omnymai) as “to seize with force the sacralizing object.” Horkos designates, then,
“not a word or an act, but a thing, a material invested with evil potency, which
confers to the commitment its obliging power” (Benveniste [1], 85–86). Horkos
is the “sacred Substance” (90), which is variously embodied in the waters of the
Styx, the scepter of the hero, or the entrails of the sacrificial victim. Following
Benveniste’s path, a great historian of Greek law, Louis Gernet, evokes in almost
the same terms the “sacred substance” with which the one who utters the oath
is put in contact (Gernet [1], 270/223: “To swear, therefore, is to enter the realm
of religious forces of the most fearsome sort.”).
In the human sciences, beginning at the end of the nineteenth century, the
idea that explaining a historical institution necessarily means tracing it back to
an origin or context that is sacred or magico-religious is so strong that when Jean
Bollack writes his article “Styx et serments” (1958) to demonstrate, contra Ben-
veniste, that the term horkos acquires its true sense only if one traces it back to its etymological origin from herkos, he does not notice that he basically maintains
the essential traits of the argument he intends to oppose:
The oath places the one who swears, through the magical force of words, in a
special relationship with the objects invoked and with the world. . . . Many of
the invoked objects, like the hearth, belong to a sacred domain. But in a broadly
sacralized universe, every object called as witness can be transformed from a
guarantor and preserver into a terrifying potency. This special relationship that
ties man to the objects invoked seems to be defined by the term horkos, which
310
HOMO SACER II, 3
designates not, as Benveniste thinks, the object on which the oath is pronounced,
but the enclosure with which it surrounds the one who swears. (Bollack, 30–31)
The sacrality is here displaced from the object to the relation, but the explanation
remains unchanged. According to an endlessly repeated paradigm, the force and
efficacy of the oath are once again sought in the sphere of the magico- religious
“forces” to which it originally belongs and which is presupposed as most archaic:
they derive from this and decline along with the decline of religious faith. Here
one presupposes the existence of a homo religiosus prior to man as we know him
historically. Yet this homo religiosus exists only in the imagination of scholars, because all the sources we have at our disposal always present to us, as we have
seen, a man who is religious and also irreligious, faithful to the oath and also
more than capable of perjury. It is this presupposition of every analysis of this
institution that I intend to call into question.
א Benveniste’s thesis concerning the horkos as a “sacred substance” derives, as the author himself suggests, from an article by Elias Bickermann, a scholar of classical antiquity who was also an excellent historian of Judaism and Christianity. The article in question,
published in the Revue des études juives in 1935, refers to the oath only under the heading of a methodological example, in the context of a critique of Gerardus van der Leeuw’s
Phenomenology of Religion, which had appeared two years earlier. The methodological principles that Bickermann lays out seem to have had a notable influence on Benveniste,
even if they actually reflect a common cultural formation. (Bickermann, who from 1933
had taught in Paris at the École pratique des hautes études and until 1942, when he was
constrained by his Hebraic origin to seek refuge in the United States—where his name
would become Bickerman—had been chargé des recherches at the Centre national de la recherche scientifique, refers explicitly to the method of Antoine Meillet, who had been Benveniste’s teacher.) The fact remains that the four methodological principles recommended
by Bickermann (abandoning the recourse to psychology to explain religious phenomena;
breaking down facts into their constitutive elements or “roots”; analyzing the function of single elements in isolation; studying their function in the phenomenon in question) are
found precisely in Benveniste. Once again, however, such an adroit scholar, examining the
oath in a note in order to exemplify his method, repeats uncritically the paradigm of the
primordiality of the sacred, which Benveniste will take up again in almost the same words:
“always and everywhere, the idea is to establish a relationship between an affirmation and something sacred. . . . The goal remains the same everywhere, viz. to establish a relationship between the affirmation and the sacred Substance” (Bickermann, 220–21/888–89).
7. I have demonstrated elsewhere (Agamben, 79–89/49–51), while discussing
the supposed ambivalence of the term sacer, the inadequacies and contradictions
connected with the doctrine of the “sacred” elaborated in the science and history
of religions between the end of the nineteenth century and the first decades of the
THE SACRAMENT OF LANGUAGE
311
twentieth century. It suffices here to recall that the defining moment in the estab-
lishment of this “scientific mythologeme,” which has negatively conditioned the
investigations of the human sciences in a particularly delicate area, is the encoun-
ter of the Latin notion of sacer with that of mana, which an Anglican missionary, Robert Henry Codrington, described in his work on the peoples of Melanesia.
Already fourteen years earlier Codrington had communicated his discovery in a
letter to Max Müller, who made use of it in the Hibbert Lectures, where the con-
cept of mana became the way in which “the idea of the infinite, of the unseen, or as we call it afterwards, the Divine, may exist among the lowest tribes in a vague
and hazy form” (Müller, 51). In the following years the notion reappeared under
various names in ethnographic studies on the American Indians ( orenda among
the Iroquois, manitou among the Algonquians, wakan among the Dakotas) until Robert Marett, in his Threshold of Religion (1909), made this invisible “force” the central category of religious experience. Despite the flimsiness of the theories
of religion of authors like Müller (who exercised a veritable dictatorship over
this nascent “science”—or, rather, as he preferred to call it—“history” of reli-
gions) and Marett, to whom we owe the notion of animism (another scientific
mythologeme that refuses to die), the idea of a “sacred power or substance,” as
terrible as it is ambivalent, vague, and indeterminate, that would be the funda-
mental category of the religious phenomenon, has exercised its influence not
only on Durkheim, Freud, Rudolf, Otto, and Mauss but also on the masterpiece
of twentieth-century linguistics that is the Vocabulaire of Benveniste.
It was necessary to wait for Lévi-Strauss’s essay of 1950 for the problem of the
meaning of terms like mana to be put on entirely new footings. In a memorable
passage Lévi-Strauss brought together these terms with common French expres-
sions like truc (thingamajig) or machin (thingamabob), which we use to designate an unknown object or one whose use we cannot explain. Mana, orenda, and
manitou do not designate something like a sacred substance or social sentiments
related to religion but a void of sense or an indeterminate value of signification,
which holds first
of all for the very scholars who make use of it: “But always and
everywhere, those types of notions, somewhat like algebraic symbols, occur to
represent an indeterminate value of signification, in itself devoid of meaning and
thus susceptible of receiving any meaning at all; their sole function is to fill a gap
between the signifier and the signified, or, more exactly, to signal the fact that
in such a circumstance, on such an occasion, or in such a one of their manifes-
tations, a relationship of non-equivalence becomes established between signifier
and signified” (Lévi-Strauss, xliv/55–56). If there is a place, adds Lévi-Strauss, in
which the notion of mana truly presents the characteristics of a mysterious or
312
HOMO SACER II, 3
secret power, it is above all in the thought of the scholars: “Mana really is mana there” (ibid., xlv/57). At the end of the nineteenth century, religion in Europe had
for all appearances become, at least for those who wanted to gather the history
and build the science of religion, something so strange and indecipherable that
they had to seek the key to it among primitive peoples rather than in their own
tradition: but the primitive peoples could only return as in a mirror the same ex-
travagant and contradictory image that these scholars had projected onto them.
א In the course of discussing the inevitable disconnection between signifier and sig-
nified, Lévi-Strauss again takes up and develops in a new way the theory of Max Müller,
who saw in mythology a sort of “disease” of consciousness caused by language. According
to Müller, the origin of mythological and religious concepts is to be found in the influence that language, in which paronyms, polysemy, and ambiguity of every kind are necessarily
present, exercises on thought. Mythology, he writes, “is in fact the dark shadow which
language throws upon thought, and which can never disappear till language becomes
entirely commensurate with thought, which it never will” (Cassirer, 4/5).
8. Another aspect of the scientific mythologeme that I have described (which
is in truth inseparable from it) is the idea that the sphere of sacredness and reli-
gion—often united to that of magic, so that one can then redouble the confu-
sion and speak of a “magico-religious” sphere—coincides with the most archaic
moment that historical research in the human sciences can prudently attempt to
recover. A simple textual analysis shows that this is an arbitrary presupposition,
set to work by the scholar at the point where he reaches, in his own sphere of
research, a documentary limit or threshold. It is as if the passage to what Franz
Overbeck called Urgeschichte and Dumézil called “fringe of ultra-history” nec-
essarily implied a blind leap into the magico-religious element, which is very
often nothing but the name that the scholar gives, more or less consciously,
to the terra incognita that lies beyond the area that the patient labor of histo-
rians is able to define. Taking the sphere of law as an example, it may be the
distinction between the religious sphere and the profane sphere, whose distinc-
tive characteristics appear to us, at least in the historical epoch, to be in some
measure defined. If he reaches in this area a more archaic stage, the scholar has
the impression that the boundaries become blurred, so he is led to hypothesize
a preceding stage, in which the religious sphere and the profane (and often
also the magical) are not yet distinct. In this sense Louis Gernet, working on
the most ancient Greek law, has designated as “prelaw” ( pré-droit) an originary
phase in which law and religion appear to be indiscernible. In the same sense
Paolo Prodi, in his political history of the oath, evokes a “primordial indistinc-
tion” in which the process of separation between religion and politics has not yet
THE SACRAMENT OF LANGUAGE
313
begun. It is essential, in cases like these, to have the wisdom not to simply and
uncritically project onto the supposed “primordial indistinction” the character-
istics that define the religious and profane spheres that are known to us and are,
precisely, the result of the patient labor of historians. Just as a chemical com-
pound has specific properties that cannot be reduced to the sum of the elements
that compose it, so also what stands before the historical division—granted that
something of the kind exists—is not necessarily the opaque and indistinct sum
of the characteristics that define its fragments. Prelaw cannot be merely a more
“archaic” law, just as what stands before religion as we know it historically is not
only a more primitive religion ( mana); it would, in fact, be advisable to bypass
the very terms religion and law and to try to imagine an x. To find the definition of this x, we must put forward every possible precaution, practicing a sort
of archaeological epoché that suspends, at least provisionally, the attribution of predicates with which we are used to defining religion and law.
What must be interrogated at this point is the threshold of indistinction
that the analysis of the researcher comes up against. It is not something that
should be incautiously projected onto chronology, like a prehistoric past for
which documents happen to be lacking, but an internal limit, the comprehen-
sion of which, by calling into question the accepted distinction, can lead to a
new definition of the phenomenon.
א The case of Mauss is a good example of how the presupposition of the sacral
system has a strong effect and yet can be neutralized, at least in part, by the exceptional attention to phenomena that defines his method. The 1902 Esquisse of a general theory of magic opens with an attempt to distinguish magical phenomena from religion, law,
and technology, with which they have often been confused. Yet Mauss’s analysis contin-
ually runs up against phenomena (for example the juridico-religious rites that include
an imprecation, like the devotio) that cannot be assigned to only one sphere. Mauss is then led to transform the dichotomous opposition religion-magic into a polar opposition, sketching out a field defined by the two extremes of sacrifice and evil spells, which necessarily presents thresholds of undecidability (Mauss, 14/27). It is on these thresholds that he focuses his labor. The result, as Dumézil has observed, is that for Mauss there are no longer magical facts on the one hand and religious facts on the other; rather, “one of
his principal concerns was to emphasize the complexity of each phenomenon and the
tendency of most of them to exceed all definition, to be situated simultaneously at various levels” (Dumézil [3], 49).
9. Let us now take the oath as it presents itself during the only epoch in which
we can analyze it, namely that for which we have documents. There it appears as
a juridical institution that includes elements that we are used to associating with
314
HOMO SACER II, 3
the religious sphere. To distinguish in it a more archaic phase in which it would
be only a religious rite, from a more modern one in which it belongs entirely to
law, is perfectly arbitrary. In reality, already in the most ancient documents in our
possession, such as the inscription on the vase of Dvenos (dated to the end of the
sixth century bce) from the Roman tradition, it appears as a promissory formula
&n
bsp; of an undoubtedly juridical character—in this specific case, the guarantee given
by the woman’s guardian to her (future) husband at the moment of the marriage
or engagement. Nevertheless, the formula, written in an archaic Latin, mentions
the gods, in fact swears by the gods: iovest deiuos quoi me mitat (“the one who
sends me”—it is the vase that speaks—“swears [by] the gods”) (Dumézil [3],
14–15). Here we have no need to presuppose as more ancient a purely religious
phase in the history of the oath, which no document in our possession attests as
such: in the most ancient sources that the Latin tradition permits us to reach, the
oath is a verbal act intended to guarantee the truth of a promise or an assertion,
which presents the same characteristics attested by the later sources and that we
have no reason to define as more or less religious, more or less juridical.
The same holds for the Greek tradition. The oath that the most ancient
sources present to us in a broad survey entails the testimony of the gods, the
presence of objects (the scepter, as the “great oath”— megas horkos—of Achilles
at the beginning of the Iliad, but also horses, the chariot, and the innards of the sacrificed animal), all of which elements we find again in the historical epoch for
oaths that certainly have a juridical nature (as in pacts between federated cities,
in which the oath is defined as “legal,” horkos nomimos; see Glotz, 749). And,
as we have seen, even the gods swear, invoking the waters of the Styx; and to
judge from what Hesiod tells us about the punishment for perjury committed
by a god, even the gods are subject to the authority of the oath. We possess,
moreover, Aristotle’s authoritative testimony informing us that the most ancient
philosophers, “who first speculated about the divine [ theologēsantas],” placed
among the first principles of the cosmos, alongside Ocean and Tethys, “the water
that serves as the gods’ oath, which they called Styx” ( Metaph. 983b32) and adds:
“For the assumption was that the most ancient thing [ presbytaton] was the most
worthy [ timiōtaton], and that the oath was the most worthy thing [ horkos de
The Omnibus Homo Sacer Page 46