oral rite, often completed by a manual rite whose form is variable. Its function
consists not in the affirmation that it produces, but in the relation that it institutes between the word pronounced and the potency invoked. (Benveniste [1], 81–82)
The oath does not concern the statement as such but the guarantee of its
efficacy: what is in question is not the semiotic or cognitive function of language
as such but the assurance of its truthfulness and its actualization.
3. All the sources and scholars seem to agree that the oath’s primary function,
in its various forms, is that of guaranteeing the truth and efficacy of language. As
Philo writes, “Now men have recourse to oaths to win belief, when others deem
304
HOMO SACER II, 3
them untrustworthy [ apistoumenoi, lacking in pistis, that is, in credibility]” (Philo of Alexandria [2], 93). And this function seems to be so necessary for human
society that, despite the clear prohibition of every form of oath in the Gospels
(Matthew 5:33–37; James 5:12), it was approved of and codified by the Church,
which made the oath an essential part of its own juridical order, legitimizing in
this way its maintenance and gradual expansion in the law and practice of the
Christian world. And when in De jure naturae et gentium [ Of the Law of Nature
and Nations] Samuel Pufendorf assembled the tradition of European law, it is
precisely in its capacity of guaranteeing and confirming not only pacts and agree-
ments among men, but also more generally language itself, that he establishes the
necessity and the legitimacy of the oath:
We proceed to examine and state the nature of an oath, which is judged to add
great strength and confirmation [ firmamentum] to our discourse and to all our
acts which have any dependence upon speech [ sermoni concipitur]; which though
we might have treated of very properly and conveniently hereafter, when we
come to explain the enforcements of pacts and covenants, yet we chose to assign
it this particular place rather than any other, because the custom of swearing is
used for the establishment and security not only of covenants, but of language
itself [ quod iureiurando non pacta solum, sed et simplex sermo soleat confirmari].
(Pufendorf, 326/333; trans. altered)
A few pages later, Pufendorf confirms the subsidiary character of the oath’s
bond, which, insofar as it confirms an assertion or promise, presupposes not
only language but, in the case of the promissory oath, the pronouncement of an
obligation: “oaths do not of themselves produce a new and peculiar obligation,
but are only applied as an additional bond [ velut accessorium quoddam vinculum]
to an obligation in its nature valid before” (ibid., 333/339).
The oath, then, seems to be a linguistic act intended to confirm a mean-
ingful proposition (a dictum), whose truth or effectiveness it guarantees. It is
this definition, which distinguishes between the oath and its semantic content,
whose correctness and implications we must verify.
א On the essentially verbal nature of the oath (even if it can be accompanied by
gestures, like raising one’s right hand) there is agreement among the majority of scholars, from Lévy-Bruhl to Benveniste, from Loraux to Torricelli. With regard to the nature of
the dictum, one is accustomed to distinguish between an assertative oath, which refers to a past fact (and hence confirms an assertion), and a promissory oath, which refers to
a future act (here a promise is confirmed). The distinction is already clearly enunciated
in Servius ( Aen. 12.816: Iuro tunc dici debere cum confirmamus aliquid aut promittimus
[“I swear” has thus been called necessary when we confirm something or make a promise]).
THE SACRAMENT OF LANGUAGE
305
Not wrongly, however, does Hobbes bring these two forms of the oath back to a single
type, essentially promissory: Neque obstat, quod iusiurandum non solum promissorium, sed aliquando affirmatorium dici possit: nam qui affirmationem iuramento confirmat, promittit se vera respondere [It is no objection that sometimes an Oath may be said to be not promissory but declarative. For in strengthening an affirmation by means of an oath, he
declares that he is giving a true reply] ( De cive 2.20/ On the Citizen, 41). The difference concerns, in fact, not the act of the oath, which is identical in the two cases, but the
semantic content of the dictum.
4. At the end of his reconstruction of the ideology of the three social functions
by means of an investigation of the epic poetry of the Indo-European peoples,
Georges Dumézil examines a group of texts (Celtic, Iranian, Vedic) in which
three evils or “scourges” ( fléaux) correspond to each of them. It is a matter, so to speak, of “functional scourges” of Indo-European society, each of which menaces
one of three fundamental categories or functions: priests, soldiers, and farmers
(in modern terms: religion, war, and economy). In one of the two Celtic texts he
examines, the scourge corresponding to the priestly function is defined as “the
dissolution of oral contracts,” that is, the repudiation or disavowal of obligations
one has assumed (Dumézil [1], 616). The Iranian and Vedic texts also evoke the
scourge in analogous terms: infidelity to the word one has given, falsehood or
error in ritual formulas.
One could think that the oath would be the remedy against this “Indo-
European scourge” that takes the form of the violation of the word one has
given and, more generally, the possibility of falsehood inherent in language. The
oath, however, proves singularly inadequate precisely for averting this scourge.
Nicole Loraux, in her chapter “Oath, Son of Discord” in The Divided City,
has lingered on a passage from Hesiod ( Theog. 231–32) in which the oath is
negatively defined solely by means of the possibility of perjury, “as if the first
had no other purpose than to punish the second and had only been created,
in the form of a terrible curse, for those oath breakers who were produced as
such by the oath itself ” (Loraux, 121–22/123). Already in the archaic epoch,
therefore, when the religious bond is supposed to have been stronger, the oath
seems to constitutively imply the possibility of perjury and to be paradoxically
intended—as Loraux suggests—not to impede falsehood but to combat per-
jury. However one understands the etymology of the Greek term for perjury
( epiorkos), about which scholars never stop debating, it is certain that in archaic and classical Greece it is taken for granted. Not only does Thucydides, describing the cities that have fallen prey to civil war, write that there is no longer any
“assurance binding enough, any oath terrible enough, to reconcile men” (3.83),
306
HOMO SACER II, 3
but the inclination of the Greeks (in particular of the Spartans) to perjury was
proverbial even in times of peace. Thus Plato advises against requiring oaths
of the parties to a trial because otherwise it would be revealed that half of the
citizens are perjurers ( Laws 12.948e). And it is significant that around the third century bce, the founders of the Stoa were discussing whether it was sufficient,
for there to be perjury, that the one who swears have, in the moment of uttering
the oath, the intention of not keeping it (this was the opinion of Cleanthes), or
&nb
sp; if it was necessary, as Chrysippus maintained, that he not in fact fulfill what he
had promised (Hirzel, 75; see also Plescia, 84). As a guarantee of an oral contract
or a promise, the oath appeared, according to all the evidence, from the very
beginning to be completely inadequate to the task, and a simple penalty for
lying would certainly have been more effective. The oath does not in any way
constitute a remedy against the “Indo-European scourge”; instead, the scourge
itself is contained within it in the form of perjury.
It is possible, then, not only that what was originally at issue in the oath was
the guarantee of a promise or of the truthfulness of an affirmation but that the
institution that we know today by that name contains the memory of a more
archaic stage, in which it was concerned with the very consistency of human
language and the very nature of humans as “speaking animals.” The “scourge”
that it had to stem was not only the unreliability of men, incapable of staying
true to their word, but a weakness pertaining to language itself, the capacity of
words themselves to refer to things and the ability of men to make profession of
their condition as speaking beings.
א The passage from Hesiod that Loraux refers to is in Theog. 231–32: “Oath [ Horkos], who indeed brings most woe upon human beings on the earth, whenever someone willfully swears a false oath.” Consistently in the Theogony (775–806), the waters of the Styx are described as “the great oath of the gods” ( theōn megan horkon) and even in this case, they function as “a great scourge for the gods ( mega pēma theoisin), for whoever of the immortals . . . swears a false oath after having poured a libation from her, lies breathless for one full year. . . . And when he has completed this sickness for a year, another, even worse trial follows upon this one: for nine years he is cut off from participation with the gods that always are, nor does he mingle with them in their assembly or their feasts for
all of nine years.”
The connection between oath and perjury seems, however, to be from the very
beginning so essential that the sources speak of a veritable “art of the oath”—in which,
according to Homer ( Od. 19.394), Autolycus excelled—which consisted in uttering oaths that, thanks to verbal tricks, could, if taken literally, signify something different from what the person to whom they were given could understand. It is in this sense that
one should understand the observation of Plato according to which Homer “praises
THE SACRAMENT OF LANGUAGE
307
Autolycus, Odysseus’ grandfather on his mother’s side, and says that ‘in swearing oaths
and thieving ( kleptosynēi th’ horkoi te) he surpassed all men’” ( Rep. 334b).
5. How should one understand the arché that is in question in an archaeo-
logical study like the one proposed here? Up to the first half of the twentieth
century, in the human sciences the paradigm of such a study had been elabo-
rated by linguistics and comparative grammar. The idea that it was possible to
go back, through a purely linguistic analysis, to more archaic stages of human
history was put forward toward the end of the nineteenth century by Hermann
Usener in his Götternamen (Names of the Gods). Asking, at the beginning of
his study, how the creation of divine names could have happened, he suggests
that in order to respond to such a question, we have no documents other than
those that arise for us out of an analysis of language (Usener, 5). But already be-
fore him comparative grammar had inspired the investigations of those scholars,
from Max Müller to Adalbert Kuhn and Émile Burnouf, who sought in the final
thirty years of the nineteenth century to found comparative mythology and the
science of religion. Just as the comparison of related linguistic forms allowed one
to go all the way back to stages of the language that were not historically attested
(those Indo-European forms, for example *deiwos or *med, that linguists are in the habit of indicating with a preceding asterisk in order to distinguish them
from words documented in historical languages), so also was it possible to go
back, through etymology and the analysis of meanings, to otherwise inaccessi-
ble stages of the history of social institutions.
It is in this sense that Dumézil was able to define his study as a work “not of a
philosopher, but of a historian of the oldest history and of the furthest fringe of
ultra-history [ de la plus vielle histoire et de la frange d’ultra-histoire] that one can reasonably seek to reach” (Dumézil [2], 14), declaring at the same time his debt
to the comparative grammar of the Indo-European languages.
The basis of the “fringe of ultra-history” that the historian here seeks to
reach, then, stands or falls with the existence of Indo-European and of the peo-
ple who spoke it. It exists in the same sense and to the same degree in which an
Indo-European form exists. Yet each of these forms, if we want to be rigorous, is
only an algorithm that expresses a system of correspondences among the exist-
ing forms in historical languages. In the words of Antoine Meillet, what we call
Indo-European is only “the ensemble of these systems of correspondences . . .
which presupposes a language x spoken by men x in a place x in a time x,” where x simply stands for “unknown” (Meillet, 324). Unless one wants to legitimate
the monstrum of a historical study that produces its own original documents,
308
HOMO SACER II, 3
one could never extrapolate from Indo-European any events that are supposed
to have happened historically. For this reason the method of Dumézil registered
a significant development with respect to the comparative mythology of the end
of the nineteenth century, when, around 1950, he acknowledged that the ideol-
ogy of the three functions (priests, soldiers, and shepherds or, in modern terms,
religion, war, and economy) “was not necessarily accompanied, in the life of a
society, by a real tripartite division of this society, on the Indian model” but that it rather represented precisely an “ideology,” something like “an ideal and, at
the same time, a way of analyzing and interpreting the forces that regulate the
course of the world and the life of men” (Dumézil [1], 15).
Similarly, when Benveniste published his Indo-European Language and Society
(1969), declaring in the preface that in his analyses “no extralinguistic presup-
positions have intruded” (Benveniste [2], 1:10/12), it was not completely clear
how the epistemological locus and the historical foundation of what he calls an
“Indo-European institution” should be understood.
It is the nature and foundation of the “oldest history” and the “fringe of
ultra-history” that an archaeology can hope to reach that we must define here to
the fullest possible extent. It is clear that the arché toward which an archaeology seeks to regress cannot be understood in any way as a given that can be situated either in a chronology (even in a broad category like “prehistoric”) or even
beyond it, in an atemporal metahistorical structure (for example, as Dumézil
ironically suggests, in the neuronal system of a hominid). It is, rather, a force
working in history, exactly as the Indo-European language expresses first of all a
system of connections among historically accessible languages; just as the child
in psychoanalysis expresses a force that continues to act in the psychic life of
the adult; and just as the “big bang,” which is supposed to have given rise to the
universe, is something that never stops transmitting its background radiation
to us. Yet unlike the “big bang,” which astrophysicists claim to be able to date,
even if only in terms of millions of years, the arché is not a given, a substance, or an event but a field of historical currents stretched between anthropogenesis
and the present, ultrahistory and history. And as such—that is, insofar as, like
anthropogenesis, it is something that is necessarily presupposed as having hap-
pened but that cannot be hypostatized into an event in a chronology—it can
eventually render historical phenomena intelligible.
Investigating the oath archaeologically will mean therefore steering the anal-
ysis of historical data, which we will essentially restrict to the Greco-Roman
sphere, in the direction of an arché stretched between anthropogenesis and the
present. My hypothesis is that the enigmatic institution, both juridical and re-
THE SACRAMENT OF LANGUAGE
309
ligious, that we designate with the term oath can only be made intelligible if it is situated within a perspective in which it calls into question the very nature
of man as a speaking being and a political animal. Hence the contemporary
interest of an archaeology of the oath. Ultrahistory, like anthropogenesis, is not
in fact an event that can be considered completed once and for all; it is always
under way, because Homo sapiens never stops becoming man, has perhaps not
yet finished entering language and swearing to his nature as a speaking being.
6. Before continuing our study, it will be necessary to clear the field of a
preliminary misunderstanding, which obstructs access to that “oldest history”
or that “fringe of ultra-history” that an archaeology can reasonably seek to reach.
Take the exemplary analyses that Benveniste dedicated to the oath first in the
article from 1948 cited above and then in Indo-European Language and Society.
In both, the essential thing is the abandonment of the traditional etymology
of the term horkos, which traced it back to herkos, meaning “enclosure, barrier, bond,” and the interpretation of the technical expression for the oath ( horkon
The Omnibus Homo Sacer Page 45