Common Knowledge

S y m p o s i u m : Ta l k i n g P e a c e w i t h G o d s , P a r t 1

WHOSE COSMOS,
WHICH COSMOPOLITICS?
Comments on the Peace Terms of Ulrich Beck

Bruno Latour

Blessed are the peacemakers. It is always nicer to read a peace proposal (like
Ulrich Beck’s) than a call for jihad (like Samuel Huntington’s). Beck’s robust and
realist form of cosmopolitanism, expressed in the lead article of this symposium,
is to be welcomed. On the other hand, peace proposals make sense only if the real
extent of the conflicts they are supposed to settle is understood. A detached and,
let us say, inexpensive way of understanding enmity, a Wilsonian indifference
to its complexity, may further infuriate the parties to a violent dispute. The prob-
lem with Beck’s solution is that, if world wars were about issues of universality
and particularity, as he makes them out to be, then world peace would have
ensued long ago. The limitation of Beck’s approach is that his “cosmopolitics”
entails no cosmos and hence no politics either. I am a great admirer of Beck’s
sociology — the only far-reaching one Europe has to offer — and have said so in
print on several occasions. What we have here is an argument among friends
working together on a puzzle that has defeated, so far, everyone everywhere.
Let me make clear from the beginning that I am not debating the useful-
ness of a cosmopolitan social science that, beyond the boundaries of nation-states,
would try to look at global phenomena using new types of statistics and inquiries.
I accept this point all the more readily since for me, society has never been the
equivalent of nation-state. For two reasons: the first is that the scientific networks

Common Knowledge 10:3

Copyright 2004 by Duke University Press

450

Published by Duke University Press

 Common Knowledge

that I have spent some time describing have never been limited to national

451
boundaries anyway: global is largely, like the globe itself, an invention of science.
The second reason is that, as disciples of Gabriel Tarde know very well, society

Talking Peace with Gods: Part 1

has always meant association and has never been limited to humans. So I have
always been perfectly happy to speak, like Alphonse de Candolle, of “plant soci-
ology” or, like Alfred North Whitehead, of “stellar societies.”1 It should also be
clear that I don’t take the expression “peace proposal” ironically. On the contrary,
it’s for me crucial to imagine another role for social science than that of a dis-
tant observer watching disinterestedly. Beck is struggling for a mixture of
research and normative intervention, and this is exactly what I mean by the new
diplomatic role of the social scientist. What is in question between us is the extent
to which we are ready to absorb dissents not only about the identity of humans
but also about the cosmos they live in.

Latour •
A historical anecdote, retold in a major paper by Eduardo Viveiros de Cas-
tro, may illustrate why Beck’s suggested approach to peacemaking is not com-
pletely up to the task.2 The main example that Beck gives is the “Valladolid con-
troversy,” the famous disputatio that Spaniards held to decide whether or not
Indians had souls susceptible of being saved. But while that debate was under
way, the Indians were engaged in a no less important one, though conducted with
very different theories in mind and very different experimental tools.3 Their task,
as Viveiros de Castro describes it, was not to decide if Spaniards had souls — that
much seemed obvious — but rather if the conquistadors had bodies. The theory
under which Amerindians were operating was that all entities share by default
the same fundamental organization, which is basically that of humans. A licuri

1. The first book that tried to describe scientific networks Strauss, who retells an episode from Oviedo’s Historia
quantitatively was written by a plant sociologist and took (supposed to have taken place earlier in Puerto Rico). The
the point of view of cosmopolitan methodology. See famous passage from Tristes tropiques (Paris: Plon, 1955),
Alphonse de Candolle, Histoire des sciences et des savants 82–83, is as follows: “Au même moment d’ailleurs, et dans
depuis deux siècles d’après l’opinion des principales académies ou une île voisine (Porto Rico, selon le témoignage d’Ovideo)
sociétés scientifiques (1873; Paris: Fayard, 1987). les Indiens s’employaient à capturer les blancs et à les faire
périr par immersion, puis montaient pendant des semaines
2. Eduardo Viveiros de Castro, “Les pronoms cos-
la garde autour des noyés afin de savoir s’ils étaient ou non
mologiques et le perspectivisme amérindien,” in Gille
soumis à la putréfaction. De cette comparaison entre les
Deleuze: Une vie philosophique, ed. Eric Alliez (Paris: Les
enquêtes se dégagent deux conclusions: les blancs invo-
Empêcheurs de penser en rond, 1998), 429 – 62 (see also
quaient les sciences sociales alors que les Indiens avaient
his essay in this symposium).
plutôt confiance dans les sciences naturelles; et, tandis que
3. It is not clear whether the two main characters of the les blancs proclamaient que les Indiens étaient des bêtes,
controversy, as retold by Beck, ever met. See Anthony les seconds se contentaient de soupçonner les premiers
Pagden, The Fall of Natural Man: The American Indian and d’être des dieux. A ignorance égale, le dernier procédé
the Origins of Comparative Ethnology (Cambridge: Cam- était certes plus digne d’hommes.” But, as Viveiros de Cas-
bridge University Press, 1982). There is a fairly bad book tro has shown, making a decisive improvement on Lévi-
where the two protagonists meet (and the film is even Strauss’s interpretation, the point was not to check whether
worse): Jean-Claude Carrière, La Controverse de Valladolid the conquerors were gods but simply whether they had
(Paris: Le Pré aux clercs, 1992). The dispute over whether bodies.
Spaniards have bodies is documented by Claude Lévi-

Published by Duke University Press

 Common Knowledge

palm, a peccary, a piranha, a macaw: each has a soul, a language, and a family
452

life modeled on the pattern of a human (Amerindian) village. Entities all have
souls and their souls are all the same. What makes them differ is that their bod-
COMMON KNOWLEDGE

ies differ, and it is bodies that give souls their contradictory perspectives: the per-
spective of the licuri palm, the peccary, the piranha, the macaw. Entities all have
the same culture but do not acknowledge, do not perceive, do not live in, the same
nature. For the controversialists at Valladolid, the opposite was the case but they
remained blissfully unaware that there was an opposite side. Indians obviously
had bodies like those of Europeans, but did they have the same spirit? Each side
conducted an experiment, based on its own premises and procedures: on the one
side to determine whether Indians have souls, and on the other side to determine
whether Europeans have bodies. The Amerindians’ experiment was as scientific
as the Europeans’. Conquistador prisoners were taken as guinea pigs and im-
mersed in water to see, first, if they drowned and, second, if their flesh would
eventually rot. This experiment was as crucial for the Amerindians as the Valla-
dolid dispute was for the Iberians. If the conquerors drowned and rotted, then
the question was settled; they had bodies. But if they did not drown and rot, then
the conquerors had to be purely spiritual entities, perhaps similar to shamans. As
Claude Lévi-Strauss summarized, somewhat ironically, the two experiments, the
Spaniards’ versus the Amerindians’: “the whites were invoking the social sciences
while the Indians had more confidence in the natural ones.”
The relevance of this anecdote should be apparent: at no point in the Valla-
dolid controversy did the protagonists consider, even in passing, that the con-
frontation of European Christians and Amerindian animists might be framed dif-
ferently from the way in which Christian clerics understood it in the sixteenth
century. At no point were the Amerindians asked what issue they took to be in
dispute, nor is Beck asking now. But asking that question is only the first step
en route to adequate complexity. Was every European in agreement with every
other? Were there not two (at least) solutions to the problem raised at Valladolid?
Indians had souls like Christians or else Indians did not — each position had its
partisans. Beck supposes there were only two solutions to the problem posed at
Valladolid about Indian souls (they have souls, they do not have souls) and he
ignores the other problem raised in South America about conquistador bodies
(they have bodies, they do not have bodies). A negotiation between Europeans
and Amerindians would thus be, at a minimum, four-sided. Bartolomé de Las
Casas, the Dominican priest, held that Europeans and Amerindians were basi-
cally the same, and he complained of the un-Christian cruelty of Christians
against their “Indian brothers.” But how would he have responded, how might
his views have modified, had he witnessed the systematic drowning of his fellow
Spaniards in a scientific experiment designed to assay their exact degree of
bodily presence? Which “side” would Las Casas, after the experience, be on? As

Published by Duke University Press

 Common Knowledge

Viveiros de Castro has persuasively shown, the question of “the other,” so central

453
to recent theory and scholarship, has been framed with inadequate sophistica-
tion. There are more ways to be other, and vastly more others, than the most tol-

Talking Peace with Gods: Part 1

erant soul alive can conceive.
How cosmopolitan is a negotiator who mediates on behalf of one or two of
the four (or more) sides to a dispute? It is hazardous, and perhaps ethnocentric
as well, to assume that enemies agree on baseline principles (the principle, for
instance, that all humans have bodies).4 I say that Beck’s stance may be ethno-
centric because his cosmopolitanism is a gentler case of European philosophi-
cal internationalism. Beck takes his key term and its definition, off the shelf, from
the Stoics and Kant. Those definitions (Beck’s, Kant’s, the Stoics’) are problem-
atic: none shows understanding that more than culture is put in jeopardy by
conflicts. The cosmos too may be at stake. Like most sociologists, Beck suffers

Latour •
from anthropology blindness. For the sociologist, nature, the world, the cosmos,
is simply there; and since humans share basic characteristics, our view of the
world is, at baseline, the same everywhere. Perversity, acquisitiveness, undisci-
plined instincts account for the fact that we do not —we rarely have — peace.
When Beck writes that Las Casas was in a state of denial over the extent of the
struggle he wished devoutly to end, Beck without realizing it is speaking of him-
self as well. Beck and Las Casas are good guys, but good intentions do not resolve
or prevent strife. I am not saying of course that Beck’s cosmopolitanism is sim-
ply a larger version of Jürgen Habermas’s humanism. The entrance to the debat-
ing room is not limited, for Beck, to rational agents able to pursue a reasonable
conversation. Beck is ready to deal with much wider conflicts. What he does not
realize, however, is that whenever cosmopolitanism has been tried out, from
Alexandria to the United Nations, it has been during the great periods of com-
plete confidence in the ability of reason and, later, science to know the one cos-
mos whose existence and solid certainty could then prop up all efforts to build
the world metropolis of which we are all too happy to be citizens. The problem
we face now is that it’s precisely this “one cosmos,” what I call mononaturalism,
that has disappeared. It’s impossible for us now to inherit the beautiful idea of
cosmopolitanism since what we lack is just what our prestigious ancestors pos-
sessed: a cosmos. Hence we have to choose, in my view, between cosmopoli-
tanism and cosmopolitics.

4. There is of course a great difference between the “cos- the cosmos. See, for instance, Étienne Tassin, Un monde
mopolitan” project of an international civil society and commun: Pour une cosmo-politique des conflits (Paris: Seuil,
what I am discussing here. The difference, as was made 2003), and Daniele Archibugi, ed., Debating Cosmopolitics,
clear in a meeting organized by Ulrich Beck at the Lon- New Left Review Debates (London: Verso, 2003). I thank
don School of Economics in February 2004, is the weight the participants at the meeting for useful insights into
given to the word cosmos. “Citizens of the world” are cos- these rather incommensurable political philosophies.
mopolitan, to be sure, but that does not mean that they
have even begun to fathom the difficulties of a politics of

Published by Duke University Press

 Common Knowledge

One way to present this argument is to contrast Beck’s use of the term cos-
454

mopolitan with that of Isabelle Stengers in her multivolume masterpiece Cos-

mopolitique.5 A Stoic or a Kantian will call cosmopolitan anyone who is a “citizen
COMMON KNOWLEDGE

of the cosmos” rather than (or before he or she is) a citizen of a particular state,
an adherent of a particular religion, a member of a particular guild, profession,
or family. Stengers intends her use of cosmopolitics to alter what it means “to
belong” or “to pertain.” She has reinvented the word by representing it as a com-
posite of the strongest meaning of cosmos and the strongest meaning of politics pre-
cisely because the usual meaning of the word cosmopolite supposed a certain the-
ory of science that is now disputed.6 For her, the strength of one element checks
any dulling in the strength of the other. The presence of cosmos in cosmopolitics
resists the tendency of politics to mean the give-and-take in an exclusive human
club. The presence of politics in cosmopolitics resists the tendency of cosmos to mean
a finite list of entities that must be taken into account. Cosmos protects against the
premature closure of politics, and politics against the premature closure of cosmos.
For the Stoics, cosmopolitanism was a proof of tolerance; cosmopolitics, in
Stengers’s definition, is a cure for what she calls “the malady of tolerance.”7
The contrast between Stengers’s understanding of cosmos and Beck’s could
not be more stark. For Beck, the word means culture, worldview, any horizon
wider than that of a nation-state. He assumes that issues of war and peace involve
only humans, each endowed with the same psychology, each knowing a language
translatable into every other language, and each possessed of only slightly con-
tradictory representations of what-there-is. For Beck, as for most sociologists and
all political scientists, wars rage because human cultures have (and defend) dif-
fering views of the same world. If those views could be reconciled or shown to dif-
fer only superficially, peace would follow automatically. This way of understand-
ing cosmos and cosmopolitics is limited in that it puts a limit to the number of entities
on the negotiating table. But if cosmos is to mean anything, it must embrace, lit-
erally, everything—including all the vast numbers of nonhuman entities making
humans act. William James’s synonym for cosmos was pluriverse, a coinage that
makes its awesome multiplicity clear. In the face of which, Beck’s calm, untaxed
coherence would be unjustifiable except that he presumes that almost all the com-
plexities of the pluriverse are out of bounds for politics, or, at any rate, that they
make no difference in the way we agree or disagree with one another. Peace, for

5. Isabelle Stengers, Cosmopolitiques, vol. 1, La guerre des sci- University Press, 1985). The main weakness of legal and
ences (Paris: La Découverte; Les Empêcheurs de penser en humanitarian forms of cosmopolitanism is to forget
rond, 1996). entirely the theory of science that has been surreptitiously
used to assemble the cosmos in a peaceful manner but
6. For Stengers as well as for me, the ability to imagine a
without due process.
political order is always directly predicated on a certain
definition of science. The link was made forcefully by sci- 7. Stengers, Cosmopolitiques, vol. 7, Pour en finir avec la
ence studies. See the now classic work of Steven Shapin tolérance (Paris: La Découverte; Les Empêcheurs de penser
and Simon Schaffer, Leviathan and the Air-Pump: Hobbes, en rond, 1997).
Boyle, and the Experimental Life (Princeton, NJ: Princeton

Published by Duke University Press

 Common Knowledge

Beck’s kind of mononaturalism, is possible because our disputes (to borrow the

455
Scholastic terms) bear upon secondary, rather than primary, qualities.
If this be peace, I must say I prefer war. By war I mean a conflict for which

Talking Peace with Gods: Part 1

there is no agreed-upon arbiter, a conflict in which what is at stake is precisely
what is common in the common world to be built. As is well known from Carl
Schmitt’s definition, any conflict, no matter how bitter, that is waged under a
common arbiter is not a war but what he calls a “police operation.” If there exists
one cosmos, already unified, one nature that is used as the arbiter for all our dis-
putes, then there are, by definition, no wars but only police operations. To use
Schmitt again: Westerners have not understood themselves as facing on the bat-
tlefield an enemy whose victory is possible, just irrational people who have to
be corrected. As I have argued elsewhere, Westerners have until now been
engaged in pedagogical wars.8 But things have changed of late and our wars are

Latour •
now wars of the worlds, because it’s now the makeup of the cosmos that is at
stake. Nothing is off limits, off the table, for dispute. The Amerindians, it is
worth recalling, did not rejoice when their side — those Europeans who thought
them fully human — triumphed at Valladolid. Yet those ungrateful Indians re-
ceived the gift of a soul that permitted their baptism and thus salvation. Why did
the European offer of both peace and eternal life not please them? The Europeans
took European cosmopolitics for granted as natural, and nature is a means to
shortcut politics before peace is authentically attainable.9 The settlements that
nature offers are reached without due process—they put 99 percent of what is up
for grabs off limits, and the result is always another round of conflict. Politics is
moot if it is not about (what John Tresch calls) “cosmograms.”10 We perhaps never
differ about opinions, but rather always about things — about what world we
inhabit. And very probably, it never happens that adversaries come to agree on
opinions: they begin, rather, to inhabit a different world.
A common world is not something we come to recognize, as though it had
always been here (and we had not until now noticed it). A common world, if there
is going to be one, is something we will have to build, tooth and nail, together.
The ethnocentrism of sociologists is never clearer than when they paper over the
threat of multiple worlds with their weak notion of cosmopolitanism.11 In Beck’s

8. This point is developed at some length in Bruno Latour, 10. John Tresch, “Mechanical Romanticism: Engineers of
War of the Worlds: What about Peace? trans. Charlotte Bigg the Artificial Paradise” (Ph.D. diss., University of Cam-
(Chicago: Prickly Paradigm, 2002). Beck’s position in the bridge, 2001).
paper I am discussing is all the more strange since he has
11. To counter that ethnocentrism is why Viveiros de Cas-
never tired, in his other books, of showing why science and
tro (see n. 2 herein) launched his monster term multinatu-
technology can no longer deliver the sort of mastery on
ralism. The difference between multinaturalism and mul-
which a quiet and sane political reason could then be
ticulturalism is that truth and reality are engaged in the
propped up. For some strange reason, he seems to have
former and never in the latter.
forgotten his own lessons.

9. Latour, Politics of Nature: How to Bring the Sciences into

Democracy, trans. Catherine Porter (Cambridge: Harvard
University Press, 2004).

Published by Duke University Press

 Common Knowledge

article here, the telltale symptom is his offhand footnote about religion. There
456

have been good historical reasons — as Olivier Christin has shown, with respect
to sixteenth-century warfare in Europe — to sequester religion when peace pro-
COMMON KNOWLEDGE

posals are advanced.12 But it is not clear that what was right four and a half cen-
turies ago is a diplomatic and respectful way to handle our newly generated wars
of religion. When men of good will assemble with their cigars in the Habermas
Club to discuss an armistice for this or that conflict and they leave their gods on
hooks in the cloakroom, I suspect that what is under way is not a peace confer-
ence at all. There are Versailles that beget Munichs that beget Apocalypse. How
is it that Beck believes religion is ignorable? Again, there is no cosmos in his cos-
mopolitanism: he seems to have no inkling that humans have always counted less
than the vast population of divinities and lesser transcendental entities that give
us life. For most people, in most places, during most eons, humans have “own-
ers,” to use Tobie Nathan’s term; and those proprietors take precedence over
humans at whatever cost. Beck appears to believe in a UNESCO koine, a socio-
logical Esperanto, that lies hidden behind stubborn defects, whether social or
psychological, in our representations. Men of good will, he would say, must agree
that gods are no more than representations. It would be pretty to think so; but,
as Nathan makes vivid in his contribution to this symposium, it is not humans
who are at war but gods. Or, at least, we should entertain the possibility that ene-
mies can be separated by disagreements that wide. Peace settlements are not, as
Stengers emphasizes, between men of good will who have left their gods (their
narrow attachments) behind but between men of ill will possessed by super- and
subhumans of ill will. A settlement reached too soon is, realistically, a grave dan-
ger. Stengers might even add that the longer it takes to reach agreement, the bet-
ter (hence her daring to utter the phrase “malady of tolerance”). Even Cromwell,
after all, while promoting the most horrendous acts of iconoclasm, had the scru-
ple to exclaim (in a sentence Stengers never tires of quoting): “I beseech ye, my
brethren, in the bowels of Christ, think that ye may be mistaken!” Where, in
Beck’s view, is the place for such a tragic warning? His cosmopolitics is much too
cosmopolite to handle the horrors of our time.
But perhaps the problem is simply that Beck explores only one dimension
of peacemaking — the traditional gradient that leads from particular to universal
and back again. Approaching peace in this way is a venerable Stoic tradition; and,
as its latest expositor, Beck develops a nice compromise between the cheap uni-
versalisms available now and some even cheaper varieties of relativism and mul-
ticulturalism. Still, the traditional gradient and the Stoic view of it are not all
there is to achieving peace: attachments are not defined solely by their expansion

12. Olivier Christin, La Paix de religion: L’autonomisation de

la raison politique au XVIe siècle (Paris: Seuil, 1997).

Published by Duke University Press

 Common Knowledge

and shrinkage along a line between universalism and particularism. We should

457
perhaps explore another dimension, perpendicular to the first — a gradient run-
ning from “naturalism” to “constructivism.” If it’s true that the traditional mean-

Talking Peace with Gods: Part 1

ing of cosmopolitanism was based on a certain definition of science, it makes
sense to see how it has been modified when another definition of science is called
for. Cosmopolitans may dream of the time when citizens of the world come to
recognize that they all inhabit the same world, but cosmopolitics are up against
a somewhat more daunting task: to see how this “same world” can be slowly com-
posed.
Confronted by a given cosmogram, the Stoic wants to know whether it is
expressive more of attachment or of detachment (whether it is local or univer-
sal in character). But a more urgent and polemical question to ask enemies might
be: “How do you differentiate between good and bad attachments?” For the

Latour •
Stoic, detachment is emancipation (and attachment is slavery). By definition a
citizen of the cosmos is free; an Egyptian, Greek, or Jew is attached (enslaved) to
his or her locale and local knowledge. To be Egyptian, Greek, or Jewish is — in
the Stoic tradition — a stigma. “Humanity” was a great and welcome discovery
and has been a great and welcome rediscovery each time that (after World War
II, notably) it has come to prominence. And yet, if all the United Nations mem-
bers were satisfied to be “just humans,” if the UNESCO lingua franca was
enough to define all inhabitants of the planet, peace would already reign. Since
there is no peace, there must be something wrong with this humanistic definition
of an emancipated human as the only acceptable member of the Club.
We face a situation in which, on the one hand, real peace is unattainable
if negotiators leave their gods, attachments, and incompatible cosmos outside the
conference room. On the other hand, a freight of gods, attachments, and unruly
cosmos make it hard to get through the door into any common space. Moreover,
humans with owners, attachments, and a cosmos (crammed with entities ignored
or ridiculed by other humans) tend not to seek new memberships in clubs. They
have reason to believe that they themselves belong to the best clubs already and
cannot fathom why others —when invited — have refused to join.13 Hence the
need for a second dimension to peacemaking, one that does not require detach-
ment from the beings (for instance, divinities) that make us exist. This second
dimension requires another protocol, another investigation, to answer another
question: Through what sort of test do you render possible the distinction between good

13. This attitude pertains, according to Philippe Descola According to Viveiros de Castro (in comments on my
and Viveiros de Castro, to all animists — as opposed to paper), Amazonian Indians are already globalized in the
totemists, naturalists, and analogists. (I am using here the sense that they have no difficulty in integrating “us” into
vocabulary developed by Descola in his essay “Construct- “their” cosmologies. It is simply that in their cosmic pol-
ing Natures: Symbolic Ecology and Social Practice,” in itics we do not have the place that “we” think we deserve;
Nature and Society: Anthropological Perspectives, ed. Descola it is not the case that “we” are global and “they” are local.
and Gísli Pálsson [London: Routledge, 1996], 82 –102.)

Published by Duke University Press

 Common Knowledge

and bad attachments? Making those distinctions necessitates, first, the abandon-
458

ment of naturalism, which is the faith in a single natural world, comprehensible

through Science — or rather, through a mistaken definition of (Western) natu-
COMMON KNOWLEDGE

ral science whose purpose has been to eliminate entities from the pluriverse. The
universal embrace of naturalism has been, for moderns, the royal road to peace.
And yet naturalism has also been the grounds on which the West has waged its
pedagogical wars. The modern West scolds the remainder of mankind: we all live
under the same biological and physical laws and have the same fundamental biological,
social, and psychological makeup. This, you have not understood because you are prison-
ers of your superficial worldviews, which are but representations of the reality to which
we, through science, have privileged access. But science is not our property; it belongs to
mankind universally! Here, partake—and with us you will be one. The problem with
this opening gambit into diplomacy is not, I rush to add, that the argument is
wrong.14 The argument is right, but it puts the cart before the horse; it begins
where it should, eventually (very eventually), end. It is possible — and from a
Western (from my Burgundian) point of view, desirable — that, in the distant
future, we come to live within a common world defined as naturalism defines it.
But to behave as if the settlement were already in place and as though it requires
no negotiation to achieve it is a sure trigger to further warfare.
The assumptions of naturalism have been shown—most recently and thor-
oughly by Philippe Descola — to be unshared by vast numbers of humans.15 But
constructivism, despite its reputation as a radical postmodern ideology, might be
more universalizable since the distinction between what is well and badly made
is more than widespread. Constructivism is a tricky word, no doubt.16 But we can
likely agree that constructivists tend to share these principles:

• the realities to which humans are attached are dependent on series of

mediations;
• those realities and their mediations are composed of heterogeneous ingre-
dients and have histories;

14. There is a large difference between “the sciences” 15. See, e.g., Descola, La Nature domestique: Symbolisme et
understood as the proliferation of entities with which to praxis dans l’écologie des Achuar (Paris: Editions de la Mai-
build the collective and “Science” as a way to eliminate son des Sciences de l’Homme, 1986), translated into En-
secondary qualities because of the postulation of primary glish by Nora Scott as In the Society of Nature: A Native
ones. Each requires a different politics. On this distinction, Ecology in Amazonia (Cambridge: Cambridge University
see my Pandora’s Hope: Essays on the Reality of Science Stud- Press, 1994).
ies (Cambridge: Harvard University Press, 1999). Con-
16. See more on this subject in Latour, “The Promises of
trary to what naturalists believe, there is little in the sci-
Constructivism,” in Chasing Technoscience: Matrix for Mate-
ences that authorizes scientists to be eliminativists. This
riality, ed. Don Ihde and Evan Selinger (Bloomington:
point is made especially well by Stengers in the brilliant
Indiana University Press, 2003), 27 – 46.
book Penser avec Whitehead: Une libre et sauvage création de
concepts (Paris: Seuil, 2002).

Published by Duke University Press

 Common Knowledge

the amount of heterogeneous ingredients and the number of mediations

459
•

necessary to sustain realities are a credit to their reality (the more medi-
ated, the more real);

Talking Peace with Gods: Part 1

• our realities are open to differing interpretations that must be considered
with caution;
• if a reality has extension (in space and time), its complex life-support sys-
tems must have been extended also;
• realities can fail and thus require careful maintenance and constant repair.

The paramount example of constructivism, on this definition, is the work

of the sciences, as I have shown numerous times.17 In the sciences, the degree
of objectivity and certainty is directly proportional to the extent of artificiality,
layering, heterogeneity, multiplicity, and complexity of mediations. The asser-

Latour •
tion sounds radical but is merely obvious: in a laboratory, no naked access to truth
is thinkable. Is a microbe visible without the mediation of a microscope? Are
microscopes found in nature or are they human fabrications? When one scien-
tist questions another, it is not to ask whether new data (new facts) have been fab-
ricated or not. The question is, “How have you proven that x is so?”— and the
emphasis is on how, by what means or mediation. The difference that counts,
when scientists meet in confidence, is not the one between fact and construction
but the one between good and bad facts. Things are different when scientists
meet the vulgum pecus: then (and only then) do they borrow arguments from the
epistemologists and join with philosophers in playing pedagogical war games,
where the Red relativists have to fight the White Knights of realism.
The example of the sciences should make clear that constructed and real are
not opposed terms and that the operative question is how to distinguish between
good constructions and bad. If these two generalizations apply to the sciences, to
which epistemologists have assigned so unique and transcendent a role, then
surely my generalizations apply to the harsh realities of cosmopolitics. And it
appears, indeed, that there is no extant or extinct way of life that has not been
passionately involved in making distinctions between good and bad fabrications.
This observation has its uses in pursuit of peace. A constructivist landing at the
mouth of the Amazon (where Amerindians, recall, were drowning Spaniards to
test if they had bodies) could have entered into conversations unimaginable to

17. In Common Knowledge alone, I have defended variations 1996): 76 – 91; “Trains of Thought: Piaget, Formalism,
on this argument in four articles and two dialogues: “On and the Fifth Dimension,” Common Knowledge 6.3 (winter
Technical Mediation — Philosophy, Sociology, Geneal- 1997): 170 – 91; “Two Writers Face One Turing Test: A
ogy,” Common Knowledge 3.2 (fall 1994): 29 – 64; “The Dialogue in Honor of HAL” (with Richard Powers), Com-
‘Pédofil’ of Boa Vista: A Photo-Philosophical Montage,” mon Knowledge 7.1 (spring 1998): 177 – 91; and “The Sci-
Common Knowledge 4.1 (spring 1995): 144 – 87; “Do Sci- ence Wars — A Dialogue,” Common Knowledge 8.1 (winter
entific Objects Have a History? Pasteur and Whitehead in 2002): 71– 79.
a Bath of Lactic Acid,” Common Knowledge 5.1 (spring

Published by Duke University Press

 Common Knowledge

the clerics at Valladolid. “Ah,” a Spanish constructivist could have said to an Ama-
460

zonian researcher, “this is how you decide matters? How horrifying. Let me sug-
gest another way to formulate and test the question at issue.” Of course, this
COMMON KNOWLEDGE

peaceful encounter is fantastic but not for the reason that a naturalist would pre-
sume. The encounter would not happen as I have described it only because of the
fundamentalism of the parties involved.
Fundamentalism is at the far end of our options from constructivism. A
fundamentalist — in science, politics, or religion —would review the list I gave
above and indignantly invert every assertion on it:

• the realities to which humans are attached are, quite simply, unmediated;
• those realities are unitary (not composed of heterogeneous ingredients)
and have no history per se;
• the amount of heterogeneous ingredients and the number of mediations
necessary to sustain a state of affairs must be debited from its degree of
reality (the less mediation, the more real);
• realities are not open to interpretation;
• realities of course have extension (in space and time) — they are by nature
universal and it is absurd to say that they require “life support”;
• realities are not susceptible to failure and thus do not require maintenance
or repair of any kind.

Realities of this description are invoked in the pages of the Wall Street Journal
as much as in the dark caves of Pakistan, and they can even come from human-
rights activists: anyone who holds that fabricated means untrue, and made means
fake, tends toward fundamentalism. Common experience in science, art, love, and
religion should prompt us to say, “the more carefully fabricated, the more real
and long-lasting.” Why we do not tend to reach that more sensible conclusion
is perhaps our old and continuing fear of idolatry, of worshiping what human
hands have made; a genuine acceptance of constructivism requires a reassessment
of the whole history of iconoclasm and critique.18
The delightful irony of the matter is that, while fundamentalism was home-
made, made in the West — the former West — it has by now become la chose du
monde la mieux distribuée. As Peter Sloterdijk has remarked, Westerners loved
globalization until les autres could reach us as easily as we could reach them. Nat-
uralizers, those in the West who appeal to a Nature Out There, unconstructed
and nonnegotiable, are now confronted by people saying the same of the Koran

18. For a systematic exposition, see Latour and Peter

Weibel, eds., Iconoclash: Beyond the Image Wars in Science,
Religion, and Art (Karlsruhe, Germany: ZKM; Cambridge:
MIT Press, 2002).

Published by Duke University Press

 Common Knowledge

and Shari’a. And when one fundamentalism butts heads with another, no peace

461
talks are possible because there is nothing to discuss: pedagogical wars are waged
to the bitter end. However, it is not the case (pace President Bush) that the pres-

Talking Peace with Gods: Part 1

ent world war is between a modern culture and an archaic one. The supposed
enemies of modernization are themselves modernizers in the extreme, using con-
ceptual tools provided them by Western fundamentalism.19 My main objection,
then, to the peace terms of Ulrich Beck is that he has not put the West’s own
native fundamentalism up for discussion. Our naturalism has failed: it was a war
plan disguised as a peace plan, and those against whom we directed it are no
longer fooled. Naturalism, like any fundamentalist ideology, amounts to a prej-
udice against fabrication. Constructivism is not, cannot possibly be, deconstruc-
tive—quite the contrary—though the former has been regularly said to be a sub-
set of the latter. Constructivism is the attitude of those who make things and are

Latour •
capable of telling good from bad fabrications, who want to compare their goods
with those of others so that the standards of their own products improve. But for
this diplomatic task to begin, a state of war has first to be declared. And once war
is declared, we can seek peace on firmer grounds: where naturalism has failed,
why not try constructivism?
The misunderstanding between Beck and myself may stem, in the end,
from differing interpretations of the present historical situation. The “first mod-
ernization,” to use his favorite expression, came with a certain definition of cos-
mopolitanism, which corresponded to the great idea that the whole earth could
actually fit snugly inside what Sloterdijk has called the “metaphysical Globe” (as
imagined by Mercator, Galileo, Descartes, Leibniz, and of course Hegel).20 The
problem is that when this version of the global was invented the world was just
beginning to be “globalized.” The Globe on which Hegel could rely to house
every event “in it” was a purely conceptual one; it was for this reason perfect, with
neither shadow nor gap. Now the planet is indeed being slowly and mercilessly
globalized, but there is no “global” anymore, no metaphysical Globe to offer
advance welcome to inhabitants and give them their rightful and predetermined
places. So the two meanings of cosmopolitanism are, so to speak, out of sync: just
when we need the Global, it has sunk deep down in the Atlantic, beyond repair.
Therefore, in my view, another definition of cosmopolitics is called for, one that

19. Compare my assumption here that Islamic fundamen- “spherology,” the questions are raised in terms of “air con-
talism is a form of modernization with, e.g., Khaled Abou ditioning” and “life support.” In his three-volume inquiry
El Fadl, Speaking in God’s Name: Islamic Law, Authority, and into the shape of spheres, he shows how and why the
Women (Oxford: Oneworld Publications, 2001), and Em- Global has existed in the past but is no longer a structure in
manuel Todd, Après l’Empire: Essai sur la décomposition du which we can freely breathe. For an excellent overview of
système américain (Paris: Gallimard, 2002). his philosophy, see Peter Sloterdijk, Ni le soleil ni la mort:
Jeu de piste sous forme de dialogues avec Hans-Jürgen Hein-
20. There is a direct connection between Beck’s cosmo-
richs (Paris: Pauvert, 2003).
politan interest and those of the great German philoso-
pher Peter Sloterdijk. The difference is that, in the latter’s

Published by Duke University Press

 Common Knowledge

does not rely on the “first modernity” dream of an already existing common
462

Sphere. It would be a tragic mistake to pursue peace by dragging in the defunct

Globe as a locus for the common world of cosmopolitanism. Since Ulrich Beck
COMMON KNOWLEDGE

does not wish to be Munich’s Hegel, he knows fairly well that the parliament in
which a common world could be assembled has got to be constructed from
scratch.

Published by Duke University Press