in the same soil (tjn)
121
What was a nonentity in §7 has become so well established in §11

(hat it has a name and a place in the most precise and most venerable
o! all branches of natural history taxonomy No sooner has Pasteur
shifted the origin of all the actions to the yeast, which thereby be-
1(unes a full-blown independent entity than he uses it as a stable ele
ment to redefine all the former practices: we did not know what we
\ ere doing before, but now we do:
All the chemists will be surprised at the rapidity and regularity of

lactic fermentation under the conditions that I have specified, that is,
when the lacticferment develops alone; it is often more rapid than the
alcoholic fermentation of the same amount of material. Lactic fer
mentation as it is ordinarily carried out takes much longer. This can
easily be understood. The gluten, the casein, the fibrin, the mem-
I>1anes, the tissues that are used contain an enormous amount of use
less matter. More often than not these become a nutrient for the lactic
lei ment only after putrefaction—alteration by contact with plant or.
immalcules—that has rendered the elements soluble and assimila-
l)le. (§12)
A slow and uncertain practice with an obscure explanation becomes

h Ime k and comprehensible set of new methods mastered by Pasteur:
ill 1long, without knowing it, cheese manufacturers had been cultivat-
np microorganisms in a medium that provides food for the ferment,
mkI l hat itself may be varied so as to vary the adaptation to an envi-
IIment of multiple ferments in competition. What was the primary
III e of a useless by-product has been transformed into food for its
11 etjuence!
lining even further, Pasteur turns this newly shaped entity into one
Ingular case” within a whole class of phenomena. The “general d r
ill lances” of this widespread phenomenon, fermentation, can now
tit lined.I
I >nt ol the essential conditions for goodfermentations is the purity of

llu U1ment, its homogeneity, its free development without any hindrance
11111willi the help of a nutrient well adapted to its individual nature.
In IIlls 1espect, it is important to realize that the circumstances of neu-
li till y ol alkalinity, of acidity, or of the chemical composition of the
II |ulds play an impoitanl pait in the predominant growth of such
md ml» a leiment, because the life ol each does not adapt itself to
llu une degiee to chlleienl stales ol the tnvironmcnt (^17)
PANDORA’S HOPE
122
By drawing on several seemingly incompatible philosophies of sci

ence, Pasteur provides a fresh solution to what is still a subject of
much controversy in epistemology, namely, how a new entity can
emerge out of an old one. It is possible to go from a nonexistent entity
to a generic class by passing through stages in which the entity is made
of floating sense data, taken as a name of action, and then, finally,
turned into a plantlike and organized being with a place within a well
established taxonomy. The circulation of reference does not take us, as
in Chapters 2 and 3, from one site of research to the next, from one
type of trace to the next, but from one ontological status to the next. Here
it is no longer just the human who transports information througli
transformation, but the nonhuman as well, surreptitiously changing
from barely existing attributes into a full-blown substance.
From Fabrication of Facts to Events

How does Pasteur’s own account of the first drama of his text modify
the commonsense understanding of fabrication? Let us say that in his
laboratory in Lille Pasteur is designing an actor. How does he do this?
One now traditional way to account for this feat is to say that Pastern
designs trials* for the actor* to show its mettle. Why is an act01
defined through trials? Because there is no other way to define an at
tor but through its action, and there is no other way to define an action
but by asking what other actors are modified, transformed, perturbed
or created by the character that is the focus of attention. This is a pi ag
matist tenet, which we can extend to (a) the thing itself, soon to hi
called a “ferment”; (b) the story told by Pasteur to his colleagues at tin
Academy of Science; and (c) the reactions of Pasteur’s interlocutoi s lo
what is so far only a story found in a written text. Pasteur is engaged it
once in three trials that should be first distinguished and then alignai
with one another, according to the notion of circulating refeienu
with which we are now familiar.
First, in the story told by Pasteur, there are characters whose com
petence* is defined by the performances* they undertake: the nenil
invisible Cinderella becomes, to the applause of the reader, the lu 10
who triumphs and becomes the essential cause of lactic fermentai ion
of which it was, at first, a useless by product. Second, Pasleui, m III
laboratory, is busy staging a new ailifiuul woild m winch to 11y out
123
lliis new actor. He does not know what the essence of a ferment is.
Pasteur is a good pragmatist: for him essence is existence and exis
tence is action. What is this mysterious candidate, the ferment, up to?
Most of an experimenter’s ingenuity goes into designing devious plots
ind careful staging that make an actant* participate in new and unex
pected situations that will actively define it. The first trial is a story: it
pertains to language and is similar to any trial in fairy tales or myths.
I lie second is a situation: it pertains to nonverbal, nonlinguistic com
ponents (glassware, yeasts, Pasteur, laboratory assistants). Or does it?
The third trial is designed to answer this very question. Pasteur under
goes this new trial when he tells his story of the Cinderella that tri
umphs against all odds and of the Prince Charming that defeats the
li.igon of chemical theory—when he has a shorter version of his pa-
pi i i ead at a meeting of the Academy on 30 November 1857. Pasteur is
now trying to convince the Academicians that his story is not a story,
lull that it has occurred independently of his wishes and imaginative
ilulily. To be sure, the laboratory setting is artificial and manmade,
ni Pasteur must establish that the competence of the ferment is its
impotence, in no way dependent on his cleverness in inventing a trial
lli 11 allows it to reveal itself. What happens if Pasteur wins this new
11111<1) trial? A new competence will now be added to his definition,
i lour will be the person who has shown, to everyone’s satisfaction,
111 yeast is a living organism, just as the second trial added a new
tmpotence to this other actant, the ferment: namely that it can trig-
1 i specific lactic fermentation. What happens if Pasteur fails? Well,
etond trial will have been a waste. Pasteur will have entertained
pieis with the tale of Cinderella-the-ferment, an amusing story
lu sure, but one which will have involved his own expectations
I 1.11lier prowess only. Nothing new will have been conveyed by
li in’s words at the Academy to modify what his colleagues say
ill linn and about the abilities of living organisms that make up
01 Id.
I In ever, an experiment is none of these three trials in isolation. It
IIk movement of the three taken together when it succeeds, or separated
it jails I lore we recognize again the movement of circulating ref-
111i we sludied in Chapter 2. The accuracy of the statement is not
iltd lo a slale ol allaiis out ihere, but to the traceability of a series
II m loi mal ions No expeimienl tan be sludied only in the labora
124
tory, only in the literature, or only in the debates among colleagues.

An experiment is a story, to be sure—and studiable as such—but a
story tied to a situation in which new actants undergo terrible trials
plotted by an ingenious stage manager; and then the stage manager, in
turn, undergoes terrible trials at the hands of his colleagues, who tesl
what sort of ties there are between the first story and the second situa
tion. An experiment is a text about a nontextual situation, later tested
by others to decide whether or not it is simply a text. If the final trial is
successful, then it is not just a text, there is indeed a real situation be
hind it, and both the actor and its authors are endowed with a new
competence: Pasteur has proved that the ferment is a living thing; the
ferment is able to trigger a specific fermentation different from that ol
brewer’s yeast.
The essential point I am trying to make is that “construction” is m
no way the mere recombination of already existing elements. In the
course of the experiment Pasteur and the ferment mutually exchange
and enhance their properties, Pasteur helping the ferment show its mol
tie, the ferment “helping” Pasteur win one of his many medals. If tin
final trial is lost, then it was just a text, there was nothing behind it lu
support it, and neither actor nor stage manager has won any additional
competences. Their properties cancel each other out, and colleague
can conclude that Pasteur has simply prompted the ferment to sn
what he wished it to say. If Pasteur wins we will find two (partially)
new actors on the bottom line: a new yeast and a new Pasteur I II 110
loses, there will be only one, and he, the Pasteur of old, will go down III
history as a minor figure together with a few shapeless yeasts and
wasted chemicals.
We need to understand that whatever we want to think or aigua
about the artificial character of the laboratory, or the literary aspi 11
of this peculiar type of exegesis, the lactic acid ferment is invented no
by Pasteur but by theferment. At least, this is the problem that the (1ial
of his colleagues, of Pasteur himself, and of the little bug in the gl 1
ware must resolve. It is essential to all of them that whatever the in i1
nuity of the experiment, whatever the perverse artificiality ol il»
setup, whatever the underdetermination or the weight of ihcoitlli (I
expectations, Pasteur manages to take himself out ol Lhe action so 1
to become an expert, that is, experitus, someone liansfoimvcU’y jh
manifestation of something not contrived by lhe loimci Paslout
125
m,iiter how artificial the setting, something new, independent of the

11ling, has to emerge, or else the whole enterprise is wasted.
11 is because of this “dialectic” between fact and artifact that, al-
Ihough no philosopher would seriously defend a correspondence the-
•i y of truth, it is nevertheless absolutely impossible to be convinced
II a purely constructivist argument for more than three minutes,
i'll, let’s say an hour, to be fair. Most philosophy of science since
11time and Kant consists in taking on, evading, hedging, coming back
i lecanting, solving, refuting, packing, unpacking this impossible
ml momy: that on the one hand facts are experimentally made up and
\ er escape from their manmade settings, and on the other hand it is
t nliai that facts are not made up and that something emerges that is
it manmade. Bears in cages pace back and forth within their narrow
11 ons with less obstinacy and less distress than philosophers and so-
I ilogists of science going incessantly from fact to artifact and back.
Ilns obstinacy and this distress come from defining an experiment
i /ero-sum game. If the experiment is a zero-sum game, if every
il|)iit has to be matched by an input, then nothing escapes from a
ilimulory that has not been previously put into it. Such is the real
il ness of common definitions of construction and fabrication:
111lever the philosopher’s list of the inputs in a setting, it always fea-
11s Ilie same elements before and after—the same Pasteur, the same
i nu ill, the same colleagues, or the same theory. Whatever the scien-
I genius, they always play with a fixed set of Lego blocks. Unfortu-
ilt ly. since it is at once fabricated and not fabricated, there is always
t m lhe experiment than was put into it. Explaining the outcome
Ilu experiment by using a list of stable factors and actors will there-
i 11ways show a deficit.
II Is lli is deficit that will then be accounted for differently by the
11 mis lealist, constructivist, idealist, rationalist, or dialectic persua-
n I ath will make up the deficit by cashing in its favorite stocks: na-
i iMil lhere,” macro- or micro-social factors, the transcendent Ego,
i h s, standpoints, paradigms, biases, or the churning blender of
11h Ians. There seems to be an endless supply of fat bank accounts
n which one can draw to complete the list and “explain” away the
In «lily of an experiment’s outcome. In this kind of solution, the
Il y Is not accounted for by modifications in the list of initial ac-
Inil by Ihe addition ol one pai amount factoi that balances the ac-
126
count. In this way, every input is balanced by an output. Nothing new
has happened. Either experiments simply reveal Nature; or alterna
tively society, or biases, or theoretical blind spots betray themselves in
the outcome, over the course of an experiment. Nothing more hap
pens in the history of science than the discovery of what was already
there, all along, in nature or in society.
But there is no reason to believe that an experiment is a zero-sum
game. On the contrary, each of the difficulties posed by Pasteur’s pa
per suggests that an experiment is an event*. No event can be accounk d
for by a list of the elements that entered the situation before its conclu
sion, before Pasteur launched his experiment, before the yeast started In
trigger the fermentation, before the meeting of the Academy. If such a
list were made, the actors on it would not be endowed with the com
petence that they will acquire in the event. On this list Pasteur is t
promising crystallographer but has not shown to anyone’s satisfacl 1011
that the ferments are living creatures; the ferment may accompany Ilu
fermentation, as Liebig allowed, but it is not yet endowed with H im
property of triggering a lactic acid fermentation different from th.il n!
brewer’s yeast; as for the Academicians, they do not yet depend on a
living yeast in their own laboratories and may prefer to remain on (li
solid foundations of chemistry they learned from Liebig instead 1I
flirting again with vitalism. This list of inputs does not have lo I*
completed by drawing upon any stock of resources, since the si oik
drawn upon before the experimental event is not the same as the on
drawn upon after it. This is precisely why an experiment is an evi ill
and not a discovery, not an uncovering, not an imposition, not a Nyn
thetic a priori judgment*, not the actualization of a potentiality*, in
so on.
This is also why the list drawn up after the experiment needs no i
dition of Nature, or society, or whatever, since all the elemenls li 1
been partially transformed: a (partially) new Pasteur, a (partially) lit
yeast, and a (partially) new Academy are all congratulating one n
other at its end. The ingredients on the first list are insuffleicnl 11
because one factor has been forgotten or because the list has nol lu l
carefully drawn, but because actors gain in their definitions llnou I
this event, through the very trials of the expeiimenl. Everyone agit
that science grows through experiment; the point is that Pasluii il
127
modified and grows through this experiment, as does the Academy,

nid yes, the yeast too. They all leave their meeting in a different state
mm the one in which they entered. As we will see in the next chapter,
1Ins may lead us to inquire whether there is a history of science, not
nlv of scientists, and whether there is a history of things, not only of
li nee.
The Second Drama: Pasteur’s Solution to the

Conflict between Constructivism and Realism
Il has not been too difficult to reconfigure the notion of construction
11l.ibrication so as to consider an experiment as an event, and not as
( 1o sum game, it is much trickier to understand how we can simul-
iii misly insist on the artificiality of the laboratory setting and also
Ilie autonomy of the entity “made up” inside the laboratory walls,
lu sure, we are helped by the double meaning of the word “fact”—
II v Inch is made and that which is not made up; “un fait est fait,” as
i Ion Bachelard put it—but a lot of conceptual work is necessary to
lu Ihe hidden wisdom of this etymology (see Chapter 9). It is easy
imdei stand why houses and cars and baskets and mugs are at once
1tiled and real, but this is of no help in accounting for the mystery
»ii nlific objects. It is not just that they are both made up and real.
Ik 1, 11 is precisely because they have been artificially made up that
>mi a complete autonomy from any sort of production, construc-
II ni fabrication. Technical or industrial metaphors are not going to
III is grasp this most puzzling phenomenon, which has taxed the
li m e of science studies for so many years. As I have often found to
(In uise, the only solution when faced with difficult philosophical
lions is to dive even deeper into some empirical sites to see how
ill 1sis (hemselves get out of the difficulty. Pasteur’s solution in this
1 is so clever that if we had followed it all along science studies
11Id have taken an entirely different course.
I 1 Inn is perfectly aware that there is a gap in his genealogy. How
1In go liom the barely visible, gray matter that sometimes appears
IIn lop ol the vessel to the plantlike, full blown substance endowed
li mill ilional needs and lathei pailicular tastes? How can he make
P AN DO R A’S HOPE
128
this crucial step? Who is responsible for the attribution of these ac
tions, and who is responsible for the endowment of properties? Is Pas
teur not giving his entity a little nudge forward? Yes, he is doing the
action, he has prejudices, he fills the gap between underdetermineel
facts and what should be visible. He “confesses” it very explicitly in
the very last paragraph of the paper:
All through this memoir, I have reasoned on the basis of the hypothesis
that the new yeast is organized, that it is a living organism, and that
its chemical action on sugar corresponds to its development and orga
nization. If someone were to tell me that in these conclusions I am
going beyond that which thefacts prove, I would answer that it is quite
true, in the sense that the stand I am taking is in a framework of ideas
[un ordre d ’idées] that in rigorous terms cannot be irrefutably demon
strated. Here is the way I see it. Whenever a chemist makes a study
of these mysterious phenomena and has the good fortune to bring
about an important development, he will instinctively be inclined to
assign their primary cause to a type of reaction consistent with the*
general results of his own research. It is the logical course of the hu
man mind in all controversial questions. (§22)
Not only does Pasteur develop a whole ontology in order to follow
the transformation of a nonentity into an entity, as we saw in the la I
section, but he also has an epistemology, and a pretty sophistical* il
one at that. Like most French scientists, he is a constructivist of the 1a
tionalist kind—against thejpositivism of his bête noire, Auguste Comic
For Pasteur facts always need to be framed and built up by aihcui
The origin of this inevitable “ordre d ’idées”is to be found in disciplhi
ary loyalties (“a chemist”), themselves tied to past investment (“con I
tent with the general results of his own research”). Pasteur roots IIII
disciplinary inertia in culture and personal history (“his own 1#
search”) as well as in human nature (“instinct,” “the logical corn sc i f
the human mind”). In his own eyes, does the confession of this pitjil
dice weaken Pasteur’s claims? Not a bit—and this is the apparent pu
adox that is so important for us to understand. The very next sent Cl U P
which I have already quoted, introduces another quite different epl If
mology, a much more classical one in which facts may be unamhigil
ously evaluated by impartial observers. In the remainder of this tint}
ter we will try to understand this gap between two opposing sentent#
which, curiously, are not taken as conliadictoiy
129
And it is my opinion, at this point in the development of my knowl

edge of the subject, that whoeverjudges impartially the results of this
work and that which I shall shortly publish will recognize with me that
lermentation appears to be correlative to life and to the organization
ol globules, and not to their death and putrefaction. (§22)
Whereas in the sentence just before this one the logical course of the
human mind precluded “impartial judgment,” especially in “contro-
»1sial questions” which cannot be “irrefutably demonstrated,” it is
inIdenly possible for the same Pasteur to convince anyone who judges
np.ii tially. Two entirely unrelated epistemologies are juxtaposed without
In slightest suggestion that there may be some difficulty here. First,
1 Is need a theory if they are to be made visible, and this theory is
•led in the previous history of the research program—it is “path de
ni lent” as economists would say—but then, facts may be judged in-
Immlently of earlier history. Once again the mystery of the two
pnsed meanings of the little word “fact” is reiterated. Is Pasteur un-
ii e of the difficulty, or are we unable to reconcile constructivism
llli empiricism as readily as he does? Whose contradiction is this—
1 Inn’s or ours?
In older to grasp how Pasteur, without giving any sign of being para-
Iat, can go from one epistemology to its polar opposite, we have to
li 1stand how he distributes activity between himself, as the experi-
11111, and the would-be ferment. An experiment, as we just saw, is
II lion performed by the scientist so that the nonhuman can be
I It lo appear on its own. The artificiality of the laboratory does not
111minier to its validity and truth; its obvious immanence is actually
1mi ce of its downright transcendence. How could this apparent
in It he obtained? Through a very simple setup that has baffled ob-
1 loi a long time and that Pasteur beautifully illustrates. The ex-
11ii lit creates two planes: one in which the narrator is active, and a
ml in which the action is delegated to another character, a nonhu-
II nit (see Figure 4.1).
11 i pel iment shifts out* action from one frame of reference to an-
1 Who is the active force in this experiment? Both Pasteur and
t 1 I Moie piecisely, Pasteur acts so that the yeast acts alone. We
I 1 I mil why it is difficult for Pasteur to choose between a
Hin llvlsl epistemology and a le.ilist one. Pasteur creates a stage in
130
a
therefore is autonomous
Ferment’s plane
o f reference
Pasteur’s plane
o f reference
Theferment is constructed by Pasteur’s hand and...
Figure 4.1 The difficulty of accounting for an experiment comes from the “shill
ing-out” that relates the scientist’s plane of reference to the object’s plane of rdt 1
ence. It is only because Pasteur has worked well and hard in his own plane thal lit
ferment is allowed to live autonomously in its own plane. This crucial connoilli 1
should not be broken.
which he does not have to create anything. He develops geslun

glassware, protocols, so that the entity, once shifted out, becomes In
dependent and autonomous. According to which of these two ton
tradictory features is stressed, the same text becomes oil lit 1
constructivist or realist. Am I, Pasteur, making up this entity becan c I
am projecting my prejudices onto it, or am I being made up and foi 11
to behave that way because of its properties? Am I, the analyst ol Pi
teur, explaining the closure of the controversy by appealing to his Ini
man, cultural, historical interests, or will I be forced to add to the b il
ance the active role of the nonhumans he did so much to shape? I lie
questions are not philosophical problems confined to the page
journals in the philosophy of science or the pitiful stakes of the sc Un
wars: they are the very questions tackled over and over by scienlll
papers, and by which they sink or swim.
The experimental scenography in Pasteur’s paper is extremely a
ied since it follows all the subtleties of the variable ontology deplo
in the text. In the same paper some experiments are backgiouml
and blackboxed while others are made the focus of attention ami a
allowed to go through changes. At first the practice of doing scliMI
is alluded to only through very stylized accounts of expei imenls will
are quickly blackboxed. In another case, human agency Is ulull
duced in a recipe like description of the piocedme that leads to II I
131
u id fermentation. But at this point there is no “trouble with experi

ments,” to use Shapin and Schaffer’s expression (Shapin and Schaffer
1 jKs). The fermentation of lactic acid is a well-known procedure
Inch Pasteur imports unchanged. He states, “Lactic acid was discov-
ird by Sheele in 1780 in soured whey. His procedure for removing
I horn the whey is still today the best one can follow” (§4); he then bl
indes the recipe. Firmly tied to practice but completely relegated
• die background, this experimental procedure defines the baseline—
1 lie fermentation—out of which the foregrounded yeast will be
1ide to appear. Without a stabilized recipe for lactic fermentation no
I I could start to show its mettle. In a single scientific paper the au-
101 may go through several philosophies of experiment with relativ-
I m constructivist moments preceded by brutal denials of the role of
IImnents and human interventions and followed by positivist dec-
ilions. Pasteur’s scenography, for instance, changes completely in
i entrai paragraphs 7 and 8, in which the main experiment is dis-
1 ul IIuman activity is back under the spotlight, and so are the
ihles that come with it:
nu I the soluble part from brewer’s yeast, by treating the yeast for
me lime with fifteen to twenty times its weight of water at the
mpeialure of boiling water. The liquid, a complex solution of albu-
11hmis and mineral material, is carefully filtered. About fifty to one
111«11cd grams of sugar are then dissolved in each liter, some chalk is
1d and a trace of the gray material I have just mentioned from a
ul 01dinary lactic fermentation is sprinkled in; then one raises
li mperature to 30 or 35 degrees centigrade. It is also good to intro-
i mirent of carbonic acid in order to expel the air from the
I wliic h isfitted with a bent exit tube immersed under water. On
11y next day a lively and regular fermentation is manifest... In
nl we have before our eyes a clearly characterized lactic fermen-
11 t( ill) all the accidents and the usual complications of this phe-
II mm whose external manifestations are well known to chem-
)
lln \iiy moment when the entity is at its weakest ontological
( 11 Ihe lust section of this chapter), shuffled among clouds of
1use data, the expei imental chemist is in full activity, extract-
tllng lilleimg, dissolving, adding, sprinkling, raising the tem-
1 Inliodiiung cat home acid, lilting tubes, and so on. But then,
132
shifting the attention of the reader, shifting out the autonomous ac

tor, Pasteur says that “we have before our eyes a clearly characterized
lactic fermentation.” The director withdraws from the scene, and the
reader, merging her eyes with those of the stage manager, sees a fei
mentation that takes form at center stage independently of any work 01
construction.
Who is doing the action in this new medium of culture? Pasteur
since he sprinkles, boils, filters, and sees. The lactic acid yeast, since il
grows fast, uses up its food, gains power (“very little of this yeast is
necessary to transform a considerable weight of sugar”), and enteis
into competition with other similar beings growing like plants on tlu
same plot of land. If we ignore Pasteur’s work, we slip into the pit ol
naive realism from which twenty-five years of science studies have
tried to extract us. But what happens if we ignore the lactic acid’s deli
gated automatic autonomous activity? We fall into the other pit, a
bottomless as the first, of social constructivism, ignoring the role ol
nonhumans, on whom all of the people we study are focusing their aI
tention, and for whom Pasteur spent months of labor designing till
scenography.
We cannot even claim that in both cases it is only the author, the lui
man author, who is doing the work in the writing of the paper, smu*
what is at stake in the text is precisely the reversal of authorship anti
authority: Pasteur authorizes the yeast to authorize him to speak in It
name. Who is the author of the whole process and who is the authoi 11
in the text are themselves open questions, since the characters and ilia
authors exchange credibilities. As we saw in the previous section, il III
colleagues at the Academy do not believe Pasteur, he will be made ilia
sole and only author of a work offiction. If the whole setup withstand
the Academy’s scrutiny, then the text itself will be in the end aullu
rized by the yeast, the real behavior of which can then be said to undi t
write the entire text.
How can we understand the artificial stagecraft of the expin mu ill
that aimed at letting the lactic acid develop alone, by its own agi ill
in a pure medium of culture? Why is it so complicated lo m o
nize that an experiment is precisely the place where this conti nil
tion is staged and resolved? Pasteur is not plagued here by lalse 10M
sciousness, erasing the traces of his own woi k as he goes along Wi il
not have to choose between two accounts ol scientific woil , In
he explicitly places both ol these two contiadicloiy icquiiemtni II
133
Ilie final paragraph of the paper. “Yes,” he says, “I went well beyond
ilie facts, I had to, but any impartial observer will recognize that lac
tic acid is made of living organisms and not of dead chemistry.” Ac-
II )owledging his activity does not, in his view, weaken his claim for
Ilie independence of the yeast, any more than seeing the threads in a
puppeteer’s hands weakens the credibility of the story enacted by
llie puppets “freely” acting in the other plane of reference. As long as
e do not understand why what appears to us as a contradiction is
iiol one for Pasteur, we fail to learn from those we study—we sim
ply impose our philosophical categories and conceptual metaphors on
llieir work.
In Search of a Figure of Speech:

Articulation and Proposition
I il possible to use these categories and figures of speech (even if it
u ms reconfiguring them again), not to obscure the scientists’ work,
til to make it both visible and capable of producing results that are
mil pendent of it? Science studies has struggled so much with this
in si ion to no avail: why tackle it again? It would be much easier, I
ue, lo stick with the older settlement and accept the results of the
illnsophy of language, without bothering to attempt to engage the
h lit with what we say about it, an attempt that seems to force us into
m i , my intractable metaphysical difficulties. Why not go back to
I llosophical common sense, and simply distinguish epistemological

lions from ontological ones? Why not limit history to people
I oueties, leaving nature immune to history altogether? Does sci-
i studies, to be understood, really require so much philosophical
II (conceptual bricolage would be a more fitting name for it)?
» not icst quietly at some happy medium and say for instance that
i I nowledge is the resultant of two contradictory forces—to use the
illilogiam of forces we all learned in primary school and David
U s veision of it as taught in Science Studies 101 (Bloor [1976]
» ) I vcm yone would be happy. We would have the power of societ-
b! i es, paiadigms, human feelings on the one hand, and those of

11 uni icalily on the other, knowledge being simply the resulting
h i il Would that not solve all the difficulties (see Figuie 4.2)?
I uloi Innately, tlieie is no going Imcl to nosh on the onions of Fgypl

134
State of affairs Resulting statement
Figure 4.2 One classical solution to the problem of experiment is to consider it a

the resultant of two forces, one representing the contribution of the empirical
world and the other the contribution of a given system of beliefs.
that the Hebrews found, in retrospect, so palatable. The safe haven ol

the modern settlement is nostalgia, a form of exoticism (see Chapl 11
9); nothing really worked in that impossibly makeshift arrangement
of contradictory positions. It is only because we are used to what w«
left behind and not to what we face now that we find the old setlli
ment more commonsensical. How unreasonable this reasonable com
promise really is.
According to the physics of the parallelogram, if no force at all cam»
from the axis that I call “biases and theories,” we would have a dim I
pristine, unfettered access to a state of affairs. What laboratory sell 11
tist would believe that for a minute? Not Pasteur, at any rate, win
knows well enough the work he puts into making a state of affairs vl I
ble, and knows that this work is what gives an accurate referent t I
the paper he presents to his Academy colleagues. But the opposite pt
sition, imputed to science studies by the science warriors, is ivh
more implausible. If there were no pull at all from the axis I call "si il
of affairs,” our statements about the world would be entirely mailt 11
of nothing but the earlier repertoire of myths, theories, paradigms I
ases that society has in stock. What laboratory scientist could hi lit
that for one minute—or what science student, for that mallei I
Pasteur in any case.
Where, in the repertoire and social prejudices of the nineteenth 111
tury, would one find anything to make up, to conjure, to slap (ogi lit
a little bug like the lactic acid ferment in Pasteui \ flasks? No mi 1 In
tion is fertile enough for that feat of fiction. Suiely a tug ol w 11 I
tween two contrary forces will not do the job No, no, the modi 1h I
135
Moment works as long as one does not think too much about it and
ipplies it unreflexively by shifting between completely contradictory
positions. Only an enormously powerful political reason—see Chap
in s 7 and 8—can explain why we attribute the label of common sense
lo such an unrealistic definition of what it is to speak truly about a
I ile of affairs. We may be uneasy about quitting our old habits of
thought, but no one can say that we are abandoning reasonable posi-
lions for extravagant claims. If anything, in spite of the furious volleys
I llie science wars, we may be slowly moving from absurdity to com
mon sense.
I lie difficulty of understanding Pasteur’s solution comes from his
mg the two statements “the ferment has been fabricated in my labo-
iloi y” and “the ferment is autonomous from my fabrication” as syn-
\ni s. More precisely, it is as if he were saying that because of his care-
il md skillful work in the laboratory, the ferment is therefore
ilonomous, real, and independent of any work he has done. Why do
li.ive so much trouble accepting this solution as common sense,
i I why do we feel obliged to protect Pasteur from committing one of
Iwo analytical crimes? Either forgetting the work he has been do-
o he can say that the ferment is “out there,” or else abandoning
nolions of nonhumans out there, so as to be able to focus our at-
111Ion on his work? To illuminate what happens in an experiment,
i uiaphor of the parallelogram of forces leaves much to be desired.
It 11 other figures of speech might be better aids for understand-
I' ist cur’s curious brand of what could be called “constructivist re
in I
I l l s begin with the metaphor of staging, which I used in the previ-
111ion, with Pasteur as the director bringing certain aspects of the
liment lo the foreground and backgrounding others outside the
ill 'lits’ glow. This metaphor has the great advantage of focusing at-
lon on the two planes of reference at once, instead of making them
In opposite directions. Although the work of the stage manager—
li it ol ,i puppeteer clearly aims at its own disappearance, direct-
il Uni ion away from what happens backstage and toward what
pi ns on the boaids, it is clearly indispensable for the performance
I iIm place Most ol the pleasuie of the audience actually comes
»tin (tctnhlmg piesence ol this othei plane which is at once con-
II lilt and happily lot got ten llowevei, with this pleasuie comes
136
the main weakness of this figure of speech. This metaphor, borrowed
from the world of art, has the unfortunate consequence of
aestheticizing the work of science and weakening its claim to truth. A1
though it may be accurate to say that a major effect of science studies
has been to render the sciences pleasurable (Jones and Galison 1998),
we are not looking for pleasure but for a truth independent of our own
making.
Comparing science with art is of course less damaging than under
standing science by using the notion of fetishism*, which we will
study in Chapter 9. When scientists are portrayed as fetishists, they
are accused of forgetting entirely the work they have just done and ol
being taken in by the seeming autonomy of the product of their own
hands. Artists, at least, can enjoy the quality of the labor even when il
vanishes from view, but there is nothing to redeem naive believei s
who forget that they are the sole cause of statements they believe to
have no other cause than a thing out there. To be sure, this figure ol
speech accounts well for the forced disappearance of any telltale trau
of labor, but, alas, it puts the laborers into a perverse position: scien
tists are seen either as clever manipulators of ventriloquous phenom
ena or as credulous magicians surprised by their own sleight of hand
We are not yet in a position to resolve this difficulty, which arises from
the fundamental definitions of action and creation used by the mod
ernists—this will have to wait until later, when I will introduce 11ip
strange concept of factish*. Can we do better and escape from art and
make-believe altogether?
Why do I portray Pasteur as someone who “gazes” at the lactic acid
ferment? Why do I use optical metaphors of seeing? The advantage of
this way of speaking is that, although it does not capture in any w«
the activity of the one who looks, it does emphasize the independent?
and autonomy of the thing to be looked at. The optical metaphoi I
used endlessly by those who say that scientists have “tinted lense
that “filter” what they “see,” that they have “biases” “distorting” tlitfll
“vision” of an object, that they have “world views” or “paradigms 01
“representations” or “categories” with which they “interpret" wlirtl
the world is like. With such expressions, however, it is utterly impo I
ble for these mediations to be anything but negative, since, in conlm I
to these expressions, the ideal of perfect vision remains that of unlp(
tered and unhampered access to the woild in the clear light ol ill
137
lnight sun of reason. Even those who sigh that “unfortunately” we

t .mnot be “totally freed” from the colored glasses of biases and preju
dices have the same imaginary goal as those who still believe that
\ e could indeed, by breaking away from all attachments to societies,
standpoints, and feelings, access the things themselves. “If only,” they
ill say, “we could do away with all these intermediary means with
liich science must abase itself in order to work—instruments, labo-
i itories, institutions, controversies, papers, collections, theories,
money [the five loops I sketched in Chapter 3]—the gaze of science
i ould be so much more penetrating . . .” If only science could exist
ilhout what science studies relentlessly shows to be its lifeblood,
how much more accurate its view of the world would be!
hut this is not at all what Pasteur alludes to when he abruptly shifts
liom the full admission of his prejudices to the full certainty that the
h 1ment is a living creature out there. The last thing Pasteur wants is to
li ive his work erased and taken for a useless distortion! How could he
move from a chair in Lille to a more powerful position in Paris if this
11e the case? On the contrary, he is extraordinarily proud of being
IIn first in history to have artificially created the conditions to make
tin lactic acid ferment free to appear, at last, as a specific entity. Far
liom opposing filters to an unmediated gaze, it is as if the more filters
it i v were the clearer the gaze was, a contradiction that the venerable
11!ual metaphors cannot sustain without breaking down.
We might then try to shift to an industrial metaphor. When, for in-
I mce, a student of industry insists that there have been a multitude of
II mslormations and mediations between the oil trapped deep in the
illogical seams of Saudi Arabia and the gas I put into the tank of my
II liom the old pump in the little village of Jaligny in France, the
him to reality of the gas is in no way decreased. On the contrary, it is
I 111y because of those many transformations, transportations, chem-
il 11'linements, and so on that we are able to make use of the real-
I ol the oil, which, without all those mediations, would remain for-
11 inaccessible to us, as safely buried as Ali Baba’s treasure. There
I llius a gieat superiority of the industrial metaphor over the optical
in o! gas over gaze, to make a horrible pun: it allows one to take
1 li Intel medial y step positively and is well in keeping with the no-
I 11ol in dilating teleience, a continuous circuit that should never be
nli 11npted il the flow of inhumation is not to bieak down. Either
PANDORA'S HOPE
138
we refuse the transformations, in which case the gas remains oil far
away, or we accept the transformations, but then we have gasoline and
not oil!
Pasteur, however, does not have any such quasi-industrial process in
mind. He does not wish to say that the lactic acid ferment is a sort of
raw material out of which, through many clever manipulations, he has
been able to refine some useful and powerful argument to convince his
colleagues, and that, if the flow of connections is not interrupted, he
will deliver the proof of what he says. The inadequacy of the gaze met
aphor does not mean that the gas one will suffice, because it breaks
down as easily as the other in the face of the bizarre nature of the phe
nomenon I want to highlight: the more Pasteur works, the more inde
pendent is the substance on which he works. Far from being a raw ma
terial out of which fewer and fewer features are conserved, it begins as
a barely visible entity and takes on more and more competences and
attributes until it ends up as a full-fledged substance! We do not sim
ply want to say that the ferment is constructed and real as all artifacts
are, but that it is more real after being transformed, as if, uncannily,
there were more oil in Saudi Arabia because there is more gas in the
tank of my car. Obviously the industrial metaphor of fabrication can
not handle that strange relation.
Metaphors having to do with roads, paths, or trails are slightly
better because they keep the positive aspect of the intermediary trans
formations without touching the autonomy of the object. If we say
that the laboratory experiment “paves the way” for the ferment to ap
pear, we obviously do not imply any negation of the existence of thal
which is eventually reached. If we point out to the soil scientists ol
Chapter 2 that the cotton thread spewed out by the Topofil Chain
“leads to” their field site, they will not consider this the exposure of a
“filter” that “distorts” their view, since without this little implement
they would be entirely unable to follow a safe path through the Am.»
zon forest. With the metaphor of trails, all the elements that were, so
to speak, vertical, interposing themselves between the gaze of the 1c
searchers and their objects, become horizontal. What the optical mela
phor forced us to take as successive veils hiding the thing, the li nil
metaphor lays down as so many red carpets that the researchers will
walk effortlessly to access the phenomenon. We thus seem able to
combine the advantage of the indusliial metaphor (that all inteimedl
139
aries are positive proofs of an entity’s reality) with that of the gaze
metaphor (that phenomena are out there, and are not the raw material
lor our conceptual refinery).
Alas, this is not yet the solution to Pasteur’s puzzle. Despite what
the metaphor of “trails” implies, phenomena are not “out there” wait
ing for a researcher to access them. Lactic acid ferments have to be
made visible by Pasteur’s work (just as Pasteur’s philosophical innova-
lion has to be made visible by my work, since this was as invisible
before my intervention as the ferment was before his!). The optical
metaphor may account for the visible but not for the “making” of
something visible. The industrial metaphor may explain why some
thing is “made” but not why it has thus become “visible.” The trail
metaphor is good at stressing the work of the scientists and their
movements, but it remains as hopelessly classical as the optical one
when it describes what the object is doing, that is, nothing at all, just
waiting for the light to fall on it, or for the trail blazed by scientists to
lead to its stubborn existence. The stage metaphor is good at pointing
out that there are two planes of reference at once, but is incapable of
locusing simultaneously on both, except by making the first plane the
i eal” backstage that allows the fiction to be believed on stage. But we
do not want more fiction and more belief; we want more reality and
more knowledge!
The weaknesses and benefits of all these metaphors are summarized
m Figure 4.3. Each metaphor contributes to our understanding of sci
ence, but each forces us to miss important aspects of the difficulties
1.used by Pasteur’s doubled epistemology. Pasteur points to an entirely
different phenomenon that should imply at least four contradictory
specifications—contradictory, that is, as long as we stick to the mod-
1mist theory of action (see Chapter 9): (1) the lactic acid ferment is
wholly independent of any human construction; (2) it has no inde
pendent existence outside the work done by Pasteur; (3) this work
liould not be taken negatively as so many doubts about its existence,
Iml positively as what makes it possible to exist; (4) finally, the experi
ment is an event and not the mere recombination of a fixed list of al-
1e.uly present ingredients.
Atcoiding to this recapitulation, experimental practice seems to be
unspeakable. It does not benefit, m public parlance, from any ready-
m ule figme of speech 1he leason loi this impossibility will appear
140
Metaphor Benefits Weaknesses
Parallelogram Explains why knowledge is Cannot focus on the two

neither just natural nor just planes at once since they are
social contradictory
Theater Shows the two planes at once Aestheticizes and shifts
toward fiction even more
Fetish Accounts for why the work is Transforms the scientist into n
forgotten dupe of his own false
consciousness
Optical Focuses attention on the Says nothing of the work and
independent thing takes all mediations as defect
to be erased
Industrial Links reality to the Takes things as raw material
transformations losing features along the way
Trail Turns every mediation into Does not modify the position
what makes possible the of the thing sitting there and
access to things undergoing no event
Articulation Stresses the independence of Is not registered in a
the thing; reveals the two commonsense metaphor;
planes at once; maintains the leads to a set of tricky
character of historical event; metaphysical difficulties
ties reality to the amount of (see Chapter 5)
work
Figure 4.3
later, in Chapter 7. It arises from the strange politics by which I h I

have been made at once completely mute and so talkative thaï M
the saying goes, “they speak for themselves”—thus providing the gi t il
political advantage of shutting down human babble with a voice 11011
nowhere that renders political speech forever empty. To escape ill
defects of all these metaphors, we have to abandon the division I»
tween a speaking human and a mute world. As long as we II I
words—or gaze—on one side and a world on the other, there In ll
possible figure of speech that can simultaneously fulfill ull Intil
specifications; hence the misrecognition from which science sludl
has suffered in the public mind.
But things may be different now that, instead of the huge vtillM
141
yap between things and language, we have many small differences be-
Iween horizontal paths of reference—themselves considered as series
<>l progressive and traceable transformations, according to the lesson
ul Chapter 2. As is usual with science studies, common sense is no
help at first and I will have to make do with my own poor resources—
ne h as another of my inscrutable doodles. What I have been groping
Inward, from the beginning of this book, is an alternative to the model
nl statements that posits a world “out there” which language tries to
itach through a correspondence across the yawning gap separating
1lie two—as we see at the top of Figure 4.4. If my solution appears
(»olly, readers should remember that I am attempting to redistribute
Ilie capacity of speech between humans and nonhumans: not a task
1li.il makes for a clear exposition! They should also remember that
i have abandoned, as largely illusory, the demarcation between on
tological and epistemological questions, which produces much of
lut passes for analytical clarity.
I 11 like to establish an entirely different model for the relations
lu Iween humans and nonhumans by borrowing a term from Alfred
01 th Whitehead, the notion of propositions* (Whitehead [1929] 1978).
I 1((positions are not statements, or things, or any sort of intermediary
I 11ween the two. They are, first of all, actants*. Pasteur, the lactic acid
I 1ment, the laboratory are all propositions. What distinguishes prop-
II ions from one another is not a single vertical abyss between words
mil Ihe world but the many differences between them, without anyone
knowing in advance if these differences are big or small, provisional or
I Imilive, reducible or irreducible. This is precisely what the word
pin positions” suggests. They are not positions, things, substances,
1 1ssences pertaining to a nature* made up of mute objects facing a
ill il 1vc human mind, but occasions given to different entities to enter
him untact. These occasions for interaction allow the entities to mod-
I Ihen definitions over the course of an event—in the present case,
ni peument.
I In key distinction between the two models is the role played by
in 11ige In the first model, the only way for a statement to have a ref-
1 in 1 is loi il to correspond to a state of affairs. But the phrase “lactic
I I lu ment" does not resemble in any way the lactic acid ferment it-
II my moie than the woid “dog” baiks or the sentence “the cat is on
I mil puns Between the statement and the stale of aflairs to which
142
THE MODEL OF STATEMENTS
THE MODEL OF PROPOSITIONS
Differences
Figure 4.4 In the canonical model—see Figure 2.20—reference is obtained by

bridging the gap between words and world by sending a statement across the
yawning abyss and assigning it the perilous task of establishing correspondence;
but if we consider neither world nor words but propositions that differ from
one another, we get another relation than correspondence; the question becomes
whether propositions are articulated with one another or not.
it corresponds, a radical doubt always sets in, since there should be a

resemblance where none is possible. The relation established between
propositions is not that of a correspondence across a yawning gap, but
what I will call articulation*. For example, Pasteur “articulates” the lac
tic acid ferment in his laboratory in the city of Lille. Of course this
means an altogether different situation for language. Instead of being
the privilege of a human mind surrounded by mute things, articula
tion becomes a very common property of propositions, in which many
kinds of entities can participate.
Although the word is used in linguistics, articulation is in no way
limited to language and may be applied not only to words but also
to gestures, papers, settings, instruments, sites, trials. Foi instance,
my friend René Boulet, in Figure 2.12, was <11liculaling the clod ol
143
earth when inserting it into the little cardboard boxes of his

“pedocomparator.” If Pasteur is able to speak truthfully about the fer
ment, it is not because he says in words the same thing as what the fer
ment is—an impossible task since the word “ferment” does not fer
ment. If Pasteur, through his clever handiwork, speaks truthfully of
the ferment, it is because he articulates entirely different relations for
the ferment. He proposes, for example, that we consider it as a living
and specific entity instead of as a useless by-product of a purely chemi
cal process. In terms of what would be demanded of a corresponding
statement, this is obviously a fallacy, a lie, at least a prejudice. That’s
exactly what Pasteur says: “I am going beyond that which the facts
prove .. . the stand I am taking is in a framework of ideas that in rigor
ous terms cannot be irrefutably demonstrated.”
Going beyond the facts and taking a stand are bad things for state
ments, since every trace of work and human agency obscures the goal
of reaching the world out there. But they are excellent things if the aim
is to articulate ever more precisely the two propositions of the lactic
acid ferment and of Pasteur’s laboratory. Whereas statements aim at a
correspondence they can never achieve, propositions rely on the artic
ulation of differences that make new phenomena visible in the cracks
that distinguish them. Whereas statements can hope at best for sterile
repetition (A is A), articulation relies on predication* with other enti
ties (A is B, C, and so on). To say that “lactic acid fermentation,” the
sentence, is like lactic acid fermentation, the thing, does not go very
far. But saying that lactic acid fermentation can be treated like a living
organism as specific as brewer’s yeast, opens up an entirely new era in
the relation of science, industry, ferments, and society in the nine
teenth century.
Propositions do not have the fixed boundaries of objects. They are
surprising events in the histories of other entities. The more articula
tion there is, the better. The terms I used in the second section of this
chapter, the name of actions* obtained through trials* during the
event* of an experiment, now take on a different meaning. All these
are ways of saying that through the artifices of the laboratory, the lac
tic acid ferment becomes articulable. Instead of being mute, unknown,
undefined, it becomes something that is being made up of many more
items, many moie aitides including papers presented at the Acad
emy! many moie i cm Iions to many more situations. There are, quite
144
simply, more and more things to say about it, and what is said by more
and more people gains in credibility. The field of biochemistry be
comes, in every sense of the term, “more articulate”—and so do the
biochemists. Actually, thanks to Pasteur’s ferment, they come into ex
istence as biochemists, instead of having to choose between biology
and chemistry as in Liebig’s day. Thus we can fulfill the four speci
fications listed above without falling into contradiction. The more
work Pasteur does, the more independent the lactic acid ferment be
comes, since it is now that much more articulate, thanks to the arti
ficial setting of the laboratory, a proposition that in no way resembles
the ferment. The lactic acid ferment now exists as a discrete entity
because it is articulated between so many others, in so many active and
artificial settings.
We will flesh out this very abstract formulation in the first section
of the next chapter. The point to be made now is that, in practice, it is
never the case that we utter statements by using only the resources ol
language and then check to see if there is a corresponding thing th.it
will verify or falsify our utterances. No one—not even the philoso
phers of language—has ever first said the “cat is on the mat” and tht n
turned to the proverbial cat to see whether or not it is sprawled on tin
proverbial mat. Our involvement with the things we speak about is til
once much more intimate and much less direct than that of the trad I
tional picture: we are allowed to say new, original things when we on
ter well-articulated settings like good laboratories. Articulation hi
tween propositions goes much deeper than speech. We speak because
the propositions of the world are themselves articulated, not the ollu 1
way around. More exactly, we are allowed to speak interestingly by whnl
we allow to speak interestingly (Despret 1996). The notion of articulait d
propositions establishes between knower and known entirely diffei t ill
relations from those in the traditional view, but it captures much moi t»
precisely the rich repertoire of scientific practice.
C H A P T E R F I V E
The Historicity of Things

Where Were Microbes before Pasteur?
Ilul,” anyone with common sense would ask with an undertone of ex-
i peration, “did ferments exist before Pasteur made them up?” There
I no avoiding the answer: “No, they did not exist before he came
ilong”—an answer that is obvious, natural, and even, as I will show,
ommonsensical! As we saw in Chapter 4, Pasteur encountered a
igue, cloudy, gray substance sitting meekly in the corner of his flasks
uni Iurned it into the splendid, well-defined, articulate yeast twirling
in ignificently across the ballroom of the Academy. That the clock
111 si 1uck twelve many times since the 1850s and her coachmen still
In 111T turned back into mice does nothing to change the fact that be-
I u 1 Pi ince Charming came along this Cinderella was a nearly invisible
II pioduct of a lifeless chemical process. Of course, my fairy tales
inn I much more helpful than those of the science warriors who
mild claim that the ferment was a part of reality “out there” all along
liU li Pasteur “discovered” with his piercing observations. No, we
I d not only to rethink what Pasteur and his microbes were doing be-
II md after the experiment but to reforge the concepts that the
1i uli m settlement has given us with which to study such events. The
lillnsophical difficulty posed by my glib response to the question
I 1 1 does not, however, reside in the historicity of ferments but in the
lllit piession “to make up.”
II we meant by “historicity” merely that our contemporary “repre-
III il I011” of mitrooiganisms dates from the mid-nineteenth century,
ill it would be no pioblem. We would have simply fallen back on the
I tdi bel ween ontological and epistemological questions that we had
MS
146
decided to abandon. To do away with this divide, we decided to grant

historicity to the microorganisms, not only to the humans discovering
them. This entails that we should be able to say that not only the mi-
crobes-for-us-humans changed in the 1850s, but also the microbes-for
themselves. Their encounter with Pasteur changed them as well. Pas
teur, so to speak, “happened” to them.
If, from another perspective, we meant by “historicity” merely that
the ferments “evolve over time” like the infamous cases of the flu virus
or HIV, there would be no difficulty either. Like that of all living spe
cies—or for that matter, the Big Bang—the historicity of a ferment
would be firmly rooted in nature. Instead of being static, phenomena
would be defined as dynamic. This kind of historicity*, however, does
not include the history of science and of the scientists. It is just an
other way to portray nature, in movement instead of as a still life
Again, the divide between what pertains to human history and whal
to natural history would not have been bridged in the slightest. Epistc
mology and ontology would remain divided, no matter how agitated
or chaotic the cosmos on either side of the gap might be.
What I want to do in this chapter, halfway through this book on ille
reality of science studies, is to reformat the question of historicity by
using the notions of proposition and articulation that I so abstractly
defined at the end of the last chapter, as the only figures of speech ahU
to fulfill all the specifications listed for Figure 4.3. What was unwoi k
able and absurd in the subject-object fairy tale may become, if not
easy, at least thinkable with the pair human-nonhuman. In the first sec
tion I will make an inventory of the new vocabulary we need in 01dt I
to extricate ourselves from the modernist predicament—still using Ilie
same example as in Chapter 4, at the risk of giving the reader an ovi I
dose of lactic acid ferment. And then, to test the usefulness of this vo
cabulary, I will shift to another canonical example from Pasteur’s lib
his debate with Pouchet over spontaneous generation—thus descend
ing from ferments to microbes.
Substances Have No History, but Propositions Do

I am going to submit a small series of concepts to a double toision h 1
as engineers do when they verify the resistance of Iheir m.Uei i.ds I lit
will be, so to speak, my laboratory dial We have now two lists nl In
THE HISTORICITY OF THINGS
147
slruments: object, subject, gap, and correspondence on the one hand;

Iuimans, nonhumans, difference, proposition, and articulation on the
other. What transformations will the notion of history undergo when
put into these two different setups? What becomes feasible or unfeasi
ble when the tension is shifted from one group of concepts to the
oilier?
Before the notion of articulation, it was impossible to answer no to
Ihe question “Did the ferments (or the microbes) exist before Pasteur”
v ilhout falling into some sort of idealism. The subject-object dichot-
1imy distributed activity and passivity in such a way that whatever was
I 11en by one was lost to the other. If Pasteur makes up the microbes,
lh.it is, invents them, then the microbes are passive. If the microbes
h ad Pasteur in his thinking” then it is he who is the passive observer
ui their activity. We have begun to understand, however, that the
pm human-nonhuman does not involve a tug-of-war between two op
posite forces. On the contrary, the more activity there is from one,
the more activity there is from the other. The more Pasteur works in
his laboratory, the more autonomous his ferment becomes. Idealism
is the impossible effort to give activity back to the humans, with-
111 dismantling the Yalta pact which had made activity a zero-sum
11ne and without redefining the very notion of action, as we will
( i m Chapter 9. In all its various forms—including of course social
imst 1uctivism—idealism had a nice polemical virtue against those
ho gi anted too much independence to the empirical world. But po-
I mu s are fun to watch for only so long. If we cease to treat activity as
II lie commodity of which only one team can have possession, it stops
I Iup; Iun to watch people trying to deprive one another of what all the
I 11 us could have aplenty.
I he subject-object dichotomy had another disadvantage. Not only
1 il .1 /ero-sum game, but there were, by necessity, only two ontolog-
il species: nature and mind (or society). This rendered any account
I 1Uill die work most implausible. How could we say that in the his-
1 ol (ciments (Chapter 4) or that of the atomic chain reaction
( h iptci 3) or lhat of the forest-savanna border (Chapter 2) there are
ill (wo (ypes of actors, nature and subjects—and that, in addition,
1) Ilung (hat one actor does not do, the second one must take over?
I 1 luu s culluie medium, loi instance: which side does it go on? Or
I m Mould's peilocomp.ilaloi ? Oi Ilalban’s calculation of the cross
148
section? Do these belong to subjectivity or to objectivity or to both?

None of the above, obviously, and yet each of these little mediations is
indispensable for the emergence of the independent actor that is nev
ertheless the result of the scientists’ work.
The great advantage of propositions is that they do not have to
be ordered into on ly tw o realm s. Of them it can be said without any
difficulty that there are m any. They unfold into a manifold, they don’t
order themselves into a duality. With the new picture I am trying to
sketch, the traditional tug-of-war is dismantled twice: not only are
there no winners and losers, but there are not even two teams. Thus i(
I say that Pasteur invents a culture medium that makes the ferment
visible, I can grant activity to a ll th ree of the elements along the way
And if I add the Lille laboratory, then I will havef o u r actors, and if I say
that the Academy has been convinced, I will hawe fiv e , and so on, with
out always worrying, terrified at the idea that I might run out of actoi s
or mix up the two reserves—and the only two—from which they
should be drawn.
To be sure, the subject-object dichotomy had one great superiority
it gave a clear meaning to the truth-value of a statement. A statement
was said to refer if, and only if, there was a state of affairs that corre
sponded to it. However, as we saw in the last three chapters, this dec I
sive advantage was turned into a nightmare when scientific practice
began to be studied in detail. In spite of the thousands of books philos
ophers of language have thrown into the abyss separating language
and world, the gap shows no sign of being filled. The mystery of refer
ence between the two—and the only two—realms of language anil
world is just as obscure as before, except that we now have an incretll
bly sophisticated version of what happens at one pole—language
mind, brain, and now even society—and a totally impoverished ver
sion of what happens at the other, that is, nothing.
With propositions, one does not have to be so lopsided and soplih
tication may be shared equally among all the contributors to the fc il
of reference. Not having to fill a huge and radical gap between tw o
realms, but merely to shift through many little gaps between slightly
different active entities, reference is no longer an all or nothing com?
spondence. As we have seen often enough, the word reference* applle
to the s ta b ility of a movement through many diffeienl implements ami
mediations. When we say that Pasteur speaks tiuthfiilly about u leal
149
state of affairs, we no longer ask him to jump from words to world.

We say something much like the “downtown expressway moving
smoothly this morning” that we hear on the radio before trying to
heat the traffic. “It refers to something there” indicates the safety,
fluidity, traceability, and stability of a transverse series of aligned in
termediaries, not an impossible correspondence between two far-
apart vertical domains. Naturally this doesn’t go quite far enough, and
I will have to show later how to recapture the normative distinction
between truth and falsity, and at less cost, with the distinction be-
Iween well-articulated and inarticulate propositions.
In any case, the sentence “The ferments existed before Pasteur made
1hem up” means two entirely different things, depending on whether
it is caught between the two poles of the subject-object dichotomy or
loaded into the series of articulated humans and nonhumans. We have
now reached the crux of the matter. This is where we will see if our
loi sion test holds up or breaks down.
111 the correspondence theory of truth, the ferments are either out
Ihe 1e or not, and if they are out there they have a lw a y s been out there,
un I if they are not there they have n ever been there. They cannot ap-
piai or disappear like the flashing signals of a lighthouse. Pasteur’s
I ilements, in contrast, either correspond or do not correspond to a
I ilo of affairs and may appear or disappear according to the vagaries
ni history, the weight of presuppositions, or the difficulties of the task.
I a c use th e su bject-object dich otom y, then th e tw o — th e only tw o —p ro ta g o -
ish la n n o t sh a re h isto ry equally. Pasteur’s statement may have a his-
Im y it appears in 1858 and not before—but the ferment cannot have
111h a history since it either has always been there or has never been
lime Since they simply stand as the fixed target of correspon-
li me, objects have no means of appearing and disappearing, that is,
I vm ying.
I Ins is the reason for the undertone of exasperation in the common-
«lislenl question raised at the beginning of this chapter. The tension
I 11ween an object with no history and statements with a history is so
it it dial when I say “Ferments of course did not exist before 1858,” I
un iltempting a task as impossible as holding the HMS B rita n n ia at
(In plei with a rope after she has started steaming away. There is no
in t in the expiession “history of science" if we cannot somehow
lui in the tension between these two poles, since we are left with
150
only a history of scientists, while the world out there remains impervi
ous to the other history—even if nature may still be said to be en
dowed with a dynamism, but that is another type of historicity alto
gether.
Fortunately, with the notion of circulating reference, there is noth
ing simpler than slackening the tension between what has and what
does not have history. If the rope holding the HMS Britannia breaks, it
is because the pier has remained fixed. But where does this fixity come
from? Only from the settlement that anchors the object of reference as
one extremity facing the statement on the other side across a yawning
gap. “Ferments exist,” however, does not qualify one of the poles—the
pier—but the whole series of transformations that make up the refer
ence. As I said, accuracy of reference indicates the fluidity and stability
of a transverse series, not the bridge between two stable points or the
rope between one fixed point and one that moves away. How does cii
culating reference help us define the historicity of things? Quite sim
ply. Every change in the series of transformations that composes the
reference is going to make a difference, and differences are all that we
require, at first, to set a lively historicity into motion—as lively as u
good lactic acid fermentation!
Although this sounds abstract it is much more commonsensical
than the model it replaces. A lactic acid ferment grown in a culture in
Pasteur’s laboratory in Lille in 1858 is not the same thing as the residue
of an alcoholic fermentation in Liebig’s laboratory in Munich in 1852
Why not the same thing? Because it is not made out of the same ai ti
cles, the same members, the same actors, the same implements, the
same propositions. The two sentences do not repeat each other. They
articulate something different. But the thing itself, where is the thing f
Here, in the longer or shorter list of elements making it up. Pasteui I
not Liebig. Lille is not Munich. The year 1852 is not the year 1858. Ik
ing sown in a culture medium is not the same as being the residue ol rt
chemical process, and so on. The reason this answer sounds funny al
first is that we still imagine the thing to be somehow at one extremity
waiting out there to serve as the bedrock for the reference. But il Ilu?
reference is what circulates through the whole series, every change III
even one element of the series will make for a change in the releientt»
It will be a different thing to be in Lille and in Munich, to be cultivait <1
with yeast and without it, to be seen under the mic lose ope and with it
pair ol glasses, and so on
151
If my slackening this tension seems like a monstrous distortion of

common sense, it is because we want to have a substance* in addition
10 attributes. This is a perfectly reasonable demand since we always
move from performances* to the attribution of a competence*. But as
we saw in Chapter 4, the relation of substance to attributes does not
have the genealogy that the subject-object dichotomy forced us to
imagine: first a substance out there, outside history, and then phe
nomena observed by a mind. What Pasteur made clear for us—what I
made clear in Pasteur’s drift through multiple ontologies—is that we
slowly moved from a series of attributes to a substance. The ferment
I)i*gan as attributes and ended up being a substance, a thing with clear
limits, with a name, with obduracy, which was more than the sum of
ils parts. The word “substance*” does not designate what “remains be
neath,” impervious to history, but what gathers together a multiplicity
ol agents into a stable and coherent whole. A substance is more like
Ilie thread that holds the pearls of a necklace together than the rock
bed that remains the same no matter what is built on it. In the same
ay that accurate reference qualifies a type of smooth and easy circu-
I ilion, substance is a name that designates the stability of an assem-
lil ige.
I liis stability, however, does not have to be permanent. The best
pi oof of this was given when, in the 1880s, to Pasteur’s great sur-
pnse, enzymology took over. The ferments as living-organisms-
1pun si Liebig’s-chemical-theory again became chemical agents that
011kl even be made through synthesis. Articulated differently they be
nne different, and yet they were still held together by a substance, a
1u substance; they now belonged to the solid house of enzymology
11111 having belonged for several decades, albeit in a different form, to
IIn solid house of the emergent biochemistry.
As we shall see, the best word to designate a substance is “institu
tion* ” It made no sense to use that word before, since it clearly came
liom Ihe vocabulary of social order and could not mean something
llu 1 Ilian the arbitrary imposition of a form onto matter. But in the
m seulement I am outlining we are no longer prisoners of the
I dull d 01 igin ol such concepts. If history can be granted to ferments,
ni» I mliahly can be granted lo institutions. Saying that Pasteur
I mud thiongh a senes of iouhni/ed gestures to produce at will a
I 1ly lue lie lei mentation that is tlcaily dilleienl fiom the other fer
I uni liions hcci and alcohol cannol pass (01 a weakening ol the
PANDORA'S HOPE
152
ferment’s claim to reality. It means, on the contrary, that we are now

talking about the ferment as a matter of fact*. The state of affairs that
the philosophy of language tried hopelessly to reach across the tiny
bridge of correspondence resides everywhere, stolid and obdurate in
the very stability of institutions. And here we have come much closer
to common sense: saying that ferments began to be firmly institution
alized in Lille in 1858 surely cannot pass for anything but a truism. And
saying that they—meaning the whole assemblage—were different in
Liebig’s laboratory in Munich a decade before, and that these kinds ol
differences are what we mean by history, certainly cannot be used as
fodder for the science wars.
So we have made some progress. The negative answer to the ques
tion that opened this chapter now appears more reasonable. Associa
tions of entities have a history if at least one of the articles making
them up changes. Unfortunately, we have solved nothing yet if we do
not correctly qualify the type of historicity that we have now distrib
uted, with such equanimity, among all the associations making up a
substance. History in itself does not guarantee that anything interest
ing happens. Overcoming the modernist divide is not the same thing
as guaranteeing that events* will take place. If we have given a reason
able meaning to the question “Did ferments exist before Pasteur?” wt
are not yet through with the modernist predicament. Its sway is not
only maintained by the polemical divide between subject and object, it
is also enforced by the notion of causality. If history has no otlu 1
meaning than to activate a potentiality*—that is, to turn into an effet I
what was already there, in the cause—then no matter how much jug
gling of associations takes place, nothing, no new thing at least, will
ever happen, since the effect was already hidden in the cause, as a pu
tential. Not only should science studies abstain from using sot it Iy
to account for nature or vice versa, it should also abstain from using
causality to explain anything. Causality follows the events and dot
not precede them, as I will try to make clear in the last section of (III
chapter.
In the subject-object framework, ambivalence, ambiguity, unttM
tainty, and plasticity bothered only humans groping their way towuitl
phenomena that were in themselves secure. But ambivalence, ambigu
ity, uncertainty, and plasticity also accompany crealuies to whu.lt the
laboratory offers the possibility of existence, a histoi ic oppoi lunily If
153
Pasteur hesitates, we have to say that the fermentation also is hesitat

ing. Objects neither hesitate nor tremble. Propositions do. Fermenta-
l ion has experienced other lives before 1858 and elsewhere, but its new
ioncrescence*, to use another term of Whitehead’s, is a unique, dated,
localized life offered by Pasteur—himself transformed by his second
gieat discovery—and by his laboratory. Nowhere in this universe—
which is not of course nature*—does one find a cause, a compulsory
movement, that permits one to sum up an event in order to explain
ils emergence. If it were otherwise, one would not be faced with an
i vent*, with a difference, but only with the simple activation of a po-
lenlial that was there all along. Time would do nothing and history
\ ould be in vain. The discovery-invention-construction of lactic yeast
I(quires that each of the articles entering its association be given the
I.it lis of a mediation*, that is, of an occurrence that is neither alto-
»i*lher a cause nor altogether a consequence, neither completely a
IIleans nor completely an end. As usual with philosophy, we eliminate
nine artificial difficulties only to encounter much trickier ones. At
Unsl these new ones are fresh and realistic—and they can be tackled
mpirically.
A Spatiotemporal Envelope for Propositions

11 I want to render the question of where the ferments were before Pas-
hin commonsensical, I have to show that the vocabulary I have out
il m (I accounts better for the history of things when they are treated
|n I like other historical events and not as a stable bedrock above
Imh social history unfolds and which is to be explained by appeal
ing lo already present causes. To do so I will use the debates between
1 «mis Pasteur and Félix Archimède Pouchet over the existence of
|iuntaneous generation. This debate is so well known that it makes a
imvement site for my little experiment in comparative historiography
(I 11ley 1972., 1974; Geison 1995; Moreau 1992; on Pouchet see Cantor
1ini) The test is simple enough: are the appearance and disappear-
nii t o( spontaneous generation highlighted more vividly with the
In 111st model or with the model of articulated propositions? Which of
(hi i Iwo accounts fares best in our torsion test?
I ( I me Inst give a sketchy histoiy of this case, which unfolds about
I 111 yuus niter the one we studied m ( haplei 4 Spontaneous genera
154
tion was a very important phenomenon in a Europe devoid of refriger

ators and ways of preserving food, a phenomenon anyone could eas
ily reproduce in his kitchen, an undisputed phenomenon made more
credible by the dissemination of the microscope. Pasteur’s denial of its
existence, on the contrary, existed only in the narrow confines of the
laboratory on the rue d’Ulm in Paris, and only insofar as he was able to
prevent, in the “swan-neck” experiment, what he called “germs car
ried by the air” from entering the culture flasks. When Pouchet at
tempted to reproduce these experiments in Rouen, the new material
culture, the new skills invented by Pasteur proved too fragile to mi
grate from Paris to Normandy, and Pouchet found spontaneous gener
ation occurring in his boiled flasks as readily as before.
Pouchet’s difficulty in replicating Pasteur’s experiments was taken
as proof against Pasteur’s claims, and thus as proof of the existence ol
the well-known universal phenomenon of spontaneous generation
Pasteur’s success in withdrawing Pouchet’s common phenomenon
from space-time required a gradual and punctilious extension of labo
ratory practice to each site and each claim of his adversary. “Finally,”
the whole of emerging bacteriology, agro-industry, medicine, by rely
ing on this new set of practices, eradicated spontaneous generation
transforming it into something that, although it had been a common
occurrence for centuries, was now a belief in a phenomenon that “had
never” existed “anywhere” in the world. This eradication, however, i e
quired the writing of textbooks, the making of historical narratives
the setting up of many institutions from universities to the Paslcui
Museum, indeed an extension of each of the five loops of science’s ui
culatory system (discussed in Chapter 3). Intense work had to be dont
to maintain Pouchet’s claim as a belief* in a nonexistent phenomenon
Indeed, intense work still has to be done. To this day, if you repm
duce Pasteur’s experiment in a defective manner by being, like me foi
instance, a poor experimenter, not linking your skills and material till
ture to the strict discipline of asepsis and germ culture learned in ml
crobiology laboratories, the phenomena making up Pouchet’s claim
will still appear. Pasteurians of course will call it “contamination,” ami
if I wrote a paper vindicating Pouchet’s claims and reviving his (null
tion based on my observations no one would publish it. But if the u»l
lective body of precautions, the standardization, the discipline leai m d
in Pasteurian laboratories weie to be interrupted, not only by me, (lie
155
bad experimenter, but by a whole generation of skilled technicians,

then the decision about who won and who lost would become uncer
tain again. A society that no longer knew how to cultivate microbes
and control contamination would have a hard time judging the claims
of the two adversaries of 1864. There is no point in history at which a
sort of inertial force can be counted on to take over the hard work of
scientists and relay it into eternity. This is another extension, this time
into history, of the circulating reference we began to follow in Chapter
1. For scientists there is no Day of Rest!
What interests me here is not the accuracy of this account but
lather the homology of the narrative of the spread of microbiological
skills with one that would have described, say, the rise of the Radical
party from obscurity under Napoleon III to prominence in the Third
Kepublic, or the expansion of diesel engines into submarines. The de
mise of Napoleon III does not mean that the Second Empire never ex
isted, nor does the emergence of diesel engines mean that they will
11si forever; nor does the slow expulsion of Pouchet’s spontaneous
generation by Pasteur mean that it was never part of nature. In the
une way that we may still, to this day, meet Bonapartists, although
Ilie1r chance of becoming President is nil, I sometimes meet spontane
ous generation buffs who defend Pouchet’s claim by linking it, for in-
1,nice, to prebiotics, that is, the early history of life, and who want to
1( wi ite history yet again, although they never manage to get their “re-
Isionist” papers published,
both Bonapartists and spontaneous-generationists have now been
Imslicd to the fringe, but their mere presence is an interesting indica-
IIon Ihat the “finally” that allowed philosophers of science, in the first
model, to definitively rid the world of entities that had been proven
1(mg, was too brutal. Not only is it brutal, it also ignores the mass of
(h I that still has to be done, daily, to activate the “definitive” version
I history. After all, the Radical party disappeared, as did the Third Re-
| nblic in June 1940, for lack of massive investments in democratic cul-
I1111 which, like microbiology, has to be taught, practiced, kept up,
In idc to sink in. It is always dangerous to imagine that at some point
ill hlsloiy, merlin is enough to keep up the reality of phenomena that
hi c been so difficult to produce. When a phenomenon “definitely”
I Is lliis does not mean that it exists forever, or independently of all
I 1 111ii e and discipline, but that it has been cnlienchcd in a costly and
156
massive institution* which has to be monitored and protected with
great care.
So, in the metaphysics of history that I want to substitute for the
traditional one, we should be able to talk calmly about relative exis
tence*. It may not be the sort of existence science warriors want for ob
jects in nature*, but it is the sort of existence science studies would
like propositions to enjoy. Relative existence means that we follow the
entities without stretching, framing, squeezing, and cutting them with
the four adverbs never, nowhere, always, everywhere. If we use these
adverbs, Pouchet’s spontaneous generation will never have been there
anywhere in the world; it was an illusion all along; it is not allowed to
have been part of the population of entities making up space and time.
Pasteur’s ferments carried by the air, however, had always been there,
all along, everywhere, and were bona fide members of the population
of entities making up space and time long before Pasteur.
To be sure, in this kind of framework historians can tell us a few
amusing things about why Pouchet and his supporters wrongly be
lieved in the existence of spontaneous generation, and why Pasteui
fumbled around for a few years before finding the right answer, but
the tracing of those zigzags will give us no new essential information
about the entities in question. Although they provide information on
the subjectivity and history of human agents, history, in such a render
ing, does not apply to nonhumans. By asking an entity to exist—01
more exactly to have existed—either nowhere and never, or always
and everywhere, the old settlement limits historicity to subjects anti
bans it for nonhumans. And yet existing somewhat, having a little reul
ity, occupying a definite place and time, having predecessors and sue
cessors, these are the typical ways of delimiting what I will call the
spatiotemporal envelope* of propositions.
But why does it seem so difficult to share historicity equally among
all the actors and to draw around them the envelope of relative exh
tence without adding or subtracting anything? Because the history of
science, like history proper, is embroiled in a moral issue that we have
to tackle first—before we can deal later, in Chapters 7 and 8, with t h e
even stronger political issue at stake. If we purge our accounts of t h e
four absolute adverbs, historians, moralists, and epistemologists a i e
afraid we may be forever unable to quality the tiuth or the falsity of
statements.
157
What do the Fafner of never-anywhere and the Fasolt of always-

everywhere assert—or more exactly roar threateningly, those two gi
ants in charge of protecting the treasure in the Nibelungen saga? That
science studies has embraced a simple-minded relativism by claiming
Ihat all arguments are historical, contingent, localized, temporal, and
thus cannot be differentiated, any of them being able, given enough
IIme, to revise the others into nonexistence. Without their help, the gi
ants boast, only an undifferentiated sea of equally valid claims will ap
pear, engulfing at once democracy, common sense, decency, morality,
and nature. The only way, according to them, to escape relativism is to
n ithdraw from history and locality every fact that has been proven
i ight, and to stock them safely in a nonhistorical nature* where they
have always been and can no longer be reached by any sort of revision.
/><marcation* between what has and what does not have a history is,
loi them, the key to virtue. For this reason, historicity is granted only
lo humans, radical parties and emperors, while nature is periodically
pm ged of all nonexistent phenomena. In this demarcationist view, his-
loiy is simply a provisional way for humans to access nonhistorical
n il lire : it is a convenient intermediary, a necessary evil, but it should
not be, according to the two treasure guards, a sustained mode of exis-
( ti<cforfacts.
I hese claims, although they are often made, are both inaccurate
uni dangerous. Dangerous because, as I have said, they forget to pay
h price of keeping up the institutions that are necessary for maintain
ing (acts durably in existence, relying instead on the cost-free inertia
I ihistoricity. But, more important, they are inaccurate. Nothing is
i In than to differentiate in great detail the claims of Pasteur and
lom hot. This differentiation, contrary to the claims of our brawny
u ii ds, is even more telling once we abandon the boasting and empty
II ilcge they want nonhumans to hold over human events. For sci-
lu ( sludies, demarcation is the enemy of differentiation*. The two giants
Ii ive like the eighteenth-century French aristocrats who claimed
h il uvil society would crash if it were not solidly supported upon
Ii Ii noble spines but were delegated to the humble shoulders of com-
i mis As it happens, civil society is better carried upon the many
Ii mldeis of the citizens than by the Atlas-like contortions of those
ill ii s ol cosmological and social 01der. It seems that the same demon-
li ilb>11 can be made foi dilfeienliahng the spaliolemporal envelopes
158
deployed by science studies in redistributing activity and historicity

among all of the entities involved. The common historians seem to do
a much better job than the towering epistemologists of maintaining
the crucial local differences.
Let us, for instance, map the two destinies of Pouchet’s and Pas
teur’s claims, to show how clearly they can be differentiated provided
they are not demarcated. Although technology as such is not an issue
here—it will be in the next chapter—it may be helpful to give a rudi
mentary model of propositions and articulations that uses some of Ihi
tools developed to follow technological projects*. Since there is no
major metaphysical difficulty in granting to diesel engines and subway
systems a relative existence only, the history of technology is veiy
much more “relaxed” than that of science as far as relative existence I
concerned. Historians of technical systems know that they can havt
their cake (reality) and eat it too (history).
In Figure 5.1, existence is not an all-or-nothing property but a 1cl 1
tive property which is conceived of as the exploration of a two
dimensional space made by association and substitution, AND and
OR. An entity gains in reality if it is associated with many others tint
are viewed as collaborating with it. It loses in reality if, on the con
trary, it has to shed associates or collaborators (human and non
human). Thus this figure does not include any final stage in which Id
toricity will be surpassed, with the entity relayed into eternity by im / tht
ahistoricity, and naturalness—although well-known phenomena lilt
blackboxing, socialization, institutionalization, standardization, and
training would be able to account for the seamless and ordinary w 1
in which they would be sustained and perpetuated. As we saw evil IU1
states of affairs become matters of fact, and then matters of com a
At the bottom of Figure 5.1, the reality of Pasteur’s airborne gei in I
obtained through an ever greater number of elements with whicli il I
associated—machines, gestures, textbooks, institutions, taxonomic
theories, and so on. The same terms can be applied to Pouchet’s clatin
which at version n + 2, time t + 2, are weak because they have losl d
most all of their reality. The difference, so important to our two gl ml
between Pasteur’s extended reality and Pouchet’s shrinking lealdy t 11
now be adequately visualized. This difference is only as big as Ilu 1111
tion between the tiny segment on the left and the long segment d ill
right. It is not an absolute demaicalion between what has nevei hui
159
lnm* 1+ 2 Assemblage of human and nonhuman elements
tiic* 5.1 Relative existence may be mapped according to two dimensions: asso-
tiion (AND), that is, how many elements cohere together at a given time, and
It titulion (OR), that is, how many elements in a given association have to be
MIIIled to allow other new elements to cohere with the project. The result is a
11 1 in which every modification in the associations is “paid for” by a move in
oilier dimension. Pouchet’s spontaneous generation becomes less and less
11 md Pasteur’s culture method becomes more and more real after undergoing
in li.uisformations.
In 11*.ind what was always there. Both are relatively real and relatively
I lent, that is, extant. We never say “it exists” or “it does not exist,”
ni Ihis is the collective history that is enveloped by the expression
I nnUneous generation, or germs carried by the air.”
EXHIBIT A
Id us assume that any entity is defined by an association profile

ni olliei entities called actors. let us suppose that those actors are
li iwn horn a list that i.inks them, lor instance, in alphabetical order.
Id us luilhei assume (hat each association, called a piogram, is
l6o
AND
(1) ABC
(2) ABCDE
(3) EFG
(4) FGH
(5) GMJ
(6) GHIJK
(7) [GHIJ] KL
tO R (8) [----- ] KLMNOPQ.
Figure A.l
counteracted by anti-programs* which dismantle or ignore the asso

ciation under consideration. Finally, let us suppose that each ele
ment, in order to move from the anti-program to the program, re
quires some elements to leave the program and some, with which
it has been already durably associated, to accompany it (Latour,
Mauguin, et al. 1992).
We shall now define two intersecting dimensions: association*
(akin to the linguist’s syntagm*) and substitution* (or paradigm*
for the linguists). To simplify, we can think of these as the AND di
mension, which will be our horizontal axis, and the OR dimension
which will be our vertical axis. Any innovation can be traced both by
its position on the AND-OR axes and by comparison with the recoi d
of the AND and OR positions that have successively defined it. If we
replace, as a convention, all of the different actors with different let
ters, we can then trace the path taken by an entity, according to a
progression such as the one in Figure A.i.
The vertical dimension corresponds to the exploration of substilu
tions, and the horizontal dimension corresponds to the number of
actors that have attached themselves to the innovation (by convoi»
tion we read these diagrams from top to bottom).
Each historical narrative can then be coded as follows: From X
point of view, between version (1), at time (1), and version (jt), il
time (2), the program ABC is transformed into ABCDE.
Then the dynamic of the narrative can be coded as follows:
To recruit F into the program, ABCD has to leave and G has to 111
ter, which yields version (3) at time (3): EFG.
After several such versions the elements that slick together tl f*
said to “exist”: they can be blackboxed together and given an idm
tity, that is, a label, as for instance is the case for the syntagm |GI lljj
l6l
after version (7), named an institution*. The elements that have
been disassociated through the different versions are said to have
lost existence.
To define an entity, one will not look for an essence, or for a corre
spondence with a state of affairs, but for the list of all the syntagms
or associations into which one element enters. This nonessentialist
definition will allow for a considerable range of variations, just as a
word is defined by the list of its usages: “air” will be different when
associated with “Rouen” and “spontaneous generation” than when
associated with “rue d’Ulm,” “swan-neck experiment,” and “germs”;
11 will mean “transport of life-force” in one case and “transport
of oxygen and transport of dust-carrying germs” in the other; but
Ilie Emperor will also be different when associated by Pouchet with
ideological support of spontaneous generation to maintain God’s
II eative power” and by Pasteur with “monetary support of laborato-
1les without any implication about the subject matters of science.”
What is the essence of air? All of these associations. Who is the Em-
peior? All of these associations.
lo make a judgment about the relative existence or nonexistence
<>l .in association, for instance “the present Emperor of France is
IMId,” one will compare this version with others and “calculate” the
lability of the association in other syntagms: “Napoleon III, Em-
pt 101 of France, has a moustache,” “the President of France is bald,”
hail dressers have no panacea against baldness,” “linguistic philoso
phai s like to use the sentence ‘the present King of France is bald.’”
I In* length of the associations, and the stability of the connections
IIn ough various substitutions and shifts in point of view, make for a
n il deal of what we mean by existence and reality.
Al first sight such an opening of reality to every entity seems to
li ly common sense, since Golden Mountains, phlogiston, unicorns,
II lit I kings of France, chimeras, spontaneous generation, black holes,
ils on mats, and other black swans and white ravens will all occupy
Ilu same space time as Hamlet, Popeye, and Ramses II. Such equa-
iilmily seems certainly too democratic to avoid the dangers of rela-
II 1 m, but this criticism forgets that our definition of existence and
1 tllly is extracted, not from a one-to-one correspondence between
in I dialed statement and a state of affairs, but from the unique sig-
1 time diawn by associations and substitutions through the concep-
I 1il p ne
has been shown so many limes by science studies, it is the col
1 history (hat allows us lo judge the iclalive existence of a phe
162
nomenon; there is no higher court that would be above the collec
tive and beyond the reach of history, although much of philosophy
was devised to invent just such a court (see Chapter 7). This sketchy
diagramming of narratives simply aims at directing our attention
toward an alternative that does not abandon the moral aims of dif
ferentiation: each relative existence has one typical envelope and
only one.
The second dimension is the one that captures historicity. Histoi y
of science does not document the travel through time of an already
existing substance. Such a move would accept too much of what the gl
ants demand. Science studies documents the modifications of the in
gredients that compose an articulation of entities. Pouchet’s spontané
ous generation, for instance, at the beginning is made of many
elements: commonsense experience, anti-Darwinism, republicanism
Protestant theology, natural history, skills for observing egg develop
ment, a geological theory of multiple creations, the equipment ol
Rouen’s natural history museum, and so on. In encountering Pasted 1
opposition, Pouchet alters many of these elements. Each alteration
substitution, or translation means a movement up or down the vt 1
tical dimension of Figure 5.1. To associate elements into a duiabli
whole, and thus to gain existence, he has to modify the list that m.il «
up his phenomenon. But the new elements will not necessarily coin it*
with the earlier ones, in which case there will be a movement down
ward on the figure—because of the substitution—and there may Ik i
shift to the left because of a lack of associations between the new I
“recruited” elements.
For example, Pouchet has to learn a great deal of the laboialoi
practice of his adversary in order to fulfill the requirements ol ill
commission nominated by the Academy of Science to adjudicaU ill
dispute. If he fails to live up to those requirements, he loses the u|
port of the Academy in Paris and has to rely more and more on 11pill
lican scientists in the provinces. His associations may be extendi tl
for instance he gains a great deal of support from the anti Bon.ip d ill
popular press—but the support he expected from the Academy v d
ishes. The compromise between associations and substitutions Is wli I
I call exploring the collective. Any entity is such an exploiadon, siu It
series of events, such an experiment, such a proposition ol wlml linl I
with what, of who holds with whom, o( who holds with wh.il ol wli
163
holds with whom. If Pouchet accepts the experiments of his adversary

hut loses the Academy and gains the popular anti-establishment press,
lus entity, spontaneous generation, will be a different entity. It is not a
single substance spanning the nineteenth century unchanged. It is a
scl of associations, a syntagm, made of shifting compromises, a para
digm*—in the linguistic sense of the word, not the Kuhnian one—ex
ploring what the nineteenth-century collective can withstand.
To Pouchet’s dismay, there seems to be no way, working in Rouen,
lie can keep all his actors united in a single coherent network: Protes-
I inlism, republicanism, the Academy, boiling flasks, eggs emerging de
1wo, his ability as a natural historian, his theory of catastrophic cre-
ilion. More precisely, if he wants to maintain this assemblage he has
Io shift audiences and give his association a completely different space
nid time. It now becomes a fiery battle against official science, Cathol-
l<ism, bigotry, and the hegemony of chemistry over sound natural his-
Iniy We should not forget that Pouchet is not doing fringe science,
lull is pushed to the fringe. At the time, it is Pouchet who seems to be
ible to control what is scientific by insisting that the “great problems”
I spontaneous generation should be tackled only by geology and
01 Id history, not by going through Pasteur’s flasks and narrow con-
*1ns.
Pasteur also explores the collective of the nineteenth century, but
I association is made of elements that, at the beginning, are largely
I Imet from those of Pouchet. He has just started to fight Liebig’s
In mical theory of fermentation, as we saw in Chapter 4. This newly
mi igmg syntagm* includes many elements: a modification of vital-
111 igninst chemistry, a reemployment of crystallographic skills such
owing and cultivating entities, a position in Lille with many con-
Iions to agrobusiness relying on fermentation, a brand-new labora-
1 experiments in making life out of inert material, a circuitous
( lo teach Paris and the Academy, and so on. If the ferments that
I Uui is learning to cultivate in different media, each with its own
1lln. ily one for alcoholic fermentation, another for lactic fermen-
II in a thud for butyric fermentation—can also be allowed to appear
ill meously, as Pouchet claims, then this will be the end of the asso-
ilton ol the entities Pasleui has already assembled. Liebig will turn
I to he light 111 saying that Pasleui legiesses lo vitalism; cultures in
I tit i medium will become impossible because of uncontiollable con
164
tamination; contamination itself will have to be reformatted to be

come the genesis of the new life forms observable under the micro
scope; agrobusiness will no longer be interested in a laboratoiy
practice as haphazard as its own, and so on.
In this sketchy description I am not treating Pasteur differently
from Pouchet, as if the former were struggling with real uncontami
nated phenomena and the second with myths and fancies. Both ti y
their best to hold together as many elements as they can in order lo
gain reality. But these are not the same elements. Anti-Liebig, and
Pouchet microorganisms will authorize Pasteur to maintain the living
cause of fermentation and the specificity of ferments, allowing him lo
control and cultivate them inside the highly disciplined and artificl il
limits of the laboratory, thus connecting at once with the Academy ol
Science and agrobusiness. Pasteur too is exploring, negotiating, trying
out what holds with what, who holds with whom, what holds wllli
whom, who holds with what. There is no other way to gain reality Bill
the associations he chooses and the substitutions he explores make loi
a different socio-natural assemblage, and each of his moves modilu
the definition of the associated entities: the air as well as the Empei 01
the use of laboratory equipment as well as the interpretation of |>u
serves (that is, preserved foods), the taxonomy of microbes as well 1
the projects of agrobusiness.
The Institution of Substance

I have shown that we can sketch Pasteur’s and Pouchet’s moves In 4
symmetrical fashion, recovering as many differences as we wish li
tween them without using the demarcation between fact and fiction I
have also offered a rudimentary map so as to replace judgments ill mm
existence or nonexistence with the comparison of the spatiolcmpni
envelopes drawn when registering associations and substitutions I
tagms and paradigms. What do we gain by this move? Why shim
anyone prefer science studies’ account of the relative existence ol «
entities over the notion of a substance existing there foievei ? Wl
should adding the strange assumption of the histoiicity ol lliing I
the historicity of people simplify the narratives ol both?
The first advantage is that we do not have to consider ceitaln till
ties such as ferments, germs, or eggs spiouting mlo existence 111 lu II
165
i.itlically different from a context made of colleagues, emperors,

money, instrument, bodily skills, and so on. The doubt about the dis
tinction between context and content, which we disputed at the end of
I liapter 3, now has the metaphysics of its ambition. Each assemblage
lli.it makes up a version in Figure A.i is a list of heterogeneous associa
tions that includes human and nonhuman elements. There are many
philosophical difficulties with this way of arguing, but, as we saw in
the case of Joliot, it has the great advantage of not requiring us to sta
inli/e either the list of what makes up nature or the list of what makes
up society. This is a decisive advantage which overcomes most of the
possible defects, since, as we will see later, nature* and society* are
llu artifacts of a totally different political mechanism, one that has
imi lung to do with the accurate description of scientific practice. The
I s Iamiliar the terms we use to describe human and nonhuman asso-
l liions are to the subject-object dichotomy, the better,
lust as historians are not forced to imagine one single nature about
Inc li Pasteur and Pouchet would make different “interpretations,”
11her are they forced to imagine a single nineteenth century impos-
n ils imprint on historical actors. What is at stake in each of the two
1 ( mblages is what God, the Emperor, matter, eggs, vats, colleagues,
ml so on are able to do. Each element is to be defined by its associa-
I »ns and is an event created at the occasion of each of those associa-
I ms This is true for the lactic acid ferment, as well as for the city of
m li. the Emperor, the laboratory on the rue d’Ulm, God, and Pas-
»ii and Pouchet’s own standing, psychology, and presuppositions.
In 111borne ferments are deeply modified by the laboratory on the
d I Jim, but so is Pasteur, who becomes Pouchet’s conqueror, and so
it air that is now differentiated, thanks to the eventful swan-neck
) 11unent, into the medium that transports oxygen on the one hand,
11lie medium that carries dust and germs on the other.
I lu second advantage, as I suggested, is that we do not have to treat
I o envelopes asymmetrically by considering that Pouchet is fum-
1 in Ilie dark with nonexisting entities, while Pasteur is slowly
mlng m on an entity playing hide-and-seek, while the historians
»» In lie the seaich with warnings like “You’re cold,” “You’re getting
1im 1 "Now you’ie hot I” We will see in Chapter 9 how this sym-
li will help us bypass the impossible notion of belief. The differ-
bt tween Pouchet and Pastern is not that the lust believes and the
166
second knows. Both Pasteur and Pouchet are associating and substi
tuting elements, very few of which are similar, and experimenting
with the contradictory requirements of each entity. The associations
assembled by both protagonists are similar simply in that each ol
them draws a spatiotemporal envelope that remains locally and tern
porally situated and empirically observable. The demarcation can be
safely reapplied to the little differences between the entities with
which Pasteur and Pouchet associate themselves, and not to the ont
big difference between believers and knowers.
Third, this similarity does not mean that Pasteur and Pouchet au*
building the same networks and share the same history. The elemen!
in the two associations have almost no intersection—apart from tin
experimental setting designed by Pasteur and taken over by Pouclu I
until he fled in the face of the harsh demands of the Academy com mi
sion. Following the two networks in detail will lead us to complehly
different definitions of the nineteenth-century collective. This mean
that the incommensurability of the two positions—an incomnun
surability that seems so important for making a moral as well as hi
epistemological judgment—is itself the product of the slow différend I
tion of the two assemblages. Yes, in the end—a local and provision il
end—Pasteur’s and Pouchet’s positions have been rendered incoin
mensurable. There is no difficulty in recognizing the differences In
tween the two networks once their basic similarity has been accept» il
The spatiotemporal envelope of spontaneous generation has limil I
sharp and as precise as those of germs carried by the air which con
taminate microbe cultures. The abyss between the claims thaï oui
two giants compelled us to admit under threat of punishment Is III
deed there, but with an added bonus: the definitive line of demat (at ion
which history stopped and natural ontology took over has disappean <I
we will see in the final chapters of this book, the implementation
this line of demarcation is now ready to be analyzed for the fn si IIn
independently of the problems of describing an event. In olhei wot I
we have freed differentiation from its kidnapping by a moial and |
litical debate that had nothing to do with it.
This advantage is important because it allows us to go on (|ii«ililyin
situating, and historicizing even the extension of a “final” icalily Wli
we say that Pasteur has defeated Pouchet, and that now gei ins cult
by the air are “everywheie,” this evciywheio can be documented 11
167
pirically. Viewed from the Academy of Science, spontaneous genera

tion disappeared in 1864 through Pasteur’s work. But partisans of
spontaneous generation persisted a long time and were convinced that
Ilicy had overthrown Pasteur’s chemical “dictatorship”—as they called
II and forced it to retreat into the fragile fortress of “official science. ”
According to them they had the field to themselves, even though Pas-
leur and his colleagues felt the same way. Now we can compare the
Iwo “extended fields” without attributing the difference to that be-
l con incompatible and untranslatable “paradigms”—in the Kuhnian
mse, this time—which would forever estrange Pasteur from Pouchet.
Kepublicans, provincials, and natural historians who have access to
Ilie popular anti-Bonapartist press maintain the extension of sponta-
Iit ous generation. A dozen laboratories of microbiology withdraw the
1si once of spontaneous generation from nature and reformat the
phenomena it was made of by the twin practices of pure medium cul-
II ilion and protection against contamination. The two are not in-
ompatible paradigms. They were made incompatible by the series
I issociations and substitutions of each of the two assemblies of pro-
11'onists. They simply began to have fewer and fewer elements in
1il mon.
Why we may find this reasoning difficult is that we imagine mi-
1o l i c s must have a substance that is a little bit more than the series of
I hislorical manifestations. We may be ready to grant that the set of
ilonnances always remains inside the networks and that they are
(mealed by a precise spatiotemporal envelope, but we cannot sup-
1 the feeling that the substance travels with fewer constraints than
1 pci lormances. It seems to live a life of its own, having been, like
1 Vi 1gin Mary in the dogma of Immaculate Conception, always al-
1I Ihoi c, even before Eve’s fall, waiting in Heaven to be implanted
line’s womb at the right time. There is indeed a supplement in the
IIon ol substance, but it is better accounted for, as I suggested in the
I o l um of this chapter, by the notion of institution*.
m h a 1(.‘working of the notion of substance is crucial because it
int to something that is badly accounted for by the history of sci-
i how do phenomena remain in existence without a law of inertia?
Ii 1 m’l we say that Pasteur was light and Pouchet was wrong?
II wi 1an say it, but only on the condition that we render very
ill md piecisely the institutional mechanisms that aie still at work
PANDORA’S HOPE
168
to maintain the asymmetry between the two positions. The solution to
this problem is to formulate the question in the following way: In
whose world are we now living, that of Pasteur or that of Pouchet? I
don’t know about you, but for my part, I live inside the Pasteurian net
work, every time I eat pasteurized yogurt, drink pasteurized milk, or
swallow antibiotics. In other words, to account for even a long-lasting
victory, one does not have to grant extrahistoricity to a research pro
gram as if it would suddenly, at some threshold or turning point, need
no further upkeep. What was an event must remain a continuing
event. One simply has to go on historicizing and localizing the ne I
work and finding who and what make up its descendants.
In this sense I participate in the “final” victory of Pasteur ovei
Pouchet, in the same way that I participate in the “final” victory of 1e
publican over autocratic modes of government by voting in the next
presidential election instead of abstaining or refusing to register, lo
claim that such a victory requires no further work, no further action
no further institution, would be foolish. I can simply say that I havi
inherited Pasteur’s microbes, I am the descendant of this event, whu It
in turn depends on what I make of it today (Stengers 1993). To claim
that the “everywhere and always” of such events covers the wholi
spatiotemporal manifold would be at best an exaggeration. SCc |*
away from the present networks, and completely different definition
of yogurt, milk, and forms of government will be generated, and till
time, not spontaneously . . . The scandal is not that science sludit
preaches relativism but that, in the science wars, those who claim (li il
the labor of keeping up the institutions of truth can be interrupted
without risk pass for paragons of morality. We will understand hit 11
how they accomplished this little trick and managed to turn the labli
of morality on us.
The Puzzle of Backward Causation

There are still, I am well aware, many loose ends in this gent 1il
ized use of the notions of event and proposition to replace expi esslon
like “discovery,” “invention,” “fabrication,” or “construction.” One 11
them is the very notion of construction borrowed fiom technical pi a
tice, which we are going to deconstruct, so to speak, in the next ill i|
ter. Another one is the glib answer that I gave at the beginning ol ill
169
chapter to the question “Did microbes exist before Pasteur?” I claimed

Iliât my answer, “Of course not,” was commonsensical. I cannot end
this chapter without demonstrating why I think it is so.
What does it mean to say that there were microbes “before” Pas-
leur? Contrary to the first impression, there is no deep metaphysical
mystery in this long time “before” Pasteur, but only a very simple opti-
1aI illusion that disappears as soon as the work of extending existence
m time is documented as empirically as its extension in space. My solu-
lion, in other words, is to historicize more, not less. No sooner had
I’asteur stabilized his theory of germs carried by the air than he rein-
li*i preted the practices of the past in a new light, saying that what
v ent wrong in the fermentation of beer, for example, was the inadver
tent contamination of the vats by other ferments:
Whenever an albuminous liquid of a suitable nature contains a sub

stance such as sugar, capable of undergoing diverse chemical trans
it>rmations dependent upon the nature of such and such a ferment,
the germs of these ferments all tend to propagate at the same time,
and usually they develop simultaneously, unless one of the ferments
invades the medium more rapidly than the others. It is precisely this
last circumstance that one determines when one uses this method of sowing
an organism that is already formed and ready to reproduce. (§16)Il1
Il is now possible, for Pasteur, to understand retrospectively what

111ining and industry have been doing all along without knowing it.
I In* difference between past and present is that Pasteur now masters
llu culture of organisms instead of unwittingly being manipulated by
Invisible phenomena. Sowing germs in a culture medium is the re-
111Uillation by Pasteur of what others before him, not understanding
li it it was, named disease, invasion, or mishap. The art of lactic acid
I 1mentation becomes a laboratory science. In the laboratory, condi-
II ms may be mastered at will. In other words, Pasteur reinterpreted the
I 1 I pi act ices of fermentation as fumbling around in the dark with en-
lllli s against which one could now protect oneself.
I low has this retrospective vision of the past been achieved? What
I i Inn did was to produce in 1864 a new version of the years 1863,
1 nj 1861, which now included a new element: “microbes fought un
it Iingly by faulty and haphazard piacticcs.” Such a relroproduction
I liislmy is a lamiliai (calme loi histonans, especially historians of
170
history (Novick 1988). There is nothing easier to understand than how

Christians, after the first century, reformatted the entire Old Testa
ment as confirmation of a long and hidden preparation for Christ’s
birth, or how European nations had to reinterpret the history of Gci
man culture after the end of World War II. Exactly the same thing
happened with Pasteur. He retrofitted the past with his own microbiol
ogy: the year 1864 that was built after 1864 did not have the same com
ponents, textures, and associations as the year 1864 produced during
1864. 1 try to make this point as simple as possible in Figure 5.2.
If this enormous work of retrofitting—which includes history tell
ing, textbook writing, instrument making, body training, and the cic
ation of professional loyalties and genealogies—is ignored, then tlu
question “Did the microbes exist before Pasteur?” takes on a paraly
ing aspect that can stupefy the mind for a minute or two. After a le\
minutes, however, the question becomes empirically answerable: Pa
teur also took care to extend his local production into other times ami
places and to make the microbes the substrate of other people’s unwll
ting actions. We now understand better the curious etymology of Ilu
word “substance,” which has been causing us so much trouble in tlu «
two chapters on Pasteur. Substance does not mean that there is a dm 1
ble and ahistorical “substrate” behind the attributes, but that it is |>o
sible, because of the sedimentation of time, to turn a new entity ini 1
what lies beneath other entities. Yes, there are substances that have ht 11
there all along, but on the condition that they are made the substi il
of activities, in the past as well as in space. So there are two pi at III il
meanings now given to the word substance*: one is the institution
that holds together a vast array of practical setups, as we saw eai lit 1
and the other is the work of retrofitting that situates a more im n l
event as what “lies beneath” an older one.
The “everywhere and always” may be reached, but it is costly an
its localized and temporal extension remains visible throughout I
may take a while before we can effortlessly juggle all these dates (an
dates of dates), but there is no logical inconsistency in talking ulu I
the extension in time of scientific networks, any more than 11ml
are discrepancies in following their extension in space. It can even I
said that the difficulties in handling these appaient parado 1 Ml
tiny compared with the smallest of those olleied by lelalivisile pli
ics. If science had not been kidnapped for entirely dilleienl ends llml
171
First dimension:
linear succession
sedimentarysuccession
oftime
I >
11
1e5.2 Time’s arrowis the resultant of two dimensions, not one: the first di-
iii usion, the linear succession of time, always moves forward (1865is a fte r 1864);
IIn setend one, sedimentary succession, moves backward (1865 occurs before
1j) Whenweaskthe question “Where was the ferment before1865?”wedonot
u It (he top segment of the column that makes up the year 1864, but only the
ui veise line that marks the contribution of the year 1865to the elaboration of
In yi.11 1864. This, however, implies no idealismor backward causation, since
nu s,ii rowalways moves irreversiblyforward.
mhl be no difficulty in describing the appearance and disappearance

I 111opositions that never stopped having a history. Now that we have
• 1111 lo see that scientific practice can be studied, we are equipped to
Uni Ilie motives behind this kidnapping and even the culprits’ hide-
nt lint beloie we can do this we still have one long detour to make,
ly ol the nuslei ol detouis Daedalus the engineer. Without he
lming lo icwoik pail ol the philosophy ol technology and pail of the
172
myth of progress, we won’t be able to shake off the moral and political
burden that the modernist settlement has so unfairly placed on the
shoulders of nonhumans. Nonhumans are born free, and everywhere
they are in chains.
EXHIBIT B
A year should be defined along two axes, not only one. The first axis
registers the linear dimension of time, that is, the succession of
years. In that sense 1864 happens before 1865. But this is not all there
is to say about the year 1864. A year is not only a figure in a series of
integers, it is also a column along a second axis that registers the sed
imentary succession of time. In this second dimension there is also a
portion of what happened in 1864 that is produced after 1864 and
made retrospectively a part of the ensemble that forms, from then
on, the sum of what happened in the year 1864.
In the case represented in Figure 5.2, the year 1864 is formed of as
many segments as there have been years since. If the year 1864 “ol
1864” contains spontaneous generation as a generally accepted plu
nomenon, the year 1864 “of 1865” includes, in addition, an intent
conflict over spontaneous generation. This conflict no longer rage
another year later, after the scientific community has definitively ai
cepted Pasteur’s theory of airborne germs. The year 1864 “of iSftft
thus includes a vestigial belief in spontaneous generation and a 111
umphant Pasteur.
This process of sedimentation never ends. If we skip forwaid Mil
years, there is still a year 1864 “of 1998,” to which has been addtd
many features, not only a rich new historiography of the dispute 1*0
tween Pasteur and Pouchet, but maybe also a complete revision of
the dispute in which, eventually, Pouchet is the winner because lie
anticipated some results of prebiotics.
What gives an appearance of depth to the question “Where w< 10
the airborne germs before 1864?” is a very simple confusion between
the first, linear dimension of time and the second, sedimentaiy di
mension. If one considers only the first dimension the answer Is m
where,” since the first segment in the column that makes lip III
whole year 1864 does not include any airborne germs. The coll 0
quence is not, however, an absurd form of idealism, since most ol ill
other sedimentary segments of the year 1864 do include nil hoi II
173
germs. It is thus possible to say, without contradiction, both “Air

borne germs were made up in 1864” and “They were there all along,”
that is, all along the vertical column that recapitulates all the compo
nents of the year 1864 produced since.
In that sense, no more fundamental objections are raised by the
question “Where were microbes before Pasteur?” than by this other
question, which nobody would even think of raising: “Where was
Pasteur before 1822 (the year of his birth)?”
I am thus arguing that the only commonsensical answer to the
question is “After 1864 airborne germs were there all along.” This so
lution involves treating extension in time as rigorously as extension
in space. To be everywhere in space or always in time, work has to be
done, connections made, retrofitting accepted.
If the answers to these apparent puzzles are so straightforward,
Ilien the question is no longer whether to take such “mysteries” seri
ously, but why people take them as deep philosophical puzzles that
would condemn science studies to absurdity.
C H A P T E R S I X
A Collective of Humans
and Nonhumans
Following Daedalus s Labyrinth
The Greeks used to distinguish the straight path of reason and

scientific knowledge, episteme, from the clever and crooked path ol
technical know-how, metis. Now that we have seen how indirect, dev I
ous, mediated, interconnected, vascularized are the paths taken I»
scientific facts, we may be able to find a different genealogy for technl
cal artifacts as well. This is all the more necessary because so much ol
science studies relies on the notion of “construction,” borrowed (min
technical action. As we are going to see, however, the philosophy ol
technology is no more directly useful for defining human and nonhn
man connections than epistemology has been, and for the same 11 I
son: in the modernist settlement, theory fails to capture practice, loi A
reason that will only become clear in Chapter 9. Technical action
thus, presents us with puzzles as bizarre as those involved in the ai Ili
ulation of facts. Having grasped how the classical theory of object Ivll
fails to do any justice to the practice of science, we are now gold
to see that the notion of “technical efficiency over matter” in no w t
accounts for the subtlety of engineers. We may then be able, finall I
understand these nonhumans, which are, I have been claiming In
the beginning, full-fledged actors in our collective; we may mult I
stand at last why we do not live in a society gazing out al a n il in 1
world or in a natural world that includes society as one ol its coni|
nents. Now that nonhumans are no longei coni used willi ohjul II
I 4
A COLLECTIVE OF HUMANS AND NONHUMANS
175
may be possible to imagine the collective in which humans are entan

gled with them.
In the myth of Daedalus, all things deviate from the straight line. Af-
lei Daedalus’s escape from the labyrinth, Minos used a subterfuge
worthy of Daedalus himself to find the clever craftsman’s hiding place
md take revenge. Minos, in disguise, heralded far and wide his offer of
1 leward to anyone who could thread the circumvoluted shell of a
nail. Daedalus, hidden at the court of King Cocalus and unaware that
Ilie offer was a trap, managed the trick by replicating Ariadne’s cun
ning : he attached a thread to an ant and, after allowing it to enter the
liell through a hole at its apex, he induced the ant to weave its way
llnough this tiny labyrinth. Triumphant, Daedalus claimed his re-
n d, but King Minos, equally triumphant, asked for Daedalus’s extra-
111ion to Crete. Cocalus abandoned Daedalus; still, this artful dodger
in maged, with the help of Cocalus’s daughters, to divert the hot water
111»m the plumbing system he had installed in the palace, so that it fell,
1 11 by accident, on Minos in his bath. (The king died, boiled like an
>g ) Only for a brief while could Minos outwit his master engineer—
I) ledalus was always one ruse, one machination ahead of his rivals.
I ).iedalus embodies the sort of intelligence for which Odysseus (of
horn the Iliad says that he is polymetis, a bag of tricks) is most famed
I )( lienne and Vernant 1974). Once we enter the realm of engineers
md craftsmen, no unmediated action is possible. A daedalion, the
old in Greek that has been used to describe the labyrinth, is some-
lilng curved, veering from the straight line, artful but fake, beautiful
ill contrived (Frontisi-Ducroux 1975). Daedalus is an inventor of con-
1iplions: statues that seem to be alive, military robots that watch
II Crete, an ancient version of genetic engineering that enables Po
li Ion’s bull to impregnate Pasiphae to conceive the Minotaur—for
liiili he builds the labyrinth, from which, via another set of ma-
liliiis, he manages to escape, losing his son Icarus on the way. De
ll id, indispensable, criminal, ever at war with the three kings who
11 Iheir power from his machinations, Daedalus is the best eponym
1 li clinique and the concept of daedalion is the best tool for pene-
1il lug 1lie evolution of what I have called so far the collective*, which
ll.ls ch.iplei I want to define more precisely. Our path will lead
mil only lluough philosophy but through what could be called a
176
pragmatogony* that is, a wholly mythical “genesis of things,” in the

fashion of the cosmogonies of the past.
Folding Humans and Nonhumans into Each Other

To understand techniques—technical means—and their place in the
collective, we have to be as devious as the ant to which Daedalus al
tached his thread (or as the worms bringing the forest to the savanna
in Chapter 2). The straight lines of philosophy are of no use when it h
the crooked labyrinth of machinery and machinations, of artifacts anil
daedalia, that we have to explore. To cut a hole at the apex of the shell
and weave my thread, I need to define, in opposition to Heideggii
what mediation means in the realm of techniques. For ïleidegge^ it
technology^Tslïëver an instrument, a mere tool. Does that mean thill
technologies mediate action? No, because we have ourselves become
instruments for no other end than instrumentality itself (Heideggi 1
1977). Man—there is no Woman in Heidegger—is possessed by teih
nology, and it is a complete illusion to believe that we can master 11
We are, on the contrary, framed by this Gestell, which is one way III
which Being is unveiled. Is technology inferior to science and pint*
knowledge? No, because, for Heidegger, far from serving as applltd
science, technology dominates all, even the purely theoretical Nil
ences. By rationalizing and stockpiling nature, science plays into Ihi*
hands of technology, whose sole end is to rationalize and stockpile 11t
ture without end. Our modern destiny—technology—appeals h
Heidegger radically different from poms, the kind of “making” that till
cient craftsmen knew how to achieve. Technology is unique, insupi Irt
ble, omnipresent, superior, a monster born in our midst which has nl
ready devoured its unwitting midwives. But Heidegger is mistake n I
will try to show why by using a simple, well-known example to ill in
onstrate the impossibility of speaking of any sort of mastery in oui 1*>
lations with nonhumans, including their supposed mastery over us
“Guns kill people” is a slogan of those who try to control the mil
stricted sale of guns. To which the National Rifle Association i i pl l t
with another slogan, “Guns don’t kill people; people kill people I h
first slogan is materialist: the gun acts by virtue of material iompt
nents irreducible to the social qualities of the gunman. On auomil 1I
the gun the law abiding citizen, a good guy, becomes dangeiou I h
177
NRA, meanwhile, offers (amusingly enough, given its political views)

.1 sociological version more often associated with the Left: that the gun
does nothing in itself or by virtue of its material components. The gun
is a tool, a medium, a neutral carrier of human will. If the gunman is a
good guy, the gun will be used wisely and will kill only when appropri-
ile. If the gunman is a crook or a lunatic, then, with no change in the
\<imitself, a killing that would in any case occur will be (simply) carried
out more efficiently. What does the gun add to the shooting? In the
materialist account, everything: an innocent citizen becomes a criminal
by virtue of the gun in her hand. The gun enables, of course, but also
Instructs, directs, even pulls the trigger—and who, with a knife in her
hand, has not wanted at some time to stab someone or something?
I ach artifact has its script, its potential to take hold of passersby and
loi ce them to play roles in its story. By contrast, the sociological ver-
lon of the NRA renders the gun a neutral carrier of will that adds noth
in' to the action, playing the role of a passive conductor, through
Inch good and evil are equally able to flow.
I have caricatured the two positions, of course, in an absurdly dia-
mcliical opposition. No materialist would really claim that guns kill
by ihemselves. What the materialist claims, more exactly, is that the
good citizen is transformed by carrying the jgun. A good citizen who,
ilhout a gun, might simply be angry may become a criminal if he
t Is his hands on a gun—as if the gun had the power to change Dr. Je-
k II into Mr. Hyde. Materialists thus make the intriguing suggestion
lli il our qualities as subjects, our competences, our personalities, de-
I« nd on what we hold in our hands. Reversing the dogma of moral-
I m l he materialists insist that we are what we have—what we have in
mi hands, at least.
As lor the NRA, its members cannot truly maintain that the gun is
i lieu Irai an object that it has no part in the act of killing. They have
I i u I nowledge that the gun adds something, though not to the moral
I id ol the person holding it. For the NRA, one’s moral state is a Pla-
I ink essence: one is born either a good citizen or a criminal. Period.
mh, the NRA account is moralist—what matters is what you are,
ni i| what you have. The sole contribution of the gun is to speed the act.
killing by fists or knives is simply slower, dirtier, messier. With a gun,
ni I Ills belter, but at no point does the gun modify one’s goal. Thus
KA sociologists make the tumbling suggestion that we can master
178
techniques, that techniques are nothing more than pliable and diligent
slaves. This simple example is enough to show that artifacts are no
easier to grasp than facts: it took us two chapters to understand Pas
teur’s doubled epistemology, and it is going to take us a long time to
understand precisely what things make us do.
The First Meaning o f Technical Mediation: Interference
Who or what is responsible for the act of killing? Is the gun no moi v
than a piece of mediating technology? The answer to these question
depends on what mediation* means. A first sense of mediation (I will
offer four) is what I will call the program of action *, the series of goo I
and steps and intentions that an agent can describe in a story like Ilie
one about the gun and the gunman (see Figure 6.1). If the agent is lui
man, is angry, wants to take revenge, and if the accomplishment of Ilie
agent’s goal is interrupted for whatever reason (perhaps the agenl I
not strong enough), then the agent makes a detour; a deviation like (lie
one we saw in Chapter 3 in the operations of conviction between Jollol
and Dautry: one cannot speak of techniques any more than of sciciue
without speaking of daedalia. (Although in English the word “technol
ogy” tends to replace the word “technique,” I will make use of built
terms throughout, reserving the tainted term “technoscience” foi N
very specific stage in my mythical pragmatogony.) Agent 1 falls b<u It
on Agent 2, here a gun. Agent 1 enlists the gun or is enlisted by U |{
does not matter which—and a third agent emerges from a fusion I f
the other two.
The question now becomes which goal the new composite ttgtltl
will pursue. If it returns, after its detour, to Goal 1, then the NRA slot
obtains. The gun is then a tool, merely an intermediary. If Agenl
drifts from Goal 1 to Goal 2, then the materialist story obtains I h
gun’s intent, the gun’s will, the gun’s script have superseded thoNtf I f
Agent 1; it is human action that is no more than an intermediary Ni I
that in the figure it makes no difference if Agent 1 and Agent 2 ai f I
versed. The myth of the Neutral Tool under complete human coni I
and the myth of the Autonomous Destiny that no human can mtt I ?
are symmetrical. But a third possibility is more commonly null 0 |
the creation of a new goal that corresponds to neilhei agent’s piogmi
of action. (You only wanted to injuie but, with a gun now in yt
179
INTERRUPTION
n
GOALi
GOAL 3
GOAL 2
FIRST MEANING OF MEDIATION : GOAL TRANSLATION

I igure 6.1 As in Figure 3.1, we can portray the relation between two agents as a
II mslation of their goals which results in a composite goal that is different from
Ilie Iwo original goals.
hand, you want to kill.) In Chapter 3 I called this uncertainty about

goals translation*. As should be clear by now, translation does not
mean a shift from one vocabulary to another, from one French word
10 one English word, for instance, as if the two languages existed inde-
| >t ndently. I used translation to mean displacement, drift, invention,
nit diation, the creation of a link that did not exist before and that to
nine degree modifies the original two.
Which of them, then, the gun or the citizen, is the actor in this situa-
llon? Someone else (a citizen-gun, a gun-citizen). If we try to compre-
Iit ml techniques while assuming that the psychological capacity of
Iminans is forever fixed, we will not succeed in understanding how
111 bniques are created nor even how they are used. You are a different
1111son with the gun in your hand. As Pasteur showed us in Chapter 4,
1nee is existence and existence is action. If I define you by what you
11ive (the gun), and by the series of associations that you enter into
In n you use what you have (when you fire the gun), then you are
1mil 1lied by the gun—more so or less so, depending on the weight of
In ol her associations that you carry.
I Ins (ranslalion is wholly symmetrical. You are different with a gun
In y*tin hand; the gun is different with you holding it. You are another
11I>j( 11 because you hold the gun; the gun is another object because it
li 1 i nleied into a lelationslup with you. The gun is no longer the gun-
!n Ilu m moi y or the gun in the diawei or the gun in the pocket, but
PANDORA’S HOPE
l8 o
the gun-in-your-hand, aimed at someone who is screaming. What is

true of the subject, of the gunman, is as true of the object, of the gun
that is held. A good citizen becomes a criminal, a bad guy becomes a
worse guy; a silent gun becomes a fired gun, a new gun becomes a used
gun, a sporting gun becomes a weapon. The twin mistake of the mate
rialists and the sociologists is to start with essences, those of subjects
or those of objects. As we saw in Chapter 5, that starting point rendei s
impossible our measurement of the mediating role of techniques as
well as those of science. If we study the gun and the citizen asproposl
tions, however, we realize that neither subject nor object (nor theli
goals) is fixed. When the propositions are articulated, they join inlo a
new proposition. They become “someone, something” else.
It is now possible to shift our attention to this “someone else,” Ilx
hybrid actor comprising (for instance) gun and gunman. We musl
learn to attribute—redistribute—actions to many more agents than
are acceptable in either the materialist or the sociological account
Agents can be human or (like the gun) nonhuman, and each can have
goals (or functions, as engineers prefer to say). Since the word “agent
in the case of nonhumans is uncommon, a better term, as we have
seen, is actant*. Why is this nuance important? Because, for exampli
in my vignette of the gun and the gunman, I could replace the gun in 111
with “a class of unemployed loiterers,” translating the individual agt lit
into a collective; or I could talk of “unconscious motives,” translalln
it into a subindividual agent. I could redescribe the gun as “what the
gun lobby puts in the hands of unsuspecting children,” translating It
from an object into an institution or a commercial network; or 1could
call it “the action of a trigger on a cartridge through the intermcdl <1
of a spring and a firing-pin,” translating it into a mechanical series of
causes and consequences. These examples of actor-actant symnuli
force us to abandon the subject-object dichotomy, a distinction IIt if
prevents the understanding of collectives. It is neither people npi gull
that kill. Responsibility for action must be shared among the vnilou
actants. And this is the first of the four meanings of mediation
T he S e c o n d M e a n in g o f T ec h n ica l M e d ia tio n : C o m p o sitio n
One might object that a basic asymmetry lingers women mid e coin
puter chips, but no computer has evei made women Common sen 0
l8l
however, is not the safest guide here, any more than it is in the sci
ences. The difficulty we just encountered with the example of the gun
icmains, and the solution is the same: the prime mover of an action
Ilecomes a new, distributed, and nested series of practices whose sum
may be possible to add up but only if we respect the mediating role of
ill ihe actants mobilized in the series.
lo be convincing on this point will require a short inquiry into the
way we talk about tools. When someone tells a story about the inven-
hon, fabrication, or use of a tool, whether in the animal kingdom or
Ihe human, whether in the psychological laboratory or the historical
ni Ihe prehistoric, the structure is the same (Beck 1980). Some agent
h is a goal or goals; suddenly the access to the goal is interrupted by
1h 11breach in the straight path that distinguishes metis from episteme.
I he detour, a daedalion, begins (Figure 6.2). The agent, frustrated,
1111ns around in a mad and random search, and then, whether by in-
i 'lit or eureka or by trial and error (there are various psychologies
i niable to account for this moment) the agent seizes upon some
nlliei agent—a stick, a partner, an electrical current—and then, so the
Imy goes, returns to the previous task, removes the obstacle, and
111neves the goal. Of course, in most tool stories there is not one but
I 0 or several subprograms* nested in one another. A chimpanzee
SI COND MEANING OF MEDIATION : COMPOSITION
no» fi 1 II llie number of subpiograms is increased, then the composite goal—

1 lilt thiek uiived line becomes the common achievement of each of the
ill hi ill by Ihe piocess o( successive lianslation
182
might seize a stick and, finding it too blunt, begin, after another crisis,
another subprogram, to sharpen the stick, inventing en route a com
pound tool. (How far the multiplication of these subprograms can
continue raises interesting questions in cognitive psychology and evo
lutionary theory.) Although one can imagine many other outcomes
for instance, the loss of the original goal in the maze of sub
programs)—let us suppose that the original task has been resumed.
What interests me here is the composition of action marked by the
lines that get longer at each step in Figure 6.2. Who performs the ac
tion? Agent 1 plus Agent 2 plus Agent 3. Action is a property of assoc1
ated entities. Agent 1 is allowed, authorized, enabled, afforded by the
others. The chimp plus the sharp stick reach (not reaches) the banana
The attribution to one actor of the role of prime mover in no way
weakens the necessity of a composition of forces to explain the action
It is by mistake, or unfairness, that our headlines read “Man fhc*N
“Woman goes into space.” Flying is a property of the whole associa
tion of entities that includes airports and planes, launch pads and
ticket counters. B-52S do not fly, the U.S. Air Force flies. Action is slm
ply not a property of humans but of an association of actants, and (III
is the second meaning of technical mediation. Provisional “actoi I ll
roles may be attributed to actants only because actants are in the pm
cess of exchanging competences, offering one another new possiblll
ties, new goals, new functions. Thus symmetry holds in the case 11
fabrication as it does in the case of use.
But what does symmetry mean? Symmetry is defined by whal I
conserved through transformations. In the symmetry betwec’ii Ini
mans and nonhumans, I keep constant the series of competence 11
properties, that agents are able to swap by overlapping with one «II
other. I want to situate myself at the stage before we can deal ly di III
eate subjects and objects, goals and functions, form and mattoi, lu It I
the swapping of properties and competences is observable and lilt I
pretable. Full-fledged human subjects and respectable objcel H
there in the world cannot be my starting point; they may be my poll
of arrival. Not only does this correspond to the notion of ai denial li I
I explored in Chapter 5, but it is also consistent with many w I
established myths that tell us that we have been made by mu In
The expression Homo faber or, better, Ilomo Jahcr jabrnalus di ni 1ll
for Hegel and Andre Leroi Gouihan (Leioi (iouilun 1993) ami ItM
183
and Bergson, a dialectical movement that ends by making us sons and

daughters of our own works. As for Heidegger, the relevant myth is
Ihat “So long as we represent technology as an instrument, we remain
held fast in the will to master it. We press on past the essence of tech-
nology” (Heidegger 1977, p. 32). We will see later what can be done
with dialectics and the Gestell, but if inventing myths is the only way
lo get on with the job, I shall not hesitate to make up a new one and
t vcn to throw in a few more of my diagrams.
The Third Meaning o f Technical Mediation:

The Folding o f Time and Space
liy is it so difficult to measure, with any precision, the mediating

II tie of techniques? Because the action that we are trying to measure is
object to blackboxing*, a process that makes the joint production of
h loi s and artifacts entirely opaque. Daedalus’s maze shrouds itself in
IIci y. Can we open the labyrinth and count what is inside?
I 11e, for instance, an overhead projector. It is a point in a sequence
I u Iion (in a lecture, say), a silent and mute intermediary*, taken for
1tilled, completely determined by its function. Now suppose the pro-
1111 bleaks down. The crisis reminds us of the projector’s existence.
Ilie repairmen swarm around it, adjusting this lens, tightening
I il bulb, we remember that the projector is made of several parts,
i 11 wilh its role and function and its relatively independent goals.
In 1e.is a moment before the projector scarcely existed, now even its
III luve individual existence, each its own “black box.” In an instant
II piojector” grew from being composed of zero parts to one to
in I low many actants are really there? The philosophy of technol-
1 need has little use for arithmetic.
I In u isis continues. The repairmen fall into a routinized sequence
i IIons, 1eplacing parts. It becomes clear that their actions are com-
I ol steps in a sequence that integrates several human gestures,
no longei locus on an object but see a group of people gathered
I 111 object. A shift has occurred between actant and mediator.
II 111( s () 1 and 6.2 showed that goals are redefined by associations
Il nonlninun actants, and that action is a property of the whole as-
I it lull not only ol those actants called human. However, as Figure
h 1 the situation is even moie confused, since the number of
184
A O ---------------- ►
Step 1 : disinterest 4
B O ---------------- ►
A Q Step 2 : interest
(interruption , detour, enlistment)
B O ----------------►
a ( K »
B CH— Step 3 : composition o f a new goal
c O — i---- ►
Step 4 o b lig a to r y passage poin t
A B C
S tep s : alignment
O— 0 — 0
D ^ ABC^ Step 6 : blackboxing
D O --------- ►
______________________________________________
THIRD MEANING OF MEDIATION:

REVERSIBLE BLACKBOXING
Figure 6 .3 Any given assembly of artifacts may be moved up or down this MU I

sion of steps depending on the crisis they go through. What we may conslih I I
routine use, as one agent (step 7) may turn out to be composed of several (Mi |i
that may not even be aligned (step 4). The history of the earlier translation ill
had to go through may become visible, until they are freed again fioni Ml
influence of the others (step 1).
actants varies from step to step. The composition of objects a I <1 if

ies: sometimes objects appear stable, sometimes they appeal agit il
like a group of humans around a malfunctioning artifact I Im ll
projector may count for one part, for nothing, for one htmdiid put
for so many humans, for no humans and each pail list'll may 1mu
for one, for zero, for many, for an object, for a gioup In IIn »
185
sleps of Figure 6.3, each action may proceed toward either the disper
sion of actants or their integration into a single punctuated whole (a
whole that, soon thereafter, will count for nothing). We need to ac
count for all seven steps.
Look around the room in which you are puzzling over Figure 6.3.
( onsider how many black boxes there are in the room. Open the black
boxes; examine the assemblies inside. Each of the parts inside the
black box is itself a black box full of parts. If any part were to break,
how many humans would immediately materialize around each? How
111 back in time, away in space, should we retrace our steps to follow
ill (hose silent entities that contribute peacefully to your reading this
h ipler at your desk? Return each of these entities to step 1; imagine
Iho time when each was disinterested and going its own way, without
bung bent, enrolled, enlisted, mobilized, folded in any of the others’
|i|o(s. From which forest should we take our wood? In which quarry
In mid we let the stones quietly rest?
lost of these entities now sit in silence, as if they did not exist, in-
I 11île, transparent, mute, bringing to the present scene their force
nul Iheir action from who knows how many millions of years past.
I In y have a p ecu lia r o n to lo g ic a l statu s, b u t d o e s th is m ea n th a t th e y
1 not act, th a t th e y d o n o t m e d ia te a c tio n ? Can w e say th a t b eca u se
I lu v e m a d e all o f th e m — a n d w h o is th is “w e ,” b y th e w a y ? n o t I,
K im ly— sh o u ld th e y b e c o n sid e r e d slaves or to o ls o r m e r e ly ev i-
iiu o l a Gestell? T h e d e p th o f o u r ig n o r a n c e a b o u t te c h n iq u e s is u n
it In unable. W e are n o t e v e n ab le to c o u n t th e ir n u m b er, n o r can w e
II w h e lh e r th e y e x ist as o b je cts o r as a sse m b lie s or as so m a n y se-
1 lu us o f sk ille d a c tio n s. Yet th e r e r e m a in p h ilo s o p h e r s w h o b e lie v e
1( m e su ch th in g s as a b ject o b je c ts . . . I f s c ie n c e stu d ie s o n c e b e-
1 il I liai re ly in g o n th e c o n s tr u c tio n o f a rtifacts w o u ld h e lp a c c o u n t
I Ik Is, it is in fo r a su rp rise. N o n h u m a n s e sc a p e th e stric tu re s o f o b -
II lly I w ice; th e y are n e ith e r o b je c ts k n o w n b y a su b je c t n o r o b je c ts
ut!| m l.Ued by a m a ste r (n or, o f co u rse, are th e y m a ste r s th e m s e lv e s ).
I lu lourth Meaning o f Technical Mediation: Crossing the

Boundary between Signs and Things
M (son loi such ignoiance is made clearer when we consider the

till nid mosl impoilanl meaning ol mediation Up to this point I
186
have used the terms “story” and “program of action,” “goal” and
“function,” “translation” and “interest,” “human” and “nonhuman,” as
if techniques were dependable denizens that support the world of dis
course. But techniques modify the matter of our expression, not only
its form. Techniques have meaning, but they produce meaning via a
special type of articulation that, once again, like the circulating refer
ence we met in Chapter 2 and the variable ontology we followed in
Chapter 4, crosses the commonsense boundary between signs and
things.
Here is a simple example of what I have in mind: the speed bump
that forces drivers to slow down on campus, which in French is called
a “sleeping policeman.” The driver’s goal is translated, by means of the
speed bump, from “slow down so as not to endanger students” in Id
“slow down and protect your car’s suspension.” The two goals are fur
apart, and we recognize here the same displacement as in our gun
story. The driver’s first version appeals to morality, enlightened disln
terest, and reflection, whereas the second appeals to pure selfishne
and reflex action. In my experience, there are many more people w h o
would respond to the second than to the first: selfishness is a dull
more widely distributed than respect for law and life—at least III
France! The driver modifies his behavior through the mediation of IIII?
speed bump: he falls back from morality to force. But from an oh
server’s point of view it does not matter through which charnu I U
given behavior is attained. From her window the chancellor sees tli il
cars are slowing down, respecting her injunction, and for her thul I
enough.
The transition from reckless to disciplined drivers has been el lei It tl
through yet another detour. Instead of signs and warnings, the campu
engineers have used concrete and pavement. In this context the im
tion of detour, of translation, should be modified to absorb, nol on I
(as with previous examples) a shift in the definition of goals and Inin
tions, but also a change in the very matter of expression. The engmi 11
program of action, “make drivers slow down on campus,” is now mil
ulated with concrete. What would the right word be to account l< 1
this articulation? I could have said “objectified” or “reified" 01 it il
ized” or “materialized” or “engraved,” but these words imply an til
powerful human agent imposing his will on shapeless mallei 1 Itil
nonhumans also act, displace goals, and conti ihute to then delmllli 11
187
As we see, it is no easier to find the right term for the activity of tech
niques than for the efficacy of the lactic acid ferments—we will under-
slund in Chapter 9 that this is because they are all factishes*. In
Ilie meantime I want to propose yet another term, delegation (see Fig-
111e 6.4).
Not only has one meaning, in the example of the speed bump, been
displaced into another, but an action (the enforcement of the speed
I iw) has been translated into another kind of expression. The engi
neers’ program is delegated in concrete, and in considering this shift
e leave the relative comfort of linguistic metaphors and enter un-
I nown territory. We have not abandoned meaningful human relations
nul abruptly entered a world of brute material relations—although
IIns might be the impression of drivers, used to dealing with negotia
ble signs but now confronted by nonnegotiable speed bumps. The
lull is not from discourse to matter because, for the engineers, the
Iiced bump is one meaningful articulation within a gamut of proposi-
IIt ms from which they are no more free to choose than the syntagms*
uni paradigms* we saw in Chapter 5. What they can do is to explore
Ilu associations and the substitutions that trace a unique trajectory
liimigh the collective. Thus we remain in meaning but no longer in dis-
111sc; yet we do not reside among mere objects. Where are we?
Ilelore we can even begin to elaborate a philosophy of techniques
( liave to understand delegation as yet another type of shifting*, in
TRANSLATION
MEANING ONE
► MEANING TWO
FOUR III MEANING OF MEDIATION : DELEGATION
in (1 I As 111 higine 6 i, lilt* m lio d u tlin n of a second agent in the path of a first
11111>1u s a pint ess ol C1«uisl.ilio n , but licit 11he* shill in meaning is much greater,
I Ik vc 1y n.lime ol (lie "meaning" has been modilied The m ailer of Ihe ex
Inn h is i hanged along ( lit way
188
addition to the one that we used in Chapter 4 to understand Pasteur’s

laboratory work. If I say to you, for instance, “Let us imagine ourselves
in the shoes of the campus engineers when they decided to install the
speed bumps,” I not only transport you into another space and time
but translate you into another actor (Eco 1979). I shift you out of the
scene you now occupy. The point of spatial, temporal, and “actorial
shifting, which is basic to all fiction, is to make the reader travel with
out moving (Greimas and Courtès 1982). You make a detour through
the engineers’ office, but without leaving your seat. You lend me, foi
a time, a character who, with the aid of your patience and imaginu
tion, travels with me to another place, becomes another actor, then ri
turns to become yourself in your own world again. This mechanism I
called identification, by means of which the “enunciator” (I) and tlu
“enunciatee” (you) both invest in the shifting delegates of ourselve
within other composite frames of reference.
In the case of the speed bump the shift is “actorial”: the “sleeping
policeman,” as the bump is known, is not a policeman, does nol it»
semble one in the least. The shift is also spatial: on the campus 10 nl
there now resides a new actant that slows down cars (or damagt
them). Finally, the shift is temporal: the bump is there night mid
day. But the enunciator of this technical act has disappeared from Ili^
scene—where are the engineers? where is the policeman? wlillt
someone, something, reliably acts as lieu-tenant, holding Ilia
enunciator’s place. Supposedly the co-presence of enunciatois ml
enunciatees is necessary for an act of fiction to be possible, but wli il
we now have is an absent engineer, a constantly present speed bum|
and an enunciatee who has become the user of an artifact.
One may object that this comparison between fictional shilling mi I
the shifts of delegation in technical activity is spurious: to be li m
ported in imagination from France to Brazil is not the same as tul In M
plane from France to Brazil. True enough, but where does the dlllil
ence reside? With imaginative transportation, you simultaneously t
cupy all frames of reference, shifting into and out of all the deli g il
personae that the storyteller offers. Through fiction, ego, hit, nunt in 1
be shifted, may become other personae, in other places, at olhi 1 IIni
But aboard the plane I cannot occupy more than one liame ol id I
ence at a time (unless, of course, I sit back and lead a novel will I
takes me, say, to Dublin on a fine June day in 1904) I am seal id III a»
object institution that connects (wo au poils tluoiigh an anlini II
189
«let of transportation has been shifted down*, not out—down to planes,

engines, and automatic pilots, object-institutions to which has been
delegated the task of moving while the engineers and managers are ab
sent (or limited to monitoring). The co-presence of enunciators and
1nunciatees has collapsed, along with their many frames of reference,
loa single point in time and space. All the frames of reference of the
i ngineers, air-traffic controllers, and ticket agents have been brought
I(igether into the single frame of reference of Air France flight 1107 to
S.10 Paulo.
An object stands in for an actor and creates an asymmetry between
disent makers and occasional users. Without this detour, this shifting
lnwn, we would not understand how an enunciator could be absent:
il her it is there, we would say, or it does not exist. But through shift
ing down another combination of absence and presence becomes pos-
il de. In delegation it is not, as in fiction, that I am here and elsewhere,
1h il lam myself and someone else, but that an action, long past, of an
Him, long disappeared, is still active here, today, on me. I live in the
nildsl of technical delegates; I am folded into nonhumans.
I he whole philosophy of techniques has been preoccupied by this
I lom. Think of technology as congealed labor. Consider the very no-
I hi ol investment: a regular course of action is suspended, a detour is
nlll iled via several types of actants, and the return is a fresh hybrid
1il c.n ries past acts into the present and permits its many investors
disappear while also remaining present. Such detours subvert the
1 h 1 ol time and space—in a minute I may mobilize forces set into
1il Ion hundreds or millions of years ago in faraway places. The rela-
1 li.ipes of actants and their ontological status may be completely
hid (led techniques act as shape-changers, making a cop out of a
111<I ol wet concrete, lending a policeman the permanence and obsti-
» ol stone. The relative ordering of presence and absence is redis-
I nit d we hourly encounter hundreds, even thousands, of absent
il <1 who are remote in time and space yet simultaneously active
I pu seul. And through such detours, finally, the political order is
I ( 1led, since I rely on many delegated actions that themselves
il t me do things on behalf of others who are no longer here, the
I t til whose existence I cannot even retrace,
di bun ol this kind is not easy to understand, and the difficulty is
nipoiiiitled by the accusal ion ol fetishism* made by critics of tech
I is we will see in Cliaplei 9 It Is us, the human makeis (so they
190
say), that you see in those machines, those implements, us under an

other guise, our own hard work. We should restore the human laboi
(so they command) that stands behind those idols. We heard till
story told, to different effect, by the NRA: guns do not act on tin'll
own, only humans do so. A fine story, but it comes centuries too Kilt
Humans are no longer by themselves. Our delegation of action to otln 1
actants that now share our human existence has developed so far ill it
a program of antifetishism could only lead us to a nonhuman woi Id M
lost, phantasmagoric world before the mediation of artifacts. The ei 1
sure of delegation by the critical antifetishists would render the shill
ing down to technical artifacts as opaque as the shifting oui lu
scientific facts (see Figure 6.4).
But we cannot fall back on materialism either. In artifacts and lec It
nologies we do not find the efficiency and stubbornness of mattei, Im
printing chains of cause and effect onto malleable humans. The spi id
bump is ultimately not made of matter; it is full of engineers and cli hi
cellors and lawmakers, commingling their wills and their story lm«
with those of gravel, concrete, paint, and standard calculations I h
mediation, the technical translation, that I am trying to understand 1i
sides in the blind spot in which society and matter exchange piopi 1
ties. The story I am telling is not a Homofaber story, in which the urn
rageous innovator breaks away from the constraints of social 01dt 1 I
make contact with hard and inhuman but—at last—objective mal li I I
am struggling to approach the zone where some, though not all, ol ill
characteristics of pavement become policemen, and some, though h I
all, of the characteristics of policemen become speed bumps I h 1
earlier called this zone articulation*, and this is not, as I hope Is in
clear, a sort of golden mean or dialectic between objectivity and ill
jectivity. What I want to find is another Ariadne’s thread nnnlli
Topofil Chaix—to follow how Daedalus folds, weaves, plots, coni 11
finds solutions where none are visible, using any expedient at band I
the cracks and gaps of ordinary routines, swapping propel lies nui I
inert, animal, symbolic, concrete, and human materials.
Technical Is a Good Adjective , Technique a Lousy Noun
We now understand that techniques do not exist as such, (liai Ilu 19

nothing that we can define philosophically 01 sociologically a III I
191
In i, as an artifact or a piece of technology. There does not exist, any
more in technology than in science, anything to play the role of the
loil for the human soul in the modernist scenography. The noun
let hnique”—or its upgraded version, “technology”—does not need to
In used to separate humans from the multifarious assemblies with
Inch they combine. But there is an adjective, technical, that we can
11 i in many different situations, and rightly so.
Icchnical” applies, first of all, to a subprogram, or a series of
11 led subprograms, like the ones discussed earlier. When we say “this
I Icchnical point,” it means that we have to deviate for a moment
11mi ihe main task and that we will eventually resume our normal
mi sc of action, which is the only focus worth our attention. A black
opens momentarily, and will soon be closed again, becoming com-
II ly invisible in the main sequence of action.
S i iond, “technical” designates the subordinate role of people, skills,
1objects that occupy this secondary function of being present, indis-
11 11>le, but invisible. It thus indicates a specialized and highly cir-
m 111bed task, clearly subordinate in a hierarchy.
I bud, the adjective designates a hitch, a snag, a catch, a hiccup in
11100th functioning of the subprograms, as when we say that
1 11 is a technical problem to solve first.” Here the deviation may
I I( ul us back to the main road, as with the first meaning, but may
1 tn the original goal entirely. Technical is no longer a mere de-
II but an obstacle, a roadblock, the beginning of a detour, of a long
III 111ion, maybe of a whole new labyrinth. What should have been a
in may become an end, at least for a while, or maybe a maze, in
I 11we are lost forever.
I In lom th meaning carries the same uncertainty about what is an
mil what is a means. “Technical skill” and “technical personnel”
lo those with a unique ability, a knack, a gift, and also to the
I lo make themselves indispensable, to occupy privileged though
loi positions which might be called, borrowing a military term,
it01 y passage points. So technical people, objects, or skills are at
lull 1101 (since the main task will eventually be resumed), indis-
ihh (since the goal is unreachable without them), and, in a way,
I I Ions myslei lous, uncertain (since they depend on some highly
I ill 1d mid si etcluly ciicumsu died knack). Daedalus the perverse
uh m the limping god me good illusdations of this meaning of
192
technical. So the adjective technical has a useful meaning that agrees

in common parlance with the first three types of mediation defined
above, interference, composition of goals, and blackboxing.
“Technical” also designates a very specific type of delegation, ol
movement, of shifting down, that crosses over with entities that have
a different timing, different spaces, different properties, diffère Ml
ontologies, and that are made to share the same destiny, thus creating
a new actant. Here the noun form is often used as well as the adjective
as when we say “a technique of communication,” “a technique foi
boiling eggs.” In this case the noun does not designate a thing, but a
modus operandi, a chain of gestures and know-how bringing aboul
some anticipated result.
If one ever comes face to face with a technical object, this is nevei
the beginning but the end of a long process of proliferating media
tors, a process in which all relevant subprograms, nested one into an
other, meet in a “simple” task. Instead of the legendary kingdom in
which subjects meet objects, one generally finds oneself in the realm
of the personne morale, of what is called the “body corporate” or the*
“artificial person.” Three extraordinary terms! As if the personality
became moral by becoming collective, or collective by becoming
artificial, or plural by doubling the Saxon word body with a Latin syn
onym, corpus. A body corporate is what we and our artifacts have hi
come. We are an object-institution.
The point sounds trivial if applied asymmetrically. “Of course,” om
might say, “a piece of technology must be seized and activated by a lui
man subject, a purposeful agent.” But the point I am making is sym
metrical: what is true of the “object” is still truer of the “subject
There is no sense in which humans may be said to exist as human
without entering into commerce with what authorizes and enabli
them to exist (that is, to act). A forsaken gun is a mere piece of malti 1
but what would an abandoned gunner be? A human, yes (a gun I
only one artifact among many), but not a soldier—and certainly not
one of the NRA’s law-abiding Americans. Purposeful action and
intentionality may not be properties of objects, but they are not pi op
erties of humans either. They are the properties of institutions, ol up
paratuses, of what Foucault called dispositifs. Only corporate boilii
are able to absorb the proliferation of mediators, to regulate then t
pression, to redistribute skills, to force boxes to blacken and dose ( )h
193
11ils that exist simply as objects, detached from a collective life, are
unknown, buried in the ground. Technical artifacts are as far from the
I Hus of efficiency as scientific facts are from the noble pedestal of ob-
|«i(ivity. Real artifacts are always parts of institutions, trembling in
lheir mixed status as mediators, mobilizing faraway lands and people,
»i ady to become people or things, not knowing if they are composed
ol one or of many, of a black box counting for one or of a labyrinth
toncealing multitudes (MacKenzie 1990). Boeing 747s do not fly, air
lines fly.
Pragmatogony: Is There an Alternative

to the Myth of Progress?
In lhe modernist settlement, objects were housed within nature and
nbjects within society. We have now replaced objects and subjects
ith scientific facts and technical artifacts, which have an entirely dif-
I»lent destiny and shape. Whereas objects could only face out at
Ilii‘ subjects—and vice versa—nonhumans may be folded into humans
IIn ough the key processes of translation, articulation, delegation,
lulling out and down. What name can we give to the house in which
IIicy have taken up residence? Not nature*, of course, since its exis-
IMice is entirely polemical, as we will see in the next chapter. Society*
ill not do either, since it has been turned, by the social scientists, into
1 I.iii y tale of social relations, from which all nonhumans have been
1 11dully enucleated (see Chapter 3). In the newly emerging paradigm,
1 have substituted the notion of collective*—defined as an exchange
nl human and nonhuman properties inside a corporate body—for the
I iniled word “society.”
W e L ive in C ollectives, N o t in S o cieties
In ibandoning dualism our intent is not to throw everything into the

une pot, to efface the distinct features of the various parts within the
i olleclive. We want analytical clarity, too, but following different lines
lli 111 Ilie one drawn for the polemical tug of war between objects and
iibjecls. I’he name of ihe game is not lo exlend subjectivity to things,
I*» lient humans like objects, to take machines loi social actors, but to
194
avoid using the subject-object distinction at all in order to talk about

the folding of humans and nonhumans. What the new picture seel s
to capture are the moves by which any given collective extends its so
cial fabric to other entities. This is what I have meant, until now, by
the provisional expression “Science and technology are what sociah r*
nonhumans to bear upon human relations.” This is the makeshift ex
pression I had forged as a substitute for the modernist expression
“Science and technology allow minds to break away from society lo
reach objective nature, and to impose order on efficient matter.”
What I’d like is one more diagram, in which we could trace, not how
human subjects can break away from the shackles of social life to im
pose order on nature or to retrieve natural laws to maintain order in
society, but how a collective of one given definition can modify ih
makeup by articulating different associations. In this impossible dm
gram I would need to follow a series of coherent moves: first, theic
would be translation*, the means by which we articulate different
sorts of matter; next, what I will call, borrowing an image from genet
ics, crossover, which consists of the exchange of properties among hu
mans and nonhumans; third, a step that can be called enrollment, by
which a nonhuman is seduced, manipulated, or induced into the col
lective; fourth, as we saw in the case of Joliot and his military clienh
the mobilization of nonhumans inside the collective, which adds fresh
unexpected resources, resulting in strange new hybrids; and, finally
displacement, the direction the collective takes once its shape, extenl
and composition have been altered by the enrollment and mobili/u
tion of new actants. If we had such a diagram, we would do away willi
social constructivism for good. Alas, I and my Macintosh have nol
been able to do better than Figure 6.5.
The only advantage of this figure is to provide a basis for the com
parison of collectives, a comparison that is completely independent ol
demography (of their scale, so to speak). What science studies lm
done over the past fifteen years is subverted the distinction betwee it
ancient techniques (the poesis of artisans) and modern (broad sea It
inhuman, domineering) technologies. This distinction was never mon
than a prejudice. You can modify the size of the half circle in Figuie
6.5, but you do not have to modify its shape. You can modify the anglt
of the tangents, the extent of the translation, the types ol cmollmciit
the size of the mobilization, the impact ol the displacement, but you
195
LIMIT OF THE FIRST

COLLECTIVE
mi‘6.5 Instead of portraying science and technology as breaking away from

1
1ict limits of a society, a collective is conceived as constantly modifying its
nil through aprocess of exploration.1
l
1 I have to oppose those collectives that deal only with social rela-
111ms and those that have been able to break away from them in order
1ileal with the laws of nature. Contrary to what makes Heideggerians
up, there is an extraordinary continuity, which historians and phi-
opliers of technology have increasingly made legible, between nu
ll 11 plants, missile-guidance systems, computer-chip design, or sub-
1 automation and the ancient mixture of society, symbols, and
1il lu that ethnographers and archaeologists have studied for genera -
I ns m the cultures of New Guinea, Old England, or sixteenth-
nlmy Burgundy (Descola and Palsson 1996). Unlike what is held by
11ad 11ional distinction, the difference between an ancient or
I 1Iiiuhve” collective and a modern or “advanced” one is not that the
1ini'i manifests a rich mixture of social and technical culture while
I it lei exhibits a technology devoid of ties with the social order.
I In dilleience, rather, is that the latter translates, crosses over, en-
II mil mobilizes more elements which are more intimately con-
tini with a more finely woven social fabric, than the former does.
I In illation between the scale ol collectives and the number of
iiluimans enlisted m then midst is ci ucial One finds, of course, lon-
1 1linns ol action 111 ''m oduli' collectives, a giealei number of
196
nonhumans (machines, automatons, devices) associated with one an
other, but one must not overlook the size of markets, the number
of people in their orbits, the amplitude of the mobilization: more oh
jects, yes, but many more subjects as well. Those who have tried to dis
tinguish these two sorts of collective by attributing “objectivity”and
“efficiency” to modern technology and “humanity” to low-tech poesis
have been deeply mistaken. Objects and subjects are made simulla
neously, and an increased number of subjects is directly related to Ilu
number of objects stirred—brewed—into the collective. The adjective
modern* does not describe an increased distance between society ami
technology or their alienation, but a deepened intimacy, a more inli I
cate mesh, between the two.
Ethnographers describe the complex relations implied by eveiy
technical act in traditional cultures, the long and mediated access to
matter that these relations suppose, the intricate pattern of myths and
rites necessary to produce the simplest adze or the simplest pot, 1e
vealing that a variety of social graces and religious mores were nett*
sary for humans to interact with nonhumans (Lemonnier 1993). Hill
do we, even today, have unmediated access to naked matter? Is our III
teraction with nature short on rites, myths, and protocols (Descol I
and Palsson 1996)? Has the vascularization of science diminished 01
increased? Has the maze of Daedalus become straighter or more con
voluted?
To believe that we have modernized ourselves would be to ignoie
most of the cases examined by science and technology studies. I low
mediated, complicated, cautious, mannered, even baroque is the lie
cess to matter of any piece of technology! How many sciences 1110
functional equivalent of rites—are necessary to prepare artifacts loi
socialization! How many persons, crafts, and institutions must be III
place for the enrollment of even one nonhuman, as we saw with Ilie
lactic acid ferment of Chapter 4, or the chain reaction of Chapter 3 ol
the soil samples of Chapter 2! When ethnographers describe our bio
technology, artificial intelligence, microchips, steelmaking, and so on
the fraternity of ancient and modern collectives is instantly obvlou If
anything, what we took as merely symbolic in the old collective I
taken literally in the new: in contexts where a few dozen people wei
once required, thousands are now mobilized; wheie shot (cuts wtM#
once possible, much longer chains ol action aie now necessaiy Nil
197
Iewer but more, and more intricate, customs and protocols, not fewer
mediations but more: many more.
The most important consequence of getting beyond the Homofaber
myth is that, when we exchange properties with nonhumans through
technical delegation, we enter into a complex transaction that per
tains to “modern” as well as to traditional collectives. If anything, the
modern collective is the one in which the relations of humans and
nonhumans are so intimate, the transactions so many, the mediations
so convoluted, that there is no plausible sense in which artifact, corpo-
1ale body, and subject can be distinguished. In order to take account of
this symmetry between humans and nonhumans, on the one hand,
md this continuity between traditional and modern collectives, on the
other, social theory must be somewhat modified.
11 is a commonplace in critical theory to say that techniques are so-
1lal because they have been “socially constructed”—yes, I know, I also
used that term once, but that was twenty years ago and I recanted it
Immediately, since I meant something entirely different from what so-
1iologists and their adversaries mean by social. The notion of a social
mediation is vacuous if the meanings of “mediation” and “social” are
not made precise. To say that social relations are “reified” in technol
ogy, such that when we are confronted with an artifact we are con-
lionled, in effect, with social relations, is to assert a tautology, and a
11y implausible one at that. If artifacts are nothing but social rela
tions, then why must society work through them to inscribe itself in
omelhing else? Why not inscribe itself directly, since the artifacts
01ml for nothing? Because, critical theorists continue, through the
mi dmm of artifacts, domination and exclusion hide themselves under
tin guise of natural and objective forces. Critical theory thus deploys a
I mlology—social relations are nothing but social relations—to which
I u Ids a conspiracy theory: society is hiding behind the fetish of tech-
1l«11u s
hut techniques are not fetishes*, they are unpredictable, not means
ill mediators, means and ends at the same time; and that is why they
h upon the social fabric. Critical theory is unable to explain why ar
id u ts enter the stream of our relations, why we so incessantly recruit
m I ouali/e nonhumans. It is not to mirror, congeal, crystallize, or
il It social 1étalions, but to lemake these very relations through fresh
111 I mu peeled souices o( action Society is not stable enough to in-
198
scribe itself in anything. On the contrary, most of the features of whal

we mean by social order—scale, asymmetry, durability, power, hierai
chy, the distribution of roles—are impossible even to define without
recruiting socialized nonhumans. Yes, society is constructed, but not so
dally constructed. Humans, for millions of years, have extended then
social relations to other actants with which, with whom, they havi
swapped many properties, and with which, with whom, they foim
collectives.
A “Servant” Narrative: The Mythical History o f Collectives
A detailed case study of sociotechnical networks ought to follow nl

this juncture, but many such studies have already been written, amI
most have failed to make their new social theory felt, as the scienu
wars have made painfully clear to all. Despite the heroic effoits ol
these studies, many of their authors are all too often misunderstood by
readers as cataloguing examples of the “social construction” of tec li
nology. Readers account for the evidence mustered in them accoidin
to the dualist paradigm that the studies themselves frequently undi I
mine. The obstinate devotion to “social construction” as an explan I
tory device, whether by careless readers or “critical” authors, seems l<
derive from the difficulty of disentangling the various meanings ol IIim
catchword sociotechnical What I want to do, then, is to peel away, ohm
by one, these layers of meaning and attempt a genealogy of their as n
ciations.
Moreover, having disputed the dualist paradigm for years, 1 In M
come to realize that no one is prepared to abandon an arbitrary bill
useful dichotomy, such as that between society and technology, il ll
is not replaced by categories that have at least a semblance of piovld
ing the same discriminating power as the one jettisoned. Of com 0 I
will never be able to do the same political job with the pair bum ill
nonhuman as the subject-object dichotomy has accomplished, sinu» ll
was in fact to free science from politics that I embarked on this st 1«n
undertaking, as I will make clear in the next chapters. In the mcnnlllil
we can toss around the phrase “sociotechnical assemblages" loir 01
without moving beyond the dualist paradigm that we wish to leave hi»
hind. To move forward I must convince the leadei that, pending ill
resolution of the political kidnapping ol science, there is an altunal
199
to Ibe myth of progress. At the heart of the science wars lies the power
ful accusation that those who undermine the objectivity of science and
Ilie efficiency of technology are trying to lead us backward into some
piimitive, barbaric dark age—that, incredibly, the insights of science
Indies are somehow “reactionary.”
In spite of its long and complex history, the myth of progress is
1»ised on a very rudimentary mechanism (Figure 6.6). What gives the
IIn ust to the arrow of time is that modernity at last breaks out of a
»on fusion, made in the past, between what objects really are in them-
( Ives and what the subjectivity of humans believes them to be, pro-
luting onto them passions, biases, and prejudices. What could be
i tiled a front of modernization—like the Western Frontier—thus
le.irly distinguishes the confused past from the future, which will be
11101 e and more radiant, no doubt about that, because it will distin-
msh even more clearly the efficiency and objectivity of the laws of na-
Itue from the values, rights, ethical requirements, subjectivity, and
polilics of the human realm. With this map in their hands, science
111lors have no difficulty situating science studies: “Since they are al
lys insisting that objectivity and subjectivity [the science warriors’
I 1 1 n s for nonhumans and humans 1are mixed up, science students are
I idmg us in only one possible direction, into the obscure past out of
lui h we must extract ourselves by a movement of radical conversion,
Past Present Future
UK 0 (t Wh.il makes ihe arrow of time thrust forward in the m odernist narra
nt pio^iess is the certainly that the past will differ from the future because
il is lo n lu sed will become d is tu u l objectivity and subjectivity will no Ion-
la mi i d up Ih e icsult ol this ic itu m ty is a lionl of modernization that al
om to distinguish slips hni I w m l h u m steps loi waul
200
the conversion through which a barbarian premodernity becomes a

civilised modernity.”
In an interesting case of cartographic incommensurability, howevei
science studies uses an entirely different map (Figure 6.7). The arrow
of time is still there, it still has a powerful and maybe irresistible thrusl,
but an entirely different mechanism makes it tick. Instead of clarifying
even further the relations between objectivity and subjectivity, time
enmeshes, at an ever greater level of intimacy and on an ever greater
scale, humans and nonhumans with each other. The feeling of time
the definition of where it leads, of what we should do, of what war we
should wage, is entirely different in the two maps, since in the one I
use. Figure 6.7, the confusion of humans and nonhumans is not only
our past but ourfuture as well. If there is one thing of which we may In
as certain as we are of death and taxation, it is that we will live tomoi
row in imbroglios of science, techniques, and society even more tightly
linked than those of yesterday—as the mad cow affair has demon
strated so clearly to European beefeaters. The difference between tlu
two maps is total, because what the modernist science warriors see ns
a horror to be avoided at all costs—the mixing up of objectivity ami
subjectivity—is for us, on the contrary, the hallmark of a civilized life
except that what time mixes up in the future even more than in the
past are not objects and subjects at all, but humans and nonhumans, and
that makes a world of difference. Of this difference the science wai
riors remain blissfully ignorant, convinced that we want to confuse
objectivity and subjectivity.
I am now in the usual quandary of this book. I have to offer an altt'i
native picture of the world that can rely on none of the resources of
common sense although, in the end, I aim at nothing but common
sense. The myth of progress has centuries of institutionalization Ik
hind it, and my little pragmatogony is helped by nothing but my mis
erable diagrams. And yet I have to go on, since the myth of progress I
so powerful that it puts any discussion to an end.
Yes, I want to tell another tale. For my present pragmatogony*, I
have isolated eleven distinct layers. Of course I do not claim for these
definitions, or for their sequence, any plausibility. I simply want In
show that the tyranny of the dichotomy between objects and subject
is not inevitable, since it is possible to envision another myth in width
it plays no role. If I succeed in opening some space loi the imaginai ion
201
I i mii e 6.7 In the alternative “servant” narrative there is still an arrow of time, but
II is icgistered very differently from Figure 6.6: the two lines of objects and sub
it 1Is become more confused in the future than they were in the past, hence the
I 11mg of instability. What is growing instead is the ever expanding scale at which
I um.ins and nonhumans are connected together.
Ilii'ii we are not forever stuck with the implausible myth of progress. If
I i mild even begin to recite this pragmatogony—I use this word to in-
Isl on its fanciful character—I would have found an alternative to the
inylli of progress, that most powerful of all the modernist myths,
Ilie one that held my friend under its sway when he asked me, in
( liapter 1, “Do we know more than we used to?” No, we don’t know
moie, if by this expression we mean that every day we extract our-
i Ives further from a confusion between facts, on the one hand, and
01 lely, on the other. But yes, we do know a good deal more, if by this
1 mean that our collectives are tying themselves ever more deeply,
nu Me intimately, into imbroglios of humans and nonhumans. Until we
111ve «in alternative to the notion of progress, provisional as it may be,
Unee warriors will always be able to attach to science studies the in-
I niions stigma of being “reactionary.”
I will build this alternative with the strangest of means. I want to
III 'blight the successive crossovers through which humans and non-
luimims have exchanged their properties. Each of those crossovers
M nils in a dramatic change in the scale of the collective, in its compo-
IIIon, and in the degree to which humans and nonhumans are en-
im lied To tell my tale I will open Pandora’s box backward; that is,
111 ling with the most lecenl types ol folding, I will try to map the laby
202
rinth until we find the earliest (mythical) folding. As we will see, con
trary to the science warriors’ fear, no dangerous regression is involved
here, since all of the earlier steps are still with us today. Far from being
a horrifying miscegenation between objects and subjects, they are sim
ply the very hybridizations that make us humans and nonhumans.
Level 11 : Political Ecology
Talk of a crossover between techniques and politics does not, in my

pragmatogony, indicate belief in the distinction between a mateilal
realm and a social one. I am simply unpacking the eleventh layei ol
what is packed in the definitions of society and technique. The elt v
enth interpretation of the crossover—the swapping of properties lu
tween humans and nonhumans is the simplest to define because il I
the most literal. Lawyers, activists, ecologists, businessmen, pohlk il
philosophers, are now seriously talking, in the context of our ecologl
cal crisis, of granting to nonhumans some sort of rights and even leg il
standing. Not so many years ago, contemplating the sky meant think
ing of matter, or of nature. These days we look up at a sociopolilh il
imbroglio, since the depletion of the ozone layer brings togellui ti
scientific controversy, a political dispute between North and Soul ti
and immense strategic changes in industry. Political representation of
nonhumans seems not only plausible now but necessary, when the no
tion would have seemed ludicrous or indecent not long ago. We u ol
to deride primitive peoples who imagined that a disorder in souci y rt
pollution, could threaten the natural order. We no longer laugh
heartily, as we abstain from using aerosols for fear the sky may (all oil
our heads. Like the “primitives,” we fear the pollution caused by out
negligence—which means of course that neither “they” nor “we" ha
ever been primitive.
As with all crossovers, all exchanges, this one mixes elements (mil
both sides, the political with the scientific and technical, and ibis ml
ture is not a haphazard rearrangement. Technologies have laugh! M
how to manage vast assemblies of nonhumans; our newest mull
technical hybrid brings what we have learned to bear on the polllkrtl
system. The new hybrid remains a nonhuman, but not only has it In I
its material and objective character, it has ncquucd pmpeilies of till
zenship. It has, for instance, the right not to be enslaved I hi (Il I
203
I »yer of meaning—the last in chronological sequence to arrive—is that

<>l political ecology or, to use Michel Serres’s term, “the natural con-
li u l”(Serres 1995). Literally, not symbolically as before, we have to
m.mage the planet we inhabit, and must now define what I will call in
Ilie next chapter a politics of things.
Level j o : Technoscience
II I descend to the tenth layer, I see that our current definition of tech
nology is itself due to the crossover between a previous definition of
01 icly and a particular version of what a nonhuman can be. To illus-
li ili*- some time ago, at the Institut Pasteur, a scientist introduced
himself, “Hi, I am the coordinator of yeast chromosome 11.” The hy-
1Id whose hand I shook was, all at once, a person (he called himself
I ) a corporate body (“the coordinator”), and a natural phenomenon
llu genome, the DNA sequence, of yeast). The dualist paradigm will
11 11 allow us to understand this hybrid. Place its social aspect on one
lili and yeast DNA on the other, and you will bungle not only the
pi il er’s words but also the opportunity to grasp how a genome be
nnes known to an organization and how an organization is natural-
1il m a DNA sequence on a hard disk.
We again encounter a crossover here, but it is of a different sort
nul goes in a different direction, although it could also be called
1lot ethnical. For the scientist I interviewed there is no question of
1ml mg any sort of rights, of citizenship, to yeast. For him yeast is a
11UIly material entity. Still, the industrial laboratory where he works
1 place in which new modes of organization of labor elicit com-
I li ly new features in nonhumans. Yeast has been put to work for
lilt lima, of course, for instance in the old brewing industry, but
11 works for a network of thirty European laboratories where its
Imme is mapped, humanized, and socialized, as a code, a book, a
1 1 mi ol action, compatible with our ways of coding, counting,
I 11 ulmg, îetaining none of its material quality, the quality of an
I li li 1 II is absorbed into the collective. Through technoscience—
llm d loi my purposes here, as a fusion of science, organization, and
In 11y (lie loi ms ol coordination learned through “networks of
11 (see 1 evel 9) aie extended lo inailiculale entities. Nonhumans
mlowid willi speech, however pnmilive, with intelligence, fore
204
sight, self-control, and discipline, in a fashion both large-scale and in

timate. Socialness is shared with nonhumans in an almost promiscu
ous way. While in this model, the tenth meaning of sociotechnical (set
Figure 6.8), automata have no rights, they are much more than mate
rial entities; they are complex organizations.
Level 9: Networks o f Power
Technoscientific organizations, however, are not purely social, be

cause they themselves recapitulate, in my story, nine prior crossovei f
of humans and nonhumans. Alfred Chandler and Thomas Hughe*
have each traced the interpenetration of technical and social factors m
what Chandler terms the “global corporation” (Chandler 1977) and
Hughes terms “networks of power” (Hughes 1983). Here again tlu
phrase “sociotechnical imbroglio” would be apt, and one could replace
the dualist paradigm with the “seamless web” of technical and soclul
factors so beautifully traced by Hughes. But the point of my little gent
alogy is also to identify, inside the seamless web, properties borrowed
from the social world in order to socialize nonhumans and propei tic
borrowed from nonhumans in order to naturalize and expand the so
cial realm. For each layer of meaning, whatever happens happens as il
we are learning, in our contacts with one side, ontological propei lit
that are then reimported to the other side, generating new, compleli ly
unexpected effects.
The extension of networks of power in the electrical industry, In
telecommunications, in transportation, is impossible to imagine willi
out a massive mobilization of material entities. Hughes’s book is e
emplary for students of technology because it shows how a technic «I
invention (electric lighting) led to the establishment (by Edison) o( M
corporation of unprecedented scale, its scope directly related to (lie
physical properties of electrical networks. Not that Hughes in any w I
talks of the infrastructure triggering changes in the superstructui 0 oil
the contrary, his networks of power are complete hybrids, though li
brids of a peculiar sort—they lend their nonhuman qualities to wlitfl
were until then weak, local, and scattered corporate bodies. The mull
agement of large masses of electrons, clients, power stations, sulmld
iaries, meters, and dispatching rooms acquiies the foimal and unlvtl
sal character of scientific laws.
205
10th meaning of “sociotechnical”
State of social State of nonhuman

relations relations
Networks
ofpower _
Crossover
-----------------------------►
Nonhumans are organizations Technoscience
Reshuffing of intimate properties
I igure 6.8 Each step in the mythical pragmatogony may be sketched as a cross-
ovei through which skills and properties learned in social relations are made rele-
\ ml for establishing relations within nonhumans. By convention, the next step
ill be understood as going in the opposite direction.
1’his ninth layer of meaning resembles the eleventh, since in both

1 1ses the crossover goes roughly from nonhumans to corporate bod-
li s (What can be done with electrons can be done with electors.) But
Ilie intimacy of human and nonhuman is less apparent in networks
n( power than in political ecology. Edison, Bell, and Ford mobilized
( 1111lies that looked like matter, that seemed nonsocial, whereas politi-
1 11 ecology involves the fate of nonhumans already socialized, so
1losely related to us that they have to be protected by delineation of
Ilien legal rights.
Level 8: Industry
Philosophers and sociologists of techniques tend to imagine that there

I no difficulty in defining material entities because they are objec-
II 1, un problematically composed of forces, elements, atoms. Only the
01 lal, the human realm, is difficult to interpret, we often think, be-
mse it is complexly historical and, as they say, “symbolic.” But when-
i u we talk of matter we are really considering, as I am trying to show
lu K, a jxukage of former crossovers between social and natural ele-
1111n(s, so that what we take to be primitive and pure terms are be-
I ill <! and mixed ones Alieady we have seen that matter varies greatly
206
from layer to layer—matter in the layer I have called “political ecol
ogy” differs from that in the layers called “technology” and “networks
of power.” Far from being primitive, immutable, and ahistorical, mat
ter too has a complex genealogy and is handed down to us through a
long and convoluted pragmatogony.
The extraordinary feat of what I will call industry is to extend to
matter a further property that we think of as exclusively social, the ca
pacity to relate to others of one’s kind, to conspecifics, so to speak.
Nonhumans have this capacity when they are made part of the assem
bly of actants that we call a machine: an automaton endowed with au
tonomy of some sort and submitted to regular laws that can be mea
sured with instruments and accounting procedures. From tools held in
the hands of human workers, the shift historically was to assemblies
of machines, where tools relate to one another, creating a massive ar
ray of labor and material relations in factories that Marx described as
so many circles of hell. The paradox of this stage of relations be
tween humans and nonhumans is that it has been termed “alienation,”
dehumanization, as if this were the first time that poor and exploited
human weakness was confronted by an all-powerful objective force
However, to relate nonhumans together in an assembly of machines,
ruled by laws and accounted for by instruments, is to grant them a soi I
of social life.
Indeed, the modernist project consists in creating this peculiar hy
brid: a fabricated nonhuman that has nothing of the character of sou
ety and politics yet builds the body politic all the more effectively
because it seems completely estranged from humanity. This famous
shapeless matter, celebrated so fervently throughout the eighteen ill
and nineteenth centuries, which is there for Man’s—but rarely
Woman’s—ingenuity to mold and fashion, is only one of many ways lo
socialize nonhumans. They have been socialized to such an extent Iha I
they now have the capacity to create an assembly of their own, an au
tomaton, checking and surveying, pushing and triggering other au
tomata, as if with full autonomy. In effect, however, the properties ol
the “megamachine” (see Level 7) have been extended to nonhumans
It is only because we have not undertaken an anthropology ol oui
modern world that we can overlook the strange and hybrid cjualily ol
matter as it is seized and implemented by industiy. We lake mal Ic 1
as mechanistic, forgetting that mechanism is one hall ol the moduli
207
definition of society*. A society of machines? Yes, the eighth mean

ing of the word sociotechnical, though it seems to designate an
unproblematic industry, dominating matter through machinery, is the
strangest sociotechnical imbroglio yet. Matter is not a given but a re
cent historical creation.
Level 7: The Megamachine
But where does industry come from? It is neither a given nor the sud
den discovery by capitalism of the objective laws of matter. We have to
imagine its genealogy through earlier and more primitive meanings of
lhe term sociotechnical. Lewis Mumford has made the intriguing sug
gestion that the megamachine—the organization of large numbers of
humans via chains of command, deliberate planning, and accounting
procedures—represents a change of scale that had to be made before
wheels and gears could be developed (Mumford 1966). At some point
in history human interactions come to be mediated through a large,
stratified, externalized body politic that keeps track, through a range
of “intellectual techniques” (writing and counting, basically), of the
many nested subprograms for action. When some, though not all, of
ihese subprograms are replaced by nonhumans, machinery and facto-
1ies are born. The nonhumans, in this view, enter an organization that
is already in place and take on a role rehearsed for centuries by obedi
ent human servants enrolled in the imperial megamachine.
In this seventh level, the mass of nonhumans assembled in cities
By an internalized ecology (I will define this expression shortly) has
been brought to bear on empire building. Mumford’s hypothesis is de
batable, to say the least, when our context of discussion is the history
ol technology; but the hypothesis makes excellent sense in the con-
lex t of my pragmatogony. Before it is possible to delegate action to
1ion humans, and possible to relate nonhumans to one another in an
milomaton, it must first be possible to nest a range of subprograms for
ml ion into one another without losing track of them. Management,
Mumford would say, precedes the expansion of material techniques.
Moic in keeping with the logic of my story, one might say that when-
m r we learn something about the management of humans, we shift that
knowledge to nonhumans and endow them with more and more organiza
tional properties I lie even numhcicd episodes I have recounted so far
208
follow this pattern: industry shifts to nonhumans the management

of people learned in the imperial machine, much as technoscienu
shifts to nonhumans the large-scale management learned through net
works of power. In the odd-numbered levels, the opposite process is
at work: what has been learned from nonhumans is reimported so as lo
reconfigure people.
Level 6: Internalized Ecology
In the context of layer seven, the megamachine seems a pure and even
final form, composed entirely of social relations; but, as we reach layc 1
six and examine what underlies the megamachine, we find the most
extraordinary extension of social relations to nonhumans: agriculluit
and the domestication of animals. The intense socialization, reeducn
tion, and reconfiguration of plants and animals—so intense that tluy
change shape, function, and often genetic makeup—is what I mean by
the term “internalized ecology.” As with our other even-numbered lev
els, domestication cannot be described as a sudden access to an objet
tive material realm that exists beyond the narrow limits of the sot ial
In order to enroll animals, plants, proteins in the emerging collect ivt
one must first endow them with the social characteristics necessai y
for their integration. This shift of characteristics results in a manmadt
landscape for society (villages and cities) that completely alters wh il
was until then meant by social and material life. In describing tilt»
sixth level we may speak of urban life, empires, and organizations, but
not of society and techniques—or of symbolic representation and In
frastructure. So profound are the changes entailed at this level th.it we
pass beyond the gates of history and enter more profoundly those ol
prehistory, of mythology.
Levels: Society
What is a society, the starting point of all social explanations, the a pi I

ori of all social science? If my pragmatogony is even vaguely sugge
tive, society cannot be part of our final vocabulary, since the tei m hml
itself to be made—“socially constructed” as the misleading expiestlnn
goes. But according to the Durkheimian inteipietation, a society I
primitive indeed: it precedes individual action, lasts veiy much longer
209
llun any interaction does, dominates our lives; it is that in which we

110 born, live, and die. It is externalized, reified, more real than our-
1Ives, and hence the origin of all religion and sacred ritual, which for
Dm kheim are nothing but the return, through figuration and myth, of
Ilie transcendent to individual interactions.
And yet society itself is constructed only through such quotidian in-
Ui actions. However advanced, differentiated, and disciplined society
Incomes, we still repair the social fabric out of our own, immanent
I nowledge and methods. Durkheim may be right, but so is Harold
( 1.11linkel. Perhaps the solution, in keeping with the generative princi
ple of my genealogy, is to look for nonhumans. (This explicit principle
I look for nonhumans when the emergence of a social feature is inex-
I>1ilable; look to the state of social relations when a new and inexplica
ble' lype of object enters the collective.) What Durkheim mistook for
IIn* effect of a sui generis social order is simply the effect of having
In ought so many techniques to bear on our social relations. It was
horn techniques, that is, the ability to nest several subprograms, that
1 learned what it means to subsist and expand, to accept a role and
II charge a function. By reimporting this competence into the
I»bullion of society, we taught ourselves to reify it, to make society
I mcl independent of fast-moving interactions. We even learned how
In delegate to society the task of relegating us to roles and functions,
nuely exists, in other words, but is not socially constructed.
nnhumans proliferate below the bottom line of social theory.
Level 4: Techniques
I IIns stage in our speculative genealogy we can no longer speak of

humans, of anatomically modern humans, but only of social pre-
liumuns. At last we are in a position to define technique, in the sense
I 1modus operandi, with some precision. Techniques, we learn from
mliK ologists, are articulated subprograms for actions that subsist (in
lim ) mid extend (in space). Techniques imply not society (that late-
I cloping hybrid) but a semisocial organization that brings together
II uihiimans fiom very different seasons, places, and materials. A bow
md mow, a javelin, a hammer, a net, an article of clothing are com-
| o (d ol pails and pieces that lequiie recombination in sequences of
Him iml space that bear no 1elation to then oiiginal settings. Tech-
210
niques are what happen to tools and nonhuman actants when they are
processed through an organization that extracts, recombines, and so
cializes them. Even the simplest techniques are sociotechnical; even at
this primitive level of meaning, forms of organization are inseparable
from technical gestures.
Level 3: Social Complication
But what form of organization can explain these recombinations? Re

call that at this stage there is no society, no overarching framework, no
dispatcher of roles and functions; there are merely interactions among
prehumans. Shirley Strum and I call this third layer of meaning social
complication (Strum and Latour 1987). Here complex interactions arc
marked and followed by nonhumans enrolled for a specific purpose
What purpose? Nonhumans stabilize social negotiations. Nonhumans
are at once pliable and durable; they can be shaped very quickly but
once shaped, last far longer than the interactions that fabricated them
Social interactions are extremely labile and transitory. More precisely,
either they are negotiable but transient or, if they are encoded (for in
stance) in the genetic makeup, they are extremely durable but difficult
to renegotiate. The involvement of nonhumans resolves the contradic
tion between durability and negotiability. It becomes possible to loi
low (or “blackbox”) interactions, to recombine highly complicated
tasks, to nest subprograms into one another. What was impossible
for complex* social animals to accomplish becomes possible loi
prehumans—who use tools not to acquire food but to fix, underline
materialize, and keep track of the social realm. Though composed only
of interactions, the social realm becomes visible and attains through
the enlistment of nonhumans—tools—some measure of durability
Level 2: The Basic Tool Kit
The tools themselves, wherever they came from, offer the only lest I
mony on behalf of hundreds of thousands of years. Many archaeolo
gists proceed on the assumption that the basic tool kit (as I call it) and
techniques are directly related by an evolution of tools into composite
tools. But there is no direct route from Hints to nuclear powei pi.ml
Further, there is no direct route, as many social thconsts piesunu
211
(here to be, from social complication to society, megamachines, net

works. Finally, there is not a set of parallel histories, the history of in-
11astructure and the history of superstructure, but only one
sociotechnical history (Latour and Lemonnier 1994).
What, then, is a tool? The extension of social skills to nonhumans.
Machiavellian monkeys and apes possess little in the way of tech
niques, but can devise social tools (as Hans Kummer has called them;
I ummer 1993) through complex strategies of manipulating and modi-
lying one another. If you grant the prehumans of my own mythology
llie same kind of social complexity, you grant as well that they may
generate tools by shifting that competence to nonhumans, by treating a
lone, say, as a social partner, modifying it, then using it to act on a
II ond stone. Prehuman tools, in contrast to the ad hoc implements of
I>1lier primates, also represent the extension of a skill rehearsed in the
II aim of social interactions.
Level 1: Social Complexity
We have finally reached the level of the Machiavellian primates, the

11 I circumvolution in Daedalus’s maze. Here they engage in social in-
li 1actions to repair a constantly decaying social order. They manipu-
I ik' one another to survive in groups, with each group of conspecifics
in a state of constant mutual interference (Strum 1987). We call this
I ilc, this level, social complexity. I will leave it to the ample literature
•I 1>t unatology to show that this stage is no more free of contact with
li «ils and techniques than any of the later stages (McGrew 1992).
An Impossible but Necessary Recapitulation
I I now I should not do it. I more than anyone ought to see that it
I madness, not only to peel away the different meanings of
a lolechnical, but also to recapitulate all of them in a single diagram,
1 II we could read off the history of the world at a glance. And yet it is
il lys sui prising to see how few alternatives we have to the grandiose
i nogiaphy o( progress. We may tell a lugubrious countertale of de-
1 (rit 1 decadence as if, at each step in the extension of science and
I i Imology, we weie stepping down, away from our humanity. This is
h il I leideggci did, and his account has the somber and powerful ap
2.12
peal of all tales of decadence. We may also abstain from telling any
master narrative, under the pretext that things are always local, histor
ical, contingent, complex, multiperspectival, and that it is a crime to
hold them all in one pathetically poor scheme. But this ban on mastei
narratives is never very effective, because, in the back of our minds,
no matter how firmly we are convinced of the radical multiplicity of
existence, something surreptitiously gathers everything into one little
bundle which may be even cruder than my diagrams—including
the postmodern scenography of multiplicity and perspective. This is
why, against the ban on master narratives, I cling to the right to tell a
“servant” narrative. My aim is not to be reasonable, respectable, 01
sensible. It is to fight modernism by finding the hideout in which sc i
ence has been held since being kidnapped for political purposes I do
not share.
If we gather in one table the different layers I have briefly outlined
one of my other excuses is how brief the survey, covering so many mil
lions of years, has been!—we may give some sense to a story in which
the further we go the more articulated are the collectives we live in
(see Figure 6.9). To be sure, we are not ascending toward a futuic
made of more subjectivity and more objectivity. But neither are we dt
scending, chased ever further from the Eden of humanity and poesis
Even if the speculative theory I have outlined is entirely false, 11
shows, at the very least, the possibility of imagining a genealogical aI
ternative to the dualist paradigm. We are not forever trapped in a boi
ing alternation between objects or matter and subjects or symbols We
are not limited to “not only . . . but also” explanations. My little 01 igln
myth makes apparent the impossibility of having an artifact that clot
not incorporate social relations, as well as the impossibility of defining
social structures without accounting for the large role played in them
by nonhumans.
Second, and more important, the genealogy demonstrates thaï II I
false to claim, as so many do, that once we abandon the dichotomy lie
tween society and techniques we are faced with a seamless web ol ( it
tors in which all is included in all. The properties of humans tuitl
nonhumans cannot be swapped haphazardly. Not only is there nn 01
der in the exchange of properties, but in each of the eleven layeis 1lie*
meaning of the word “sociotechnical” is clarified if we consider the c
change: that which has been learned from nonhumans and leimpoi led
213
State of social State of nonhuman

relations Crossover relations
1st
Social
complexity 1 Social tools
pliability
durability Basic tool kit 2nd
Social
ird articulation
complication
extemalization Techniques 4th
<;th Society domestication
Internalized
reification
ecology 6th
large-scale
th Megamachine 1 management
automaton Industry 8th

Networks of extension
yth rearticulation
power
object-institutions Technoscience 10th
•<-----------------------
n th
Political
ecology i politics of nature
I Igmc 6.9 If the successive crossovers are summed up, a pattern emerges: rela-
II ns among humans are made out of a previous set of relations that related
11 mliumans to one another; these new skills and properties are then reused to pat-
I 111 new types of relations among nonhumans, and so on; at each (mythical) stage
IIn i ale and the entanglement increase. The key feature of this myth, is that, at
lilt I1n.1l stage, the definitions we can make of humans and nonhumans should re-
ipllul.ilc all the earlier layers of history. The further we go, the less pure are the
hull ions of humans and nonhumans.
lulu Ilie social realm, that which has been rehearsed in the social
I dm and exported back to the nonhumans. Nonhumans too have a
III loi y I hey are not material objects or constraints. Sociotechnical1
I dlllcicnt Irom sociotechnical6 or7 or8 or". By adding superscripts
t in able to qualify the meanings of a term that until now has been
It |u Ussly confused In place of the gieat vertical dichotomy between
»It ly and techniques, llicie Is conceivable (in fact, now, available) a
PANDORA'S HOPE
214
range of horizontal distinctions between very different meanings ol

the sociotechnical hybrids. It is possible to have our cake and eal il
too—to be monists and make distinctions.
All this is not to claim that the old dualism, the previous paradigm
had nothing to say for itself. We do indeed alternate between stales
of social and states of nonhuman relations, but this is not the saim
as alternating between humanity and objectivity. The mistake of Ilu
dualist paradigm was its definition of humanity. Even the shape of hu
mans, our very body, is composed to a great extent of sociotechnical
negotiations and artifacts. To conceive of humanity and technology
as polar opposites is, in effect, to wish away humanity: we au1
sociotechnical animals, and each human interaction is sociotechnical
We are never limited to social ties. We are never faced only with ol)
jects. This final diagram relocates humanity right where we belong
in the crossover, the central column, the articulation, the possibility ol
mediating between mediators.
But my main point is that, in each of the eleven episodes I liavw
traced, an increasingly large number of humans are mixed with an III
creasingly large number of nonhumans, to the point that, today, lliM
whole planet is engaged in the making of politics, law, and soon, 1111
pect, morality. The illusion of modernity was to believe that the mow
we grew, the more separate objectivity and subjectivity would bcconu*
thus creating a future radically different from our past. After the p 11it
digm shift in our conception of science and technology, we now I no
that this will never be the case, indeed that this has never ban til
case. Objectivity and subjectivity are not opposed, they grow toge Ilui
and they do so irreversibly. At the very least, I hope I have convinu I
the reader that, if we are to meet our challenge, we will not meel It I
considering artifacts as things. They deserve better. They deserve fill
housed in our intellectual culture as full-fledged social actoi s I )o Ilie
mediate our actions? No, they are us. The goal of our philosophy
cial theory, and morality is to invent political institutions (hill util mI
sorb this much history, this vast spiraling movement, this lubyillllh
this fate.
The nasty problem we now have to deal with is that, unloi lundi el
we do not have a definition of politics that can answei lhe* spulfl 4
tions of this nonmodern history. On the conitmy, evuy 111 I
definition we have of politics comes liom ihe modelmst Ntlllvm 11
215
and from the polemical definition of science that we have found so

wanting. Every one of the weapons used in the science wars, includ
ing the very distinction between science and politics, has been handed
down to the combatants by the side we want to oppose. No wonder we
11ways lose and are accused of politicizing science! It is not only the
piactice of science and technology that epistemology has rendered
opaque, but also that of politics. As we shall soon see, the fear of mob
Mile, the proverbial scenography of might versus right, is what holds
Ilie old settlement together, is what has rendered us modern, is what
lias kidnapped the practice of science, all for the most implausible po
lit ical project: that of doing away with politics.
C H A P T E R S E V E N
The Invention of the Science Wars

The Settlement o f Socrates and Callicles
“If Right cannot prevail, then Might will take over!” How often haw
we heard this cry of despair? How sensible it is to cry for Reason III
this way when faced with the horrors we witness every day. And yt I
this cry too has a history, a history that I want to probe because doln
so may allow us to distinguish science from politics once again anti
maybe to explain why the Body Politic has been invented in such a w I
as to be rendered impossible, impotent, illegitimate, a born bastaul
When I say that this rallying cry has a history, I do not mean thaï ll
moves at a fast pace. On the contrary, centuries may pass without aI
fecting it a bit. Its tempo resembles that of Fermat’s theorem, 01 pi lie
tectonics, or glaciations. Witness for instance the similarity between
Socrates’ vehement address to the Sophist Callicles, in the famous (lid
logue of the Gorgias, and this recent instance by Steven Wei nIwig III
the New York Review of Books:
Our civilization has been powerfully affected by the discovery (h ll
nature is strictly governed by impersonal laws . . . We will need lu
confirm and strengthen the vision of a rationally understand ihld
world if we are to protect ourselves from the irrational tendcm le
that still beset humanity. (August 8,1996,15)
And here is Socrates’ famed admonition: geômetriasgar amt l<i I
In fact, Callicles, the expert’s opinion is that co opeiation, love 01
der, discipline, and justice bind heaven and earth, gods and nit n 1h ll
why they call the universe an ordered whole, my li lend, rnllui III III d
disorderly mess or an unruly shambles It seems to me (liai loi dll
216
THE INVENTION OF THE SCIENCE WARS
217
your expertise in the field, you’re overlooking this point. You’ve

Iailed to notice how much power geometrical equality has among gods
and men, and this neglect of geometry has led you to believe that one
should try to gain a disproportionate share of things. (507e-5o8a)
What these two quotations have in common, across the huge gap of
11nturies, is the strong link they establish between the respect for im-
| u 1sonal natural laws, on the one hand, and the fight against irratio-
inlity, immorality, and political disorder, on the other. In both quota-
llons the fate of Reason and the fate of Politics are associated in a
Ingle destiny. To attack Reason is to render morality and social peace
Impossible. Right is what protects us against Might; Reason against
Ivil warfare. The common tenet is that we need something “inhu-
m m”—for Weinberg, the natural laws no human has constructed; for
1u 1«îles, geometry whose demonstrations escape human whim—if we
ml to be able to fight against “inhumanity.” To sum up: only inhu-
in m11y will quash inhumanity. Only a Science that is not made by
in m will protect a Body Politic that is in constant risk of being made
I l lie mob. Yes, Reason is our rampart, our Great Wall of China, our
1 igmot Line against the dangerous unruly mob.
lins line of reasoning, which I will call “inhumanity against inhu-
lii inily,” has been attacked ever since it began, from the Sophists,
N ilnsl whom Plato launches his all-out assault, all the way to the mot-
I y mg of people accused of “postmodernism” (an accusation, by the
i ns vague as the curse of being a “sophist”). Postmoderns of the
1 I 11id of the present have tried to break the connection between the
I 1oveiy of natural laws of the cosmos and the problems of making
I Body Politic safe for its citizens. Some have claimed that adding in-
iim inily lo inhumanity has simply increased the misery and the civil
lllit mil that a staunch fight against Science and Reason should be
III It d lo protect politics against the intrusion of science and technol-
Sl ill others, who are targeted publicly today and with whom, I am
II It) s.iy, I am often lumped by mistake, have tried to show that
111 nie*. Ihe violence of the Body Politic, is everywhere polluting the
llll ol Science, which becomes every day more human, all too hu-
Mll md eveiy day mote adulterated by the civil strife it was sup-
d lo assuage Olheis, like Niet/sche, have shamelessly accepted
III It position and claimed, against (lie degeneiate and moralistic
218
Socrates, that only violence could bend both the mob and its retinue ol
priests and other men of ressentiment, among whom, I am sorry to say,
he included scientists and cosmologists like Weinberg.
None of these critiques, however, has disputed simultaneously the
definition of Science and the definition of the Body Politic that it im
plies. Inhumanity is accepted in both or in at least one of them. Only
the connection between the two, or its expediency, has been disputed
In this chapter and the next I want to go back to the source of what 1
call the scenography of the fight of Right against Might, to see how it
was staged in the first place. I want, in other words, to attempt the ar
chaeology of the Pavlovian reflex that makes any lecture in science
studies trigger these questions from the audience: “Then you want
force alone to decide in matters of proof? Then you are for mob rule
against that of rational understanding?” Is there really no other way?
Is it really impossible to build up other reflexes, other intellectual re
sources?
To go some way toward this genealogy, no text is more appropriate
than the Gorgias, especially in the lively translation by Robin Wa
terfield (Oxford University Press, 1994), since never was the genealogy
more beautifully set up than in the acrimonious debate between Soc
rates and Callicles, which has been commented on by all the lalci
Sophists from Greece and then from Rome, as well as, in our time, by
thinkers as different as Charles Perelman and Hannah Arendt. I am
not reading the Gorgias as if I were a Greek scholar (I am not, as will
become painfully clear) but as if it had been published a few mont hi
ago in the New York Review of Books as a contribution to the raging Sd
ence Wars. Fresh as in 385 b . c ., it deals with the same puzzle as the one
besetting the academy and our contemporary societies today.
This puzzle can be stated very simply: the Greeks made one invent
tion too many! They invented both democracy and mathematical
demonstration, or to use the terms Barbara Cassin comments on so
beautifully, epideixis* and apodeixis* (Cassin 1995). We are still sling
gling, in our “mad cow times,” with this same quandary, how to have a
science and a democracy together. What I call the settlement between
Socrates and Callicles has made the Body Politic incapable of swallow
ing the two inventions at once. More fortunate than the Gieeks, we
maybe able, if we rewrite this settlement, to profit at last liom both
To revisit this “primal scene” of Might and Right, I am all aid we
219
have to follow the dialogue in some detail. The structure of the story is
clear. Three Sophists in turn oppose Socrates and are defeated one af-
ler the other: Gorgias, a bit tired from a lecture he just gave; Polus, a
hit slow; and finally the harshest of the three, the famous and infa
mous Callicles. At the end, Socrates, having discouraged discussion,
speaks to himself and makes a final appeal to the shadows of the after-
world, the only ones able to understand his position and to judge it—
with good reason, as we shall see.
In my commentary I will not always follow the chronological order
ol the dialogue and will focus mainly on Callicles. I want to point out
Iwo features of the discussion that, in my view, have often been over
looked. One is that Socrates and his third opponent, Callicles, agree
on everything. Socrates’ invocation of reason against the unreasonable
people is actually patterned on Callicles’ request for “an unequal share
ol power.” The second feature is that it is still possible to recognize
m lhe four protagonists’ speeches the dim trace of the conditions of fé
tu ily * that are proper to politics and that both Callicles and Socrates
(.is characters in Plato’s puppet show at least) have tried their best
II>erase. This will be the focus of Chapter 8, in which I will try to show
Ilui the Body Politic could behave very differently if another defini-
11011 of science and of democracy were provided. A science freed at last
IImn its kidnapping by politics? Even better, a polity freed at last from
Ils delegitimation by science? It is certainly, everyone would admit,
v 01 ih a try.
Socrates and Callicles versus the People of Athens

The Demotic Hatred
V e «ire so used to opposing Might and Right and to looking in the

I in {'ins for their best instantiations that we forget to notice that Socra-
li .uid Callicles have a common enemy: the people of Athens, the
mwd assembled in the agora, talking endlessly, making the laws
il Iheir whim, behaving like children, like sick people, like animals,
hilling opinions whenever the wind changes direction. Socrates ac-
11 1s Goigias and then Polus of being the slaves of the people, or of
Ik lug like Callicles, uiuble lo ullei olher words than those the raging
mwd puls m his mouth Ihil Cnllules loo, when it is his turn lo talk.
220
accuses Socrates of being enslaved by the people of Athens and of foi

getting what makes noble masters superior to the hoi polloi: “You pre
tend that truth is your goal, Socrates, but in actual fact you steel
discussions towards this kind of ethical idea—ideas which are unso
phisticated enough to have popular appeal, and which depend entirely
on convention, not on nature” (482e).
The two protagonists compete in trying to avoid being branded
with that fatal accusation: resembling the people, the common people,
the menial manual people of Athens. As we will see, they soon disagree
on how best to break the majority rule, but the goal of breaking the
rule of the crowd remains beyond question. Witness this exchange, in
which a condescending and tired Callicles seems to lose the contes I
over how much distance one should keep from the demos:
c a l l ic l e s :I can’t explain it, Socrates, but I do think you’re making
your points well. All the same, I’m feeling what people invariably feel
with you: I’m not entirely convinced.
s o c r a t e s : It’s the demotic love residing in your heart which is resisting
me, Callicles. (513c)
Obviously the love of the people is not stifling Socrates’ breath 1 1It*
has a way to break the rule of majority that no obstacle can restrain
What should we call what resists in his heart if not “demotic hatred’ I
If you make a list of all the derogatory terms with which the common
crowd is branded by Callicles and Socrates, it is hard to see which ol
them despises it most. Is it because assemblies are polluted by women
children, and slaves that they deserve this scorn? Is it because they ui e
made up of people who work with their hands? Or is it because they
switch opinions like babies and want to be spoiled and overfed like it
responsible children? All of that, to be sure, but their worst quality, loi
our two protagonists, is even more elementary: the great constitutive
defect of the people is that there are simply too many of them A
rhetorician, then,” says Socrates with his tranquil arrogance, “isn I
concerned to educate the people assembled in lawcourts and so on
about right and wrong; all he wants to do is persuade them. I mean I
shouldn’t think it’s possible for him to get so many people to uiuli 1
stand such important matters in such a short time” (455a).
Yes, there are too many of them, the questions are too impoil ml
221
megala pragmata], there is too little time [oligô chronô]. Are these not,
however, the normal conditions of the Body Politic? Is it not to deal
with these peculiar situations of number, urgency, and priority that
Ihe subtle skills of politics were invented? Yes, as we shall see in Chap-
ler 8, but this is not the tack that Socrates and Callicles take. Horror-
.11uck by the numbers, the urgency, and the importance, they agree on
mother radical solution: break the majority rule and escape from it. It
Is at this juncture that the fight between Might and Right is being in-
vented, the commedia dell’arte scenography that is going to entertain so
many people for so long.
because of the clever staging by Plato (so clever that it continues
t ven today in the campus amphitheatres) we have to distinguish be-
Iween two roles played by Callicles, so that we don’t attribute to the
Sophists the position in which Socrates is trying to corner them—a
position they kindly accept because Plato is holding all the puppet
lungs of the dialogue at once. Believing what Plato says of the Soph-
I Is would be like reconstituting science studies from the science war-
i loi s’ pamphlets! I will thus call the Callicles playing the role of a foil
loi Socrates the straw Callicles. The Callicles that retains features of
IIn1precise conditions of felicity invented by the Sophists and still visi-
Iik in the dialogue, I will call the positive, or the historical, or the an-
I topological Callicles. While the straw Callicles is a strong enemy of
llii1demos and the perfect counterpart for Socrates, the anthropologi-
ii I Callicles will allow us to retrieve some of the specificities of politi-
il li uth-saying.
How Best to Break the Majority Rule
( ill it les’ solution is well known. It is the age-old aristocratic solution,

| 11soiled in a crisp and naive light by the Nietzschean blond brute de-
11 tided from a race of masters. But we should not be taken in by what
II ippens on the stage. Callicles is not for Might understood as “mere
ii u ’ bul for something, on the contrary, that will make might weak.
I li Is looking for a might mightier than might. We should follow with
mu piecision the tricks that Callicles employs, because, in spite of
III neei mg remarks, it is on the bad guy that the good guy, Socrates,
I going to pattern his copycat solution to the same problem: for both,
222
beyond the conventional laws made for and by the mob, there is an
other natural law reserved for the elite, which makes the noble souls
unaccountable to the demos.
In a visionary anticipation of certain aspects of sociobiology
Callicles appeals to nature above manmade history:
But I think we only have to look at nature to find evidence that it is
right for better to have a greater share than worse, more capable than
less capable. The evidence for this is widespread. Other creatures
show, as do human communities and nations, that right has been de
termined as follows: the superior person shall dominate the inferior
person and have more than him . . . These people act, surely, in con
formity with the natural essence [kata phusin] of right and, yes. I’d
even go so far as to say that they act in conformity with natural law
[kata nomon ge tès phuseôs], even though they presumably contravem
our man-made laws. (483c-e)
As Socrates and Callicles immediately see, however, this is nol it
sufficient definition of Might, for a simple and paradoxical reason
Callicles who appeals to the superior natural law is nonetheless physl
cally weaker than the crowd. “Presumably you don’t think that two
people are better than one, or that your slaves are better than you, ju t
because they’re stronger than you” (489d), says Socrates ironically. "()|
course,” says Callicles, “I mean that superior people are better. Haven 11
been telling you all along that ‘better’ and ‘superior’ are the same, III
my opinion? What else do you think I’ve been saying? That law toll
sists of the statements made by an assembly of slaves and assorted oihtr
forms of human debris who could be completely discounted if it wei t n I
for the fact that they do have physical strength at their disposal” (4891 )
We should be careful here not to introduce the moral argument (lull
will come later, and we should focus only on Callicles’ way of estupln
the rule of the majority. His appeal to irrepressible natural law exaill
resembles the “inhumanity to quash inhumanity” with which I still tfil
this chapter. Stripped of its moral dimension, which will be udiki]
later in the dialogue in the interests of staging, not of logic, Cnlliiltf
plea becomes a moving appeal to a force stronger than the demoi mill
force of the assembled people, a force beautifully defined by S o u til?
when he summarizes Callicles’ position:
Here’s your position, then a single 1lever pci son is nlinn I
so c r a tes:
hound to he superior to ten thousandJools, political powet should be III
223
and they should be his subjects; and it is appropriate for someone with
political power to have more than his subjects. Now, I’m not picking
on the form of words you used, but that, I take it, is the implication of
what you’re saying—of a single individual being superior to ten thou
sand others.
: Yes, that’s what I mean. In my opinion, that’s what natural
c a l l ic l e s
right is—for an individual who is better (that is, more clever) to rule
over second-rate people and to have more than them. (490a)
Thus when Might enters the scene in the person of the Nietzschean
t aHides, it is not as the Brownshirts smashing their way through the
I iboratories—as in the nightmares epistemologists have when they
llnnk of science studies—it is as an elitist and specialized expertise
111caking the neck of mob rule and imposing a Right superior to all the
uuiventional property rights. When Might is invoked on the stage it is
not as a crowd against Reason, it is as one man against the crowd,
tgainst myriad fools. Nietzsche has deftly drawn the moral of this par-
i(lo\ in his famous advice: “One should always defend the strong
I *mist the weak.” Nothing is more elitist than the nightmarish Might.
I lie model employed by Callicles is of course nobility, the aristo-
II 11ic upbringing to which Plato himself, as has been so often noticed,
1 es his virtue. Nobility gives an ingrained quality and a native status
lli il makes masters different from the hoi polloi. But Callicles shifts
line lassie pattern considerably by complementing upbringing with an
ippeal to a law that is superior to the law. Elites are defined not only
I» Iheir past and their ancestors but also by their connection to this
1 il m al law that does not depend on the “social construction” made by
I ives We are so used to laughing when Callicles falls into all the traps
II by Socrates that we fail to see how similar are the roles both offer
m 11repressible natural law that is not manmade: “What do we do
till Ihe best and strongest among us?” asks Callicles.
Wi capture them young, like lions, mould them, and turn them into
at1 s by chanting spells and incantations over them which insist
lli il Ihey have to be equal to others and that equality is admirable
nul 1ight But I’m sure that if a man is born in whom nature is strong
• Hough, he’ll shake off all these limitations, shatter them to pieces, and
In Ins lieedotn; he’ll trample all our regulations, charms, spells, and
1natural laws into Ihe dusl, this slave will rise up and reveal himself
1 oui maslei, and then natural tn*ht /to tes phuseos dikaion] will bla/e
ItMill (^Hu* /|H/|b)
224
This sort of sentence has done a lot for Callicles’ bad reputation,
and yet it is the same irrepressible urge that even bad education can
not spoil will “shake off’ irrationality and “blaze forth” when Socrates
defeats his ten thousand fools. If you remove from Callicles the cloal
of immorality, if you make him swap offstage his brutish and hail y
wig for the virginal white cloth of Antigone, we will be forced to no
tice that his plea possesses the same beauty as hers against Creon, ovet
which so many moral philosophers have shed so many tears. Both say
that deformation by the “social construction” cannot stop the natu
ral law from “blazing forth” in the hearts of naturally good people. In
the long run, the noble hearts will triumph over manmade conven
tions. We despise the Callicleses and we praise the Socrateses and tin?
Antigones, but this is to hide the simple fact that they all wish to stand
alone against the people. We complain that without Right the war ol
all against all will take over, but we fail to notice this war of two, Sou «
tes and Callicles, against all the others.
With this little warning in mind, we can now listen to Socrates’ so
lution with a different ear. On the stage, to be sure, he has a field day
when ridiculing Callicles’ appeal to an unlimited Might: “Would you
go back to the beginning, though, and tell me again what you and
Pindar mean by natural right? Am I right in remembering that accoi d
ing to you it’s the forcible seizure of property belonging to inferior pi 0
pie by anyone who is superior, it’s the dominance of the worse by t 1im
better, and it’s the unequal distribution of goods, so that the élite have
more than second-rate people?” (488b).
The entire audience screams in horror when confronted with (III
threat of Might swallowing the rights of ordinary citizens. But how I
Socrates’ own solution technically different? Again, let the pailnil
stay on the stage for a moment in plain clothes, without the impi essl 1»
garments of morality, and listen carefully to Socrates’ definition ol
how to resist the same assembled crowd. This time it is the pool Poll!
who suffers the sting of the numbfish:
The trouble, Polus, is that you’re trying to use on me the kind of 1he
torical refutation which people in lawcourts think is successful lime
too, you see, people think they’re proving the oilier side wioug II
they produce a large number of eminent witnesses m suppoil ol (he
points they’re making, but iheir opponent comes up will) only a sin
225
gle witness or none at all. This kind of refutation, however, is com

pletely worthless in the context of the truth [outos de o elegchos oudenos
axios estinpros tèn alètheian], since it’s perfectly possible for someone
to be defeated in court by a horde of witnesses with no more than ap
parent respectability who all testify falsely against him. (47ie-472a)
How often his position has been admired! How many voices have
quivered in commenting on the courage of one man against the
hordes, like Saint Genevieve stopping Attila’s throngs with the sheer
light of her virtue! Yes, it is admirable, but no more than Callicles’ ap
peal to a natural law. The goal is the same, and even Callicles, in his
ildest definition of forceful domination, never dreams of a position
ol power as dominant, as exclusive, as undisputed as the one Socrates
u quests for his knowledge. It is a great power to which Socrates ap-
pi .ils, comparing it to the physician’s knowledge of the human body
nice it can enslave all the other forms of expertise and know-how:
Nicy don’t realize that this kind of expertise should properly be the
Ion1inant kind, and should be allowed a free hand with the products
il .ill those other techniques because it knows—and none of the others
lues—which food and drink promotes a good physical state and
Inch doesn’t. That’s why the rest of them are suited only for slavish,
on diary, and degrading work, and should by rights be subordinate to
11lining and medicine” (5i7e-5i8a).
11ulh enters and the agora is emptied. One man can triumph over
11yone else. In the “context of truth,” as in the “context of aristoc-
l 11y,” ihe hordes are defeated by a force—yes a force—superior to the
1 pi 11.11ion and physical force of the demos and to its endless and use-
I piactical knowledge. When Might comes onstage, as I said above,
I 1 not as a crowd but as one man against the crowd. When Truth en-
1 Ilie scene, it is not as one man against everyone else, it is as an im-
| 1 nn.il, transcendent natural law, a Might mightier than Might. Ar-
iiiiu nls prevail against everything else because they are rationally
ikIi This is what Callicles has missed: the power of geometrical
jiiilily “You neglected geometry, Callicles!” The young man will
11 1ecover from the blow.
I li it Callicles and Sociales aie acting like Siamese twins in this dia-
iii is made explicit by Plato’s many parallels between his heroes’
I 1 olulions Sociales computes C allules’ slavish attachment to the
PANDORA'S HOPE
226
demos with his own slavish attachment to philosophy: “I love

Alcibiades the son of Cleinias, and philosophy, and your two loves ai 1
the Athenian populace and Demus the son of Pyrilampes . . . So rathoi
than expressing surprise at the things I’ve been saying, you should
stop my darling philosophy voicing these opinions. You see, my
friend, she is constantly repeating the views you’ve just heard from nu*
and she’s far lessfickle than my other love. I mean, Alcibiades says dil
ferent things at different times, but philosophy’s views never changt
(48id-482a).
Against the capricious people of Athens, against the even moie
whimsical Alcibiades, Socrates has found an anchor that allows him to
be right against everyone else’s vagaries. But this is also, in spite ol
Socrates’ sneering remark, what Callicles thinks of natural laws: the y
protect him against the vagaries of the assembled people. There is, In
be sure, a big difference between the two anchors, but this should
count in favor of the real anthropological Callicles, not Socrates: llu
good guy’s anchor is fastened in the ethereal afterworld of shadow
and phantoms, whereas Callicles’ anchor is at least gripping the solid
and resisting matter of the Body Politic. Which one of the two anchm
is better secured? Incredible as it seems, Plato manages to make us lu
lieve that it is Socrates’!
The beauty of the dialogue, as has often been noticed, lies mainly In
the opposition between two parallel scenes, one in which Callith
mocks Socrates for being unable to defend himself in the tribunal ol
this world, and the other at the end, when Socrates mocks Callicles loi
being unable to defend himself in the afterworld tribunal of I lack
Round one:
Socrates, you’re neglecting matters you shouldn’t neglect. Look al

the noble temperament with which nature has endowed you! YtI
what you’re famous for is behaving like a teenager. You couldn’t do
liver a proper speech to the councils which administer justice, or mal 0
a plausible and persuasive appeal. . . The point is that if you or any ol
your sort were seized and taken away to prison, unjustly accused ol
some crime, you’d be incapable—as I’m sure you’re well awaie ol
doing anything for your self. With your head spinning and mouth gapitu(
open, you wouldn’t know what to say. (485e 486b)
A terrible situation indeed for a Greek to be loll speechless by ill

unfair accusation in the midst of the ciowd Nolice that Callkles dot
227
not admonish Socrates for being too lofty, but for being an impotent,
lowly, and idiotic teenager. Callicles has a resource of his own that
comes from an ancient aristocratic tradition: an inherited talent for
.peech which allows him to find just the right thing to say against the
<onventions created by “second-rate citizens.”
10 find a retort to that one, Socrates has to wait until the end of the
dialogue, and must abandon his dialectic of questions and answers to
lell a crepuscular tale. The final round:
I think it’s a flaw in you that you won’t be able to defend yourself
when the time comes for you to undergo the trial and the assessment
which I’ve just been talking about. Instead, when you come to be
judged by that son of Aegina [Rhadamanthys] and he seizes you and
lakes you away, your head will spin and your mouth will gape there in
that world just as much as mine would here, and the chances are that
someone will smash you in the face and generally abuse you as if you
were a nobody without any status at all. (5266-5273)
Abeautiful effect on the stage, to be sure, with naked shadows pacing a

11ipier-mâché inferno and artificial fumes and fog lingering in the air.
Ibil a bit late, Socrates,” the historical and anthropological Callicles
( ould have retorted, “because politics is not about the naked dead liv
ing in a world of phantoms and judged by half-existing sons of Zeus,
bill about clothed and living bodies assembled in the agora with their
I il ns and their friends, in the bright sun of Attica, and trying to de
li le, on the spot, in real time, what to do next.” But the straw
( ilhcles, by now, through a happy coincidence, has been shut down
b I’lato. So much for the dialectical method and the appeal to “the
1immunity of free speech.” When the time of retribution has come,
in 1.lies speaks alone in the much despised epideictic way (465e).
11 is a pity that the dialogue ends with such an admirable but empty
ippeal to the shadows of politics, because Callicles could have shown
lit it even his selfish and extravagant claim to hedonism, which made
blm so contemptible to the theater crowd, is also used by Socrates to
li line his way of dealing with the people: “And yet, my friend, in my
pinion, it’s preferable for me to be a musician with an out-of-tune
I u 01 a chon leader with a cacophonous choir, and it’s preferable for
most t veryone in the unnld lo (md my beliefs misguided and wrong,
111he 1 than lor just one pi rson me lo conliachct and clash with my
/ (4«±b c)
228
“Perish the people of Athens,” claimed the straw Callicles, “pio

vided I have a good time, and forcibly seize as much as I can from tlu*
hands of the second-rate human debris!” In what sense is Socrates’ ap
peal less selfish? “Perish the whole world, provided I am in agreement
not only with one other person”—as, we shall see, he has earlier said
to Polus—“but with myself!” Knowing that Plato willfully misrepie
sents Callicles’ and Gorgias’s position, whereas he presents SocralP
as having the last word and responding seriously, who is the moip
dangerous—the agoraphobic mad scientist, or the “blonde brute 0!
prey”? Who is the more deleterious for democracy, Right or Might
All through the dialogue the parallelism between the solutions of tliP
two sparring partners is inescapable.
And yet it is also completely invisible, as long as we keep our eyes nit
the stage. Why? Because of the definition of knowledge that Sou alp
forcibly imposes over Callicles' definition. This is where the symim 11
is broken; this is what makes Callicles exit to the sound of boos it
matter how many Nietzscheans will later try to push him back out
the boards. QED ; TKO.
The Triangular Contest of Socrates,

the Sophists, and the Demos
In the three dialogues of the Gorgias, Might and Right never appi'd
as comparable; later we will see why. What remain commensuial
enough to be disputed are the relative qualities of two types of c* |
knowledge: one in the hands of Socrates, the other in the hands of ll
rhetoricians (a word invented, it seems, in the Gorgias). What I I
yond question for both Socrates and the straw Sophists is that 11
expert knowledge is necessary, either to make the people of At lu II I
have in the right way or to keep them at bay and shut their molli I
They no longer consider the obvious solution to the problem lie Pill
the agora, the one we will explore in Chapter 8, although it Is Ht111J I
ent in the dialogue, at least as a negative template: the assemblt d I
Politic, in order to make decisions, cannot rely on expel t kilo I
alone, given the constraints of number, totality, urgency, and pit I
that politics imposes. Reaching a decision without appealing to a lia
ral impersonal law in the hands of expel ts lequues a dhsunillftl
knowledge as mulliiaiious as the multitude ilsell I ht knowhtlgt
229
whole needs the whole, not the few. But that would be a scandal for
( .1Hides and for Socrates, a scandal whose name has been the same at
ill periods: democracy.
So here again the disagreement of the partners is secondary to their
i mnplete agreement: the contest is about how to shut the mouths
ni Ihe people faster and tighter. On this ground, Callicles is going to
|i ise fast. After agreeing, with a common paternalism, that experts are
in eded to “look after a community and its citizens” (513e), the two ar-
iie over what sort of knowledge will be best. Rhetoricians have one
I pe of expertise and Socrates has another. One is epideictic, the other
ipodeictic. One is employed in the dangerous conditions of the agora,
I In other in the quiet and remote one-to-one conversation Socrates
Imi sues with his disciples. At first glance it looks as if Socrates should
1e «il this game, since it is of no use at all to have a method for
Img the citizens of the agora that is itself agoraphobic and operates
nly on a one-to-one basis. “I’m content,” Socrates confesses naively
I’olus, “if you testify to the validity of my argument, and I canvass
\ loi your vote, without caring about what everyone else thinks” (476a).
it politics is precisely about “caring for what everyone thinks.” Can-
i mg for only one vote is worse than a crime, it is a political mistake.
hen Callicles admonishes Socrates for this infantile behavior, he
1mid win the day: “Even a naturally gifted person isn’t going to de-
I ip 111I0 a real man, because he’s avoiding the heart of his community
I Ihe {hick of the agora, which are the places where, as Homer tells
i nun ‘earns distinction.’ Instead he spends the rest of his life
1ml of sight, whispering in a corner with three or four young men,
In 1 llun giving open expression to important and significant ideas”
d 1)
luis (he dialogue, logically, should end up with only one scene, in
h Si mates is sent back to his campus corner, philosophy being
Hid lo a useless specialized obsession, with no relation to what
il nun" do to “earn distinction” with “important and significant
1 I his is what rhetoric will do. But this is not what we did when
I Invi 11Icd the power of Science, with a capital S, over and over
II With the “context of truth” that Socrates is bringing to the fore,
I h mmph becomes impossible. It is a very subtle trick, but it is
1 h to level sc* the logical com sc* ol the dialogue and to make Soc
hi wheic lie should have lost
hit I the supplant nl piovidt d by npodeichc icasonmg that makes
230
it so much better than the natural laws invoked by the Sophists against
the conventions of “slaves and assorted human debris”? This kind («I
reasoning is beyond dispute:
so c r ates: But can knowledge be either true or false?
g o r g i a s : Certainly not.
s o c r a t e s : Obviously, then, conviction[pistis] and knowledge[epistèrm/
are not the same. (454d)

The Sophists’ transcendence is beyond convention, but not beyond
dispute, since the questions of being superior, more natural, better
born, better bred open another swarm of discussions, as can be wit
nessed even today—no matter how many Bell Curves one throws inti
the pot. Callicles has invented a way to discount the crowd’s physical
weight and number, but not to escape altogether from the site ol ill
chock-full agora. Socrates’ solution is much stronger. The fabulous P
cret of mathematical demonstration that he has in his hands is th il It
is a step-by-step persuasion that forces one to assent no matter wli il
Nothing, though, makes this way of reasoning able to adjust to tin p
tremely harsh conditions of the agora, where it should be as usclul I
borrow an old feminist slogan, as a bicycle to a fish. So a bit tin I
work is needed for Socrates to be able to make use of this weapon 11
first has to disarm everyone else, or at least make them believe they m
thoroughly disarmed: “So we’d better think in terms of two kind
persuasion, one of which confers conviction without understand!/!
men pistin parchomenon aneu tou eidenai], while the other confoi s km
edge [epistèmèf (454e).
Epistèmè, how many crimes have been committed in your nami I (
this the whole history hinges. So venerable is this opposition th
contrary to the obviously rigged fight of Might and Right, we nil
lose our nerve at this point and fail to see how bizarre and illogli il I
argument is. The whole difference between the two kinds ol pi 1
sion relies on two innocuous little words: “without undoist imlli
But understanding of what? If we mean the understanding ol the
specific conditions of felicity for political discussion that is nun
urgency, and priority—then Socrates is certainly wrong II tin Ilit
it is the apodeictic reasoning of causes and consequences, Ilu tf
that is “without understanding,” meaning that it lads to I ilo« I
account the pragmatic conditions ol deciding wlial to do m I h
231
Iluck of the agora with ten thousand people talking all at once. On his
own, Socrates cannot replace this pragmatic knowledge in situ, with
Ins unsituated knowledge of demonstration. His weapon is mind-
Itoggling, mouth-shutting, but a useless deterrent in the context of the
igora. He needs help. Who is going to give him a hand? The foils in-
»nled by Plato, who, as usual, conveniently fall into the trap like ideal
11aw men.
I he dialogue could not work and make Socrates triumph against all
nids if the puppet Sophists did not share Socrates’ loathing for the
I ills and gimmicks with which common people go about their daily
uismess. So when Socrates makes a distinction between real knowl-
Ige and know-how the (straw) Sophists don’t protest, since they have
lln same aristocratic contempt of practice: “There’s absolutely no ex
il ise involved in the way it [cookery] pursues pleasure; it hasn’t con-
lilt 1ed either the nature of pleasure or the reason why it occurs ... All
I 11lie technical cook] can do is remember a routine which has become
mined by habituation and past experience, and that’s also what it re-
011 to provide us with pleasant experiences” (soia-b).
musingly enough, this definition of merely practical know-how, al
l o t igh uttered with scorn, would today fit what psychologists, prag-
1ill sis, and cognitive anthropologists would call “knowledge.” But
I ey point is that this distinction itself has no other content than Soc-
tlisdain for the common people. Socrates is here on very thin ice.
1 distinction between knowledge and practical know-how is both
1111 11lows him to appeal to a mouth-shutting superior natural law
I 11so what is enforced by the very action of shutting the mouths of
lui thousand people who go about their business every day “with-
I I Mowing what they do.” If they knew what they were doing, the
tin lion would be lost. So if this absolute demarcation is not im-
d by sheer force—the true task of epistemology over the ages—the
nli I ol truth” cannot be brought to bear on the impossibly delete-
1 11mosphere of public debate. This is one of the rare cases in his-
111 which “sheer force” has been applied. To enforce this divide
il do wc* have? Only Socrates’ word for it—and the meek retreat of
1 11 I’olus, and Callicles into acceptance of Socrates’ definition,
hilly staged in Plato’s theatrical machinery. That’s quite a few
III Ions loi an unconditional appeal to an unconstructed “imper-
1I I 1
PANDORA'S HOPE
232
As Lyotard showed some time ago, and as Barbara Cassin (Cassin

1995) has recently demonstrated so forcefully, distinguishing the two
forms of knowledge and setting up the absolute difference be
tween force and reason requires a coup deforce—the one that expels the
Sophists from philosophy and the common people from rigorous
knowledge. Without this coup, the expert knowledge of demonstra
tion could not take over the precise, subtle, necessary, distributed, in
dispensable knowledge of the members of the Body Politic who takt
it upon themselves to decide what to do next in the agora. Epistenu
will not replace pistis. Apodeictic reasoning will remain important, ol
course, even indispensable, but in no way bound to the question of hou
best to discipline the multitude. As in the birth of all political regimes
undisputed legitimacy resides in an original bloody coup. In this casi
and this is the beauty of the play, the blood that is shed is Socrates’own
That sacrifice makes the move even more irresistible and the legill
macy even more indisputable. By the end, there won’t be a dry eye k 11
in the theater . . .
The Sophists are no match for this dramatic move, and after accept
ing, first, that expert knowledge is necessary to replace that of tin
poor ignorant multitude, and second, that the knowledge of demon
stration is absolutely, not relatively, different from the skills and gun
micks of the common people, they have to confess that their form ol
expertise is empty. How silly Gorgias’s boasting now sounds: “Does 11 I
that simplify things, Socrates? Rhetoric is the only area of expel 11 1
you need to learn. You can ignore all the rest and still get the better ol IIn
professionals” (459c).
We will see in the next chapter that this apparently cynical answt 1
is in fact a very precise definition of the nowprofessional nature ol po
litical action. However, if we agree to overlook this point and il wi»
start to accept the contest and pit the specialized knowledge of sc it n
tists against the specialized knowledge of rhetoricians, then sophist 1
is immediately turned into an empty manipulation. It is like inliodui
ing a race car into a marathon; the new machine renders the slow* I
runners ridiculous.
s o c r a t e s : Faced with phenomena like the one you’ve mentioned, il

comes across as something supernatural, with enormous power
g o r g i a s : You don’t know the half of il, Sociales! Almost evei y attorn
233
plishment falls within the scope of rhetoric . . . Often in the past,

when I’ve gone with my brother or some other doctor to one of their
patients who was refusing to take his medicine or to let the doctor op
erate on him or cauterize him, the doctor proved incapable of persuad
ing the patient to accept his treatment, but I succeeded, even though I
didn’t have any other expertise to draw on except rhetoric. (456a-b)
Iwen for sentences like that we need centuries of Pavlovian training

to read them as cynical, because what the real Gorgias alludes to here
Is the impotence of specialists to make the people as a whole make
lough decisions. The real Gorgias points out an extraordinarily subtle
I ill, one that Socrates does not want to understand (although he prac-
Iu os it so cleverly); the puppet Gorgias is made to say that no knowl-
<«Ige at all is necessary. After their staged defeat, the rhetoricians are
Iuitling their own heads on the chopping block. Having accepted that
1luioric is an expertise, then having found it empty, they are now ex-
|><lied from knowledge altogether, and their skills branded as mere
ll.illery” (502d), one of the many obscure types of popular know-how
limn which rhetoric cannot be distinguished. “Well, in my opinion,
iimgias, it doesn’t involve expertise; all you need is a mind which is
«ul at guessing, some courage, and a natural talent for interacting with
Ioplc. The general term I use to refer to it is ‘flattery,’ and this strikes
nu .is a multifaceted activity, one of whose branches is cookery. And
II it I’m saying about cookery is that it does seem to be a branch of
Ih *11ise, but in fact isn’t: it is a knack, acquired by habituation [ouk estin
1hue, all’empeiria kai tribè]” (4ô3a-b).
I lie most moving feature, which will deserve all our attention later,
IIi.i I even in this famous coup de grâce Socrates is still complimenting
In Im ic. IIow can we not consider as positive qualities being “good at
in ssmg,” having “courage,” knowing “how to interact with people”—
it inily not skills that Socrates lacks in spite of his claims to the con-
1nyi> I'or that matter, what is so bad about being as talented as a
ml I I myself prefer a good chef to many bad leaders! And yet Socra-
li is won. The weakest has turned the tables on the strongest. The
1 I logical that is, the “happy few,”—have won over the “universal”
It that is, everyone minding the whole Body Politic at once. Socra-
who by his own confession is the least adapted to rule over the
oplo iules ovei them at least limn the conveniently fai away place
234
of the Isles of the Blessed: “I think,” he says, wrapping his words in

three degrees of irony, “I’m the only genuine practitioner of politics in
Athens today, the only example o f a true statesman” (52id).
And it is true: no tyranny has been longer lasting than that by this
sacrificed, dead man over the living, no power more absolute, no reign
more undisputed.
The defeat of the (straw) Sophists is nothing compared with that ol
the common people of Athens, as can be seen by a summary of the ar
gument so far. The “human debris and assorted slaves” are the great
absent ones, without even a chorus to defend their common sense as
in classic tragedies. When we start reading this most famous din
logue carefully, we discover not only a fight between Callicles (that is
Might) and Socrates (Right), but two overlapping disputes, only the
first of which has been commented on ad nauseam. One dispute, as in
a puppet show, pits the wise sage against the blond brute, and is so
beautifully staged that the little kids scream in terror that Might will
beat down Right. (As we saw earlier, it makes no difference at all if Ilit
plot is reworked later by a Nietzschean scriptwriter and now pi!
the beautiful and sunny Callicles, head of the race of masters, again I
the black Socrates, degenerate scion of a race of priests and men ol
ressentiment. We, the kids, are still supposed to scream, this time ill il
Right will beat down Might and turn it into a weak and meek sheep )
But there is a second fight going on silently, offstage, pitting the pi o
pie of Athens, the ten thousand fools, against Socrates and Calluli
allied buddies, who agree on everything and differ only about the faste I
way to silence the crowd. How can we best reverse the balann ol
force, close the mouths of the multitude, put an end to the disouli 11
democracy? Will it be through the appeal to reason, geometry, pin
portion? Or will it be through aristocratic virtue and upbringing? So»
rates and Callicles are alone against the crowd, and each ol t lu II
wants to dominate the mob and obtain a disproportionate shaie ol 11
ther this world’s or the other world’s laurels.
The fight of Might and Right is rigged like a game of calth Mil I
hides the settlement between Callicles and Socrates, each agieiln I
serve as the other’s foil. In order to avoid the fall into Might Id II
accept unconditionally the rule of Reason such has been the i nil I
version. The later version is the same in revet se in Older to «void I ill
ing into Reason, let us unconditionally agiee to lall into tin mil
of Might. But in the meantime, silent and mule, pu/ lid in
235
flabbergasted, the people of Athens remain offstage, waiting for their

masters to sort out the best way to reverse their “physical force,”
which could be “entirely discounted” if there were not so many of
lhem. Yes, there are too many, too many to be taken in anymore by
Ihis childish story of the cosmic dispute between Might and Right.
The hands of the puppeteers are too visible now, and the scandal of
seeing Socrates and Callicles, the arch-rivals, arm in arm, is an experi
ence as enlightening for the little kids as seeing the actors of Hamlet
ill ink together laughingly at the pub after the curtain has fallen.
Such an experience should leave us older and wiser. Instead of a dra
matic opposition between force and reason, we will have to consider
Ihi ce different kinds of forces (or three different kinds of reasons—the
i hoice of words adding, from now on, no decisive nuance): the force of
Socrates, the force of Callicles, and the force of the people. It is a
II ilogue we have to deal with, and no longer a dialogue. The absolute
i onlradiction between the two famous protagonists is now displaced
Inlo a more open contest between two tugs of war: one between the
Iv o heroes and the other, not yet recognized by philosophers, be-
I e'en the two heroes pulling on the same side of the rope and the ten
llmusand average citizens pulling on the other side. The principle of
Ilu1excluded middle that seems so strong in the burning choice be-
I een Might and Right—“choose your camp fast or all hell will break
loose!”—is now interrupted by a third party, the assembled people of
Iliens. The excluded middle is the Third Estate. It sounds better in
I II ne h : le tiers exclu c’est le Tiers Etat! The philosopher does not escape
hi>111 ihe Cave, he sends the whole demos down into the Cave to feed
nly on shadows!
When we hear about the danger of mob rule, we will now be able to
i I c(uielly: “Is it Callicles’ solitary rule that you mean, or that of the
ill i less assembly of ‘human debris and assorted slaves’?” When we
il Ihe little red-flag word “social,” we will be able to disentangle two
lllhicnl meanings: the one that designates the power of Callicles’
I I hi against Socrates’ Reason, and the one that designates the never-
I ilisuibed crowd resisting the attempts of both Socrates and
illli les to exert a solitary form of power over them. Two weak, na-
I mil anogant men on the one hand; the City of Athens on the
IIn i cIllicit on, women, and slaves included. The war of two against
II lhe siiange wai ol ihe duo Hying to make us believe that without
ll would be (lie wai ol all agamsl all
C H A P T E R E I G H T
A Politics Freed from Science

The Body Cosmopolitic
Napoleon’s mother used to sneer at her emperor son’s fits of ragi

“Commediante! Tragediante!” We could mock in the same way llu m
two races of masters, the one descended from Socrates, the other boni
Callicles. On the comedy side we have the fight between Might and
Right; on the tragedy side we have the absolute distinction betwu II
epistèmè and pistis, this coup deforce whose origin is cleansed by tin*
blood of one martyr. But we can also turn our eyes to the Third l;s( tic
and extract from the Gorgias the trace of another voice, which is ml
ther comedy nor tragedy but plain prose. Plato is close enough to IIn*
benighted time when politics was respected for what it was, that I
before the advent of the scenography set up in common by Sou all
and Callicles, which I have defined as “inhumanity against inhum ill
ity.” Much as an archaeologist would do with the Delphic Tolos 01 ill
statue of Glaucus unearthed by Rousseau, we can thus reconsliml
out of the ruins of the dialogue the original Body Politic before il w I
smashed to pieces—except that I will use the same myth as Roussi ail
for exactly the opposite goal, that is, to free politics from an execs i (
reason.
Here is Rousseau in the foreword to the Discourse on the Origin of I
equality: “The human soul, like the statue of Glaucus which turn ill
sea and storms had so much disfigured that it resembled a wild lit I I
more than a god ... by now we perceive in it, instead of a being alwa
acting from certain and invariable principles, instead of that luiivi III
and majestic simplicity which its author had impiessed upon it unlit
216
A POLITICS FREED FROM SCIENCE
237
ing but the shocking contrast of passion that thinks it reasons, and an
understanding grown delirious.”1
By unwinding the adventures of Reason, we can imagine how it was
before it turned into an unlivable chimera, a monstrous Big Animal
whose unrest horrifies the masters even today. Needless to say, this
is an attempt at an archaeology-fiction: the invention of a mythical
lime when political truth-saying would have been fully understood, a
world that was later lost through the accumulation of mistakes and de
generation.
How Socrates Reveals the

Virtue of Political Enunciation
In Chapter 7 we noticed many of the specifications of political debate.
Io reconstruct the virtual image of the original Body Politic, we sim
ply have to take positively the long list of negative remarks made by
I’l.ilo: they show in reverse what is missed when one converts what
w.is, until then, the distributed knowledge of the whole about the
hole into an expert knowledge held by a few. Through this bit of ar-
1haeology-fiction we can thus be privileged witnesses to two phenom-
111.1 at once: the specification of the conditions of felicity proper to
politics, and their systematic destruction by Plato, who turns them
Into ruins. We thus witness at once the iconoclastic gesture that de
ll oys our much-treasured ability to deal with one another and the
I (militions of its possible reconstruction.
I he dialogue is very explicit about this iconoclasm, since Socrates
II lively confesses: “In my opinion, you see, rhetoric is a phantom of a
In inch of statesmanship [politikès morious eidolon]” (463d). That is ex
it Ily what he and his buddies have done: they have turned a fleshy,
losy living Body Politic that kicked and bit into “a phantom,” by ask
ing 11 to feed on a diet of expert knowledge on which no such organ-
1 in could survive. They have turned it into an eidolon without realiz
ing ill.it by smashing it they deprived us of one part of our humanity.
1 Rousseau, D i u o n r s c o n th e O rig in o f h u q u a lity , tians. Lester G. Crocker (New York:
I 111 Hooks, k/(i )
238
As Gorgias rightly points out, the first specification of political
speech is that it is public and does not take place in the silent isolation
of the study or of the laboratory:
g o r g ia s : When I say there’s nothing better, Socrates, that is no more

than the truth. It [rhetoric] is responsible for personalfreedom and en
ables an individual to gain political power in his community.
s o c r a t e s : Yes, but what is it?
g o r g i a s : I’m talking about the ability to use the spoken word to per
suade—to persuade the jurors in the courts, the members of the Coun
cil, the citizens attending the Assembly—in short, to win over any and
every form of public meeting of the citizen body. (452d-e)
As we just saw, this very specific condition of speaking to all the dil
ferent forms of assemblies essential to Athenian life (courts, councils,
assemblies, burials, ceremonies: all sorts of private and public meet
ings), is denied by Socrates and turned into a defect, whereas Socrates'
weakness, his inability to live in the agora—although he spends all his
time in it and seems to enjoy himself immensely!—is vaunted as his
highest quality:
I ’m no politician, Polus. In fact, last year I was on the Council, thanks

to the lottery, and when it was the turn of my tribe to form the execu
tive committee and I had to put an issue to the vote, I made afool of my
self by not knowing the procedure for this. So please don’t tell me to ask
the present company to vote now either... My expertise is restricted
to producing just a single witness in support of my ideas—the person
with whom I’m carrying on the discussion—and I pay no attention to
large numbers of people; I only know how to ask for a single person’s
vote, and lean’t even begin to address people in largegroups. (4736-4743)
Tough luck, because “addressing large numbers” and “paying alien

tion” to what they mean, think, and desire is exactly what is being dc
bated under the despised label “rhetoric.” If Socrates is so proud ol
“not being a politician,” why is he teaching those who know belUl
and why does he not remain in the confines of his own selfish, spec lui
ized, expert discipline? What business do agoraphobics have in IIn?
agora? This is what Callicles (the real Callicles, the historical, aiillno
pological one whose negative presence can still be delected in the di I
logue) rightly points out:
239
In actual fact, philosophers don’t understand their community’s legal

system, or how to address either political or private meetings, or what
kinds of things people enjoy and desire. In short, they are completely
out of touch with human nature. When they do turn to practical activity,
then, in either a private or a political capacity, they make ridiculous
fools of themselves—just as, I imagine, politicians make fools of them
selves when they’re faced with your lot’s discussions and ideas.
(484d-e)
But Callicles’ derision, although it accurately underlines the quali
ties required from a leader, is itself made useless by his own appeal to
an expert knowledge of rhetoric that is content to know nothing at all,
to just be manipulative. Yet when he defines the goal of his aristocratic
friends, he paints an accurate portrait of the real qualities that Socra
tes entirely lacks: “The superior people I mean aren’t shoemakers or
cooks : above all, I’m thinking of people who’ve applied their cleverness
10 politics and thought about how to run their community well. But
cleverness is only part of it; they also have courage, which enables
them to see their policies through to thefinish without losing their nerve and
giving up” (49ia-b).
It is precisely this courage to see “through to the finish,” that Socra
tes will misrepresent so unfairly when he destroys the subtle mecha
nism of representation by polluting it with the question of an absolute
morality. To see a political project through, with the crowd, for the
c1owd, in spite of the crowd, is so stunningly difficult that Socrates
(lees from it. But instead of conceding defeat and acknowledging the
specificity of politics, he destroys the means of practicing it, in a sort
ol scorched-earth policy the blackened wreckage of which is still visi
ble today. And the torch that set the public buildings ablaze is said to
be that of Reason!
The second specification that can be recovered from the wreckage
is that political reason cannot possibly be the object of professional
I nowledge. Here the ruins have been so deformed by Plato’s icono-
1l.islic obstinacy that they have been made as barely recognizable as
Ihose of Carthage. And yet this is what most of the dialogue turns
uound, as all the commentators have noticed: the question, it ap
peal s, is to decide what soi t of knowledge rhetoric is. At first, though,
11 seems veiy cleai that politics is not about professionals telling the
pi ople wlut to ilo (loiglas says, "I assume you’ie aware that it was ei
240
ther Themistocles or Pericles, not the prof essionals, whose advice led to
those dockyards you mentioned, and to Athens’ fortifications and the
construction of the harbours” (455d-e).
The protagonists agree that what is needed is not knowledge as such
but a very specific form of attention to the whole Body by the whole
Body itself. This is what Socrates recognizes under the name of a
good and ordered cosmos in the qualities required of the expert techni
cians (demiourgos): “Each of them organizes the various components he
works with into a particular structure and makes them accommodait
andfit one another until he’s formed the whole into an organized and or
dered object”(5036-5043).
But then, as usual, every time a condition of felicity is clearly articu
lated it is perverted into its opposite by Socrates, who, as Nietzsche rc
marked, has King Midas’s hands except that he turns gold into mud
The nonprofessional nature of the knowledge of the people by lilt
people turning the whole into an ordered cosmos and not “a disoi
derly shambles” becomes, through a subtle shift, the right of a few
rhetoricians to win over real experts even if they know nothing. Wlial
the Sophists meant was that no expert can win in the public agora be
cause of the specific conditions of felicity that reign there. After Sou a
tes’ translation, this sensible argument becomes the following absuid
one: any expert will be defeated by an ignorant person who knows only
rhetoric. And of course, as usual, the Sophists kindly oblige Socratt
by saying the ridiculous thing they have long been accused of saying
this is the great advantage of the dialogue form that epideixis lacks
so c r ates: Now, you claimed a while back [456b] that a rhetorician

would be more persuasive than a doctor even when the issue was
health.
g o r g i a s : Yes, I did, as long as he’s speaking infront of a crowd.
s o c r a t e s : By “in front of a crowd” you mean “infront of non experts
don’t you? I mean, a rhetorician wouldn’t be more persuasive than i

doctor in front of an audience of experts, of course.
g o r g i a s : True. (459a)
Socrates triumphs. Yet again, Gorgias is insisting on the veiy piolt

lem that still besets us today and that no one has ever been abit ti
solve, certainly not Plato and his Republic. Politics is about tit ilin
with a crowd of “non experts,” and ibis situation cannot possibly hi
241
the same thing as experts dealing with experts in the inner recesses of
their special institutions. So when Plato is making his famous joke
about a cook and a physician pleading for votes in front of an assem
bly of spoiled brats (52.2), it takes very little talent to twist the story to
Socrates’ embarrassment. This funny scene works only if the crowd
of Athens is made up of spoiled kids. Even putting Socrates’ aristo
cratic scorn aside, nowhere does it state, if the story is read carefully,
that it pits a serious expert against a populist flatterer. Rather, it stages
a controversy between two specialists, the cook and the physician, talk
ing to an assembly of grown men about either short-term or long-term
strategy, the outcome of which neither of them knows, and through
which only one party is going to suffer, namely the demos itself.
Here again Socrates’ use of a pleasant story hides the dramatic con
dition of felicity for what it is to speak in real time, in real life, and in
Iull scale about things that no one knows for sure and that affect every
one. About how to fulfill this pragmatic condition he does not have the
slightest suggestion, and yet the only solution that the non-experts
had in hand—that is, listening in the agora to both the short-term
1ook and the long-term physician before running the risk of making a
decision together that will have legal consequences—is smashed into
pieces. We in Europe, who do not know which beefsteak to eat because
<>l ihe many controversies we read about every day in our newspapers
be Iween cooks and physicians about mad cows infected or not by
pi ions, would give several years of our life to recover the solution that
Socrates simply ignores.
The third condition of felicity is similarly important and similarly
Ignored. Not only does political reason deal with important matters,
I 1Uen up by many people in the harsh conditions of urgency, it also
( mnot rely on any sort of previous knowledge of cause and conse
quence. In the following passage, which I discussed earlier, the misun-
tli 1standing is already clear:
Rhetoric is an agent of the kind of persuasion [peithous demiurgos]

which is designed to produce conviction, but not to educate the peo
ple, about matters of right and wrong . . . A rhetorician, then, isn’t
loncemed to educate the people assembled in lawcourts and so
on about light and wiong, all he wants to do is to persuade them
(mslikosl I mean, I shouldn’t lluul it’s possible for him to gel so
242
many people to understand [didaxai] such important matters in such a
short time. (4546-4553)
The “demiurge of persuasion” does exactly what the “didactic” urge

cannot: it deals with the very conditions of urgency with which poli
tics is faced. Socrates wants to replace pistis with a didacticism that is
fit for professors asking students to take exams on things known in ad
vance and rehearsed by training and rote exercises, but that is not fit
for the trembling souls who have to decide what is right and wrong
on the spot. Socrates recognizes this readily: “I think it’s a knacl
[empeirian],” he says of rhetoric, “because it lacks rational understanding
either of the object of its attention or of the nature of the things it dis
penses (and so it can’t explain the reason [aitian] why anything hap
pens), and it’s inconceivable to me that anything irrational involves is
pertise [ego de technèn ou kalô oanè alogon pragma/ ’ (465a).
How accurate is this definition of what is being destroyed! It is as II
we were seeing at once the venerable statue of politics and the ham
mer that breaks it into pieces. How moving to see, by returning to IIk*
past, how close all these Greeks still were to the positive nature of I111
democracy that remains their wildest invention. Of course “it does nnl
involve expertise,” of course “it lacks rational understanding”; (In*
whole dealing with the whole under the incredibly tough constrain I
of the agora must decide in the dark and will be led by people as blind
as themselves, without the benefit of proof, of hindsight, of foresiglil
of repetitive experiment, of progressive scaling up. In politics thou I
never a second chance—only one, this occasion, this kairos. Thoi e 1
never any knowledge of cause and consequence. Socrates has a good
laugh at the ignorant politicians, but there is no other way to do poli Ili
and the invention of an afterworld to solve the whole question Is v
actly what the Sophists laugh at, and rightly so! Politics imposes (III
simple and harsh condition of felicity: hic est Rhodus, hic est salins
Here too, after Gorgias points out the real-life conditions m with It
the demos has to reach a decision through rhetoric “I repeat tit tl i(
effect is to persuade people in the kinds of mass meetings whu h happt 1
law courts and so on; and I think its province is right and u nu
(454b)—Socrates requires from rhetoric something it cannot pos ihl
deliver, a rational expertise about right and wiong What could wotk
efficiently with a relative difference between bad and good cannol lu I I
243
water if an absolute foundation is required of it, as Socrates demands:

“Do you think . . . that all activity aims at the good, and that the good
should not be a means towards anything else, but should be the goal of
every action? . . . Now, is just anyone competent to separate good plea
sures from bad ones, or does it always take an expert?” (499e-sooa).
And Callicles swallows the hook! “It takes an expert,” he responds, a
technicos. From then on, there is no solution and the Body Politic be-
comes impossible. If there is one thing that does not require an expert,
.ind cannot be taken out of the hands of the ten thousand fools, it is de-
1iding what is right and wrong, what is good and bad. But the Third
I state has been turned, by Socrates and by Callicles, into a barbaric
population of unintelligent, spoiled, and sickly slaves and children,
who are now waiting eagerly for their pittance of morality, without
which they would have “no understanding” of what to do, what to
i hoose, what to know, what to hope. Yes, “morality is a phantom of
statesmanship,” its idol. And yet, at the same time that Socrates ren
ders the task of politics impossible by asking from the people a knowl-
1dge of causes that is totally irrelevant, he defines it accurately:
I here’s nothing which even a relatively unintelligent person would
I ike more seriously than the issue we’re discussing—the issue of how to
live one’s life? The life you’re recommending to me involves the manly
in hvities of addressing the assembled people, rhetorical training, and the
I md of political involvement you and your sort are engaged in” (500c).
Nothing is more moving in the Gorgias than the passage in which
Son ates and Callicles, after agreeing on the relevance of statesman-
hip, destroy, one after another, the only practical means by which a
II owd of blind people fumbling in the dark could get the light to help
Ihem decide what to do next: “So these are the qualities which that ex-
( //<ut rhetorical expert of ours will be aiming for in all his dealings
ilh people’s minds, whether he’s talking or acting, giving or taking.
I h II (onstantly be applying his intelligence to find ways for justice, self-
1onliol, and goodness in all its manifestations to enter his fellow citi-
t ns minds, and for injustice, self-indulgence, and badness in all its
m imlestations to leave” (504d-e).
I Ins is what they agree on. This high-minded definition of politics,
1 we will see, is common sense, but only as long as it is not deprived of
ill (lie ways and means that mal e it ellective. And yet this is what Soc-
1lit s is going to do, with the stiaw Callie les following suit obediently.
244
In a denigration of Athens’s beauties that is worse than the city’s plun

der by the Persians or the Spartans because it comes from within, they
are going to persuade themselves that every art aims at nothing but
corruption. As usual with hearts full of demotic hatred, the loathing
for popular culture “blazes forth” every time they talk of politics.
“There’s absolutely no expertise involved in the way it pursues pleasure,
it hasn’t considered either the nature of pleasure or the reason why it
occurs” (501a).
About what do they talk so irreverently? Cookery first, and then the
skills of the greatest playwrights, the greatest sculptors, the greatest
musicians, the greatest architects, the greatest orators, the greatest
statesmen, the greatest tragedians. All of these people are dumped be
cause they don’t know what they know in the didactic fashion that
Professor Socrates wants to impose on the people of Athens. Stripped
of all its artistic means to express itself to itself, this most sophist I
cated demos appears this way in the eyes of its disappointed teacher
“So we’re faced here with a kind of rhetoric which is addressed to the
assembled population of men, women, and children all at once—slave*
as well as free people—and it’s a kind of rhetoric we find we can ’tap
prove of. I mean, we did describe it as flattery”(sold).
Was it simply being flattered to go to the tragedies, to hear the 01a
tions, to listen to poetry, to watch the Panathenean’s pageantry, to
vote with one’s own tribe? No, these were the only means by which
the demos could accomplish this most extraordinary feat: to represent
itself publicly to the public, to render visible what it is and what It
wants. All the centuries of arts and literature, all the public spaces
the temples, the Acropolis, the agora—that Socrates is denigrating one
by one, were the only ways the Athenians had invented to seize them
selves as a totality living together and thinking together. We see lu le
the dramatic double bind that turns the Body Politic into a sc hi o
phrenic monster: Socrates appeals to reason and reflection—but tlun
all the arts, all the sites, all the occasions where this reflexivity till 1
the very specific form of the whole dealing with the whole, are deemed
illegitimate. He decries the knowledge of politics for its inability to nil
derstand the causes of what it does, but he severs all the leedbiuk
loops that would make this knowledge of the cause piactical No won
der Socrates was called the numbfish! What he paralyzes with his ch 1
trie sting is the very life, the vei y essence of the Body Politic I low ni 11
245
sible was the Athenian demos to invent this derided institution of

ostracism, this very intelligent way to get rid of those who want to get
rid of the people!
In this passage the two partners switch off, one by one, each of the
hundreds of fragile and tenuous lamps, plunging the demos in a dark
ness much more profound than it was before they started to “en
lighten” it—an odious self-annihilation that we cannot mock as a bad
show happening on a stage, because it is not Socrates and Callicles
who blind themselves; it is we, in the streets, who are deprived of our
only fragile lights. No, there is no reason to laugh, because the con
tempt for politicians is still today what creates the widest consensus in
academic circles. And this was written, twenty-five centuries ago, not
by a barbaric invader, but by the most sophisticated, enlightened, lit
erate of all writers, who all his life gorged himself on the wealth and
beauty that he so foolishly destroys or deems irrelevant for producing
political reason and reflection. This sort of “deconstruction,” not the
slow iconoclasm of the present-day sophists, is worth our indignation,
because it parades as the highest virtue, and, as Weinberg claims, as
our only hope against irrationality. Yes! If there has ever been a form
of “higher superstition,” it is seen in the dialogue in Socrates’ fury for
destroying idols and invoking afterworldly, extraterrestrial phantoms.
In a sort of blinding rage, the two sparring partners start killing not
only the arts that make reflexivity possible but each of the slightly less
blind leaders whose experience was crucially important to the practi-
1aI politics of Athens: Themistocles, and Pericles himself. This sinister
loi m of iconoclasm does not occur without a concession by Socrates:
I’m actually not criticizing them in their capacity as servants of the

slate. In fact, I think they were better at serving the state than current
politicians are . . . However, it’s more or less true to say that they
were no better than current politicians as regards the only responsibil
ity a good member of a community has—that is, altering the commu
nity’s needs rather than going along with them, and persuading, or
even forcing, their fellow citizens to adopt a course of action which
would result in their becoming better people. (5i7b-c)
Hill Socrates, as we will see, has deprived the statesmen of all the
imans to obtain this “alteialion,” this “betterment,” this “forcing
Iuih 11011," and so the only (lung that is lelt is either a slavish attach
246
ment to what the people think or a mad flight into a fanciful after
world in which only professors and good pupils would exist. With his
inadequate benchmark Socrates takes upon himself the incredible task
of passing judgment on all of those who, contrary to what he claims,
have led the politics of Athens: “Well, can you name a single rhetorician
from the past who’s supposed to have been instrumental, from his
very first public speech onwards, in changing the Athenian people from
the terrible state they’d been in before to a better one?” (503b).
To which the only devastating answer is that no one has: “It follows
from this argument, then, that Pericles was not a good statesman” (5i6d)
And the straw Callicles agrees, taking with him the real anthropologl
cal Callicles, and Gorgias, and Polus, who of course would have
screamed in indignation against this iconoclasm. Instead of defending
the great invention of a rhetoric adapted to the subtle conditions ol
that other great invention, democracy, the straw Callicles shamefully
accepts Socrates’ judgment.
Among the smoking ruins of those institutions, only one man 111
umphs: “I’m the only genuine practitioner of politics in Athens today
the only example of a true statesman” (52id). One man against all I In
hide the megalomaniacal dimension of this insane conclusion anollu 1
folly is added. After mocking rhetoric for providing only “a phantom
of statesmanship,” Socrates provides an even paler picture. He rule
indeed, but as a shadow, over a demos of shadows: “They’d [the soul |
better be judged naked, stripped of all this clothing—in other woul
they have to be judged after they’ve died. If the assessment is to be Jab
the judge had better be naked as well—which is to say, dead so (h il
with an unhampered soul he can scrutinize the unhampered soul of N
freshly dead individual who isn’t surrounded by his friends and relallt t
and has left those trappings behind in the world” (523e).
How right Nietzsche was to put Socrates at the head of his hit 11si t 1
“men of ressentiment. ”A beautiful scene indeed, this last judgment Iml
totally irrelevant to politics. Politics is not about “freshly dead” pit
pie, but about the living; not about ghoulish stories of the aflei woi Id
but about gory stories of this world. If there is one thing politics dot»
not need, it is yet another afterworld of “unhampered souls " Wli ll
Socrates does not want to consider is that these attachments Ilit»
“friends and relatives,” these “trappings,” aie exactly what oblige 11
to pass judgment now, in the bright sun of Athens, not m the ciepu 111
247
lar light of Hades. What he does not want to realize is that if, by some
nightmarish miracle, all of Athens were made of Socrateses who had,
like him, shed their wise pistis for his didactic knowledge, none o f the
problems of the city would have even begun to be solved. An Athens
made of virtuous Socrateses will be no better off if the Body Politic is
deprived of its specific form of rationality, this unique circulating vir-
tue, which is like its blood.
How Socrates Misconstrues the Work

Done by the Body Politic upon Itself
Socrates’ project is tantamount to replacing the blood of a healthy

body with a transfusion from an altogether different species: it can be
done, but it is too risky to be done without the informed consent of
Ilie patient. If I am using irony and indignation, it is to counterbalance
Ilie old habit that makes us either share Socrates’ demotic hatred or
i mbrace, without further ado, Callicles’ definition of politics as “mere
lorce.” What I want to do with this burlesque style is to focus our at-
Untion on the middle position, that of the Third Estate which does
not ask either for reason or for cynicism. Why is it necessary to make a
II mice between these two positions, even if this choice paralyzes the
body Politic? As with all choices of this sort, it is because iconoclasm
II is broken a crucial feature of action (see Chapter 9). An operator that
as crucial to the common sense of the common people has been
luined into an irrelevant choice—as irrelevant as the epistemologist’s
imteasing question in Chapter 4, “Are the facts real or are they fabri-
i iled?” If we want to speak less polemically, we can say that Socrates’
misiepresentation of the Sophists depends on a category mistake. He
ipplies to politics a “context of truth” that pertains to another realm.
I lie stunning beauty of the Gorgias is that this other context is
li 111ly visible in the very lack of comprehension Socrates displays for
I1.1l it is to re-present the people. I am not talking here about the
modem notion of representation that will come much later, and that
III 11sell be infused with rationalist definitions, but about a com-
|i|i lely ad hoc sort of activity that is neither transcendent nor imma-
III nt but moie closely lesembles a fermentation through which the
I (ople biews ilsell toward a decision never exactly in accordance
lib ilsell, and never led 01 commanded 01 directed from above:
PANDORA' S HOPE
248
“Please tell me, then, which of these two ways of looking after the
state you’re suggesting I follow. Is it the one which is analogous to the
practice of medicine and involves confronting the Athenian people anil
struggling to ensure their perfection? Or is it the one which is analo
gous to what servants do and makes pleasure the point of the opei a
tion? Tell me the truth, Callicles” (521a).
We can ignore for now the childish pleasure Plato takes in making
Callicles answer that it is the second, and focus instead on the reason
for that choice. The choice is as brutal as it is absurd: either head on
confrontation, the teacher’s way, or slavish obsequiousness, the Soph
ist’s way. No teacher, and indeed no servant, has ever behaved like
this—and of course no Sophist either. The choice is so bizarre that It
can be explained only by Socrates’ attempt to bring in a foreign re
source, which makes him ask a totally irrelevant question. We know
where it comes from. Socrates applies to politics a model of geomcli I
cal equality that requires a strict conformity to the model since what I
in question is the conservation of proportions through many diffcmil
relations. Thus the faithfulness of a representation is judged by ll
ability to transport a proportion through all sorts of transformai ion
Either it transports it without deformation, and it is deemed act male
or it transforms it, and it is deemed inaccurate.
As we raw in Chapter 2, in practice the nature of this transformai Ion
is precisely to lose information on its way and to redescribe it in a 1 I
cade of re-representations, or circulating reference, whose precise 111
ture has been as difficult to grasp as that of politics. But tin 111n
like Plato offered only a theory of how demonstration progus til
not of its practice. Thus they could use the idea of a propoilion 1111
problematically maintained through different relations as a hi nth
mark against which to judge all the others. Equipped with this Nl hi
dard, Socrates is going to calibrate every utterance ol llu* |»o l
Sophists: “So that’s the course any young member of the commmill
we’re imagining must follow if he’s wondering how to have a gn at tit
of power and avoid being at the receiving end of wrongdoing Ile inn I
train himself from an early age to share the dictator’s likes and di I
and he must find a way to resemble the dictator as closely as ht tt
(5iod).
Since Socrates voluntarily ignoies all the conditions ol It lit it 1
listed above, when he evaluates the quality ol an lit lei ante it 1 oil ill
249
basis of the resemblance between the source (here the dictator who
represents the spoiled people) and the receptor (here, the young men
thirsty for power): “You’re so incapable of challenging your loved
ones’ decisions and assertions that if anyone were to express surprise at
the extraordinary things they cause you to say once in a while, you’d prob
ably respond—if you were in a truthful mood—by admitting that it’s
only when someone stops them voicing these opinions that you ’ll stop echoing
them” (48ie-482a).
Politics is conceived by Socrates as an echo chamber, and there
should be no difference between represented and representing except
Ihe slight delay that is imposed by the nymph Echo’s narrow band
width. The same is true for obedience to the master. Once the order is
ultered, everyone applies it without deformation or interpretation. No
wonder the Body Politic becomes a rather impossible animal: what
ever it says, it is always the same thing. Echo for representation, echo
lor obedience, minus a little bit of static. No invention, no interpreta
tion. Every perturbation is judged a mistake, misrepresentation, mis
behavior, betrayal. Imitation for Socrates is necessarily total, either
when Callicles repeats what the people say, or when Socrates himself
1epeats what his true love, philosophy, makes him say (482a), or when
l.itesmen force the people to change their bad ways for better ways
(so?a). With this benchmark it is easy to see, in Socrates’ eyes at least,
Ili.it Pericles never improved anyone else and that Callicles simply fol
lows the populace: “Now you’re terribly clever, of course, but all the
une I’ve had occasion to notice that you’re incapable of objecting to any
thing your loved ones say or believe. You chop and change rather than con-
nulnt them. If in the Assembly the Athenian people refuse to accept
in idea of yours, you change tack and say what they want to hear, and your
hi luviour is pretty much the same with that good-looking lad of
I' 1ilampes’” (48id-e). (Let us remember that in this passage Socrates
mu pares his two loves, Alcibiades and philosophy, with Callicles’
I o, Ihe Athenian populace and his minion.)
I veil here, however, Callicles’ behavior—the real Callicles, not the
11 iw one is perfectly adapted to the ecological conditions of the
ipoi a Par from believing in a “diffusionist” model of information that
III liavel unadulteiated no matter what, he uses an excellent “model
il It mslalion" that obliges him lo "change tack” when people “refuse
10 hull his ideas " One can say that Callicles does not hold lo truth
250
when he “chops and changes” only if truth-telling is defined as being con
vinced alone in the afterworld. But if the conditions of felicity are, as
Callicles so aptly defined them above, for courageous statesmen “to
see their policies through to the finish without losing their nerve and
giving up,” then there is no other way than to negotiate one’s posi
tion until every one of those who are party to the deal is convinced
In a democracy this means everyone. In the agora there is never
any echo, but rumors, condensations, displacements, accumulations,
simplifications, detours, transformations—a highly complex chemis
try that makes one stand for the whole, and another chemistry, equally
complex, that (sometimes) makes the whole obey one.
Socrates misjudges the great positive distance between what the rep
resented and the representing are saying, because he judges it accoi d
ing to either slavish resemblance or total difference, the only two mod
els he is able to imagine. This is true for representation as well as loi
obedience. When the citizens repeat what the Body Politic is about, 01
when they obey the law, none of them slavishly transports without dt
formation a piece of information. Socrates’ dream of replacing ull
their subtle translations with a strict didactic form of reasoning, like*
the multiple-choice exams teachers enjoy so much, shows his com
plete ignorance of what it is to be collectively convinced about mallei
for which no one has the definite answer. The Sophists in particulai
had worked out many little tricks and a great treasure of lore to dt il
with the peculiarity of what cannot be considered an echo chamlui
or a schoolroom—but their expertise, after Plato’s onslaught, is lull I
waste. The proof is that even here I employ the words “trick" ami
“lore” to describe an accurate form of knowledge, so powerful is llu»
shadow cast on political reasoning by the notion of information wll li
out deformation—the sort of transportation devised as the theoietkul
justification for geometrical demonstration (see Chapter 2).
Our dialogue catches the specific form of political distance ml
handed, so to speak—that is, just when the deed of destruction Is lu»
ing committed. Later, when the iconoclasts have won the day and Ilu»
dust has settled, the people will be completely unaware that a liugt»
and beautiful statue once stood there. Witness the extraordmai y 11
therly advice that Socrates gives to Callicles, which accuialcly di lint»
the proper form of transcendence within which Callicles is still opt I
ating and which Socrates is quashing befoie oui veiy eyes
251
If you’re under the impression that anyone is going to hand you the
kind of expertise which will enable you to be a political force here while
you’re not assimilated to our system of government (whether this means
that you’re better or worse than it), I think you’ve been misled,
Callicles. If you’re to achieve any kind of meaningfulfriendly relation
ship with the Athenian people.. . then it’s not just a matter of imperson
ation: you have to be inherently similar to them. In other words, it’s
someone who can abolish differences between you and them [ostis ouv se
toutoi omoiotaton apergasetai] who can turn you into a rhetorician and
the kind of politician you aspire to be, because everyone enjoys hearing
their own characteristicpoints of view in a speech and resents hearing any
thing unfamiliar—unless you tell me otherwise, my friend. (si3a-c)
The real anthropological Callicles would have told him otherwise,

1f Plato had not held the stylus and turned him into a straw man. “Not
only mimesis is sufficient but a complete and total assimilation to
Ihe nature of everyone [ou gar mimètèn dei einai all’ autophuôs omoin
loutois].” Never was political reasoning defined so precisely as by the
one who rendered it forever impossible. Autophuôs says it all, defining
wilh incredible precision that strange form of transcendence and that
even stranger kind of reflexivity that remain completely immanent,
since, far from the foolish dreams of transparent representation, Soc
iales endows the Sophists with the power to “grow by themselves”
inlo what all the others are doing and willing. Yes, this is the mysteri
ous quality of politics—which has become a mystery to us but which
politicians fortunately preserve with great skill, hidden in their de-
pised tricks and lore.
lb read Callicles’ calling as immanence, as “assimilation” that
ibolishes difference,” is to miss the very specific form of transcen-
ilmce that occurs when the whole represents itself reflexively to the
hole, through the mediation of one who takes it upon himself (or
lu 1self) to be everyone else—exactly the sort of thing that Socrates is
o incapable of doing that he flees from the agora with one or
I o young men and fulminates against Athens from the safe and
uouexislent standpoint of Hades. By reading this alchemy as represen-
I il ton, we miss it exactly as much as Socrates did—and this is the
it il advantage of the Sophists They offered a dark definition of the
I oily Politic’s “leimentation," instead of the mythically clear self-
II p u sentation that was invented in the modernist peiiod. Manipula
252
tions, differences, combinations, tricks, rhetorics, contribute to that

slight difference between the Body and itself. Neither organic bliss noi
rationalist transparency: this was the knowledge of the Sophists, ex
pelled from the Republic by the philosopher king.
We are not faced here with one transcendence. Reason, against the
immanence of populist leaders, but with two transcendences, one ad
mirable to be sure, that of geometrical demonstration, and the othei
exactly as admirable although utterly distinct, which obliges the whole
to deal with itself without the benefit of guaranteed information
Viewed from Socrates’ remote standpoint, the aim of politics is as im
possible as the bootstrapping of Baron von Munchausen. The demos,
deprived of knowledge and of morality, needs outside help in ordei
to stand up, and Socrates, generously enough, offers to give it a hand
But if this help were accepted it would not raise the people one inch
The specific transcendence it needs to bootstrap itself is not that ol a
lever coming from the outside, but much more like the kneading ol a
dough—except that the demos is at once the flour, the water, the
baker, the leavening ferment, and the very act of kneading. Yes, a
fermentation, the sort of turmoil that has always seemed so terrible
to the powerful, and that has nonetheless always been transcendent
enough to make the people move and be represented.
As I said in the previous chapter, the Greeks made one invention
too many, either geometry or democracy. But it is a matter of histoi I
cal contingency that we have inherited this impossible Body Politic
Nothing in principle, except a lack of nerve, obliges us to choose be
tween the two inventions and to forgo part of our rightful heritage 11
Socrates had not, by mistake, tried to substitute one type of demon
stration, geometry, for another, mass demonstration, we would be able
to honor the scientists without despising the politicians. It is true that Ilu
skills of politics are so difficult, so strenuous, so counterintuitive, and
require so much work, so many interruptions, that, to paraphiase
Mark Twain’s, “there is no extremity to which a man will not go to
avoid the hard work of thinking politically.” But the mistakes ol out
forefathers do not prevent us from sorting out their deeds and adopt
ing their good qualities without their defects.
Before we can conclude and restore the two transcendences at once
with the fragile plausibility of this archaeology fiction, we have to un
derstand one last bit about the dialogue. Why is it so ollen (al en us n
253
discussion about morality? I want to argue that, in spite of the moving

commentaries by moral philosophers, the ethical questions debated
by Socrates and Callicles are so many red herrings. Every time the
rhetoricians say something to prove that Socrates’ requirements are
totally irrelevant to the task at hand, Socrates reads it as proof that
Sophists are uninterested in moral standing. With admirable irony, he
delivers, for instance, the following challenge: “Is there anyone—from
here or elsewhere, from any walk of life—who was previously bad
(that is, unjust, self-indulgent, and thoughtless), but who has become,
thanks to Callicles, a paragon of virtue?” (515a).
We should not hasten to answer that politics and morality are, of
course, two different things, and that, naturally, no one has asked
Callicles to turn all citizens into “paragons of virtue”—because if we
concede this, we still accept the Machiavellian definition of politics as
being unconcerned with morality. This would be to live under Callicles’
and Socrates’ settlement, to take politics as the degraded exercise of
conserving power a little longer without any hopes for betterment.
This would be playing right into Socrates’ hands, because such a disre
gard for morality is exactly what he wants for the people of Athens
without him, and what Machiavelli will later overesteem as a positive
definition of political cleverness—although Machiavelli’s own posi-
lion is, of course, not a wholly immoral one.
In fact Plato’s perversity is much greater than this. If by morality we
mean efforts to provide the Third Estate with ways and means by
which to represent themselves to themselves in order to decide what
lo do next in matters about which there is no definite knowledge, then
Socrates is exactly as immoral as Callicles, as I showed earlier, since
each is competing with the other over how best to break the major-
ily rule. If anything, Socrates is much worse, since, as we have just wit
nessed, he systematically destroys what makes representation
1Ificient; whereas Callicles, in spite of Plato’s rewrite, still presents,
even through his blunders, a vague reminiscence of proper political
si ills the real Sophists being dimly visible through their straw coun-
lei parts.
Actually, Socrates’ crime is mind-boggling because he manages, by a
III I le shift, to take away horn the Third Estate exactly the sort of moral
Ik liavior that evcryom ai*K(s on, and then to turn this behavior into an
impossible Iasi Ibut urn be iKinmplishcd only by following his own
254
impossible requirements—the whole thing ending, as we know, in tlu*

afterworld of shadows. Quite a feat! And one that, in my view, should
be met with grinding of teeth rather than cheers of admiration.
Gorgias, the first to enter the scene, is easily paralyzed by the echo
chamber argument. Exit poor Gorgias. Then Polus is the first to fall
into the ethical trap. The question raised by Socrates appears so irrelo
vant that it works perfectly to divert attention from his own misun
derstanding of political representation: “It follows that wrongdoing Is
the second worst thing that can happen; the worst thing in the woild
the supreme curse, is to do wrong and not pay the penalty for it
(479d); “I also claim that to steal, enslave, burgle, and in short to do
any kind of wrong against me and my property, is not only worse loi
the wrongdoer than it is for me, the target of his wrongdoing, but I
also more contemptible” (508e).
We need an enormously long conditioning to see this question a
crucially important. Even if morality were taken as nothing more than
a sort of basic ethological aptitude of primates in groups, it would st III
be pretty close to such an assessment. The only thing Socrates adds to
turn this into a “big question” is the strict and absolute order of pi loi
ity that he imposes between suffering wrongdoing and doing it. In 1
actly the same way as the absolute difference between knowledge and
know-how was imposed by a coup deforce for which we had only Sin
rates’ words (see Chapter 7), the absolute difference between what 1 v
ery moral animal believes and what Socrates’ higher morality requii t
is to be imposed by force.
Actually something else is needed, and that is, as usual, the sti ivy
Sophists’ slavish behavior. It is Polus who makes us believe that wc ai p
dealing here with a revolutionary statement: “If you’re serious, and ||
what you’re saying really is the truth, surely human life would hi*
turned upside down, wouldn’t it? Everything we do is the opposite n|
what you imply we should be doing” (481c). It is great luck for Sou alp
that Plato hands him foils like this one, because, without the Soph I |
indignation, what Socrates says and what the common people a
would be undistinguishable. As is usual with revolutionary speeihp
there is no safer way to make a revolution than to say that you ai y
making one!
What is so extraordinary is that Socrates, very late in the dialogue
recognizes the obvious commonsense natuie of what he has spun 1
much strenuous elloil to piovc “All I’m saying is what I always n ly |
255
myself don’t know the facts of these matters, but I ’ve never met anyone,
including the people here today, who could disagree with what I’m say
ing and still avoid making himself ridiculous”(509a). Is this not a clear
confession that all this long debate with Polus on how to rank moral
behavior was never doubted by anyone for any length of time? Every
one is relatively bound by the Golden Rule. It is only if you want to
Iurn it into an absolute demarcation between suffering and doing evil
Ihat it can fail to enlighten you. Exit Polus.
The same paralyzing trick is going to work on poor Callicles who,
alter appealing, as we saw, to natural laws against conventional laws,
is immediately turned into someone who demands unlimited hedo
nism! This smokescreen is very efficient at hiding how close Socrates’
solution is to Callicles’ own. And here again, after a lengthy acrimoni
ous disputatio in which Callicles conveniently plays the unrestrained
beast of prey—as if beasts of prey were themselves unrestrained! as if
wolves behaved like wolves, and hyenas like hyenas!—Socrates can
didly confesses the basic ethological nature of the morality he, like ev-
ei y slave, child, or, for that matter, chimpanzee (De Waal 1982), relies
on: “We shouldn’t refuse to restrain our desires, because that condemns
us to a life of endlessly tryingto satisfy them. And this is the life of a pred
atory outlaw, in the sense that anyone who lives like that will never be on
i>ood terms with anyone else—any other human being, let alone a god—
since he’s incapable of co-operation, and co-operation is a prerequisite for
p tendship”{soje).
1don’t know about the gods, about whom ethological knowledge is
Inn, but I am confident that even Shirley Strum’s baboons and Steve
(ilickman’s hyenas, if they could read Plato, would applaud this de-
11îplion of relative morality in social groups (Strum 1987). The amus
ing thing is that no one ever said the opposite except the straw Callicles
is portrayed by Plato! The mythology of the war of all against all that
llueatens to engulf civilization if morality is not enforced is told only
by those who have withdrawn from the people the basic morality that
(a lability has imposed for millions of years on animals in groups.
I Ins should be obvious but is not—because, unfortunately, moral phi
losophy is a narcotic as addictive as epistemology, and we cannot eas
ily I ick the habit of thinking that the demos lacks morality as totally
is it lacks cpislcnut knowledge Even Socrates’ admission that what
lit says is common sense and in no way 1evolutionary is not enough.
I vt n Callules' sneci Ing m u 111 that questions of moiahly aie totally
256
irrelevant to the discussion of political rhetoric is not enough: “Whal

I’ve been thinking about is the adolescent delight you take in seizing on
any concession someone makes to you, even if he means it as a joke
Do you really think that lo r anyone else would deny that there are bettei
and worse pleasures?” (499b).
No one denies what Socrates says! No matter what the evidence,
moral philosophers portray the Gorgias as the magnificent fight of a
high-minded Socrates offering the people a goal too high for them to
achieve. It is a fight, yes, but one fought by Socrates to impose on the
people a definition of morality that they always possessed, minus the
ways to apply it (Nussbaum 1994)! What Socrates does to the demos
of Athens is something as blatantly absurd as if a psychologist, let’s say
from America, went to China, and working under the very chauvinisl
conceit that “Chinese people all look alike,” decided to paint big num
bers on them so as to make them recognizable at last. With whal
glares will he be met when he arrives with his brush and his bucket ol
paint and his candid psychological explanation? Can we think that Ilu
inhabitants of the huge city of Shanghai will welcome this new way ol
recognizing one another, because for centuries they have been unahh
to tell one another apart? Of course not, they will jeer the psychologisl
away and rightly so, and “his head will spin and his mouth will gapt
there in that world”! Yet Socrates’ use of the morality question in Ilu
Gorgias is based on exactly the same sort of vast misunderstanding
The Chinese do recognize one another without the use of big pain UJ
numbers. The demos is endowed with all the morality and all Ilu
reflexive knowledge it needs in order to behave itself.
Conclusion: Socrates’Deal and Death
If we bring together all the successive moves that Plato makes Sou aIf
play on the stage, we have a very tricky juggling act:
In the first move, Socrates takes away from the people of Al he 11
their basic sociality, their basic morality, their basic know how, with II
no one before had ever denied they possessed.
Then, in a second move, stripped of all their qualities, the people u P
portrayed as children, as beasts of prey, as spoiled slaves leady to ul
tack one another at their slightest whim. Sent down to the Cim
grasping only at shadows, they begin a wai ol all against all
257
Third move: something needs to be done to keep this horrifying

mob at bay and set up order against their disorder.
It is at this point that, with trumpet flourishes, the solutions arrive,
Reason and Morality. That is the fourth move. But when these are
handed back to the people by Socrates, from the exotic realm of geo
metrical demonstration, the people cannot recognize what has been
taken from them, because there is one thing added and one thing miss
ing! What has been added during the passage in the realm of shadows
is an absolute requirement that renders morality and know-how in
efficient. What has been subtracted is all the practical mediations
through which the people could turn their relative knowledge and rel
ative morality to good use in the specific conditions of the agora.
Fifth move: Professor Socrates writes on the blackboard his trium
phant equation: politics p lu s absolute morality m in u s practical means
equals the Impossible Body Politic.
Sixth move, the most dramatic: since the Body Politic is impossible,
let us send the whole thing to hell! The d eu s ex m a ch in a descends and
the three judges of Hades condemn everyone to death— ex cep t Socrates
and “a few other souls”!2 Clap, clap, clap . . .
Let me be naughty (just one last time, I promise) and explain the
seventh move, which is the epilogue of this show, which will take
place once the crowd has gone home. Is there another explanation, in
Ilie end, for this very famous and fair trial, through which the people
of Athens forced Socrates to poison himself? To be sure, it was a polit
ical mistake, because it made a martyr out of a mad scientist—but it
might have been, at least, a healthy reaction against Socrates’ most un-
l.iir trial of the demos. Was it not fair for someone who wanted to
judge naked shadows from the superior seat of eternal justice to be
sent to the Isles of the Blessed by the living and fully clothed citizens
ol Athens? But as we shall now see, this tragicomedy had a great ad
vantage over the later ones: that only one character shed his blood,
md he was not part of the public.l
l "Onasionally, however, IRhadmanthys] comes across a different kind of soul, one

liith has lived a life ol moi al inlegnly, and which belonged to a man who played n o
u l m / n i h l n l i f t 01 lo u philosophei who minded his o w n business and remained d e -
ti< h < < / f r o m l / u n i f s I liiouglintil III lilt ( { /f ih i)
258
Science Wars? What about Peace?

Let’s now abandon the irony and the rage that were needed to press
away the poison and extract the honey. We can now salvage from the
Gorgias the powerful definition of real politics to which epistemic
knowledge and absolute morality are so obviously irrelevant. The cate
gory mistake is now clear enough. Socrates’ and Callicles’ settlement
can no longer prevent us from liking scientists as much as politicians
Contrary to what Weinberg asserts after Plato, there are many possible
settlements other than the one I described as “inhumanity to quash in
humanity.” A slight change in our definition of science and in out
definition of politics will be sufficient, at the end of this chapter, to
show the many ways we can now go.
A Science Freedfrom the Politics o f Doing A w ay with Politics
First let’s see, briefly, how the sciences can be freed from the burden ol
making a type of politics that shortcuts politics. If we now cal ml
read the Gorgias, we recognize that a certain specialized form of u 1
soning, epistèmè, was kidnapped for a political purpose it could not
possibly fulfill. This has resulted in bad politics but in an even woi y
science. If we let the kidnapped sciences escape, then two dillmnl
meanings of the adjective “scientific” become distinguishable again ul
ter being lumped together for so long.
The first meaning is that of Science with a capital S, the ideal ot
the transportation of information without discussion or déformai Ion
This Science, capital S, is not a description of what scientists do. Ion 1»
an old term, it is an ideology that never had any other use, in the ( pi
temologists’ hands, than to offer a substitute for public discussion ll
has always been a political weapon to do away with the constraint 11
politics. From the beginning, as we saw in the dialogue, it was lailou I
for this end alone, and it has never stopped, through the ages, lu In
used in this way.
Because it was intended as a weapon, this conception of Scieiu t ill
one Weinberg clings to so forcefully, is usable neither to “mal 1 lin
manity less irrational” nor to make the sciences better. It has only on
use: “Keep your mouth shut!” the “you” designating, mleitslln I
enough, other scientists involved in conlioveisies as much as Ilu p«
259
pie in general. “Substitute Science, capital S, for political irrationality”

is only a war cry. In that sense, and that sense only, it is useful, as we
can witness in these days of the Science Wars. However, this
definition of Science No. 1, I am afraid, has no more use than the
Maginot Line, and I take great pleasure in being branded as
“antiscientific” if “scientific” has only this first meaning.
But “scientific” has one other meaning, which is much more inter
esting and is not engaged in doing away with politics, not because it is
apolitical or because it is politicized, but because it deals with entirely
different questions, a difference that is never respected when Science
No. l is taken, by its friends as well as by its foes, as all there is to say
about science.
The second meaning of the adjective “scientific” is the gaining of ac-
i ess, through experiments and calculations, to entities that at first do
not have the same characteristics as humans do. This definition may
seem odd, but it is what is alluded to by Weinberg’s own interest in
impersonal laws.” Science No. 2 deals with nonhumans, which in the
I>cginning are foreign to social life, and which are slowly socialized in
mir midst through the channels of laboratories, expeditions, institu-
lions, and so on, as recent historians of science have so often de-
si 1ibed. What working scientists want to be sure of is that they do not
make up, with their own repertoire of actions, the new entities to
which they have access. They want each new nonhuman to enrich
Iheir repertoire of actions, their ontology. Pasteur, for example, does
tit>1 “construct” his microbes; rather his microbes, and French society,
II e changed, through their common agency, from a collective made up
ol, say, x entities into one made up of many more entities, including
mit lobes.
I lie definition of Science No. 2 thus alludes to the maximum possi
ble distance between standpoints as different as possible and to their in
mate integration into the daily life and thoughts of as many humans
i 1xissible. To do justice to this scientific work, Science No. 1 is totally
In idequale, because what Science No. 2 needs, contrary to Science
o 1, is lots of controversies, puzzles, risk-taking, imagination, and a
\ isculaii/ation” with the rest of the collective as rich and as com-
pli ns possible. Naluially, these many points of contact between hu-
nnns and nonluunans uie unthinkable either if by “social” we mean
( tllulos’ pine huilai loue, 01 il by "leason" we mean the mouth
2ÔO
shutting Science No. 1. We recognize here, by the way, the two enemy
camps between which science studies is trying to gain a foothold:
those from the humanities who think we give too much to the
nonhumans; and those from some quarters of the “hard” sciences who
accuse us of giving too much to the humans. This symmetrical accusa
tion triangulates with great precision the place where we in science
studies stand: we follow scientists in their daily scientific practice in
the No. 2 definition, not in the No. 1, politicized definition. Reason—
meaning Science No. 1—does not describe science better than cyni
cism describes politics.3
So freeing science from politics is easy—not, as has been done in the
past, by trying to isolate as much as possible the autonomous core o(
science from the deleterious pollution by the social—but by liberating
Science No. 2 as much as possible from the political disciplining thal
went with Science No. 1 and that Socrates introduced into philosophy
The first solution, inhumanity against inhumanity, relied too much on
a fanciful definition of the social—the mob that has to be silenced and
disciplined—and on an even more fanciful definition of Science No. 1,
conceived as a type of demonstration with no other goal than to bring
in the “impersonal laws” to stop controversies from boiling over. The
second solution is the best and fastest way to free science from poll
tics. Let Science No. 2 be represented publicly in all its beautiful origi
nality, that is, as what establishes new, unpredictable connections be
tween humans and nonhumans, thus profoundly modifying what tin
collective is made of. Who defined it most clearly? Socrates—and hei e
I want return to the passage I started with and make amends for hav
ing ironized so much at the expense of this master of irony: “In lad
Callicles, the experts’ opinion is that co-operation, love, order, disc 1
pline, and justice bind heaven and earth, gods and men. That’s why the y
3. A third meaning of “scientific” could be added, which I will call l o g i s t i c s bcutu « ll

is directly connected to the n u m b e r of entities one wants to access and to socialize Jil I
as there is a logistical problem to be solved if 20,000 fans are simultaneously liying I
park near a baseball stadium, there is a logistical problem to be solved if masses ol d ilil
have to be transported over a great distance, treated, soiled, “paiked,” summail til
and expressed. Much of the common usage of the adjective “scientific” lelers to (III I
gistical question. But it should not be confused with the other two, especially nnl will
science as access to nonhumans. Science No. 3 ensuies dial last and sale commiiuli l
tions of data aie established, it does no! cnsuie dial something sensible is cai 1it d o cl
“(iaihage in, gaihagc oui" as the compulci motto goes
26l
call the universe an ordered whole, my friend, rather than a disorderly

mess or an unruly shambles [kai to olon touto dia tauta kosmon kalousin, 0
etaire, ouk akosmian oude akolasian]”(so7e-5o8a).
Far from taking us away from the agora, Science No. 2—once clearly
separated from the impossible agenda of Science, capital S—redefines
political order as that which brings together stars, prions, cows, heav
ens, and people, the task being to turn this collective into a “cosmos”
instead of an “unruly shambles.” For scientists such an endeavor
seems much more lively, much more interesting, much more adapted
to their skill and genius, than the boring repetitive chore of beating
the poor undisciplined demos with the big stick of “impersonal laws.”
This new settlement is not the one Socrates and Callicles agreed on—
“appealing to one form of inhumanity to avoid inhumane social be
havior”—but something that could be defined as “collectively making
sure that the collective formed by ever vaster numbers of humans and
nonhumans becomes a cosmos.”
For this other possible task, however, we not only need scientists
who will abandon the older privileges of Science No. 1 and at last take
up a science (No. 2) freed from politics, we also need a symmetrical
transformation of politics. I confess that this is much more difficult,
because, in practice, very few scientists are happy in the artificial
straitjacket that Socrates’ position imposes on them, and they would
be very happy to deal with what they are good at, Science No. 2. But
what about politics? To convince Socrates is one thing, but what about
Callicles? To free science from politics is easy, but how can we free
politics from science?
F reein g P o litic s f r o m a P o w e r /K n o w le d g e
th a t M a k e s P o litic s Im p o ssib le I
I be paradox that is always lost on those who accuse science studies of

polilicizing science is that it does exactly the reverse, but that, in doing
so, it meets another, much stronger opposition than that of episte-
mologists or of a few disgruntled scientists. If the battle-lines of the
so called Science Wars were drawn in any plausible way, the people
III e us who aie said to “fight” science would be heartily supported by
ballalions liom the social sciences 01 the humanities. And yet here too
II is exactly the leveise Science No 1 is a scandal to sociologists and
humanists alii e because it tot illy subvcils (he definition ol the social
262
they work with—whereas it is common sense to the scientists, who ai 0

worried of course, but only at seeing their unwieldy Science No. 1
taken away from them. The opposition from those who believe in the
“social” is a lot more acrimonious than our (on the whole) friendly ex
changes with our contradictors from the scientific ranks. How is this
possible?
In this too the settlement between Socrates and Callicles can en
lighten us, although this is much harder to comprehend. As we saw
earlier, when deciphering the tug-of-war between Reason and Foiu
on the one hand and the demos on the other, there are two meanings
of the word “social.” The first, Social No. 1, is used by Socrates against
Callicles (and accepted by the straw Callicles as a good definition ol
force); the other, Social No. 2, should be used to describe the specdii
conditions of felicity for the people representing itself, conditions thill
the Gorgias reveals so well even as Socrates smashes them to pieces.
I want to indicate here, as I did in Chapter 3, that the two mean
ings of “social” are as different as Science No. 1 and Science No. 2. No
wonder: the ordinary notion of the social is patterned on the same
rationalist argument as that of Science with a capital S—it is a tran
portation without deformation of inflexible laws. It is called “powi 1
instead of “epistèmè,” but this makes no difference, because while
epistemologists speak of the “power of demonstrations” sociologist
are happy to use their most famous recent motto: “Knowledge
Power.” The damning irony of the social sciences is that when they u e
this Foucaldian expression to exert their critical skill they in effect say
without realizing it: “Let the agreement of Socrates (Knowledge) mill
Callicles (Power) stand and triumph over the Third Estate!” No ci it k il
slogan is less critical than this one, no popular flag is more ehtl I
What makes this argument difficult to grasp is that natural and soi I il
scientists are both behaving as if Power were made of another mat I«I
altogether than Reason—hence the supposed originality of scpaiulln
them and then reuniting them with a mysterious slash. The critics Hit1
taken in by Socrates’ and Callicles’ show. Power and Reason me on
and the same, and the Body Politic built by one or by the ol lu I I
shaped with the same clay; hence the uselessness of the slash, wlilill
heightens the interest for the players, and for the ciilics in llieli In
seats, while boring the audience to teai s.
It seems that after the Gorgias political philosophy nevei iccovtit I
the full right it once possessed to think over its spec die condition 11
263
felicity and to build the Body Politic with its own flesh and blood. The
factish*, once smashed, can be patched up but never made whole
again. Barbara Cassin, to be sure, has shown beautifully how the sec
ond Sophistics won against Plato and reestablished rhetoric over phi
losophy. But this millennium of Pyrrhic victories counted for naught
once, in the seventeenth century, another treaty again linked Science
and Politics into a common settlement—especially after Machiavelli
fell into Socrates’ trap and defined politics as a cleverness entirely
freed from scientific virtue. Hobbes’s Leviathan is a rationalist Beast
through and through, made of arguments, proofs, cogs, and wheels. It
is a Cartesian animal-machine which transports power without discus
sion or deformation.
Again, Hobbes was used as a foil against reason, much as Callicles
was used against Socrates, but the common settlement is even clearer
m the seventeenth century than twenty centuries earlier: natural laws
and indisputable demonstrations now make for a rationally founded
politics. The conditions of felicity for the slow creation of a consensus
m the harsh conditions of the agora disappeared underground. There
is even less genuine politics in Hobbes than in Socrates’ appeal from
an afterworld. The only difference is that Socrates’ Body Politic has
I>een called back from the dead, to become a Leviathan of this world, a
monster and a half, composed only of “unhampered” individuals half
dead, half-alive, “without trappings, without clothes, without relatives
mid friends” (523c)—a scenography altogether more ghoulish than the
one imagined by Plato.
Things don’t get any better when the Body Politic, to escape from
I lobbesian cynicism, is given another transfusion of Reason by Rous-
eau and his descendants. The impossible surgery started by Socrates
( ontinues on an even bigger scale: more Reason, more artificial blood,
bul less and less of this very specific form of circulating fluid that is
Ilie essence of the Body Politic, and for which the Sophists had so
many good terms and we so few. The Body Politic is now supposed to
be liansparent to itself, freed from the manipulations, dark secrets,
i level ness, tricks of the Sophists. Representation has taken over, but a
it piesentation understood in the very terms of Socrates’ demonstra
tion By pretending to clean Glaucus’s statue of all its later deforma
tions, Rousseau makes the Body Politic even more monstrous.
Should I go on with the sml stoiy ol how to transform a once
lit ill hy Body into nil even moie unvl ihle mid dangerous monstei? No,
264
no one wants to hear more horrific stories, all in the name of Reason
Suffice it to say that when a “scientific politics” is finally invented,
then even worse monstrosities come hard and fast. Socrates had only
threatened to leave the agora alone, and only his blood was shed at llu
end of his strange attempt at rationalizing politics. How innocent il
looks to children of our century! Socrates could not have imagined
that scientific programs could later be invented to send the whole ol
the demos into the afterworld and to replace political life with llu
iron laws of one science—and economics at that! The social scientc
in most of their instantiations represent the ultimate reconciliation ol
Socrates with Callicles, since the brute force advocated by the lalUi
has become a matter of demonstration—not through geometntal
equality, of course, but through new tools such as statistics. Every sin
gle feature of our definition of the “social” now comes from Social!
and Callicles, fused into one.
I have said enough to make clear why Power/Knowledge is not a so
lution but yet another attempt to paralyze what is left of the Body Poll
tic. To take Callicles’ definition of Power and use it to deconsliml
Reason and to show that, instead of the demonstration of truths, Urn
son involves only the demonstration of force, is simply to reverse Ilie
twin definitions that have been devised to make politics unthinkable
Nothing has been achieved, nothing analyzed. It is black and while III
stead of white and black. The strong hand of Callicles simply lal t
over from the weakening hand of Socrates the rope used in the lug ol
war against the demos, and later, as the slash indicates, Socrates’ h mil
will take over from the tired hand of Callicles! Admirable collaboirt
tion, indeed, but not one that will reinforce the Third Estate, the pi o
pie pulling on the other end of the rope. To sum up the argument 0111 U
again, there is not a single trait in the definition of Reason (lui I
not shared by the definition of Force. Thus nothing is gained by II
ing to alternate between the two or to expand one at the expenst ( I
the other. Everything will be gained, however, if we turn our alU lit li I
toward the sites and situations against which the twin lesouuc I
Force/Reason have been devised: the agora.
It is often said that twentieth-century people's bodies aie Inin I
cated by sugar, slowly poisoned by a fabulous excess of caiboliyili llu
unfit for organisms that have evolved for eons on a sugar pool ilUl
This is a good metaphor for the Body Politic, slowly poisoned by it I il
ulous excess of Reason. I low ill adapted was the cmc* ol Piolessoi Si
265
rates is now, I hope, clear, but how much worse is that of the physi
cian qua physicist Weinberg, who wants to cure the people’s supposed
irrationality by bringing in even more “impersonal laws” in order to
eliminate even more thoroughly the damnable tendency of the mob to
discuss and to disobey. The older settlement had great appeal in the
past, even the recent past, because it seemed to offer the fastest way to
transform the unruly shambles of gods, heaven, and men, into an or
dered whole. It seemed to provide an ideal shortcut, a fabulous acceler
ation, as compared with the slow and delicate politics of producing
politics through political means, in the way we learned—and then,
alas, unlearned—from the Athenian people. But it has now become ap
parent that instead of adding order this older solution adds disorder
as well.
In the story of the dispute between the cook and the physician with
which Socrates amused the public so much, there was some plausibil
ity in this idea of kicking out the cook and letting the physician dictate
what we should eat and drink. It no longer applies in our “mad cow”
limes, when neither the cook nor the physician knows what to tell
the assembly, which is no longer made up of spoiled brats and “as
sorted slaves” but of grown-up citizens. There is a Science War, but it
is not the one that pits descendants of Socrates against descendants of
t allicles in the rerun of that tired old show: it is the one between “un-
Iuly shambles” and “cosmos.”
Ilow can we mix Science No. 2, which brings an ever greater num
ber of nonhumans into the agora, with Social No. 2, which deals with
the very specific conditions of felicity that cannot be content with
IIansporting forces or truth without deformation? I don’t know, but I
im sure of one thing: no shortcuts are possible, no short-circuits,
md no acceleration. Half of our knowledge may be in the hands of sci-
I nl ists, but the other, missing half is alive only in those most despised
ol all people, the politicians, who are risking their lives and ours in
si icntifico-political controversies that nowadays make up most of our
d.iily bread. To deal with these controversies, a “double circulation”
I I is lo flow effortlessly again in the Body Politic: the one of science
( No 2) freed from politics, and the one of politics freed from science
(No 1). The task of today can be summed up in the following odd sen
ti ncc* Can we loam lo III e scientists as much as politicians so that at
la\( we can bcnclil (mm the C■1ec*l s' two inventions, demonstration
md demouncy?
C H A P T E R N I N E
The Slight Surprise of Action

Facts, Fetishes, Factishes
What a surprise! I seem to have accomplished my task, to have di

mantled the old settlement that held sway over us. The hideout of lilt
kidnappers has been exposed and the nonhumans set free—free, ill il
is, from the squalid fate of providing cannon fodder for the political
wars against the demos while clothed in the drab uniform of “objet 1s
This was a perverse politics indeed, the one that aimed at erasing II
own conditions of felicity and rendering the Body Politic forever Iin
possible.
And yet it is still as if I have achieved nothing. In the previous til »p
ters I multiplied movements that do not follow the straight pall» uf
reason. I proposed many terms to map circuitous moves: labyilnlli
translation, shifting out, shifting down. I made great use of metaplioi
of vascularization, transfusion, connection, and entanglement. In I»
sure, every time I presented an example my description seemed plan I
ble when it followed the complicated detours made by accurale I ul
efficient artifacts, virtuous politics. And yet every time I looked Hi H
crucial moment, for the term that would allow me to jump, in a nIii I
bound, over construction and truth, words failed me. This is not Jil I
the usual inadequacy of general words for the particulars of e pell
ence. It is as if scientific practice, technical practice, and political pi n
tice led into entirely different realms than those of theoiy ol sclent
theory of techniques, theory of politics. Why is it that we until I
readily recover for our ordinary speech what is so tanlah/mgly of lei I
by practice? Why is it that associations of humans and nonlium in til
ways become, once claiified, iedified, and sliaightcned out oil»
THE SLIGHT SURPRISE OF ACTION
267
thing so utterly different: two opposing sides in a war between sub
jects and objects?
Something is missing. Something has been escaping us, chapter af
ter chapter: a way of negotiating a peaceful passage between object
and subject, a way of ending this battle without escalating the
firepower even more. We need a means of bypassing this standoff alto
gether, a vehicle, a figure of speech that, instead of breaking the subtle
language of practice with the intimidating choice “Is it real or is it fab
ricated? You have to choose, you fools!” would provide a different
move, a different register for practice. One thing is sure: once theory
has made its analytical cut, once the noise of the breaking bones has
been heard, it is no longer possible to account for how we know,
how we construct, how we live the Good Life. We are left to try and
patch back together subjects and objects, words and world, society
and nature, mind and matter—those shards that were made to render
any reconciliation impossible. How can we recover our freedom of
passage? How can we be trained again to make this swift, elegant,
efficient “passing shot,” as tennis players say? Why should it be so
difficult when everywhere it seems so easy, so widespread? It seems so
t ommonsensical when we listen to the lessons of practice, and yet
so contradictory, twisted, and obscure when we hear the lectures of
llieory.
Where is the solution? At the point of the break itself I want to at-
lempt in this chapter to make us aware of the very act of smashing
pi actice into bits. Contrary to what the pragmatists believed (and this
Is, in my view, why their philosophies never took hold in the public
mind), the difference between theory and practice is no more a given
Ilian the difference between content and context, nature and society. It
Is a divide that has been made. More exactly, it is a unity that has been
11ac lured by the blow of a powerful hammer.
In the settlement pictured in Figure 1.1, there is one box we have not
Ioik lied yet, and that is the one labeled “God.” I am not alluding here
lo Ihe moderns’ pathetic notion of a God-of-beyond—a supplement of
ou I (or those who have no soul—but to God as the name given to a
llieoiy of action, mastery, and creation that served as the foundation
lot (he old model nist settlement. We have interrogated facts and arti-
luls, we have seen how difficult it is lo understand them as being
misleied 01 tonsilucted, lull we have not yel inquired inlo maslery
268
and construction themselves. This is what I want to do now, because I
know full well that, without this, no matter how good we are at do
scribing the intricacies of practice we will be immediately attacked ni
iconoclasts who want to destroy science and morality. I, an icono
clast?! Nothing irritates me more than being presented as provocative
or even critical. Especially when such an accusation—or, worse, such a
compliment—comes from those who have broken all our figures ol
speech, from all those descendants of Socrates, one of the first icon
breakers in the lengthy genealogy of idol-smashers who have made* u
modern. The bitter irony is that iconophiles like me are forced to dt
fend ourselves against iconoclasts. How can this be done? By des Itoy
ing them and taking our revenge, adding more debris to the debris U11
by critiques? No, by another means. By suspending the crushing blow
of the hammer.
Let’s start, not at the beginning of this long history as we jusl 11ItI
with Socrates, but at its very end. We will take as our exemplar a 1U
ter-day iconoclast, one of the courageous critics the moderns have ftrlil
around the world to extend the reach of reason, who learn the h ml
way why they should have suspended their critical gesture instead
Two Meanings of Agnosticism

His name is Jagannath, and he has decided to break the spell ol ta 11*
and untouchability by revealing to the pariahs that the sacred sally a
the powerful stone that protects his high-caste family, is nothing to I
afraid of (Ezechiel and Mukherjee 1990). When the pariahs aie gllll
ered in the courtyard of his family estate, the well-meaning iconot 11 I
to the horror of his aunt, seizes the stone and, crossing the foihiddtl
space that separates the Brahmins from the untouchables in the colli
pound that they share, carries the object to be desecrated by Ilit* |M I
slaves. Suddenly, in the middle of the courtyard, in the bla/ing HI
Jagannath hesitates. It is this very hesitation that I wanl to list a IH
starting point:
Words stuck in his throat. This stone is nothing, but 1 bave Nt I III
heart on it and I am reaching it for you: touch it; touch Ihe vulni'ia
ble point of my mind; this is the lime ol evening piayei, (mu h ill
nandadeepa is burning still Those standing behind me | Ills auul HHI
269
the priest] are pulling me back by the many bonds of obligation.
What are you waiting for? What have I brought? Perhaps it is like
this: this has become a saligram because I have offered it as stone. If
you touch it, then it would be a stone for them. This my importu
nity becomes a saligram. Because I have given it, because you have
touched it, and because they have all witnessed this event, let this
stone change into a saligram, in this darkening nightfall. And let the
saligram change into a stone. (101)
But the pariahs recoil in horror:
Jagannath tried to soothe them. He said in his everyday tone of a
teacher: “This is mere stone. Touch it and you will see. If you don’t,
you will remain foolish forever.”
He did not know what had happened to them, but found the entire
group recoiling suddenly. They winced under their wry faces, afraid
to stand and afraid to run away. He had desired and languished for
this auspicious moment—this moment of the pariahs touching the
image of God. He spoke in a voice choking with great rage: “Yes,
touch it!”
He advanced towards them. They shrank back. Some monstrous
cruelty overtook the man in him. The pariahs looked like disgusting
creatures crawling upon their bellies.
He bit his underlip and said in a firm low voice: “Pilla, touch it!
Yes, touch it!”
Pilla [an untouchable foreman] stood blinking. Jagannath felt
spent and lost. Whatever he had been teaching them all these days
had gone to waste. He rattled dreadfully: “Touch, touch, you TOUCH
11'!” It was like the sound of some infuriated animal and it came tear
ing through him. He was sheer violence itself; he was conscious of
nothing else. The pariahs found him more menacing than Bhutaraya
11he demon-spirit of the local god]. The air was rent with his
si 1earns. “Touch, touch, touch.” The strain was too much for the pa-
1whs. Mechanically they came forward, just touched what Jagannath
was holding out to them, and immediately withdrew.
Ixhausted by violence and distress Jagannath pitched aside
Ilie saligram. A heaving anguish had come to a grotesque end. Aunt
could be human even when she treated the pariahs as untouchables.
I le had lost his humanity for a moment. The pariahs had been mean
ingless things to him I le hung his head. He did not know when the
pm lahs had gone Dili I ness had lallen when he came to know that he
was all by hnnsell Disgusted with lus own poison he began to walk
270
about. He asked himself: when they touched it, we lost our human
ity—they and me, didn’t we? And we died. Where is the flaw of it all,
in me or in society? There was no answer. After a long walk he came
home, feeling dazed. (98-102)
Iconoclasm is an essential part of any critique. But what does the

critic’s hammer smash? An idol. A fetish. What is a fetish*? Some
thing that is nothing in itself, but simply the blank screen onto which
we have projected, erroneously, our fancies, our labor, our hopes and
passions. It is a “mere stone,” as Jagannath tries to convince himscK
and the pariahs. The difficulty, of course, lies in explaining how a fet
ish could be at once everything (the source of all power for the believ
ers), nothing (a simple piece of wood or stone); and a little bit ol
something (that which can reverse the origin of action and make one
believe that, through inversion, reification, or objectification, the ol)
ject is more than the product of one’s own hands). Yet somehow the
fetish gains in strength in the hands of the anti-fetishists. The more you
want it to be nothing, the more action springs back from it. Hence the
disquietude of the well-intentioned iconoclast: “This has become «
saligram because I have offered it as a stone.”
What has the courageous iconoclast broken? I contend that it is not
the fetish that has been destroyed, but instead a way of arguing and at I
ingthat used to make argument and action possible and that I now want to
recover (“when they touched it, we lost our humanity—they and me
didn’t we? And we died”). This is the most painful aspect of anil
fetishism: it is always an accusation. Some person, or some people, aie
accused of being taken in—or worse, of cynically manipulating credo
lous believers—by someone who is sure of escaping from this Mill
sion and wants to free the others as well: either from naive belief 01
from being manipulative. But if anti-fetishism is clearly an accusal lot
it is not a description of what happens with those who believe or 110
manipulated.
Actually, as Jagannath’s move beautifully illustrates, it is the ci it lull
thinker who invents the notion of belief and manipulation and pHtjul
this notion upon a situation in which the fetish plays an enlnely dllkl
ent role. Neither the aunt nor the priest ever consideied the sallgiwiil
as anything but a mere stone. Never. By making it into the powcitul
object that must be touched by the punahs, Jagannath 11nnsul>sl mil
271
ates the stone into a monstrous thing—and transmutes himself into a

cruel god (“more menacing than Bhutaraya”)—while the pariahs are
transmogrified into “crawling beasts” and mere “things.” Contrary to
what the critics always imagine, what horrifies the “natives” in the
iconoclastic move is not the threatening gesture that would break
their idols but the extravagant belief that the iconoclast imputes to
them. How could the iconoclast demean himself to the point of believ
ing that we, the natives, should believe so naively—or manipulate so
cynically, or fool ourselves so stupidly? Are we animals? Are we mon
sters? Are we mere things? This is the source of their shame, mistak
enly read by the critic as the horror these naive believers should feel
when faced with the desecrating gesture that exposes, or so the critic
believes, the emptiness of their creed.
In reality the hammer strikes sideways, landing on something other
than what the iconoclast wanted to break. Instead of freeing the pari
ahs from their abject condition, Jagannath destroys his own humanity,
and that of his aunt, along with the humanity of those he believed he
was liberating. Somehow humanity depended on the undisturbed pres
ence of this “mere stone.” Iconoclasm does not break an idol, but de
stroys a way of arguing and acting that was anathema to the icono
clast. The only one who is projecting his feelings onto the idol is he,
the iconoclast with a hammer, not those who by his gesture should be
freed from their shackles. The only one who believes is he, the fighter
of all beliefs. Why? Because he (I use a masculine pronoun, and that
serves him right!) believes in the feeling of belief*, a very strange feel
ing indeed, one that may not exist anywhere but in the iconoclast’s
mind.
As we saw in Chapter 5, belief, naive belief, is the only way for the
iconoclast to enter into contact, violent contact, with the others—ex
actly as epistemologists had no other way of contrasting Pasteur and
Pouchet than to say that the latter believed and the former knew. Be
lief, however, is not a psychological state, not a way of grasping state
ments, but a polemical mode of relations. It is only when the statue is
lut by the violent blow of the iconoclast’s hammer that it becomes a
potential idol, naively and falsely endowed with powers that it does
not possess the pi oof being, for the critic, that it now lies in pieces
and nothing happens Nothing but the indignant bewilderment of
those who loved Ihe statue those who weie accused of being taken in
272
by its power and who now stand “liberated” from its sway—but as tin*
novel shows, what lies in ruins in the middle of the desecrated family
temple is the humanity of the icon-breaker.
Before it was smashed to bits, the idol was something else, not 11
stone mistaken for a spirit or any such thing. What was it? Can we 1e
trieve a meaning that would bring the broken pieces back togethci I
Can we, like archaeologists, repair the damage inflicted by time, timl
greatest of all iconoclasts? We can begin by dusting off the broken
shards that we use in our language today, forgetting that they wm
once joined.
“Fetish” and “fact” can be traced back to the same root. The fai t I
that which is fabricated and not fabricated—as I discussed in Chaplt 1
4. But the fetish too is that which is fabricated and not fabricated 1
There is nothing secret about this joint etymology. Everyone say
it constantly, explicitly, obsessively: the scientists in their laboraloiy
practice, the adepts of fetishist cults in their rites (Aquino and Bai m
1994). But we use these words after the hammer has broken them In
two: the fetish has become nothing but an empty stone onto width
meaning is mistakenly projected; the fact has become an absolute u 1
tainty which can be used as a hammer to break away all the delusion
of belief.
Now let us try to glue the two broken symbols together again, to 1c
store the four quadrants of our new repertoire (see Figures 9.1 and
9.2). As we saw in Chapter 4, the fact that is used as a solid hamim I
is also fabricated, in the laboratory, through a long and complex nego
tiation. Does the addition of its second half, its hidden history, its lab
oratory setting, weaken the fact? Yes, because it is no longer solid
and sturdy like a hammer (bottom left of Figure 9.1). No, because ll
is now, so to speak, threadlike, more fragile, more complex, mill
vascularized (see Chapter 3) and fully able to generate circulating lef
erence, accuracy, and reality (left side of Figure 9.2). It can still lu»
used, but not by an iconoclast and not to shatter a belief. A somcwli ll
subtler hand is required to seize this quasi-object, and a somewhat dlf
ferent program of action should be implemented with it.
1. One of the inventors of the word “fetishism” links il lo another etymology f a t 1

(De Brosses 1760, is), but all the diclionai les link il lo the I’oitugueNt* pa I
f a n u m . f a r i
participle of “fahiicale." On the conceptual hisloiy of I lie teim see IMel/ 1991 l a i n
1992, and the last mal mg cm |uuy in compaialive anlluopology ol Sell illtr 199
273
Facts Fetishes
1 2
E ither If fabricated If fabricated

f a b r ic a te d ... then illusory then illusory
3 4
Real as long Powerful only as

... or
as they are seen long as they seem
un fabricated
as not made autonomous
KNOWLEDGE BELIEF
Figure 9.1 In the canonical division o f fact and fetish, each o f the tw o divided
Iunctions (knowledge and belief) can be exposed by the question: Is it fabri
cated or is it real? The question implies that fabrication and autonomy are contra
dictory.
What about the other shard? What happens to the fetish? It is said
quite clearly that it has been fabricated, made, invented, devised. None
of its practitioners seems to need the belief in belief to account for its
efficacy. Everyone is willing to spell out quite frankly how it was made.
I)oes the acknowledgment of this fabrication in any way weaken the
claim that the fetish acts independently? Yes, because it is no longer
an irresistible ventriloquous phenomenon, an inversion, a reification,
an echo, in which the maker is taken in by what it has just created
(bottom right of Figure 9.1). No, because it can no longer be seen as a
naive belief, as a mere rétroprojection of human labor onto an object
Ihat is nothing in itself. It is not breakable and fragile like a belief wait
ing for the iconoclast’s hammer. It is now sturdier, much more
1ellexive, richly invested within a collective practice, reticulated like
blood vessels (right side of Figure 9.2). Reality, not belief, is entangled
in ils filaments. If lhe hammer’s blows threaten it with destruction,
Ihey will bounce bad liom (his yielding but resilient network.
II we add (o the lac Is the 11 lain ic at ion in the laboratory, and if we
uld to the fetishes Ilie. h e pile It 11id 11 (lexive lain ication by their mak
PANDORA'S HOPE
274
Facts Fetishes
1 2
When facts When fetishes
Because they are
are well are well
fa b ric a te d ...
fabricated... fabricated...
3 4
...th e y a llow ...they are what
...facts are
reality to be make us act
autonomous
autonom ous rightly
FACTISHES
Figure 9.2 Once the fabrication is seen as the cause o f autonomy a n d reality loi
both facts and fetishes, the vertical division between knowledge and belief ol 1 ig
ure 9.1 disappears; it is replaced by a new transversal question: W hat is it to l.ibi I
cate w e ll so as to make autonomy possible?
ers, the two main resources of the critique disappear: the hammei a
well as the anvil (I did not say the hammer and the sickle!). Appeal ing
in their stead is that which had been broken by iconoclasm, and had
always been there; that which always has to be carved anew and I
necessary for acting and arguing. This is what I call the factish* We
can retrieve the factish from the massacre of facts and fetishes when
we explicitly recover the actions of the makers of both (top of Figme
9.2). The symmetry of the two broken symbols is put back into plaie
If the iconoclast could naively believe that believers exist who ai e na
ive enough to endow a stone with spirit (bottom right of Figure 9 1 )//
was because the iconoclast also naively believed that the very facts he em
ployed to shatter the idol could exist without the help of any human atft m y
(bottom left of Figure 9.1). But if human agency is restored in hot
cases (top of Figure 9.2), the belief that was to be shattered disip
pears, along with the shattering fact. We enter a world that wt* had
never left, except in dreams—the dreams of reason—a world whcitf NC
guments and actions are everywhere facilitated, permitted, and ajJoule
by factishes.
The notion of factish is not an analytical category that can be added
to others by means of a clear and crisp discourse, since clanly ol ill
course results from drawing upon the deepest obsiunly, being lotted
to choose between constructivism and lealily (the veilical and hoi I
275
zontal axes of Figure 9.1), ushering us to the Procrustean bed in which

the modernist settlement wants us all to slumber: Are scientific facts
real or are they constructed? Are fetishes beliefs that are projected on
idols or are these idols “really” acting? Although these questions are
commonsensical enough, and seem necessary for any analytical clar
ity, they are, on the contrary, the questions that render all of the asso
ciations between humans and nonhumans totally opaque. If there is
one thing that obscures the saligram’s function, it is asking whether or
not it is a “mere” stone, a powerful object or a social construction.
But if one refuses to answer the question “Is it real or is it con
structed,” a serious problem can arise. Answering with the agnostic’s
“no comment,” can easily be confused with a cynical acceptance of
the falsity of all human representations. This, as I said at the end of
Chapter 1, is where science studies flirts dangerously with its polar op
posite, postmodernism. The solution of the factish is not to ignore
the choice, as many postmoderns do, by saying, “Yes, of course, con
struction and reality are the same thing; everything is just so much il
lusion, storytelling, and make believe. Who would be so naive, nowa
days, as to dispute such trivia?” The factish suggests an entirely
different move: it is because it is constructed that it is so very real, so
autonomous, so independent of our own hands. As we have seen over
and over, attachments do not decrease autonomy, but foster it. Until
we understand that the terms “construction” and “autonomous real
ity” are synonyms, we will misconstrue the factish as yet another form
ol social constructivism rather than seeing it as the modification of the
entire theory of what it means to construct.
Another way to say this is to point out that the modernists and
postmodernists, in all their efforts at critique, have left belief, the un
touchable center of their courageous enterprises, untouched. They be
lieve in belief. They believe that people naively believe. There are thus
Iwo forms of agnosticism. The first one, so dear to the critics’ hearts,
consists of a selective refusal to believe in the content of belief—usu-
dly God; more generally fetishisms and such things as saligrams;
inoie recently popular culture; and eventually scientific facts them
selves. In this definition of agnosticism the thing to be avoided at all
costs is being taken in. Naivete is the capital crime. Salvation always
comes Irom revealing (lie laboi that hides behind the illusio of auton
omy and independence the sli mgs th.it hold the puppets up. But I will
276
define agnosticism, not as the doubting of values, powers, ideas, truths,

distinctions, or constructions, but as the doubt exerted against this
doubt itself, against the notion that belief could in any way be whal
holds any of these forms of life together. If we do away with belief (in
beliefs) then we can explore other models of action and mastery. Be
fore we can do this, however, we have to take a last quick look at the
modern critic.
A Sketch of the Modern Critic

There is some difficulty in my speaking as though only the iconoclast
is a naive believer, as if he and he alone projects feelings onto objects
and forgets that the facts he makes in the laboratory are not the prod
ucts of his own hands. How could he and he alone be naive, immersed
in bad faith, and blinded by false consciousness? Am I not displaying
a lack of charity here, or worse, a lack of reflexivity? It is true that the
modern iconoclast does not believe more naively in his double con
struction of facts and fetishes than any of the others believed in Ilie
idols that the iconoclast destroyed to “free” them from their chain
Something else is at stake in his obsession, a different wisdom, which
to be sure, is not that of the factish, but is a wisdom all the same, no
matter how tortuous it appears. Let us consider one last time the e
traordinary power of the modern iconoclast in his native habitai
when he is not being self-conscious, that is, before he stops belli
modern, when he still possesses his pristine and unspoiled exotic ism
in the very moment at which he tries, like Jagannath, to desec 1at 1»
what he believes to be a mere stone that common people endow with
nonexistent powers !
Is the modernist critic imprisoned and enchained by his delusion 11
and muddled belief? On the contrary; the belief that others believe I rt
very precise mechanism that allows the human an extraordinal y iU
gree of freedom. By removing human agency twice, it makes it possible
at no cost, to free the passage for action, to clear the path by dish it 1
grating entities into mere beliefs and solidifying opinions and po I
tions into hard facts. No one has ever had so much freedom, hot dnni
is precisely what permits and justifies the iconoclast’s sliokes I ill
freedom from what? Freedomfrom caution and care, as I will disc 11 III
the next section.
277
We see now that the iconoclast is not free from factishes, because he
cannot escape the human agency that manufactures facts in the labo
ratory; nor is he free to do away with entities by confining them in in
ternal states of a mind endowed with an imagination and a “deep” un
conscious. In this regard modernists are like everyone else: everyone
everywhere has need of factishes to act and to argue. There is only one
nonmodern humanity—and in this sense, yes, I believe in a universal-
ist anthropology. But the main cunning of the critical modernist lies in
his ability to use the two sets of resources at once: on the one hand the
factishes, like everyone else, and on the other the apparently contra
dictory theory that radically distinguishes facts (which no one has
made), from fetishes (which are totally nonexistent objects, simply be
liefs and internal representations)—see the two columns of Figure 9.1.
This is what makes the modernist a true anthropological curiosity, this
is his unique and incommensurable “genius” which allows compara
tive anthropology to recognize this culture among all the others.
How can you recognize a modernist? Let us very quickly list the
items on the modernist’s psycho-social profile.
Modernists are iconoclasts. They have all the rage and violence and
power that allows them to break the factishes and to produce two ir
reconcilable enemies: fetishes and facts.
Modernists are freed, by this very act of shattering, from the chains
that bind all other cultures, since they can, at will, pump out of exis
tence whatever entities restrict their action, and pump into existence
whatever entities will enhance or accelerate their action (at least this is
the way they used to understand the “other cultures,” as if these were
“blocked,” or “limited,” or “paralyzed”).
Modernists, protected by this iconoclasm, can then proceed like ev
eryone else to produce, inside the insulated wombs of their “laborato-
1ies, ” as many factishes as they want. To them even the sky is not a
limit. New hybrids can be launched endlessly since there are no conse
quences attached to them. The inventiveness, originality, and juvenile
«11dor of the moderns can flourish unfettered. “This is only practice,”
lliey can say, “it has no consequence; theory will remain safe for ever.”
Modernists behave like the Carthaginians, who say, as they sacrifice
Iheir own childien to Baal, “They are only calves, only calves, not chil-
dien!” (Senes 1987)
I ligh above them, watt lung III e piolcctive goddesses, the sharp cut
278
distinctions between subject and object, science and politics, facts and
fetishes, render forever invisible the complicated and rather bizarre
means by which all of these categories are mixed. Above, subjects and
objects are infinitely distant, especially in theories of science. Below,
subjects and objects are intermingled to the greatest extreme, espe
cially in the practice of science. Above, facts and values are kept
infinitely far apart. Below, they are confused and redistributed and
tossed around endlessly. Above, science and politics never mix. Below,
they are continually remade anew from top to bottom.
Notice the construction that makes factishes thrice invisible: above,
they have disappeared, replaced by a clear and radiant theory whose
blinding light is fueled by a complete and constant distinction be
tween fact and fiction; below, the factishes are there—how could they
not be?—but they are hidden, invisible, mute, since only silent and
babbling practice* can account for that which is strictly forbidden
above. To be sure, actors constantly speak about “that,” the vast caul
dron at the heart of all their projects, but in a shattered and hesitant
language that only fieldwork can retrieve, and that never threatens the
opposed discourse of theory. Finally, an absolute distinction keeps the
top of the setup separate from the bottom part. Of course the factishi'i
of the modern exist, but their construction is so strange that although
they are active everywhere, visible to the naked eye, they remain invh
ible and impossible to register.
Naturally, however, the moderns are conscious, reflexive, and e
plicit about this threefold construction. We are not dealing here with rt
“superego” of theory, obsessively silencing the “id” of practice. If they
were not conscious, we would need another conspiracy theory, un
other psychoanalysis, to account for the belief in belief, to explain the
modernists’ belief in illusio, and to deny to the moderns, and only to
them, the right to be like everyone else, that is, to be free from belli I
in the firm hands of factishes—and I, for one, would be foiced In
become the iconoclast who would reveal the harsh reality of piuithe
behind the veil of theory.
How do we know the moderns are aware that they have never bi en
modern? Because, far from keeping the facts separate from the In lion
and the theory of this separation from the practice ol mediation, Ilie
endlessly, obsessively fix up, repair, and ovcieome these bioken li 1
ments. They use everything at hand to show that subjeils anil objul
279
should be reconciled, patched up, overtaken, “aufhebunged.” Modern

ism never stops repairing, and patching up again, and being desperate
about not being able to fix it because, despite all this repair work,
modernists never abandon the shattering gesture that started it all, the
one that created modernity in the first place. So desperate are they
that, after having shattered all the other cultures, they start to envy
them and to devise, under the name of exoticism, the muséographie
cult of the whole, complete, organic, wholesome, unspoiled, un
touched, unmodernized savage! To the modern they add an even more
bizarre invention, the premodern*.
We can now sketch the ideal psychosocial type of the modern, the
model of the critique. As an iconoclast, the modern breaks the idols,
all of them, always, fiercely. Then, protected by this gesture, in the si
lent practice opened up for him like a huge underground cavity, he
can get his kicks, with all the juvenile enthusiasm of the inventor, from
mixing up all sorts of hybrids without fearing any of the conse
quences. No fear, no past, only more and more combinations to try.
But then, terrified by a sudden realization of the consequences—how
could a fact be just a fact with no history, no past, and no consequence,
a “bald” fact instead of a “hairy” one?—he suddenly shifts from brave
iconoclasm and youthful ardor to fits of guilt-ridden bad conscience :
and this time he destroys himself, in endless ceremonies of atone
ment, looking everywhere for the broken fragments of his creative de
struction, gathering them back into huge and fragile bundles.
The strangest thing is that these godless, fetishless creatures are
viewed by all the others as having terrifying protectors and gods ! And
the other cultures cannot decide when the moderns are at their most
terrifying: is it when they crush the idols and burn them in autos-da-
Je? Is it when they innovate freely in their laboratories without the
slightest worry about the consequences? Or is it when they go around
beating their chests and tearing out their hair, desperately inflicting
penance on themselves for the sins they have committed, trying to re
cover in their museums, films, retreats, and self-help books the whole
ness of the lost paradise? “The pariahs found him more menacing
Iban Bhutaraya” which means that the freedom fighter now has the
power of three gods on his side instead of one: the menacing head of
the master Bialmun, the menacing force of modernization, and the
P A N D O R A ’S HOP E
28o
power of the local god. Whether the struggle for modernization suc
ceeds or not, it always seems to be the pariahs who end up losing.
Yes indeed, the moderns are interesting characters, well worth the
attention of comparative anthropologists!
Another Theory of Action and Creation

Now that we have turned the modernist repertoire from a resource to
a topic for study, now that we have portrayed the guilt-ridden icono
clast as one interesting but peculiar type in one culture among others,
is it possible to imagine a model for the practice of politics that would
not rely so heavily on the model of the critique? This is a difficult
question because the scenography of activism has been so powerfully
based on iconoclasm that it seems that if you do away with the icono
clasm you immediately fall into one of a very few models of reaction
ary politics. If one is neither modem nor premodern, is not the only
alternative left that of being antimodern? How can the number ol
models for political action be multiplied; how can we undo the cm
rent definitions of “reactionary” versus “enlightened” politics? One
way is to modify the scenography of politics itself, as I attempted to do
in Chapters 7 and 8. Another path, which I took in Chapter 6, is to ol
fer an alternative to the idea of progress that still makes use of Ihe
traditional arrow of time. The possibility that I want to outline now
requires us to consider what sort of life we would lead if we lived un
der the protection of factishes again—no longer caught between hut
and fetishes. At least three things would change profoundly: the dell
nition of action and mastery; the dividing line between a physic (I
world “out there” and a mental world “in there”; and the definitions ol
care and caution along with the public institutions that would exhibit
them.
Action and M astery
What is it that iconoclasm breaks, and what is it that factishes allow 11

to restore? A certain theory of action and of mastery. Once the h 1111
mer has fallen, shattering the world into facts on one side and Irtish?
on the other, nothing can stop the dual question liom being posed did
you construct the thing yourself, or is it autonomous? This teasth
sterile, and boring question p.ualy/ed the field ol science studies cm
281
turies before it even began. When a fact is fabricated, who is doing the
fabrication? The scientist? The thing? If you answer “the thing,” then
you are an outdated realist. If you answer “the scientist,” then you are
a constructivist. If you answer “both,” then you are doing one of those
repair jobs known as the dialectic, which seem to patch up the dichot
omy for a while, but only hide it, allowing it to fester at a deeper level
by turning it into a contradiction that has to be resolved and over
come. And yet we have to say that it is both, obviously, but without the
assurance, certainty, or arrogance that seems to go with the realist or
the relativist answer or with a clever oscillation between the two. Lab
oratory scientists make autonomous facts. That we have to hesitate
between two versions of this simple “making do” (fait-faire) proves
that we have been hit by a hammer that has broken the simple and
straightforward factish into two parts. The shock of critical intelli
gence has rendered us stupid.
Things change entirely, as we saw in Chapter 4, when we listen to
what is said by practicing scientists without adding or withdrawing
anything. The scientist makes the fact, but whenever we make some
thing we are not in command, we are slightly overtaken by the action:
every builder knows that. Thus the paradox of constructivism is that it
uses a vocabulary of mastery that no architect, mason, city planner, or
carpenter would ever use. Are we fooled by what we do? Are we con
trolled, possessed, alienated? No, not always, not quite. That which
slightly overtakes us is also, because of our agency, because of the
clinamen of our action, slightly overtaken, modified. Am I simply re
stating the dialectic? No, there is no object, no subject, no contradic
tion, no Aufhebung, no mastery, no recapitulation, no spirit, no alien
ation. But there are events*. I never act; I am always slightly surprised
by what I do. That which acts through me is also surprised by what I
do, by the chance to mutate, to change, and to bifurcate, the chance
that I and the circumstances surrounding me offer to that which has
been invited, recovered, welcomed (Jullien 1995).
Action is not about mastery. It is not a question of a hammer and
shards, but one of bifurcations, events, circumstances. These subtle-
lies are difficult to retrieve once iconoclasm has struck, because facts
and tools aie now f11mly in place, suggesting the model for Homofaber
that can nevei, alia that, be displaced and reworked. But, as we saw in
t haptcr 6, no human agtnl has evei built, constructed, or fabricated
inytlung, not even a Mone tool not even a basket, not even a bow, by
282
using the repertoire of action invented for Homo faber. Homo faber is
man’s fable, a Homo fabulosus through and through, a retrospective
projection into our fantastic past of a definition of matter, humanity,
mastery, and agency that dates entirely from the modernist period,
and that uses only a quarter of its repertoire—the world of inert au
tonomous matter. We cannot account for laboratory practice by fall
ing back on a modernist definition of technical construction—or, even
less, one of social construction.
Why is it so difficult to retrieve other theories of action? Because it
is crucially important to the modernist ethos to demand a choice be
tween that which you fabricate—as a free and naked human—and that
which is a fact out there, made by no one. The whole work of the mod
ern has been to render these two agents, the human and the object,
unfit for any other role than that of opposing each other. No wonder
they cannot be used for anything else! It is a simple question of ergo
nomics: they aren’t suited for any other job.
But the idiom changes immediately once the two halves are brought
together again. Facts are fabricated; we make facts, that is, there is a
“fait-faire.” Of course the scientist does not make up facts—who
has ever made up anything? This is another fable, symmetrical with
that of Homofaber and dealing, this time, with the fancies of the mind
I do not deny that people have minds—but the mind is not a woi Id
creating despot that makes up facts to suit its fancy. Thought is seized
modified, altered, possessed by nonhumans, who, in their turn, given
this opportunity by the scientists’ work, alter their trajectories, desll
nies, histories. Only modernists believe that the only choice to lie
made is between a Sartrean agent and an inert thing out there, a root
on which to vomit. Every scientist knows in practice that things have a
history too; Newton “happens to” gravity, Pasteur “happens to" Ihe
microbes. “Intermingle,” “bifurcate,” “happen,” “coalesce,” “ncgoll
ate,” “ally,” “be the circumstances o f’: these are some of the veibs (hat
signal the shift in attention from the modernist to the nonmodoml I
idiom.
What is at stake here is mastery. In making the world the pioduct of
individuals’ thoughts and fancies and in talking about construction a
though it involved the free play of fancy, model nists believe they m ike
the world in their image, just as God made them in his 1’liis In a
strange and rather impious description ol God As i( God weie mil Ivl
283
of His Creation! As if He were omnipotent and omniscient! If He had

all these perfections, there would be no Creation. As Whitehead so
beautifully proposed, God, too, is slightly overtaken by His Creation,
that is, by all that is changed and modified and altered in encountering
Him: “All actual entities share with God this characteristic of self
causation. For this reason every actual entity also shares with God the
characteristic of transcending all other actual entities, including God”
(Whitehead [1929] 1978, 223, my italics). Yes, we are indeed made in
the image of God, that is, we do not know what we are doing either.
We are surprised by what we make even when we have, even when
we believe we have, complete mastery. Even a software programmer is
surprised by her creation after writing two thousand lines of software;
should God not be surprised after putting together a much larger
package? Who has ever mastered an action? Show me a novelist,
a painter, an architect, a cook, who has not, like God, been sur
prised, overcome, ravished by what she was—what they were—no lon
ger doing.
And do not tell me they were “possessed,” “alienated,” or “domi
nated” by outside forces. They never exactly say so. They say that these
others have been modified, altered, taken over, in the circumstances of
the action, by the unfolding of the event. Mastery, domination, or re
capitulation is not the way to think of such instances. No nonmodern
wants to have to deal with that sort of God or that sort of Man.
Factishes bring with them a quite different definition of God, of hu
man agency, of action, of nonhumans. No model of political action can
be offered as an alternative to the model of the critique until we mod
ify our anthropology of creation, that is, until we retrieve the anthro
pology practiced by the modernists even while they believed them
selves to be modern, and while they always explicitly said, in practice,
that they were not.
An Alternative to Beliefs
Is il really possible to be agnostic in the sense I have defined? Is

not belief in belief what allows the distinction between a world “out
lim e” and a palace of ideas, imagination, fancies, and distortions
m llieie”? I low could we suivive without this distinction between
(pislemological and ontological questions? Into what soit of obscu
284
rantism would we fall if we could no longer make the sharp distinc
tion between the contents of our heads and the world outside our
minds? And yet the price paid for obtaining this semblance of com
mon sense is extraordinarily high. We are so used to living under the
sway of anti-fetishism, so accustomed to taking for granted the abyss
between the wisdom of practice and the lessons of theory, that we
seem to have entirely forgotten that this most cherished analytical
clarity was reached at the price of an incredibly costly invention: one
physical world “out there” versus many mental worlds “in there.” How
did this come about?
If, as common sense would have it, there are no factishes, but only
fetishes, which are nothing but pieces of wood and mute stones, where
can all those things that believers believe in be located? There is no so
lution but to push them into the minds of believers or into their fecund
imaginations, or to embed them even deeper, in a rather perverse and
crooked unconscious. Why not leave them where they were, that is,
among the multiplicity of nonhumans? Because there is no longer any
room for nonhumans or for any multiplicity. The world itself has been
stuffed beyond capacity, thanks to the other, simultaneous move that
transformed factishes into facts. If no human agency is at work 01
has been at work—in the manufacture of facts, if there are no limits ol
cost, information, networks, or manpower for the production, expan
sion, and maintenance of facts, then nothing, absolutely nothing stops
them from proliferating everywhere, continuously, filling in every last
corner of the world—and at the same time unifying the many
worlds into a single homogeneous one. The notions of matter, a me*
chanical universe, a mechanical world-picture, a natural world: these
are the simple consequences of the rupture between the two meaning
of “fact”—that which is fabricated, that which is not fabricated Ihil
the notions of belief, mind, interior representation, illusion: Ilust
are merely the consequences of having split the factish in two lint
which is fabricated, that which is not fabricated.
It is hard to decide which came first. Was the notion of an intei loi
mind invented as a repository for all the entities squeezed out ol flit»
world, or did the belief in beliefs empty the world, leaving loom loi
“factoids” to proliferate like rabbits in Australia? What is coitam I
that with the destruction of the means ol argumentation and at lion
that factishes enabled, with the removal ol human agency horn lilt»
285
fabrication of facts and from the fabrication of fetishes, two fabulous
reservoirs were invented, one fo r epistemology, one fo r ontology. These
subjects endowed with an inside are as strange as these objects rele
gated to an outside. Indeed, the notion of an inside divided from an
outside is very strange and is, in its own right, a fabulous innovation.
With one stroke the iconoclast sets in motion the most powerful suc
tion pump ever devised. Whenever entities are obstacles to his action
they can be pumped out of existence, emptied of all reality until they
are nothing but hollow beliefs. Whenever there is a deficit of certain,
positive mechanical entities to render his actions steady and beyond
objection, they can be pumped into existence by the score: now there
are stones everywhere “out there,” in the only world there is, matched
by many naive beliefs about saligrams “in there,” inside the believers’
minds. With this device, powered by the opposition between episte
mology and ontology, the iconoclast is able to empty the world of all
its inhabitants by turning them into representations, while filling it up
with continuous mechanical matter.
But what happens when this pump has stalled, when there is no lon
ger an inside mind into which, under the name of fancy or belief, one
can squeeze every entity, and when there is no longer an outside world
made of ahistorical, inhuman causes “out there”? The first thing to go,
naturally, is the very difference between inside and outside. This does
not mean that everything is now outside, but simply that the entire
scenography of outside and inside has evaporated.
What appears in its place is, at first, as we witnessed in Exhibit A
in Chapter 5, a bewildering array of entities, divinities, angels, god
desses, golden mountains, bald kings of France, characters, controver
sies about facts, propositions in all possible phases of existence. The
stage will be so crowded with this heterogeneous crew that one may
start to worry, and get nostalgic for the good old modern age when
the pump was still at work, sucking all beliefs out of existence and re
placing them with sure and safe and certain objects of nature. But for
tunately these entities do not ask for the same kinds of ontological
specifcations. They cannot be ordered, to be sure, into beliefs and reali-
tios, but they can be ordered, and very neatly, according to the types of
existence they claim
Jagannath’s stone, lot instance, does not claim to be a spirit as in the
letislust mode, not does it claim to be the symbol for a spirit projected
286
onto the stone, as in the anti-fetishist version. As Jagannath realizes
clearly when he fails to desecrate the saligram, this stone is what
makes him, his family, and the untouchables human, what holds them
in existence, that without which they would die. Understood accord
ing to the fact-fetish dichotomy, the stone immediately becomes a
spirit, that is, a transcendent entity that obeys the same specifications
as an object of nature except that it is invisible. In practice, however,
the stone is a factish and does not claim to be a spirit, to be invisible; it
never ceases to be, even for the aunt and the priest, a “mere stone.” It
simply asks to be that which protects humans against inhumanity and
death, the thing that, when removed, turns them into monsters, ani
mais, things (Nathan and Stengers 1995).
The problem is that this way of arguing—granting ontological con
tent to beliefs—runs counter to the whole deontology of the social sci
ences. “When the sage points at the Moon,” says the Chinese proverb,
“the fool looks at his fingertip.” Well, we have all educated ourselves to
be fools! This is our deontology. This is what a social scientist learns at
school, mocking the unwashed who naively believe in the Moon. We
know that when actors speak about the Virgin Mary, divinities
saligrams, UFOs, black holes, viruses, genes, sexuality, and so foil!»
we should not look at the things thus designated—who could be so na
ive nowadays?—but should look instead at the finger, and from then*
following down the arm along the nerve fibers, to the mind of the lu
liever, and from there down the spinal cord to the social structuies
the cultural systems, the discursive formations, or the evolution
ary bases that make such beliefs possible. The anti-fetishist bias is 10
strong that it seems impossible to argue against it without heating
the indignant screams: “Realism! Religiosity! Spiritism! Reaction!
We should now imagine a scene that would enact Jagannath’s trauma
but in reverse: the nonmodern thinker wants to touch the conitnl
of beliefs again, and the modernist and postmodernist critics, hoi 101
stricken, scream, “Don’t touch them!! Don’t touch them! Anathoin il
And yet we, the science students, have touched them, and nolliln
happened except that the dreams of social constructivism tlhip
peared! Through a transfiguration exactly opposed to that o!
Jagannath, when we touched subjects and objects they suddt nl
turned into humans and nonhumans.
After centuries of detachment, our attention is now tm mng liai I li
287
the fingertip, and from it to the Moon. The simplest explanation for
all the attitudes of humanity since the dawn of its existence is proba
bly that people mean what they say, and that, when they designate an
object, this object is the cause of their behavior—not a delusion to be
explained by a mental state. Here again we should understand that the
situation has changed entirely since the advent of science studies. It
was feasible to be anti-fetishist when facts could be used as destructive
weapons against beliefs. But if we now speak of factishes, there exist
neither beliefs (to be fostered or destroyed) nor facts (to be used as a
hammer). The situation has become much more interesting. We are
now faced with many different practical metaphysics, many different
practical ontologies.
By granting ontology back to nonhuman entities, we can begin to
tackle the major question at issue in the science wars. The modernist
Enlightenment, in its republican ideal at least, became, for a while, a
popular movement. It struck a chord in all the oppressed around the
world. When facts were accommodated into our collective existence,
great clouds of delusion, oppression, manipulation were dissipated.
But since then the models offered by the critique have ceased to be
popular. They now run against the very grain of what it is to be human
and to believe. Facts have gone too far, attempting to transform every
thing else into beliefs. The burden of supporting all these beliefs be
comes unbearable when, as in the postmodern predicament, science it
self is submitted to the same doubt. It is one thing to attack beliefs
when we are fortified by the certainties of science. But what are we
to do when science itself is transformed into a belief? The only solu
tion is postmodern virtuality—the nadir, the absolute zero of politics,
aesthetics, and metaphysics. The engine of virtuality, however, is in
postmodern heads, not in the worlds surrounding them. Virtuality is
what everything else turns into when belief in belief has run amok. It
is lime to stop the little salt-mill’s grinding, before everything left be
comes bitter.
Could we not say, quite simply, that people are tired of being ac-
1used of believing in nonexistent things, Allah, jinns, angels, Mary,
(iaia, gluons, retroviruses, rock ‘n’ roll, television, laws, and so on?
I he nonmodern intellectual does not take Jagannath’s position, day
iltei day bi mgmg new sahgiams to desecrate and then throwing them
isnle, discounted to discovei that only he, the desecrator, the icono
288
clast, the liberator, believes in them and that everyone else—ordinary
pariahs, average laboratory scientists—has always lived under a com
pletely different definition of action, in the hands of factishes of to
tally different shapes and functions.
Care and Caution
What did the factish do before it was broken by the anti-fetishisl’s

blow? To say that it mediated action between construction and auton
omy is an understatement, and relies too heavily on the ambiguity ol
the term mediation*. Action is not what people do, but is instead the
“fait-faire,” the making-do, accomplished along with others in an
event, with the specific opportunities provided by the circumstances
These others are not ideas, or things, but nonhuman entities, or, as I
called them in Chapter 4, propositions*, which have their own onto
logical specifications and populate, along their complex gradients, t
world that is neither the mental world of psychologists nor the physl
cal world of epistemologists, although it is as strange as the first and u
real as the second.
What the factishes were good at was articulating caution and publh
ity. They publicly declared that care should be taken in the man ipu 11
tion of hybrids. When they tried to break the fetishes, the iconoclast
broke the factishes instead. As I have said, these rampages are win!
have given the moderns their fabulous energy, invention, and créât Iv
ity. They are no longer held by any constraints, any responsibility IIip
broken halves of the factish, nailed above the threshold of the mod
ernist temple, protect them against all the moral implications ol wli it
they do, and they can be all the more inventive since they belli VP
themselves to be wallowing in “mere practice.” What has boon ip
moved by the hammer is care and caution.
Of course, action did have consequences, but these came laid, lit 11
ally after thefact, and under the subservient guise of unexpected con p
quences, of belated impact (Beck 1995). Modernist objects woi e bald
aesthetically, morally, epistemologically—but the ones pioduud b
the nonmoderns have always been hairy, networky, lhizomolil t I h
reason one should always beware of factishes is that thou ion p
quences are unforeseen, the moral order liagile, the social one un Irt
ble. This is just what modernist tacts have shown us ovoi and ovt I P
289
cept that, for the modern, consequences are nothing but an

afterthought. It is only after the desecrating ceremony that Jagannath
realizes that no one ever believed the saligram to be anything but
a stone, and that the only inhumanity was that which he, the free
thinker, produced by destroying the idol. When the aunt and the
priest screamed “Beware! Beware!” they did not mean, as he thought,
that they were afraid he would break the taboo, but that they were
afraid he would break the factish that kept care and caution under at
tentive public consideration (Viramma, Racine, et al. 1995).
How strange it is to realize that the blows of the iconoclast’s ham
mer always missed their target. Are we not the inheritors of all
the iconoclastic gestures of our history? Of Moses striking down the
Golden Calf (Halbertal and Margalit 1992.)? Of Plato breaking up the
shadows of the Cave to honor this highest of all the idols, the Idea—
eidon—itself? Of Paul sending all the pagan idols packing? Of the
great wars of the Byzantine era between iconoclasts and iconodules
(Mondzain 1996)? Of the Lutherans deciding what should and what
should not be painted (Koerner 1995)? Of Galileo shattering the an
tique cosmos? Of the revolutionaries tearing down the ancien régime?
Of Marx denouncing the illusions of commodity fetishism? Of Freud
turning the fetish into a stopper that closes off the horrifying dis
covery of what is always missing? Of Nietzsche, the philosopher with
a hammer, smashing every idol, or, more accurately, tapping them
gently to hear how hollow they sound? To believe the opposite, to re
nounce this pedigree, this prestigious genealogy, would be to accept
the grave accusation of becoming archaic, reactionary, even pagan.
I low could such an absurd position lead to another model for politics?
First of all, “paganism,” “archaism,” and “reaction” are dangerous
Ilungs, but only when used as foils for modernization. There is, as an-
lln opology has been teaching us lately, no such thing as an archaic
pi imitive culture to which one could return. This has never been any-
llung but an exotic fantasy of reactionary racism. The same is true
ol paganism, and of reactionary politics, itself an invention of mod-
II ni/ers. “Reactionary” is a dangerous and unstable word (Hirschman
1991), but it might be understood as simply the wish to bring care
nul caution back into the fabrication of facts and to make the salutary
Bewaie!" audible again in the depths of the laboratories—including
those ol the science students In (hat sense, only the modernists want
290
to drag us back to an earlier time and an earlier settlement, and this

nonmodern precaution appears commonsensical enough, perhaps
even progressive—if we accept that progress means stepping into an
even more entangled future, as we saw in Chapter 6.
Second, becoming nonmodern again necessarily implies a rework
ing of our genealogy and of our ancestry. Idolatry may have been, all
along, a misplaced target for monotheism. The fight against icons
may have been the wrong battle for Byzantine emperors to wage. The
Protestant Reformation probably chose the wrong target in fighting
Catholic piety. Irrationalism may have been the wrong target for sc I
ence; commodity fetishism the wrong target for Marxism; divin il y
the wrong target for psychiatry; realism the wrong target for social
constructivism. Each time the error is the same, and comes from fht
naive belief in the others’ naive belief. The modernists have always had a
hard time understanding themselves because of their iconoclasm and
because of the anxiety that icon-breaking brings. To study iconoclasm
anthropologically, as part of the moderns’ total way of life, as the. Ii
ideal psychosocial type, necessarily modifies its effect and its impact
The knife no longer has a cutting edge, the hammer is too heavy W e
must rethink the will to iconoclasm, our most venerable virtue, since
its targets are no longer viable: we will not modernize the world, "w e
meaning the tiny cult of “nonbelievers” at the tip of the Westei n pc n
insula.
Third, and more important, setting aside the iconoclast’s hummel
allows us to see that we have always been involved in cosmopollllt
(Stengers 1996). It is only through an extraordinary shrinking ol t he
meaning of politics that it has been restricted to the values, intei t I
opinions, and social forces of isolated, naked humans. The gieul ud
vantage of letting facts merge back into their disheveled netwoi ks an I
controversies, and of letting beliefs regain their ontological weight I
that politics then becomes what it has always been, anthropologic«II
speaking: the management, diplomacy, combination, and negothlli ll
of human and nonhuman agencies. Who or what can withstand win III
or what? Thus another political model is offered, not one that su k I
add a supplement of soul, or asks citizens to adjust their values to til
facts, or drags us back to some archaic tribal galheiing, but one (liai
entertains as many practical ontologies as lheie aie luclishes
The role of the intellectual is not, then, to grab a hammei and bleak
291
beliefs with facts, or to grab a sickle and undercut facts with beliefs
(as in the cartoonish attempts of social constructivists), but to be
factishes—and maybe also a bit facetious—themselves, that is, to protect
the diversity of ontological status against the threat of its transfor
mation into facts and fetishes, beliefs and things. No one is asking
Jagannath to be content with his high-caste rank and to maintain the
status quo. But, at the same time, no one is asking him to debunk
the sacred family stones or to set the others free. In the long history
of the model of the critique, we underestimated the meaning of free
dom, the freedom that comes from adding human agency twice: to the
fabrication of fetishes and to that of facts. We seem to have missed
something along the way. It may be time to retrace our steps; the risk
of appearing reactionary may be smaller than that of being modernist
at the wrong time and in the wrong way.
The subject-object dichotomy has lost its ability to define our hu
manity because it no longer allows us to make any sense of an impor
tant little adjective: “inhuman.” What is inhumanity? Look at how
strange it was in the modernist era. To protect subjects from fall
ing into inhumanity—subjectivity, passions, illusions, civil strife, delu
sions, beliefs—we needed the firm anchor of objects. But then objects
also began generating inhumanity, so that in order to protect objects
from falling into inhumanity—coldness, soullessness, meaningless
ness, materialism, despotism—we had to invoke the rights of subjects
and “the milk of human kindness.” Inhumanity was thus always the
inaccessible joker in the other stack of cards. Surely this cannot pass
lor common sense. It is certainly possible to do better, to locate inhu
manity somewhere else: in the gesture that produced the subject-
object dichotomy in the first place. This is what I have tried to do by
suspending the anti-fetishist’s urge. The green field of humanity is not
(ar off on the other side of the fence, but close at hand, in the move
ment of the factish.
In the Tel Aviv Diaspora Museum one can see a medieval illumina-
Iion in which Abraham’s gesture, interrupted by the hand of God,
ums at the helpless Isaac standing on a pedestal; the child strikingly
lesembles an idol about to be broken. This bloodiest of all cities is
founded on an inteirupted human sacrifice. Is not one of the many
( mises of this bloodshed the strange contradiction there is in suspend
ing human sau dites while tanymg out the destruction of idols with
292
self-righteousness and glee? Should we not abstain from this destruc

tion of humanity too? Whose hand should restrain us before we carry
out the critical gesture? Where is the ram that could be used as a sub
stitute for the critical mode of reasoning? If it is true that we are all
descendants of Abraham’s suspended knife, what sort of people will we
become when we also abstain from destroying factishes? Jagannath
was left pondering: “When they touched it, we lost our humanity
they and me, didn’t we? And we died. Where is the flaw of it all, in me
or in society? There was no answer. After a long walk he came home,
feeling dazed.”
Conclusion
What Contrivance Will Free
Pandora’s Hope ?
What have we achieved through this admittedly strange and bumpy

exploration of the reality of science studies? One point at least should,
by now, be established: there exists only one settlement, which con
nects the questions of ontology, epistemology, ethics, politics, and the
ology (see Figure 1.1). There is thus no longer much sense in pursuing
in isolation questions like “How can a mind know the world outside?”
“How can the public participate in technical expertise?” “How can we
elevate ethical barriers against the power of science?” “How can we
protect nature from human greed?” “How can we build a livable polit
ical order?” Very quickly inquiries into these matters stumble over so
many aporias, since the definitions of nature, society, morality, and the
Body Politic were all produced together, in order to create the most
powerful and most paradoxical of all powers : a politics that does away
with politics, the inhumane laws of nature that will keep humanity
1rom falling into inhumanity.
It should also now be clear that science studies does not occupy a
position inside this old settlement, no matter how hard the science
wai riors try to contain it within the narrow confines of modernism.
Science studies does not say that facts are “socially constructed”; it
does not spur the masses to smash their way through the laboratories;
it does not claim that humans are forever cut off from the outside
woi Id and locked in the cells ol theii own viewpoints; it does not wish
to go back to the licit tulhcntic, and humane picmodern past. What
y
294
is most bizarre to the eyes of the social scientists is that science studies
is not even critical, debunking, or provocative. By shifting attention
from the theory of science to its practice*, it has simply happened, by
chance, upon the frame that held together the modernist settlement.
What, at the level of theory, looked like so many different and uncon
nected questions to be taken seriously but independently, revealed
themselves, when daily practice was scrutinized, as being tightly inter
twined.
Then everything followed quite logically. Since an enormous num
ber of conundrums had been attached to the theory of science, once
we shifted our attention to practice all those classical topics became
shaky as well. Hence the bouts of megalomania that, from time to
time, seem to agitate science studies—some of them probably émanai
ing from my own word processor. Is it our fault if so many cherished
values—from theology to the very definition of the social actor, from
ontology to the very conception of what a mind is—have been hooked
upon a theory of science that a few months of empirical inquiries ai e
enough to put into serious doubt? This does not mean that all these is
sues are not important, or that these values should not be defended; on
the contrary, it means that they have to be fastened with a rather stui
dier nail and joined to the fate of somewhat loftier ends.
I am well aware that the most contentious aspect of this search
for an alternative to the old settlement is having done away with the*
subject-object dichotomy altogether. Since the beginning of model
nity, philosophers have tried to overcome this dichotomy. My claim Is
that we should not even try. All attempts to reuse it positively, nega
tively, or dialectically have failed. No wonder: it is made not to be ovei
come, and only this impossibility provides objects and subjects willi
their cutting edges. Through inquiries, anecdotes, myths, legends, It*
tual studies, and more than a little bit of conceptual bricolage, 1 have
sought in this book to offer a more plausible explanation for the si ul»
bornness of this divide: the object that sits before the subject and Ilie
subject that faces the object are polemical entities, not innocent me la
physical inhabitants of the world.
The object is there to protect the subject from drifting into mini
manity; the subject is there to protect the object from di ill mg min In
humanity. But the protective shield of fjelishes has disappeaicd md
the Body Politic has been rendeied impotent Humaneness has Ik*
CONCLUSION
295
come irretrievable, since it is always to be found on the other side of

that great unbridgeable gap. Once inside this huge, solemn, and beau
tiful architecture, no one can say a word about objects without it being
used immediately to staunch some trace of subjectivity somewhere
else: one cannot utter a word about the rights of subjectivity without it
being seized upon to humble the power of science or to counteract the
soullessness of nature. As modernity unfolded, subjectivity and objec
tivity became concepts of resentment and revenge. Not a trace of their
liberating youthfulness can be found in them any longer. Science has
been so thoroughly politicized that neither the aims of politics nor
those of the sciences have remained visible. Even their common des
tiny has been erased. The science wars are only the latest episode in
this polemical use of objectivity—and not the last, I am afraid.
I have attempted to substitute another pair—that of humans and
nonhumans—for the subject-object dichotomy, which I have left un
touched. Instead of overcoming the divide, I have kept the settlement
where it was and headed off in a different direction, digging occasion
ally beneath the huge megaliths when it was expedient to do so: be
neath, not above. I deserve no credit for doing so, since I was simply
following practice, not theory. How, for instance, could I have con
sidered, without an enormous distortion, Pasteur as a subject facing
an object, the lactic acid ferment (Chapter 4)? The very subtle process
of delegation that allows Pasteur to fabricate fact would have been
squashed flat if I had tried to locate it in the scenography of modern
ism. I would have had to answer the questions bellowed by the new
Fafner and Fasolt we met in Chapter 5: “Is the ferment real or is it fab
ricated?”
And it would have been even worse if I had answered “both,” be
cause the truth—the nonmodernist truth—is that facts are neither real
nor fabricated, but escape altogether from this comminatory choice
invented to render the Body Politic impossible. To make it through
ihis difficult pass, they would have needed a little help from their
lactishes; but these facilitators have all been broken in two by the
iconoclastic gesture of the critical modernists. It is not an easy thing to
escape the sway of the old predicament. If readers find this book to be
ciudely patched together, I respectfully ask them to remember the
lumdieds ol fiagmenls among which I found delegation, translation*,
ai liculalion*, and the olhei concepts I’ve ti ied to rehabilitate, lying on
296
the floor, smashed to pieces, deconstructed to dust! Better to have
them badly restored by a clumsy curator but in full working order
than to leave them there broken and useless . ..
So we have made some progress. There is one modernist settlement,
and there is at least one alternative to it, which does not represent its
fulfillment, its destruction, its negation, or its end. This is the only
thing that can be asserted with some semblance of certainty. What a
lively and sustainable alternative may be, I don’t know. However, if
we try to replace each of the fixtures of the old settlement—the boxes
of Figure 1.1—we can note some specifications for the task that lies
ahead.
The easiest and quickest thing to replace will be the entire artifact of
epistemology. The idea of an isolated and singular mind-in-a-vat look
ing at an outside world from which it is thoroughly cut off, and trying
nonetheless to extract certainty from the fragile web of words spun
across the perilous abyss separating things from discourse, is so im
plausible that it cannot hold up much longer, especially since psychol
ogists themselves have already redistributed cognition beyond recog
nition. There is no world outside, not because there is no world at all,
but because there is no mind inside, no prisoner of language with
nothing to rely on but the narrow pathways of logic. Speaking truth
fully about the world may be an incredibly rare and risky task for a sol
itary mind steeped in language, but it is a very common practice foi
richly vascularized societies of bodies, instruments, scientists, and in
stitutions. We speak truthfully because the world itself is articulated
not the other way around. That there was once a time when a war
could be waged between “relativists,” who claimed that language re
fers only to itself, and “realists,” who claimed that language may occa
sionally correspond to a true state of affairs, will appear to our descen
dants as strange as the idea of a fight over sacred relics.
Second, there is clearly a space in which the sciences can unfold
without being kidnapped by Science No. 1. Scientific disciplines aie
born free, everywhere they are in chains. I see no reason remaining fol
scientists, researchers, or engineers to prefer the old settlement. I'ph
temology was never intended to protect them, it was always a war mu
chine—a Cold War machine, a Science War machine. The expiesslon
“socializing nonhumans to bear upon the human collective" seems (o
me a perfectly acceptable, albeit cleaily piovisional, solution, one lb il
CONCLUSION
\ 297
shelters the practice of the sciences and respects the many

vascularizations they need to thrive. It is certainly better, in any case,
than being submitted to this double bind: “Be absolutely discon
nected”; “Be absolutely certain of what you say in words about the
world out there.” That this twin injunction could have passed for com
mon sense under the pretense of fighting “relativism” will, I am con
vinced, seem odd a few years from now, once circulating reference has
been provided to every household, like gas, water, and electricity.
Third, and more important because it concerns many more people,
the conditions of felicity for politics may begin to unfold as well, now
that they needn’t be constantly interrupted, shortcut, quashed, and
thwarted by the continual injection of inhumane laws of nature. More
exactly, nature* now appears as what it always was, namely, the most
comprehensive political process ever to gather into one superpower
everything that must escape the vagaries of the society “down there."
An objective nature facing a culture is something entirely different
from an articulation of humans and nonhumans. If nonhumans aie
to be assembled into a collective, it will be the same collective, and
within the same institutions, as the humans whose fate the sciences
have brought nonhumans to share. Instead of this bipolar power
source—nature and society—we will have only one, clearly identifiable
source of politics for humans and nonhumans alike, and one clearly
identifiable source for new entities socialized into the collective.
The word “collective” itself at last finds its meaning: it is that which
collects us all in the cosmopolitics envisaged by Isabelle Stengers. In
stead of two powers, one hidden and indisputable (nature) and the
other disputable and despised (politics), we will have two different tasks
in the same collective. The first task will be to answer the question: How
many humans and nonhumans are to be taken into account? The sec
ond will be to answer the most difficult of all questions: Are you ready,
and at the price of what sacrifice, to live the good life together? That
this highest of political and moral questions could have been raised,
for so many centuries, by so many bright minds, for humans only with
out the nonhumans that make them up, will soon appear, I have no
doubt, as extravagant as when the Founding Fathers denied slaves and
women the vote.
The fourth and moie difficult specification has to do with mastery.
We have exchanged masleis many limes; we have shifted from the
298
God of Creation to Godless Nature, from there to Homofaber, then to

structures that make us act, fields of discourse that make us speak,
anonymous fields of force in which everything is dissolved—but we
have not yet tried to have no master at all. Atheism, if by that we mean a
general doubt about mastery, is still very much in the future; and so is
anarchism, in spite of the disingenuousness of its beautiful slogan
“neither god nor master,” since it always had one master, man!
Why always replace one commander with another? Why not recog
nize once and for all what we have learned over and over again in this
book: that action is slightly overtaken by what it acts upon; that it
drifts through translation; that an experiment is an event which offers
slightly more than its inputs; that chains of mediations are not the
same thing as an effortless passage from cause to effect; that trans
fers of /«formation never occur except through subtle and multiple
transformations ; that there is no such thing as the imposition of cat
egories upon a formless matter; and that, in the realm of techniques,
no one is in command—not because technology is in command, but
because, truly, no one, and nothing at all, is in command, not even
an anonymous field of force? To be in command, or to master, is a
property of neither humans nor nonhumans, nor even of God. It was
thought to be a property of objects and subjects, except that it nevci
worked: actions always overflowed themselves, and gnarled entangle
ments always ensued. The ban on theology, so important in the staging
of the modernist predicament, will not be lifted by a return to the God
of Creation but, on the contrary, by the realization that there is no
master at all. That religion too was seized by modernists as oil for then
political war machine, that theology debased itself by agreeing to play
a role in the modernist settlement and betrayed itself even to the pom I
of talking about nature “out there,” the soul “in there,” and society
“down there,” will, I hope, serve as a source of bewilderment foi (lu
next generation.
It is certainly in the forward movement of time’s arrow that the (11
ture settlement can do better than the modernist one. History win
never at ease in the house of modernity. Either, as we saw in Chaplei 5
it had to be limited to humans, nature out there escaping it allogetlu 1
or, as we saw in Chapter 6 , it had to appear under the deeply itnpioh \
ble guise of progress, progress itself being conceived as an imrcas< In
detachment that freed the objectivity ol naluie, the elliuency ol tocli
CONCLUSION
299
nology, and the profitability of the market from the imbroglios of a
more entangled past. Detachment! Who could now believe for a sec
ond that science, technology, and the market lead us forward to less
entanglement, fewer imbroglios than in the past? No, the parenthesis
of progress is now coming to a close—but, contrary to the doubts that
beset the postmodern sensibility, there is no need to despair, no need
even to abandon the arrow of time.
There is a future, and it does differ from the past. But where once it
was a matter of hundreds and thousands, now millions and billions
have to be accommodated—billions of people, of course, but also bil
lions of animals, stars, prions, cows, robots, chips, and bytes. The only
feature that kept time moving forward in modernism and made it sus
pend itself in postmodernism was the definition of object, subject, and
politics, which has now been redistributed. That there was a decade
when people could believe that history had drawn to a close simply be
cause an ethnocentric—or better yet, epistemocentric—conception of
progress had drawn a closing parenthesis will appear (indeed, already
appears) as the greatest and let us hope the last outburst of an exotic
cult of modernity that has never been short on arrogance.
Unfortunately, as we have learned so painfully in this century, wars
have devastating effects, since they force every camp to stoop to the
level of its adversary. War has never been a situation in which to think
subtle thoughts, but rather has always offered licence for taking short
cuts, seizing any expedient at hand, and riding roughshod over all the
values of discussion and argumentation. The Science Wars have been
no exception. Just when a long and thoughtful peace was needed to re
assemble the broken factishes and to reinvent a politics of united hu
mans and nonhumans, a call to arms was heard from the Right and
from the Left, and “truth squads” were sent from campus to campus to
fumigate the hornets’ nests of science studies. I have nothing against a
good fight, but I would like to be able to choose my terrain, my wit
nesses, and my weapons—I want, above all, to decide for myself what
my war aims are. This is what I have tried to achieve in this book.
If I have not answered the science warriors’ arguments term for
term or even cited their names—it is because the science warriors
too often waste their lime attacking someone who has the same name
as mine, who is said to defend all the absurdities I have disputed for
twenty live yeais (hat science is socially constructed; that all is dis
300
course; that there is no reality out there; that everything goes; that sci
ence has no conceptual content; that the more ignorant one is the
better; that everything is political anyway; that subjectivity should be
mingled with objectivity; that the mightiest, manliest, and hairiest sci
entist always wins provided he has enough “allies” in high places; and
such nonsense. I don’t have to come to the rescue of that homonym!
Let the dead bury the dead, or as my mentor Roger Guillemin used to
say less grandly, “Science is not a self-cleaning oven so there is nothing
you can do about the layers of artifacts incrusted on its walls.”
Instead of this shadowboxing, I have decided to behave as if the
science wars were a respectable intellectual issue, not a pathetic dis
pute over funding fueled by campus journalists. According to my own
cartography, it is true indeed that everything to do with progress,
value, and knowledge is at stake here. In the powerful words of
Isabelle Stengers (1998), if we were really setting out to debunk sci
ence’s claims to know the world out there, everyone would admit that
“this means war,” a world war, even—at least a metaphysical one. It is
a battle that is worth fighting only if there are clearly two settlements
in opposition. The modernist one, which, in my eyes at least, is now
clearly behind us (though it was for many decades our most cherished
source of light, defended by giants before it fell to the care of dwarfs),
and another that is still in the offing. If anyone wants to wage this war,
they will now know on what grounds I stand, what values I am ready
to defend, and what simple weapons I expect to wield.
But I am pretty sure that when we meet on that front line, as will)
my friend who asked me the question that triggered this book, “Do
you believe in reality?” we will all be without weapons, dressed in ci
vilian clothes, since the task of inventing the collective is so formida
ble that it renders all wars puny by comparison—including, of course
the science wars. In this century, which fortunately is coming to a
close, we seem to have exhausted the evils that emerged from the open
box of the clumsy Pandora. Though it was her unrestrained curiosity
that made the artificial maiden open the box, there is no reason to
stop being curious about what was left inside. To retrieve the I lope
that is lodged there, at the bottom, we need a new and rather convo
luted contrivance. I have had a go at it. Maybe we will succeed with
the next attempt.
G L O S S A R Y
B I B L I O G R A P H Y
I N D E X
Glossary
actor, actant: The great interest of science studies is that it offers,

through the study of laboratory practice, many cases of the emergence of an
actor. Instead of starting with entities that are already components of the
world, science studies focuses on the complex and controversial nature of
what it is for an actor to come into existence. The key is to define the actor by
what it does—its performances*—under laboratory trials*. Later its compe
tence* is deduced and made part of an institution*. Since in English “actor” is
often limited to humans, the word “actant,” borrowed from semiotics, is
sometimes used to include nonhumans* in the definition.
actualization A term from the philosophy
of a p o t e n t i a l i t y :
of history, especially the work of Gilles Deleuze and Isabelle Stengers. The
best example is the pendulum whose movement is entirely predictable
from its initial position; letting the pendulum fall adds no new information. If
history is conceived in this way, there is no event*, and history unfolds in
vain.
a nti-programs : See programs of action.
apodeixis: See epideixis.
articulation: Like translation*, this term occupies the position left
empty by the dichotomy between the object and the subject or the external
world and the mind. Articulation is not a property of human speech but an
ontological property of the universe. The question is no longer whether or not
statements refer to a state of affairs, but only whether or not propositions* are
well articulated.
a s s o c iai io n , s u b s i i ru! i o n ; s y n t a g m , p a r a d ig m : T h e s e t w o p a ir s
o l te n u s te p la c e th e o b s o le te d is tin c tio n b e tw e e n o b je c ts a n d su b je c ts. In lin
g u istic s n sy n la g m is th e s i t o l w o l d s t h a t c a n b e a s s o c i a t e d in a s e n t e n c e
to t
GLOSSARY
304
(“the fisherman goes fishing with a basket” thus defines a syntagm), while a
paradigm is all the words that can be substituted in a given position in the
sentence (“the fisherman,” “the grocer,” “the baker” form a paradigm). The
linguistic metaphor is generalized to formulate two basic questions: Associa
tion—which actor can be connected with which other actor? Substitution—
which actor can replace which other actor in a given association?
belief: Like knowledge, belief is not an obvious category referring to a
psychological state. It is an artifact of the distinction between construction
and reality. It is thus tied to the notion of fetishism* and is always an accusa
tion leveled at others.
blackboxing: An expression from the sociology of science that refers to
the way scientific and technical work is made invisible by its own success.
When a machine runs efficiently, when a matter of fact is settled, one need fo
cus only on its inputs and outputs and not on its internal complexity. Thus,
paradoxically, the more science and technology succeed, the more opaque and
obscure they become.
Any site where inscriptions* are combined
c e n t e r of c a l c u l a t i o n :
and make possible a type of calculation. It can be a laboratory, a statistical in
stitution, the files of a geographer, a data bank, and so forth. This expression
locates in specific sites an ability to calculate that is too often placed in the
mind.
chain of t r a n s l a t io n : See translation.
circulating reference: See reference.
collective: Unlike society*, which is an artifact imposed by the mod
ernist settlement*, this term refers to the associations of humans and
nonhumans*. While a division between nature* and society renders invisible
the political process by which the cosmos is collected in one livable whole, the
word “collective” makes this process central. Its slogan could be “no reality
without representation.”
competence: See name of action.
complex vs. c o m p l i c a t e d : This opposition circumvents the tradi
tional opposition between complexity and simplicity by focusing on two
types of complexity. One, complication, deals with series of simple steps (u
computer working with o and 1 is an example); the other, complexity, deals
with the simultaneous irruption of many variables (as in primate interactions,
for example). Contemporary societies may be more complicated but less com
plex than older ones.
GLOSSARY
305
concrescence: A term employed by Whitehead to designate an event*

without using the Kantian idiom of the phenomenon*. Concrescence is not an
act of knowledge applying human categories to indifferent stuff out there but
a modification of all the components or circumstances of the event.
An expression borrowed from the theory of
c o n d i t i o n s of felicity:
speech acts to describe the conditions that must be fulfilled for an act of lan
guage to have meaning. Its opposite is conditions of infelicity. I extend the
definition to regimes of articulation such as science, technology, and politics.
context, co n ten t: Terms borrowed from the history of science to situ
ate the familiar puzzle of internalist* vs. externalist* explanations in science
studies.
COPERNICAN revolution: Introduced by Kant, this has become a cliche
in philosophical writings. Originally it meant the shift from geo to
heliocentrism. Paradoxically, Kant uses it to mean not a decentering of the hu
man position in the world but a recentering of the object around the human
ability to know. The expression “counter-Copernican revolution” thus com
bines two metaphors, one from astronomy and one from political upheaval,
to refer to the movement away from all sorts of anthropomoiphism, including
the sort invented by Kant. Politics does not have to be made through natuic*,
and objects may be freed as nonhumans from the obligation to shoilcut due
political process.
co sm opolitics: An ancient word from the Stoics to express an
affiliation to no city in particular but to humanity in general. The concept ac
quired a deeper meaning through its use by Isabelle Stengers to mean the new
politics that is no longer framed inside the modernist settlement* of nature*
and society*. There are now different politics and different cosmos.
demarcation vs. d i f f e r e n t i a t i o n : Normative philosophy of science
has devoted much effort to finding criteria to demarcate science from
parascience. To distinguish between this normative enterprise and the one of
this book, I use the word “differentiation” instead. Differentiation does not
require the generation of one normative distinction between science and
nonscience, but allows for many differences, making possible a much finer
normative judgment that does not rely on the weaknesses of the modernist
settlement*.
d i c t u m , m o d u s : Terms of rhetoric to distinguish the part of the sen
tence that does not change (the dictum) from the part of the sentence that
modifies (hence the name “modus”) the truth value of the dictum. In the sen
tence "I believe the eaith is getting wanner,” “I believe” is the modus.
GLOSSARY
306
differentiation: See demarcation.

envelope: An ad hoc term invented to replace “essence” or “substance”
and provide actors* with a provisional definition. Instead of opposing entities
and history, content* and context*, one can describe an actor’s envelope, that
is, its performances* in space and time. There are thus not three words, one
for the properties of an entity, another for its history, and a third for the act of
knowing it, but only one continuous network.
EPiDEixis, apodeixis: Terms of Greek rhetoric summarizing the whole
debate between philosophers and sophists. Etymologically both mean the
same thing, demonstration, but the first has drifted to refer to what the soph
ist does—a flourish of words—and the other to describe a mathematical or at
least rigorous demonstration.
event : A term borrowed from Whitehead to replace the notion of discov
ery and its very implausible philosophy of history (in which the object re
mains immobile while the human historicity of the discoverers receives all
the attention). Defining an experiment as an event has consequences for the
historicity* of all the ingredients, including nonhumans, that are the circum
stances of that experiment (see concrescence).
FACTISH, fetishism : Fetishism is an accusation made by a denunciator;
it implies that believers have simply projected onto a meaningless object their
own beliefs and desires. Factishes, in contrast, are types of action that do not
fall into the comminatory choice between fact and belief. The neologism is a
combination of facts and fetishes and makes it obvious that the two have a
common element of fabrication. Instead of opposing facts to fetishes, and in
stead of denouncing facts as fetishes, it is intended to take seriously the role of
actors* in all types of activities and thus to do away with the notion of belief*
fetishism : Seefactish.
historicity: A term borrowed from the philosophy of history to refer
not just to the passage of time—1999 after 1998—but to the fact that some
thing happens in time, that history not only passes but transforms, that it is
made not only of dates but of events*, not only of intermediaries* but of
mediations*.
immutable m o b il e : See inscription.
in scriptio n: A general term that refers to all the types of liansloiuia
tions through which an entity becomes materialized into a sign, an au Inve \
document, a piece of paper, a trace. Usually but not always insuiptions aie
two dimensional, superimposable, and combinable I hey aie always mobile
GLOSSARY
307
that is, they allow new translations* and articulations* while keeping some
types of relations intact. Hence they are also called “immutable mobiles,” a
term that focuses on the movement of displacement and the contradictory re
quirements of the task. When immutable mobiles are cleverly aligned they
produce the circulating reference*.
institution : Science studies has devoted much attention to the institu
tions that make possible the articulation* of facts. In common usage, “institu
tion” refers to a site and to laws, people, and customs that continue in time. In
traditional sociology, “institutionalized” is used as a critique of the poor qual
ity of overly routinized science. In this book the meaning is thoroughly posi
tive, since institutions provide all the mediations* necessary for an actor* to
maintain a durable and sustainable substance*.
intermediary: See mediation.
internal referent: See referent.
internalist ex pla na tio n s, externalist ex p l a n a t io n s: In the
history of science these terms refer to a largely obsolete dispute between
those who claim to be more interested in the content* of science and those
who focus on its context*. Although this distinction has been used for decades
to settle the relations between philosophers and historians, it has been totally
dismantled by science studies because of the multiple translations* between
context and content.
invisible college: An expression devised by sociologists of science to
refer to the informal connections among scientists as opposed to the formal
structure of university affiliations.
The general drift of science studies is to make matters
m a t t e r of f a c t :
of fact not, as in common parlance, what is already present in the world, but a
rather late outcome of a long process of negotiation and institutionalization.
This does not limit their certainty but, on the contrary, provides all that is
necessary for matters of fact to become indisputable and obvious. To be indis
putable is the end point, not the beginning as in the empiricist tradition.
mediatio n vs. i n t e r m e d i a r y : The term “mediation,” in contrast with
“intermediary*,” means an event* or an actor* that cannot be exactly defined
by its input and its output. If an intermediary is fully defined by what causes
it, a mediation always exceeds its condition. The real difference is not be
tween leahsls and relativists, sociologists and philosophers, but between
those who îecogni/e in the many entanglements of practice* mere intermedi-
ai les and those who ieiogm/e mediations
GLOSSARY
308
MODERN, POSTMODERN, NONMODERN, PREMODERN: Loose terms that

take on more precise meanings when the conceptions of science they entail
are taken into account. “Modernism” is a settlement* that has created a poli
tics in which most political activity justifies itself by referring to nature*.
Thus any conception of a future in which science or reason will play a larger
role in the political order is modernist. “Postmodernism” is the continuation
of modernism except that confidence in the extension of reason has been
abandoned. The “nonmodern,” in contrast, refuses to shortcut due political
process by using the notion of nature, and replaces the modern and
postmodern divide between nature and society with the notion of the collec
tive*. “Premodernism” is an exoticism due to the invention of belief*; those
who are not enthusiastic about modernity are accused of having a culture and
only beliefs, not knowledge, about the world.
modus: See dictum.
NAME OF action: An expression used to describe the strange situa
tions—such as experiments—in which an actor* emerges out of its trials*.
The actor does not yet have an essence. It is defined only as a list of effects—or
performances—in a laboratory. Only later does one deduce from these perfor
mances a competence, that is, a substance that explains why the actor behaves
as it does. The term “name of action” allows one to remember the pragmatic
origin of all matters of fact.
nature: Like society*, nature is not considered as the commonsense ex
ternal background of human and social action but as the result of a highly
problematic settlement* whose political genealogy is traced throughout the
book. The words “nonhumans*” and “collective*” refer to entities that have
been freed from the political burden of using the concept of nature to short
cut due political process.
nonhuman: This concept has meaning only in the difference between
the pair “human-nonhuman” and the subject-object dichotomy. Associations
of humans and nonhumans refer to a different political regime from the war
forced upon us by the distinction between subject and object. A nonhuman is
thus the peacetime version of the object: what the object would look like if it
were not engaged in the war to shortcut due political process. The pair hu
man-nonhuman is not a way to “overcome” the subject-object distinction but
a way to bypass it entirely.
paradigm: See association.
performance: See name of action.
GLOSSARY
309
phenomenon: In Kant’s modernist solution, a phenomenon is the meet

ing point of things in themselves—which are inaccessible and unknowable
but whose presence is necessary to avoid idealism—and the active involve
ment of reason. None of these features is kept in the notion of proposition*.
practice: Science studies is not defined by the extension of social expla
nations to science, but by emphasis on the local, material, mundane sites
where the sciences are practiced. Thus the word “practice” identifies types of
studies that are exactly as far from the normative philosophies of science as
they are from the usual efforts of sociology. What has been revealed through
the study of practice is not used to debunk the claims of science, as in critical
sociology, but to multiply the mediators* that collectively produce the sci
ences.
pragmatogony: A neologism invented by Michel Serres on the same
template as “cosmogony” to mean a mythical genealogy of objects.
predication: A term of rhetoric and logic meaning what happens in the
activity of definition when, to avoid a tautology, a term is necessarily defined
through the use of another term. This entails for each definition a transla
tion*, the one being obtained through the mediation* of the other.
Terms from the sociology of
p r o g r a m s of a c t i o n , a n t i - p r o g r a m s :
technology which have been used to give technical artifacts their active and
often polemical character. Each device anticipates what other actors, humans
or nonhumans, may do (programs of action), but these anticipated actions
may not occur because those other actors have different programs—anti
programs from the point of view of the first actor. Hence the artifact becomes
the front line of a controversy between programs and anti-programs.
project: The great advantage of technology studies over science studies
is that the former deals with projects that are obviously neither objects nor
subjects nor any combination of the two. A large part of what is learned in the
study of artifacts is then reused to study facts and their history.
proposition: I do not use this term in the epistemological sense of a
sentence that is judged to be true or false (for this I reserve the word “state
ment”), but in the ontological sense of what an actor offers to other actors.
The claim is that the price of gaining analytical clarity—words severed from
world and then reconnected by reference and judgment—is greater and pro
duces, in the end, more obscurity than granting entities the capacity to con
nect to one another through events*. The ontological meaning of the word
has been elaboialed by Whitehead
GLOSSARY
310
reference, referent: Terms from linguistics and philosophy that are

used to define, not the scenography of words and the world, but the many
practices that end up in articulating propositions*. “Reference” does not des
ignate an external referent that will be meaningless (that is, literally without
means to achieve its movement), but the quality of the chain of transforma
tion, the viability of its circulation. “Internal referent” is a term from
semiotics to mean all the elements that produce, among the different levels of
signification of a text, the same difference as the one between a text and the
outside world. It is connected to the notion of shifting*.
relati ve e x i s t e n c e : As a consequence of the positive meaning of rela
tivism*, the insistence on the emergence of actors, the pragmatic and rela
tional definition of action, and the importance given to envelopes*, it is possi
ble to define existence not as an all-or-nothing concept but as a gradient. This
allows for much finer differentiations* than the demarcation between exis
tence and nonexistence. It also makes it possible to avoid using the notion of
belief*.
relativism: This term does not refer to the discussion of the
incommensurability of viewpoints—which should be called absolutism—but
only to the mundane process by which relations are established between
viewpoints through the mediation* of instruments. Thus insisting on relativ
ism does not weaken the connections between entities, but multiplies the
paths that allow one to move from standpoint to standpoint. Science studies
has elaborated a new solution to replace the simpleminded distinction be
tween local and universal.
settlement: Shorthand for the “modernist settlement,” which has
sealed off into incommensurable problems questions that cannot be solved
separately and have to be tackled all at once: the epistemological question of
how we can know the outside world, the psychological question of how a
mind can maintain a connection with an outside world, the political question
of how we can keep order in society, and the moral question of how we can
live a good life—to sum up, “out there,” “in there,” “down there,” and “up
there.”
shifting i n , shifting out, shifting d o w n : Terms from semiotics
to designate the act of signification through which a text relates different
frames of reference (here, now, I) to one another: different spaces, dillerent
times, different characters. When the reader is sent from one plane of leler
ence to another, it is called shifting out; when the reader is biought back to
the original plane of reference, it is called shilling in, when the mallei ol ex
pression is entirely changed, it is called shilling down These shills lesull in
GLOSSARY
311
th e production o f an internal referent*, a depth o f vision, as if on e is dealing

w ith a differentiated w orld.
s o c i e t y : T he w ord does not refer to an en tity that exists in itse lf and is

ruled by its ow n laws by opp osition to other entities, such as nature; it m eans
the result o f a settlem ent* that, for political reasons, artificially divides things
betw een the natural and the social realms. To refer not to the artifact o f so ci
ety but to the m any connections betw een hum ans and nonhum ans*, I use the
w ord “collective*” instead.
s u b s t a n c e : This w ord designates w hat “lies beneath” properties; science

studies has not attem pted to do away w ith the notion o f substance altogether
but to create a historical and political space in w hich new ly em erging entities
are slow ly provided w ith all their m eans, all their institutions*, to be slow ly
“substantiated” and rendered durable and sustainable.
s u b s t i t u t i o n : See association.
s y n t a g m : See association.
s y n t h e t i c a p r i o r i j u d g m e n t : An expression em ployed by Kant to

solve the problem o f the fecundity o f know ledge w h ile at th e sam e tim e in sist
ing on th e prim acy o f hum an reason in the shaping o f know ledge. As opp osed
to analytical a p r io r i judgm ents, w h ich are tautological and sterile, and syn
th etic a p o s te r io r i judgm ents, w h ich are fecund and m erely em pirical, these
judgm ents are sim ultaneously a p r io r i and synthetic. W hen on e deals w ith ar
ticulated propositions* this classification b ecom es obsolete, since neither fe
cundity— events*—nor logic has to be allocated betw een the object and sub
ject poles.
t r a n s l a t i o n : Instead o f opp osin g w ords and th e w orld, science studies,

by its insistence on practice*, has m ultiplied the interm ediary term s that fo
cus on the transform ations so typical o f the sciences; like “inscription *” or
“articulation*,” “translation” is a term that crisscrosses the m odernist settle
m ent*. In its linguistic and material connotations, it refers to all the displace
m ents through other actors w h ose m ediation is indispensable for any action
to occur. In place o f a rigid op p osition betw een context* and content*, chains
o f translation refer to the w ork through w hich actors m odify, displace, and
translate their various and contradictory interests.
triai s :
In their emerging state, actors* are defined by trials, which can be
experiments of various soils in which new performances* are elicited. It is
ihiough Inals that actois aie clcimcd
Bibliography
Alder, K. 1997. E n g in eerin g th e R evo lu tio n : A r m s a n d E n lig h te n m e n t in F rance,

1763-1815. Princeton: Princeton University Press.
Apter E., andW. Pietz, eds. 1993. F etish ism a s C u ltu ra l D isco u rse. Ithaca: Cornell
University Press.
Aquino, P. d., and J. F. P. d. Barros. 1994. “Leurs noms d’Afrique en terre
d’
Amérique.”N o u v elle revu e d e th n o p s y c h ia tr ie (24): 111-125.
Beck, B.B. 1980. A n im a l T ool B eh a vio r: The Use a n d M a n u fa c tu re o f T ools b y A n i

m a is. NewYork, London: Garland STPMPress.
Beck, U. 1995. E co logical P o litic s in a n A g e o f R isk. Cambridge: Polity Press.
Bensaude-Vincent, B. 1986. “Mendeleev’s Periodic System of Chemical Ele
ments.”B ritish J o u rn a l f o r th e H is to r y a n d P h ilo so p h y o f S cien ce 19: 3-17.
Bloor, D. [1976] 1991. K n o w le d g e a n d S o c ia l Im a g e ry , 2d ed. with newforeword.
Chicago: University of Chicago Press.
Callon, M. 1981. “Struggles and Negotiations to Decide What Is Problematic
and What Is Not: The Sociologies of Translation.” In K. D. Knorr,
R.Krohn, and R.Whitley, T h e S o c ia l P ro cess o f S c ie n tific In v e stig a tio n , 197
220. Dordrecht: Reidel.
Cantor, M. 1991. “Félix Archimède Pouchet scientifique et vulgarisateur.”
Thèse de doctorat, Université d’Orsay.
Cassin, B. 1995. L e ffet so p h istiq u e . Paris: Gallimard.
Chandler, A. D. 1977. T h e V isible H a n d : T h e M a n a g e r ia l R ev o lu tio n in A m e r u a n

B u sin ess. Cambridge, Mass. : Harvard University Press.
Conant, J. B.1957. P a s te u r s S tu d y o f F e rm e n ta tio n . Cambridge, Mass.: Ilarvatd
University Press.
De Brosses, C. 1760. D u C u lte d e s d ie u x f e tic h e s . Paris: Fayard, Corpus des
Oeuvres de Philosophie.
DeWaal, F. 1982. C h im p a n ze e P olitics: P o w e r a n d S ex a m o n g A pes New Y0

1k
Harper and Row.

BIBLIOGRAPHY
313
Descola, P., and G. Palsson, eds. 1996. N a tu r e a n d S o ciety: A n th r o p o lo g ic a l P e r

s p e c tiv e s. London: Routledge.
Despret, V. 1996. N a issa n c e d ’u n e th é o rie éth o lo g iq u e. Paris: Les Empêcheurs de
penser en rond.
Détienne, M., and J. P. Vernant. 1974. Les R u ses d e l ’in tellig en ce. L a m é tis d e s
G recs. Paris: Flammarion Champs.
Eco, U. 1979. T he R ole o f th e R e a d e r : E x p lo ra tio n s in th e S e m io tic s o f Texts. Lon
don: Hutchinson; Bloomington: Indiana University Press.
Eisenstein, E. 1979. The P r in tin g P ress a s a n A g e n t o f C hange. Cambridge: Cam
bridge University Press.
Ezechiel, N., and M.Mukherjee, eds. 1990. A n o th e r I n d ia : A n A n th o lo g y o f Con

te m p o r a r y I n d ia n F iction a n d P o etry. London: Penguin.
Farley, John. 1972.. “The Spontaneous Generation Controversy—1700 1860
The Origin of Parasitic Worms,”J o u rn a l o f th e H is to ry o f B iology 95
125.
-
---
---
---
---
-1974. T he S p o n ta n e o u s G e n e ra tio n C o n tro v e rsy f r o m D e s c a r te s to O p a rin
Baltimore: Johns Hopkins University Press.
Frontisi-Ducroux, F. 1975. D é d a le . M y th o lo g ie d e l ’a r tis a n en G r e ie A n a e im e

Paris: Maspéro-La Découverte.
Galison, P. 1997. Im a g e a n d Logic: A M a te r ia l C u ltu re o f M ic ro p h y sic s. Chicago:
University of Chicago Press.
Geison, G. 1974. “P a s te u r .” D ic tio n a r y o f S cien tific B iograph y, ed. C. Gillispie,
351-415. NewYork: Scribner.
-
---
---
---
---
-1995. The P r iv a te S cien ce o f L ou is P a steu r. Princeton: Princeton Univer
sity Press.
Goody, J. 1977. The D o m e s tic a tio n o f th e S a v a g e M in d . Cambridge: Cambridge
University Press.
Greimas, A. J., and J. Courtès, eds. 1982. S e m io tic s a n d L a n g u a g e: A n A n a ly tic a l

D ic tio n a ry . Bloomington: Indiana University Press.
Hacking, I. 1983. R e p re se n tin g a n d I n terv e n in g : I n tr o d u c to r y T opics in th e P h ilo so

p h y o f N a tu r a l Science. Cambridge: Cambridge University Press.
-
---
---
---
---
-1992. “The Self-Vindication of the Laboratory Sciences.” In S cien ce as
P r a c tic e a n d C u ltu re, ed. A. Pickering, 29-64. Chicago: University of Chi
cago Press.
Halbertal, M., and A. Margalit. 1992. Id o la try . Cambridge, Mass.: Harvard
University Press.
Heidegger, M. 1977. T he Q u estio n C o n cern in g T ech n ology a n d O th e r E ssays. New
York: Harper and Row.
Hirschman, A . O. 1991. T he R h eto ric o f R ea ctio n : P erversity, F u tility, J e o p a rd y
C am biidge, Mass.: Harvard University Press

BIBLIOGRAPHY
314
Hirshauer, S. 1991. “The Manufacture of Bodies in Surgery.” S o c ia l S tu d ie s o f

S cien ce 21(2): 279-320.
Hughes, T. P. 1983. N e tw o r k s o f P o w e r : E le c trifica tio n in W estern S ociety, 1 8 8 0 -

1930. Baltimore: Johns Hopkins University Press.
Hutchins, E. 1995. C o gn ition in th e W ild. Cambridge, Mass:. MIT Press.

Iacono, A. 1992. Le f é tic h is m e . H isto ire d 'u n co n cep t. Paris: PUF.
James, W. [1907] 1975. P r a g m a tis m a n d T h e M e a n in g o f T ru th . Cambridge,

Mass. : Harvard University Press.
Jones, C, and P. Galison, eds. 1998. P ic tu r in g Science, P ro d u c in g A rt. London:
Routledge.
Jullien, F. 1995. T he P r o p e n s ity o f T h in gs: T o w a r d a H is to r y o f E ffica cy in C h ina.

Cambridge, Mass. : Zone Books.
Koerner, J. L. 1995. “The Image in Quotations: Cranach’s Portraits of Luther
Preaching.”In S h o p T alk: S tu d ie s in H o n o r o f S e y m o u r Sliue, 143-146. Cam
bridge, Mass. : Harvard UniversityArt Museums.
Rummer, H. 1993. Vies d e sin ges. M o e u rs e t s tr u c tu r e s so c ia le s d e s b a b o u in s

h am adryas. Paris: Odile Jacob.
Latour, B., and P. Lemonnier, eds. 1994. D e la p r é h is to ir e a u x m issile s

b a lis tiq u e s —l ’in te llig e n c e so c ia le d e s tech n iq u es. Paris: LaDécouverte.
Latour, B., P.Mauguin, et al. 1992. “ANote on Socio-technical Graphs.”S o c ia l
S tu d ie s o f S cien ce 22(1): 33-59; 91-94.
Law, J., and G. Fyfe, eds. 1988. P ic tu rin g P o w e r: V isu al D e p ic tio n s a n d S o c ia l R e la

tion s. London: Routledge.
Lemonnier, R, ed. 1993. T ech n o lo g ica l C hoices: T ra n sfo rm a tio n in M a te r ia l C ul

tu re s sin ce th e N eo lith ic. London: Routledge.
Leroi-Gourhan, A. 1993. G estu re a n d Speech. Cambridge, Mass.: MIT Press.

Lynch, M., and S.Woolgar, eds. 1990. R e p re se n ta tio n in S cien tific P ra ctice. Cam
bridge, Mass. : MITPress.
MacKenzie, D. 1990. I n v e n tin g A c c u ra c y : A H is to ric a l S o cio lo g y o f N u c le a r M issile

G u idan ce. Cambridge, Mass. : MITPress.
McGrew,W. C. 1992. C h im p a n ze e M a te r ia l C u ltu re: I m p lic a tio n s f o r H u m a n E vo

lu tio n . Cambridge: Cambridge University Press.
McNeill, W. 1982. T he P u r s u it o f P o w e r: Technology, A r m e d Force a n d S o c iety sin ce

a .d . 1000. Chicago: University of Chicago Press.
Miller, P. 1994. “The Factory as Laboratory.”Scien ce in C o n tex t 7(3): 469 496.

Mondzain, M.-J. 1996. Im a g e, icône, écon om ie. Les so u rc es b y z a n tin e s de
l ’im a g in a ir e c o n te m p o ra in . Paris: Le Seuil.
Moore, A. W., ed. 1993. M e a n in g a n d R eference. Oxford: Oxford Univeisily
Press.
M o r e a u , R. 1992. “L e s e x p e r ie n c e s d e P a s t e u r s u r les g e n e r a t i o n s s p o n t a n é e s
Le p o in t de vue d ’u n m icro b io lo g iste P ieinieie p aitie I.» l i n ( l ’u n
BIBLIOGRAPHY
315
m ythe; D euxièm e partie: les conséq uences.” L a v ie d e s scien ce s 9(3): 231-

260; 9(4): 287-321.
M um ford, L. 1967. T h e M y th o f th e M a c h in e : T ech n ics a n d H u m a n D e v e lo p m e n t.
N e w York: Harcourt, Brace and W orld.
Nathan, T., and I. Stengers. 1995. M é d e c in s e t so rc iers. Paris: Les Empêcheurs de
penser en rond.
N ovick, P. 1988. T h a t N o b le D r e a m : T h e “O b je c tiv ity Q u e stio n ” a n d th e A m e ric a n
H is to r ic a l P rofession . Cambridge: Cambridge U niversity Press.
N ussbaum , M . 1994. T h e ra p y o f D esire: T h e o ry a n d P ra c tic e in H e lle n istic E th ics.
Princeton: Princeton U niversity Press.
Ochs, E., S. Jacoby, et al. 1994. “Interpretive Journeys: How Physicists Talk
and Travel through Graphic Space.” C o n fig u ra tio n s 2(1): 151-171.
Pestre, D. 1984. P h y siq u e e t p h y s ic ie n s en F rance, 1 9 1 8 -1 9 4 0 . Paris: E ditions des
Archives C ontem poraines.
Pickering, A. 1995. The M a n g le o f P ra c tic e : T im e, A gen cy, a n d S cience. Chicago:
University of Chicago Press.
Ruellan, A., and M. D osso. 1993. R e g a rd s s u r le sol. Paris: Foucher.
Schaffer, S. 1997. “Forgers and Authors in the Baroque Economy.” Paper
presented at the meeting “What Is An Author?” Harvard University,
March.
--------- 1992. “A Manufactory of OHMS, Victorian Metrology and Its Instru
mentation.” In In v isib le C on n ection s, ed. R. Bud and S. Cozzens, 25-54.
Bellingham, Wash. : SPIE Optical Engineering Press.
------- 1994. “Empires of Physics.” In E m p ire s o f P h ysics, ed. R. Staley. Cam
bridge: Whipple Museum.
Serres, M. 1987. S ta tu e s. Paris: François Bourin.
--------- 1993. L ’o rig in e d e la g é o m é tr ie . Paris: Flammarion.
------- 1995. T h e N a tu r a l C o n tra ct, trans. E. MacArthur and W. Paulson. Ann
Arbor: U niversity o f M ichigan Press.
Shapin, S., and S. Schaffer. 1985. L e v ia th a n a n d th e A ir-P u m p : H o b b e s, B oyle, a n d
th e E x p e r im e n ta l Life. Princeton: Princeton U niversity Press.
Star, S. L., and J. Griesemer. 1989. “Institutional Ecology, ‘Translations,’
and Boundary Objects: Amateurs and Professionals in Berkeley’s Mu
seum of Vertebrate Zoology, 1907-1939.” S o c ia l S tu d ie s o f S cien ce 19: 387-
420.
Stengers, 1. 1993. L ’in v e n tio n d e s scien ce s m o d e rn e s. Paris: La Découverte.
------- 1996. C o sm o p o litiq u e s, tome 1: L a G u erre d e s scien ces. Paris: La
D écouverte et Les Em pêcheurs de penser en rond.
1998. “The Science Wars: What about Peace?” In Baudoin Jurdant, ed.,
Im p o stu res m l c l l a tu elles l c \ m a le n te n d u s d e TaJJaire S okal, 268 292. Paris:
I a Decouvei le
BIBLIO GRAPHY
316
Strum, S. 1987. A lm o s t H u m a n : A J o u rn ey in to th e W o rld o f B a b o o n s. N e w York:

Random House.
Strum, S., and B. Latour. 1987. “The M eanings o f Social: From Baboons to H u
m an s.” In fo rm a tio n s u r les S cien ces S o c ia le s /S o c ia l S cien ce I n fo r m a tio n 26:
783-802.
Tufte, E. R. 1983. T h e V isu al D is p la y o f Q u a n tita tiv e In fo rm a tio n . Cheshire,
Conn. : Graphics Press.
Viramma, J. Racine, and J.-L. Racine. 1995. Une v ie p a r ia . Le r ire d e s a sse rv is, p a y s
ta m o u l, In d e d u Su d. Paris: Plon-Terre hum aine.
Weart, S. 1979. S c ie n tists in P o w e r. Cambridge, M ass.: Harvard U niversity
Press.
W hitehead, A. N. [1929] 1978. P ro cess a n d R e a lity : A n E ssa y in C osm ology. N ew
York: Free Press.
Index
Abraham, 291-292 Archaeology-fiction, 236-237, 252

Absolute certainty, 4-8, 12-13, 15, 17, 21-22 Arendt, Hannah, 218
Absolutism, 20 Ariadne, 175,190
Abstraction, 48-50, 54 Art, 136
Accusations, 270 Articulation, 51,162,169,174,182,186-
Actors/actants/action, 122-123, 298, 303; 187,190, 193, 214, 295, 303; as meta
name of action, 119-120,143, 309; prop phor, 140; and propositions, 142-144,
ositions as actants, 141; programs of ac 146-147,149-151,158,180; and collec
tion, 159-160,178, 272, 310; and techni tives, 212
cal mediation, 178-190; and mastery, Artifacts, 112,125,165,178,184-185,190-
280-283 193, 196-197, 212, 214, 300. See also
Actualizations of potentiality, 126,152, Facts
303 Associated stimuli, 5
Aegina, 227 Associations, 158-167,179,183,198, 304
Agnosticism, 275-276, 283 Atheism, 298
Agrobusiness, 163-164 Athens, 11-12, 219-220, 226, 228, 234-
Alcibiades, 226, 249 235, 238, 240-241, 244-249, 251, 253,
Alcohol, fermentation of, 150-151,163,169 256-257, 265
Alienation, 206 Atlases, 101
Alliances, 100, 103-104 Atomic bomb, 82-83, 88
Allier, Jacques, 85-86 Attila the Hun, 225
Amazonia, 24-79 Automatons, 206-207
Amplification, 71-72 Autonomization, 100,102-103
Analytical judgments, 312 Autonomous Destiny myth, 178
Analytic philosophy, 48-49
Anthropology, 24, 26, 47, 67, 80,196, 206, Bachelard, Gaston, 127
231, 277, 280, 283, 290 Basic tool kit, 210-211
Antifetishism, 190, 270, 284, 286-288, 291. Bedrock, 40-41
See also Fetishism Belief, 154, 290-291, 304; in reality, 1-23;
A ntim odernism , 280 and knowledge, 16, 165 166; and fads/
Anti progiam s oi action, 160, 310. See also fetishes, 270 276; in beliefs, 276, 278;
Programs ol action alleinalive lo, 283 288
Apodeixis, 218 229 230, i u , 306 Bcigson, Ilcmi, 183
31
INDEX
318
Berzelius, Jôns Jakob, 117 nonhumans, 16,18, 20,174-176,180,

Big Bang, 146 193-202, 261, 296-297; and transla
Big Science, 100 tion, 193-195; and articulation, 212
Biochemistry, 144 Collège de France, 81, 83-84, 95,101
Blackboxing, 70,130,183-185,191-193, 304 Colors, standard, 58-61
Bloor, David, 133 Commissariat à l’Énergie Atomique, 88
Boa Vista, Brazil, 25-27, 30, 34, 47, 53, 55- Competences, 119-120,122,138,151,177,
57, 61, 68-69, 71- 73, 76-78, 79,101 209, 309
Body, reattachment to mind, 8-9 Complexity/complication, 210-211, 305
Body Politic, 216-219, 2.21, 226, 228, 232- Composition, 180-183,192
233, 236-237, 243-244, 247, 249-252, 257, Comte, Auguste, 128
262-266, 293-295 Conant, J. B., 115
Bonapartists, 155,162,167 Concrescence, 153, 305
Botany, 25-28, 32, 34, 36, 38-40, 42, 47, 51, Conditions of felicity, 219, 230, 240-242,
71, 76 248, 250, 262-263, 265, 266, 297, 305
Boulet, René, 28, 40-43, 46-47, 49-68,142, Construction, 114-115,124-125,127,168,
147 174, 223-224, 2.75, 280-282, 288. See
Brazil, 1, 4, 11, 24-79,188 also Fabrication
Byzantium, 289-290 Constructivism, 5-7, 21, 30, 79,115,125,
128-129,135,147, 194, 197, 274, 281,
Callicles, 11-12,14-16,18, 21-22, 216-231, 286, 290-291
234-236, 238-239, 243, 245-251, 253, 255, Content, 27, 91, 99,107-110, 305
258-265 Context, 27, 80, 91-92, 99,104,109-110,
Cartesian coordinates, 32, 34, 47, 49, 56 165, 305
Carthage, 239, 277 Copernican Revolution, 6,16,101, 305
Cassin, Barbara, 218, 232, 263 Correspondence, 69, 73,113,125,141-144,
Causality, 152 149
Centers of calculation, 55, 304 Cosmology, 19
Centre National de la Recherche Cosmopolitics, 16,18, 290, 305-306
Scientifique, 81, 83, 85, 98,101,103 Cosmos, 240
Certainty, 4-8, 12-13, 15, 17, 21-22 Counter-Copernican revolutions, 16, 305
Chains of translation, 27, 69-74, 79, 91-92, Critics, modern, 276-280
298, 313 Crossover, 194-195, 202-203, 205, 213
Chandler, Alfred, 204 214
Chauvel, Armand, 25-29, 40, 44, 46, 56, Curie, Marie, 81, 86
61-63, 67-68, 78 Curie, Pierre, 81
Chemistry, 116-124,126,128,131,133, 144,
150-151,162-163 Daedalia,175-176,178,181
Circulating reference, 24, 53, 71-74, 80, Daedalus, 171,175-176,183, 190 191, 19ft,
103,113,122,150, 248, 272, 297, 311 211
Classification, 34, 36, 38-39, 51 Darwin, Charles, 9-10, 21, 39, 105, 162
Cocalus, 175 Dautry, Raoul, 83 85, 88 89, 92, 97 99
Collections, 34, 36, 38-40 104, >78
Collective experimentation, 20 De Gaulle, Chai les, 88
Collectives, 3, 7, 98, 108, 259, 287, 304; and Delegation, 185 190, 192 191, 29s
society, 111, 114 , >93 >98; exploialion Deleu/e, Gilles, 303
of, 161 163; oi humans and
INDEX
319
Demarcation/differentiation, 141,157-158, Factishes, 16,136,187, 263, 274-278, 280-

164,166, 306 281,283-284, 288-292, 294,307
Democracy, 218-219, 2.22, 228-229, 234, Facts, 125,127,129,133,140,157,178, 293,
242, 246, 252, 265 295; scientific, 93-94, 99,109,112,174,
Descartes, René, 4-5, 8-9, 263 193, 281; and fetishes, 272-278, 281-
Despotic Ego, 6-7,15 282, 284-291. See also Artifacts
Deuterium, 83-85, 95 Fermat, Pierre de, 216
Diagrams, 54-57, 64-68, 78 Fermentation, 115-124,126,129-133,135-
Diaspora Museum, 291 139,141-153,156,163,169,171
Dictum, 93-94, 306 Fermi, Enrico, 93
Didacticism, 242, 247, 250 Fetishes/fetishism, 16,136,189-190,197,
Differentiation. See Demarcation/differ 270, 272-278, 280, 284-291
entiation Filizola, Héloïsa, 26, 28, 40, 44, 46-47,
Displacement, 194-195 49, 56, 61-63, 67-68
DNA, 203 Fission, nuclear, 82, 88-89, 94, m
Dualist paradigm, 198, 212, 214 Force, 10-12, 21, 259-264
Durkheim, Emile, 208-209 Forests, 25-27, 30-32, 34, 36, 38-42, 51,
53- 58, 63, 65-67, 70, 76-78
Earthworms, 66, 74, 76 Foucault, Michel, 192, 262
Ecology, 202-203, 208 France, 2, 25, 74, 81-86, 88-90, 93, 95, 98,
Edison, Thomas A., 204-205 107,111,128,137,155,161,186,188,195
Elites, 223 Freedom, 276
Empiricism, 4-5, 26, 77, 115,129,153, 166- Freud, Sigmund, 289
167, 169-170
Enrollment, 194-195 Galileo Galilei, 289
Enucleation of society, 109-112 Garfinkel, Harold, 209
Envelopes, 119,156-159,161, 166-167 Garimperos, 27, 30, 46
Enzymology, 151 Genevieve, Saint, 225
Epideixis, 218, 240, 306 Geography, 26, 28, 61
Episteme, 174,181, 230-232, 236, 255, 258, 262 Geometrical equality, 11-12, 217, 248,
Epistemology, 13-14, 19, 22-23, 26, 53, 68, 264
79, 93, 98, 109-111, 115, 122, 128-129, 133, Geometry, 42-43, 49, 54~55, 67, 217, 252
141, 145-146, 158, 166, 174, 178, 215, 223, Geomorphology, 47
231, 247, 2.55, 258>271, 2.83, 285, 288, 293, Germany, 82-86, 88, 93, 95,107
296 Germs. See Microbes
Euclidean geometry, 42-43 Gestell, 176, 183,185
Events, 126, 143, 152 153, 166, 281, 306 Glaucus, 236, 263
Existence. See Relative existence Glickman, Steve, 255
Experimental protocols, 46-47 Goal translation, 178-180
Experiments, 20,124 127, 129 132,135, God, 1, 4-5,14, 267, 275, 282-283, 291,
162, 166, 298 298
Expei Is, 228, 240 243 Gorgias, 219, 228, 230-233, 238-240, 242,
Externalist explanations, 85, 91 92, 308 246, 254
Greece, ancient, 11 12,174-175, 218-220,
I a lu n a tio n , 114 115, 122, 125, 127, 135, 138, 226 228, 234 235, 238, 240 242, 244-
2 2 2 5 280 282, 284 ,V( «Ad ( 011 249, 251 253, 256 257,265
slim (ion (iiulirmin Kogei 300
INDEX
320
Halban, Hans, 81, 83-86, 96,107,147 Joliot, Frédéric, 81-90, 93-100,104,107-

Haraway, Donna, 4 108,111-112,114,165,178,194
Harvey, William, 80,107 Jussieu, Joseph de, 30
Hegel, G. W. F., 182
Heidegger, Martin, 3,176,183,195, 211 Kant, Immanuel, 5-7,16, 21, 43 - 4 4 , 57, 72,
Historicity, 145-146,150,152,156,158,161, 101,125, 305, 309, 312
164,177,168, 307 Knots, 99-100,106-108
Hobbes, Thomas, 263 Know-how, 174,181, 231, 233, 254, 256-
Homer, 229; Iliad, 175 257
Homofaber, 182, 190, 197, 281-282, 298 Knowledge, 7, 9,19, 39, 53, 58, 69, 71-72,
Horizons, 40-43, 64-66 133,174, 247, 254, 258, 262, 264, 304;
Hughes, Thomas, 204 and belief, 16,165-166; for the people,
Humanism, 3, 17-19, 23, 261 228-232, 240, 242-244, 252, 255, 257;
Humanities, 21, 261 and facts/fetishes, 273-274
Humboldt, Alexander von, 30 Kowarski, Lew, 81, 83, 96, 98, 107
Hume, David, 5,125 Kummer, Hans, 211
Icarus, 175 Laboratory practice. See Practice

Iconoclasm, 268, 270-274, 276-281, 285, Lactic acid, fermentation of, 115-119,
289-290 121-124,126,131-133,136-139,141- 144,
Idealism, 147,172 150-151,153,163,165,169
Immutable mobiles, 102, 307 Language, 69-70, 72, 9 6 -9 7 , 133,141-142,
Impersonal laws, 228, 259-261, 265 144,148,152, 296
Industrial metaphors, 137-140 Laugier, André, 81
Industry, 205-207 Leroi-Gourhan, André, 182
Inhumanity, 13,15, 217-218, 222, 236, 258, Liebig, Justus von, 115-117,126,144, 150-
260-261, 291, 293-294 152,163
INPA, 26, 76 Life-world, 9
Inscriptions, 28-29, 54, 65, 67, 307 Lille, France, 122,137,142,148,150,152,163
Institutions, 103,151,156,160,167,170, Links, 99-100,106-108
297, 307 Lyotard, Jean-François, 232
Institut Pasteur, 203
Instruments, 99-102,176, 206 Machiavelli, Niccolô, 211, 253, 263
Interference, 178-180,192 Machines, 206-208
Intermediaries. See Mediation/intermedi- Manaus, Brazil, 26, 34, 46, 74, 76,101
aries Maps, 28-30, 42, 78
Internalist explanations, 85, 91-92, Marx, Karl, 182, 206, 289
308 Marxism, 41, 290
Internalized ecology, 208 Mastery, 280-283, 297-298
Internal referent, 56, 311 Materialism, 176-177,180,190
Invisible college, 308 Mathematics, 54-56, 218, 230
Isaac, 291 Matters of fact, 152, 308
Mediating technology, 178 190
Jagannath, 268-271, 276, 285-287, 289, Mediation/inlermediaiies, 7, 56 57. 73.
291-292 79, 101, 148, 153. 197. 214. 278, 288, 298,
James, William, 73- 74, 79, 113 308, technical, 178 190
INDEX
321
Megamachines, 207-208 165,172, 185-186, 189-190, 211-215,

Mendeleev, Dmitri, 5, 78 259-260, 265, 266, 275, 282-286, 290,
Metaphors, 133-140,187 298-299, 309; in collectives, 16,18, 20,
Metis, 174,181 174-176,180,193-202, 261, 296-297;
Microbes, 145-147,154-155,158, 164-170, symmetry with humans, 182; in
172-173 pragmatogony levels, 202-211
Microbiology, 154-155,167,170 Nonmodernism, 21-22, 215, 282-283,
Midas, 240 286, 290, 295, 308-309
Might versus Right, 20-21, 216-236 Norsk Hydro Elektrisk, 83, 85
Mind, 8-9,14, 282 Norway, 83-84, 87, 99,107
Mind-in-a-vat, 4-10,12-16, 21,113, 296 Nuclear physics, 80-112
Ministry of Armaments, 83, 88-90, 92
Minos, 175 Objectivity of science, 3,174,199
Mobilization: of world, 99-102; and col Objectification, 13,15-16, 23
lectives, 194-195 Odysseus, 175
Mob rule, 7,10-17, 23, 217-218, 223, 234- Ontology, 14, 93, 98,115,118,120,122,
235, 260, 265, 293 128,131,133, Hi, 145- 147,151,166,185-
Modernism, 21-22,196-197,199-200, 206, 186,189, 204, 259, 283, 285-288, 290-
212, 214-215, 251, 268-269, 275-283, 291, 293
286-291, 293-299, 308-309 Optical metaphors, 136-140
Modernist settlement, 14, 96,134-135,145, ORSTOM, 25-26, 28, 56
174.193, 214, 275, 296, 298, 300, 311 Outside world, 13-16,113,141, 296
Modus, 93, 306
Morality, 6 , 11, 13- 14 . 22,168,177, 222, Pandora, 201, 300
224, 239. 242-243, 252-258, 293, 297 Paradigms, 91-92, 96,109,160,162,164,
Moses, 289 167,187,193,198, 212, 214, 304
Mumford, Lewis, 207 Parallelogram metaphor, 134-135,140
Munsell code, 58-60, 63 Pariahs, 268-271, 279-280
Paris, France, 53, 73, 78, 90, 95,137, 154,
Name of action, 119-120,143, 309 163
Napoleon I, 236 Pasteur, Louis, 16, 93,115-173,178-179,
Napoleon III, 155,161,164 188, 259, 271, 295
National Rifle Association, 176-178,190, Paul, Saint, 289
192 Pedocomparators, 47~49. 51. 53~59. 67
Naturalists, 39-40 Pedogenesis, 40, 66
Nature, 9-10,14,125-126,141,153,156-157, Pedology, 25-28, 32, 40-68, 71, 75-76
164.193, 297, 309 Perelman, Charles, 218
Nature, 82, 96, 98 Performances, 119,122,151,167, 309
Networks of power, 204-205 Pericles, 12, 240, 245-246, 249
Neutral Tool myth, 178 Phenomena, 71-72,121,151,154- 155, 309-
Neutrons, 82-85, 87-90, 93- 95, 99.107 310
Newton, Isaac, 105, 282 Phenomenology, 9-10
Nietzsche, Friedrich, 217, 221, 223, 228, Physics, 80-112
234, 240,246,289 Pistis, 232, 236, 242, 247
Nonhumans, 3, 15, 20 22, 85, 96, 99,112, Plato: Gorgias, 10-12,14-16,18-19, 216-
115, 118, 122, )29 130, 135. 141. 147. «57. 265; Republic, 240, 252, 289
INDEX
322
Platonism, 49, 60,110,177 of discourse, 30; traceability of, 46-47,

Plutonium, 111-112 69; internal, 56, 311; standardized, 59-
Political ecology, 202-203 61; translation of, 70-71
Politics, 13-14, 21,140, 214, 219, 227, 280, Relative existence, 156,159,161,164, 311
293, 295, 297; and science, 84-88, 215- Relativism, 4,17, 20, 58, 73,156-157,168,
218, 258-265; freed from science, 236- 296-297, 311
265 Research, 20-22
Polus, 219, 224, 228-229, 231, 238, 246, Retrofitting, 170,173
254-255 Reversible blackboxing, 183-185
Postmodernism, 21-22, 217, 275, 286, 299, Rhadamanthys, 227, 257
308-309 Rhetoric, 96, 229, 232-233, 238-244, 246,
Pouchet, Félix Archimède, 146,153-168, 253, 256
172, 271 Right versus Might, 20-21, 216-236
Power, 204-205, 262, 264 Rousseau, Jean-Jacques: Discourse on the
Practice, 3-4, 15, 24, 32, 53, 56, 174, 267- Origin of Inequality, 236-237, 263
268, 278, 282, 294, 310
Pragmatogony, 176,178,193, 200-202, Saligrams, 268-270, 275, 285-286
207-208, 310 Samples, 44, 46, 51, 54-55, 61, 63, 74
Predication, 310 Sandoval, 29, 43-45, 61-62, 64, 74, 76
Premodernism, 200, 279-280, 293, 308- Sâo Paulo, Brazil, 26, 28, 46,189
309 Sartre, Jean-Paul, 282
Preservation, 34, 36, 38-39 Savannas, 24-27, 32, 34, 39~42, 47, 51, 53-
Programs of action, 159-160,178, 272, 57, 63, 65-66, 70, 76-78
310 Schaffer, S., 131
Projects, 158, 310 Science: cumulative nature of, 1-2,17;
Propositions, 288, 310-311; and state objectivity of, 3,174,199; left to itself,
ments, 141-144; and articulation, 142- 9-10; as invading everything, 9-10;
144,146-147,149-151,158,180; having and relativity, 17; and research, 20-22;
history, 146-150,153; envelope for, 153, as both realist and constructivist, 30,
158 79; and society, 84, 87, 91-92, 95,112;
Protocols, 46-47, 49, 51, 61, 196-197 and politics, 84-88, 215-218, 258-265;
Psychology, 13-14, 22, 231 content/context of, 91-92, 99,104,
Public representation, 100,105-106 107-110; and art, 136; and technology,
174-176,191,194,196,198-199, 214-
Radioactivity, 81-82, 98,111 215, 298-299; and reason, 217; and
Radium, 81, 86 rhetoric, 229
Realism, 30, 64, 67, 73, 79,109, H5,127, Science studies, 2-4, 6-7,13-14, 69, 78
135,158, 274, 290 79, 80, 94, 96, 114,137,146,158,161,
Reality, belief in, 1-23 173,174,185,199-200, 223, 261, 280,
Reason, 10-12, 21, 216-217, 223, 234-235, 293-294; originality of, 17 23; uniting
237, 239, 252, 257, 259-264 science and society, 84 87, 90 92, 112,
Reduction, 70-72 152; and content of science, 107 109;
References/referents, 32, 34, 43, 56, 64, 99, and language, 133 134; and relativism,
133,148-149, 312; circulating, 24, 53, 71- 156 157, 168
74, 80,103,113,122, 150, 248, 272, 297, Science wais, 1, 15, 17 19, 107 109, 111
311; scientific, 26, 30, 36,130; relerenl 112, 130, 134 ns, 152, 199, 202, 216,
INDEX
323
218, 258-259, 261, 265, 287, 293, 2.96, Substitutions, 89 90. 158, 160 162, 164,
299-300 166-167, 304
Scientific facts, 93-94, 99,109,112,174, Surveys, 100
193, 281 Symmetry, 182
Scientific institutions, 103 Syntagms, 160-161,163-164,187, 304
Scientific professions, 102-103, 296 Synthetic judgments, 126, 312-313
Scientific references, 26, 30, 36,130 Szilard, Leo, 82, 85, 90, 95, 98,107
Scientists, 17-19
Serres, Michel, 42, 203, 310 Tagging, 34, 36, 39. 46, 49
Setta-Silva, Edileusa, 25-34, 36, 39, 42, 46, Taxonomy, 121-122,164
49, 53, 56, 67-68, 76 Technical mediation, 178-190
Settlements, 13, 21,113, 150, 261-263, 267, Techniques, 176,178-179,186,190-192,
295, 311-312; modernist, 14, 96,134-135, 197, 209-210, 213, 298
145,174,193, 214, 275, 296, 298, 300, 311; Technology, 174-176,183,189-192,194,
old, 114, 133, 215, 265, 266, 289, 293-294, 196,198-199, 214-215, 298-299
300 Technoscience, 203-204
Shapin, S., 131 Teresopolis, Brazil, 1, 4,11
Shifts/shifting, 129-130,187-190,193, 312 Thales, 29
Siodmak, Curt : Donovan’s Brain, 4 Theater metaphors, 135-136,139-140
Sites, 101 • Themistocles, 240, 245
Social complication, 210 Theology, 14, 22, 293, 298
Social Darwinism, 11 Theories, 129,174, 294
Social relations, 197-198, 208-214 Tools, 210-211
Social sciences, 21,112,193, 261, 264 Topofils, 28, 43-44, 46, 49, 58, 138,
Society, 6,13-15,19,126,152,164,174, 190
207-209, 297, 312; and science, 84, 87, Traceability of data/references, 46-47,
91-92, 95,112; enucleation of, 109-112; 69, 78
and collectives, 111,114,193-198 Trail metaphors, 138-140
Sociobiology, 222 Transcendental Ego, 6, 72,125
Sociotechnology, 198, 203-204, 214 Transformations. See Translations
Socrates, 11-12,14-20, 22, 216-265, 268 Translation model, 91-92, 249
Soil studies. See Pedology Translations, 53, 58, 61, 64-65, 68,110,
Sophists, 21, 216-219, 221, 228, 230-232, 137-138,147,150,179, 295, 298, 313;
234, 240, 242, 247-248, 250-254, 263 chains of, 27, 69-74, 79, 91-92, 298,
Spontaneous generation, 153-156,159-163, 313; operation of, 87-98; and collec
166-168,171-172 tives, 193-195
Standardization, 58-61 Trials, 122-124, 143, 313
Statements, 48, 58,134,141-144.148 Truth, 64, 97,115, 125, 136,149, 225, 231,
Statesmen, 243, 245-246, 250 264-265, 296
Slengers, Isabelle, 16, 297, 300, 303, 305 Twain, Mark, 252
Stoics, 305 Two-culture gap, 17-18, 22-23
Stiucturalism, 38
Sli uctural pedology, 43 Union Minière du Haut Katanga, 81, 83
Slium, Shirley, 210, 255 87, 90, 98
Subpiogiams, 181 182,191 192,207 Universality, 6 8
Substances, 151. ifii, 164, 167, 170, 312 Uianium, 81 82, 86, 88 90
I NDEX
324
Waterfield, Robin, 218 Yeast, fermentation from, 115-116,118,

Weart, Spencer, 84-86, 90 120-121, 124,126-127,129, 131-133,143.
Weinberg, Steven, 2x6-218, 258-259, 265 150,153
Wenner-Grenn Foundation, 2
Whitehead, Alfred North, 141,153, 305- Z ero-sum gam e, 114, 125-127, 147
306, 311
World War II, 82-83, 88, 107, 111
Bruno Latour knew that when he opened up the Pandora’s box of sci--||g
jence it would be difficult to predict the results. In an incident tha^fe.
begins the book a scientist friend asked Latour point-blank: Do you
'believe in reality? Latour presents his meticulous response in Pandora Rag
■Hope. He seeks to shift the discussion onto an entirely different t e r r a i n ^
from where it is usually conducted—back to earth. This book offers a jig
Remarkable argument for understanding the reality of science in practi-WH
ical terms and the kind of realism pursued by science studies.
this book Latour gives us his most extensive concrete presentation

i r ^ ^ h i s position since Science in Action. Through base studies of scientists inr'^ r '|^
^ c ^ f f h e Amazon analyzing soil and in Pasteur’s lab studying the 1er-menia-|j#*Kl
^ 5 ^ t i o n of lactic acid, he shows us the myriad steps by which events in th e S r B I
îjigËflmaterial world are transformed into items of scientific know ledge.#
r^gL^ Through many examples in the world of technology, we see how the ’ i
^ r j ^m aterial and human worlds come together and are reciprocally trans-j g f o uj
- " ^form ed in this process.
n S r'
T. ' ’-Why, Latour asks, did the idea of an independent reality, free of hum an^SES
-^Interaction, emerge in the first place? His answer to this question, h a r k ^ B a ja
^ ^ F* 4ng back to the debates between Might and Right as narrated by Plato
y.^Tbhis Gorgias, points to the real stakes in the so-called science wars: th ë 3 5 |p i|
perplexed submission of ordinary people before the warring forces
T ^ ^ c la im a n ts to the ultimate truth. He suggests that the way out of o u t^ § j|!s |
'dj-^%eeming dilemma is to fetch the hope that legend tells us Pandora le ft^ M rl
j^ l^ ly in g on the floor of her box.
rakflg J jg p g
iitfH^-Bruno Latour is Professor at the Center for the Study of Innovation atfS P ^^
^fflrapbe School of Mines, Paris. He is the author of Aramis, or the Love p/fBP'ÿ'
Technology; We Have Never Been Modern; Science in Action; and The4 ÊB&
^ ^ ^ I Pasteurization of France (all from Harvard).
Harvard University Press

Cam bridge, M assachusetts
Lcm don, England
Www. hup. harvard, edu
ICOuer design by Lisa Clark

ST
yiiu.stration by Eric Dinyer, courtesy of Graphistock

Latour Pandora's Hope

Uploaded by

Copyright:

Available Formats

Latour Pandora's Hope

Uploaded by

Document Information

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

Latour Pandora's Hope

Uploaded by

Copyright:

Available Formats

Reality of Science Studies

HARVARD UNIVERSITY PRESS

Library of C ongress Cataloging - in -P ublication D ata

Second printing, 2000

1. “Do You Believe in Reality?” 1

( 'niu lusion 293

Words and phrases that I use in a technical sense

“Do You Believe in Reality?”

instruments, steeped in know-how, closely connected to a larger and

TIk*Strange Invention of an “Outside” World

intellectual history that should at least be sketched. Without this de­

world could directly send us enough information to produce a stable

food, and the inaccessible things-in-themselves to which the world

the realism to which we science students have become attached, and

into damnation suspended, since we no longer have a mind dealing

The Fear of Mob Rule

need to rely on another inhuman resource, the objective object un-

ity to be the answer to a question of belief asked by a brain-in-a-vat

neers began to swap properties. In Chapter 4 we see how Pasteur

The Originality of Science Studies

inside one’s own standpoint, its overemphasis on reflexivity, its mad­

The only way to understand the reality of science studies is to follow

. machinent of the savanna. Is the forest advancing like Birnam Wood

ai e four scientists whose gaze is able to dominate two maps of the

t.•milled against an otherw ise uniform ly green background. On this

and refer to it in the future. In order to be able to say that Afulamata

perfect intermediary between hardware (since it shelters) and soft­

forest. Let us first review what we know of that superiority before

In the naturalist’s collection things happen to plants that have never

•mic would be explained by the infrastructure, to use an old M arxist

Itied by a collection of coordinates. Had Kant used this instrument,

of triangles. Equipped only w ith the a priori form s of intuition, to use

As with Edileusa’s specimens, most of the analyses cannot be per­

board, the pedocomparator belongs to “things.” But in the regularity

ol a pedolibrary. All the transects have become compatible and

Look at René in the photograph: he is master of the phenomenon

1tier of Boa Vista.

Relations between Vegetation Dynamics and the Differ­

limiti i'honzon © .Homonte de fe ie s deminhocas (#.<U dejections d t vers de lerre)

Figure's. Coupe du tran se ct •)

I f.*10up of researchers. On each occasion a new phenomenon is

clay to prepare for growth in the twenty-meter-wide strip of land. But

I hr num ber is a reference that is quickly understandable and repro-

u>servation o f traces that establish a reversible route that makes it

molding figurines? No, it is to extract another judgment, one that no

state so radical that now a sign appears in place of a thing. Here it is no

minor in com parison w ith the transform ations through w hich we

li\e people and by passing through successive geometrical construc-

Here we are on much m ore fam iliar terrain—rhetoric, discourse,

11 iiu 1 10 I he “s.illulionisl’s” (James [1907] W75) conception of the feat of cor-

Another property is revealed by the comparison of my two

established centers of calculation. Instead of growing from two fixed

l'he canonical view

The intermediaries which in their concrete particularity form a

From a pedological point of view, admitting that the forest is advanc

2. that this transformation would begin in the savanna in a 15- to

Unfortunately, I will not be able to follow the next expedition.

room transformed into an office. On the wall, Radambrasil’s map of

loi Ihat matter. Through successive stages they link us to an aligned,

Science’s Blood Flow

intellectual history that should at least be sketched. Without this de

inside one’s own standpoint, its overemphasis on reflexivity, its mad

perfect intermediary between hardware (since it shelters) and soft

As with Edileusa’s specimens, most of the analyses cannot be per

Relations between Vegetation Dynamics and the Differ

By drawing on several seemingly incompatible philosophies of sci

And it is my opinion, at this point in the development of my knowl

u id fermentation. But at this point there is no “trouble with experi