The Omnibus Homo Sacer

by Giorgio Agamben

  us to suppose, one understands what the gods do not as a testimony but as the

  giving of a guarantee. As with testimony, here no standing surety can technically

  take place, either at the moment of the oath or after: it is presupposed as already

  accomplished with the utterance of the oath (Hirzel, 27).

  326

  HOMO SACER II, 3

  The oath is, then, a verbal act that accomplishes a testimony— or a guar-

  antee—independently by the very fact that it has taken place. The formula of

  Pindar cited above acquires here its full meaning: karteros horkos martys estō Zeus,

  “as a mighty oath, may Zeus be our witness”: Zeus is not a witness of the oath,

  but rather oath, witness, and god coincide in the utterance of the formula. As

  in Philo, the oath is a logos that is necessarily accomplished, and this is precisely the logos of God. The testimony is given by language itself and the god names a

  potentiality implicit in the very act of speech.

  The testimony that is in question in the oath must therefore be understood

  in a sense that has little to do with much of what we normally understand by this

  term. It concerns not the verification of a fact or an event but the very signifying

  power of language. When in the discussion of the oath given by Hector to Achil-

  les ( Il. 22.254–55), we read that the gods “are the fittest witnesses [ martyroi] and guardians of all covenants [ episcopoi harmoniaōn],” the “joining together” (such

  is the original meaning of the term harmonia, which comes from the vocabulary

  of carpentry) of which the gods are witnesses and guardians can only be that

  which unites words and things, that is, the logos as such.

  א A gloss of Hesychius ( horkoi: desmoi sphragidos) defines oaths as “bonds of the seal”

  (or sealing, if one prefers the reading sphragideis). In the same sense in fragment 115 of Empedocles one speaks of an “eternal decree of the gods, sealed with great oaths” ( plateessi katesphrēgismenon horkois). The bond that is in question here can only be the one that links the speaker to his speech and, at the same time, words to reality. Hirzel rightly calls attention to the fact that the divine testimony is invoked not only by the promissory oath but also by the assertorial, in which it does not seem to have any meaning, unless what

  is in question here is meaning itself, the very signifying force of language.

  15. If we leave the problem of the intervention of the gods as witnesses in

  order to turn our attention to that of their role in the curse, the situation is no

  less confused. That the curse performed an important function in the polis is

  proved by the fact that, in a perfect analogy with Lycurgus’s thesis on the oath,

  Demosthenes mentions (20.107)—however scandalous it might seem to us—

  curses ( arai) alongside the people and the laws ( nomoi) among the guardians of the constitution ( politeia). Similarly, Cicero, evoking the bonds among men that

  it is impossible not to fulfill, names both curses and fides ( Verr. 5.104: ubi fides, ubi exsecrationes, ubi dexterae complexusque? [What signify his promises? What do

  the curses that he will heap on him? What do the pledges of friendship and mu-

  tual embraces?]). But what is a curse, and what can its function be here? Already

  from the terminological point of view the situation is far from clear. The terms

  that designate it, both in Greek and in Latin, seem to have opposed meanings:

  THE SACRAMENT OF LANGUAGE

  327

  ara (and the corresponding verb epeuchomai) mean, according to the lexicons, both “prayer” (and “to pray”) and “imprecation, curse” (and “to imprecate, to

  curse”).The same can be said for the Latin terms imprecor and imprecatio, which are the equivalent of both “to augur” and “to curse” (even devoveo, which means

  “to consecrate,” is equivalent to “to curse” in the technical sense in the case of a

  devotio to the infernal gods). The entire vocabulary of the sacratio is, as is well known, marked by this ambiguity, the reasons for which I have sought to reconstruct elsewhere.

  Once again, interpretations of the curse uncritically repeat the paradigm of

  the primordiality of the magico-religious and limit themselves to going back to

  a no more specified “numinous power” (see the entry Fluch in the Reallexicon fur Antike und Christentum, 1161) or evoking religion as “a practical auxiliary for the efficacy of law” (Ziebarth, 57). Thus Louis Gernet, in his article “Le droit pénal

  de la Grèce antique” (Penal Law in Ancient Greece), can write:

  The curse has played an important role in the origins of law: it sometimes sanc-

  tions the law or substitutes for it, as we can see in a catalogue of public impre-

  cations published in the fifth century in the city of Teos, where it is formulated

  against an entire series of offences having to do with the security of the State

  and the very subsistence of the city. Naturally, it is in religious life and in the

  practice of the sanctuaries that its use was perpetuated above all; but there it could

  only be a matter of an extremely ancient tradition. The curse presupposes the

  collaboration of religious forces: these (which, in principle, are not represented

  in a personal form) are in some way condensed by the incantatory power of the

  oral rite, and they act on the guilty and those around him by drying up in him

  the source of all life: the imprecation exercises its fatal effect even on the soil, on

  what is born from it and nourished by it. At the same time as and by the very fact

  that it is a devotio, it is an exclusion from the religious community constituted by society: it manifests itself through an interdiction in the proper sense and, in its concrete application, it is a putting outside the law. (Gernet [2], 11–12)

  Only the prestige of the paradigm of the originarity of the magico-religious fact

  can explain how a sensible scholar like Gernet, repeating the old arguments

  of Ziebarth, can be satisfied with such a palpably insufficient interpretation in

  which not only—in clear contrast with the fact that the curse is fully attested

  in the oath in the historical epoch—are mythical presuppositions like the “in-

  cantatory force of the oral rite,” the “religious force,” and their “lethal effects”

  taken for granted, but it does not even become clear whether it is an institution

  in itself or is instead identical with the devotio and, in the last instance, with the oath itself, which would then constitute a derivation of it.

  328

  HOMO SACER II, 3

  It will be helpful, therefore, to put in parentheses, at least provisionally,

  the traditional definitions—which see the curse as an invocation directed at

  the gods so that, in order to punish perjury, they are transformed from wit-

  nesses into avengers—and instead ask ourselves what is effectively at stake in

  the curse, in other words, what is the immanent function that the curse has in

  the oath. According to the common opinion, the gods (or, to be more precise,

  their names) are mentioned in the oath twice: once as witnesses of the oath and

  a second time, in the curse, as punishers of perjury. In both cases, if we leave

  aside mythical definitions, which seek an explanation outside of language, we

  can see that what is at stake is the relationship between words and facts (or

  actions) that defines the oath. In one case the name of the god expresses the

  positive force of language, namely the just relation between words and thing
s

  (“as a mighty oath, may Zeus be our witness”). In the second case it expresses a

  weakness of language, namely the breaking of this relation. To this double pos-

  sibility there corresponds the twofold form of the curse, which, as we have seen,

  generally presents itself also as a blessing: “If I fulfill this oath without violating

  it [ euorkounti], may good things be granted to me. But if I violate it and perjure myself [ epiorkounti], may the opposite befall me” (Glotz, 752; Faraone, 139). The

  name of the god, which signifies and guarantees the juncture between words and

  things, is transformed into a curse if this relation is broken. What is essential,

  in every case, is the co-originarity of blessing and curse, which are constitutively

  copresent in the oath.

  א It suffices to read the very ample entry for Fluch in the Reallexicon fur Antike und Christentum (which seeks to make up for the very small space dedicated to the problem in the Pauli-Wissowa and the Daremberg-Saglio, in which the curse is treated only fleetingly in the article on the devotio by Bouché-Leclercq) to realize that the critical literature has not made much progress with respect to the above-cited article of Erich Ziebarth

  or that of George Hendrickson (1926). Christopher Faraone’s recent study is focused on

  the difference between oaths that contain both blessings and curses (in general directed

  toward the private sphere) and oaths accompanied only by curses (most often reserved

  for the public sphere). In every case, beyond the traditional explanation, which sees the

  oath as a recourse to religious power to guarantee the efficacy of the law, the oath-curse connection remains uninterrogated.

  16. Ziebarth has demonstrated, with ample documentation, the consubstan-

  tiality of the curse to Greek legislation. Its function was so essential that the

  sources speak of a veritable “political curse,” which always confirms the efficacy

  of the law. In the preamble of the laws of Caronda one thus reads: “It is necessary

  to observe [ emmenein] what has been proclaimed, but the one who transgresses

  THE SACRAMENT OF LANGUAGE

  329

  is subjected to the political curse [ ara politikē]” (Stobaeus Florigelium 44, 40; quoted in Ziebarth, 60). Similarly, Dio Chrysostom of Prusa (80.8) informs us

  that the Athenians had put down ( ethento, in the strong sense of the term, as in

  nomon tithenai, to put down a law) in the laws of Solon a curse that extended

  even to children and descendants ( paides kai genos). Ziebarth has traced the presence of the “political” curse in the legal apparatuses of all the Greek cities, from

  Athens to Sparta, from Lesbos to Teos and Chios and finally to the Sicilian col-

  onies (Tauromene). It concerns even questions that have no “religious” element

  at all, as in Athens’s prohibition of exporting agricultural products other than

  oil (Ziebarth, 64). Moreover, before every assembly the kēryx, the town crier,

  solemnly pronounced the curse against anyone who had betrayed the people or

  violated their decisions. “This means,” comments Ziebarth, “that the entire con-

  stituted legal order, according to which the demos is sovereign, is sanctioned by means of a curse” (ibid., 61). Not only the oath, but also the curse—in this sense

  it is rightly called “political”—functions as a genuine “sacrament of power.”

  It is possible, in this perspective, as William Fowler has already discerned

  (Fowler, 17), to consider the formula sacer esto, which appears in the system of

  the Twelve Tables, as a curse. It is not, however, as Fowler holds, to be treated as

  the production of a taboo but as the sanction that defines the very structure of

  law, its way of referring to reality ( talio esto / sacer esto) (Agamben, 31/22). The enigmatic figure of the homo sacer, which is still a topic for debate (and not only among historians of law), seems less contradictory in this light. The sacratio that has struck him—and that renders him both killable and unsacrificeable—is only

  a development (perhaps carried out for the first time by the plebs led by the

  tribune) of the curse by means of which the law defines its scope. In other words

  the “political” curse marks out the locus in which, at a later stage, penal law will be established. It is precisely this peculiar genealogy that can somehow make

  sense of the incredible irrationality that characterizes the history of punishment.

  א It is in the perspective of this technical consubstantiality of law and curse (present

  even in Judaism—cf. Deuteronomy 21:23—but very familiar to a Jew who lived in a

  Hellenistic context) that one must understand the Pauline passages in which a “curse of

  the law” ( katara tou nomou—Galatians 3:10–13) is spoken of. Those who want to be saved through works (the execution of precepts)—this is Paul’s argument—“are under a curse

  [ hupo katara eisin]; for it is written, ‘Cursed is everyone who does not observe and obey

  [ emmenei, the same word that one finds in the law of Caronda] all the things written in the book of the law.’” Subjecting himself to the judgment and curse of the law, Christ

  “redeemed us from the curse of the law by becoming a curse for us—for it is written,

  ‘Cursed is everyone who hangs on a tree.’” The Pauline argument—and, therefore, the

  330

  HOMO SACER II, 3

  very meaning of redemption—can be understood only if it is situated in the context of

  the mutual belonging, in a juridical and not only religious sense, of law and curse.

  17. How should we understand this double valence (benediction and male-

  diction) of the divine names in oath and perjury? There is an institution that

  has always lived in such close intimacy with perjury and the curse that it is often

  confused with them. It can perhaps furnish us with the key for a correct inter-

  pretation of them. I am speaking of blasphemy. In his study “La blasphémie et

  l’euphémie” (“Blasphemy and Euphemism,” originally a lecture held at a col-

  loquium that was, significantly, devoted to the name of God and the analysis

  of theological language), Benveniste often refers to the proximity among blas-

  phemy, perjury, and oath (evident in French in the paronym juron: jurer):

  Outside of worship, society demands that the name of God be invoked in a sol-

  emn circumstance, which the oath is. For the oath is a sacramentum, an appeal

  to God, the supreme witness of truth, and a devotion to divine chastisement in

  case of lying or perjury. It is the most serious commitment that man can enter

  into, and the most serious breach he can commit, because perjury relates not to

  the justice of men, but to the divine sanction. For this reason the name of the

  god must figure in the formula of the oath. In blasphemy as well, the name of

  God must appear, because blasphemy, like the oath, calls God as witness. The

  swearword [ juron] is an oath, but an oath of outrage. (Benveniste [4], 255–56)

  Benveniste underscores, moreover, the interjectory nature proper to blasphemy,

  which, as such, communicates no message: “The formula pronounced in blas-

  phemy does not refer to any objective situation in particular; the same swear-

  word is pronounced in entirely different circumstances. It expresses only the

  intensity of a reaction to these circumstances. It does not refer to a second or

  third person. It transmits no message, it opens no dialogue, it gives rise to no

  response, and the presence of an int
erlocutor is not even necessary” (ibid.). It

  is therefore quite surprising that to explain blasphemy, the linguist puts aside

  the analysis of language and, in one of his rare appeals to the Hebraic tradition,

  refers to “the biblical interdiction against pronouncing the name of God” (ibid.,

  254). Blasphemy is, certainly, an act of speech, but it is precisely a matter of

  “substituting the name of God with its outrage” (ibid., 255). The interdiction

  does not in fact have a semantic content as its object, but the simple pronuncia-

  tion of the name, that is, a “pure vocal articulation” (ibid.). Immediately after, a

  citation from Freud introduces an interpretation of blasphemy in psychological

  terms: “the interdiction of the name of God holds in check one of the most in-

  tense desires of man: that of profaning the sacred. As is well known, the sacred

  THE SACRAMENT OF LANGUAGE

  331

  inspires ambivalent behaviors. Religious tradition has wanted to retain only the

  divine sacred and exclude the cursed sacred. Blasphemy, in its own way, seeks to

  reestablish this totality by profaning the very name of God. One blasphemes the

  name of God, because all that God possesses is his name” (ibid.).

  Coming from a linguist accustomed to working exclusively on the patri-

  mony of the Indo-European languages, the appeal to biblical data is at least

  odd (as is the psychological explanation of a linguistic fact). If it is true, in fact,

  that in the Judeo-Christian tradition blasphemy consists in taking the name of

  God in vain (as in modern forms of the type: nom de Dieu! sacré nom de dieu!

  “by God!”), the blasphemous utterance of the name of God is just as common

  in the classical languages, which are quite familiar to linguists in exclamatory

  forms of the type: edepol, ecastor, by Pollux, by Castor (Greek: Nai ton Castora), edi medi (by Dius Fidius), mehercules, mehercle. It is significant that in all these cases the formula of imprecation is identical to that of the oath: nai and ma introduce the oath in Greek; in Latin edepol and ecastor are also formulas for an oath, exactly like the English “by God” (Festus is, moreover, perfectly aware

  of the derivation of these exclamations from the oath: Mecastor et mehercules