The Omnibus Homo Sacer

by Giorgio Agamben

  It is authority which is exercised at the same time as protection for somebody

  who submits to it, in exchange for, and to the extent of, his submission” (Ben-

  veniste [2], 1:118–19/97–98).

  Thus the strong link between the two Latin terms fides and credere, which was to assume such importance in the Christian sphere, becomes comprehensible. Meillet showed that fides became a substitute in everyday usage for the

  ancient verbal substantive *kred, which expressed a very similar notion. Credere in fact meant, originally, “to give *kred,” to put one’s faith in someone from

  whom one expects protection and, in this way, to link oneself with him in faith

  (each person often shaking the other’s right hand: dextrae dextras iungentes fidem

  obstrinximus [Livy 23, 9, 3]).

  Beyond regulating personal relationships, fides performed an important

  function in international public law, in the special relationships that were es-

  tablished by means of fides between cities and peoples. In a war the enemy city

  could be defeated and destroyed by force ( kata kratos) and its inhabitants killed

  or reduced to slavery. But it could also happen, on the contrary, that the weaker

  city could have recourse to the institution of the deditio in fidem. That is to say, it could capitulate, submitting itself unconditionally to the fides of the enemy, thus obligating the victor in some sense to a more benevolent form of control.

  This institution was also called pistis by the Greeks ( dounai eis pistin, peithesthai) and fides by the Romans ( in fidem populi Romani venire or se tradere). And we also encounter here the same connection between faith and oath: the cities and

  people who mutually bound themselves in the deditio in fidem exchanged sol-

  emn oaths to sanction this relationship.

  THE SACRAMENT OF LANGUAGE

  321

  The fides is, then, a verbal act, as a rule accompanied by an oath, with which

  one abandons oneself completely to the “trust” of someone else and obtains, in

  exchange, that one’s protection. The object of the fides is, in every case, as in the oath, conformity between the parties’ words and actions.

  Dumézil has shown that, when in Rome the history of the monarchical pe-

  riod was, little by little, constructed retrospectively and assumed a definite form,

  the fides, which assumed an important role in public and private life, became

  divinized and was associated with the figure of Numa, to whom the foundation

  of the sacra and the leges is attributed (Dumézil [4], 202/198). Fides thus becomes a goddess, for whom a temple on Capitoline Hill is constructed around 250; but,

  just as in the case of Deus Fidius, of whom it is disputed whether he is in ori-

  gin distinct from Jove and who, like Mitra, was a sort of “personified contract”

  (ibid.), here religion does not precede the law but rather follows it.

  With the fides, exactly as with the oath, we thus find ourselves in a sphere

  in which the problem of the genetic relationship between religion and law has

  to be taken up again on new foundations. It does no good, in light of the com-

  plexity of these institutions, which seem to be at once moral, religious, social,

  and juridical, to appeal, as some do, to the category of prelaw (Imbert, 411). The

  fact that the institutions in question are not juridically sanctioned (impunity of

  perjury in the most ancient epoch, absence of legal recourse for the creditor who

  has trusted the fides of the debtor) does not mean that they must be considered

  religious rather than juridical; it means rather that in them the investigation

  has hit upon a limit, which obliges us to reconsider our definitions of what is

  juridical and what is religious.

  א One of the commonplaces of the theory of the oath is that the fact that legal

  sanction is lacking in the ancient epoch is a sign that it belongs to the religious sphere, insofar as the punishment of perjury was left to the gods. Scholars continue to cite the

  dictum of Tacitus, deorum iniurias dis curae (“wrongs done to the gods are the gods’

  concern” [ Ann. 1.73]), without taking notice of the juridico-political context from which it is drawn. Rubrius has been accused before Tiberius of “having violated by perjury the

  numen of Augustus” (it is a matter, then, of a particular type of oath “by the genius of the emperor,” which became common in the imperial age). The question is not whether

  perjury in general is more or less punishable but whether Rubrius must be accused,

  because of his perjury, of lèse majesté. Tiberius prefers in this moment not to resort to a charge of which, as Tacitus informs us, he will later make ferocious use and affirms

  sarcastically that “as to the oath, the thing ought to be considered as if the man had

  deceived Jupiter. Wrongs done to the gods are the gods’ concern [ deorum iniurias dis

  curae].” In no way is it a matter, according to the words of a rash commentator, of an

  “ancient principle of Roman law” but of the sarcasm of an emperor whose scant religious

  322

  HOMO SACER II, 3

  piety is well known ( circa deos et religiones negligentior [in regard to the gods, and matters of religion, he exhibited great indifference]—Suetonius Tib. 69). This is confirmed by the fact that the other case in which we find the same principle enunciated is clearly later and refers to the same problem of the applicability of the offence of lèse majesté to an oath on the numen principis (even here the response of the emperor is negative and, probably referring to the dictum of Tiberius, it is suggested that iusiurandi contempta religio satis deorum ultorem habet [the god’s vengeance is enough for the one who has contempt for the oath]— Codex iuris 4, 1, 2; quoted in Schied, 333).

  It is incorrect to claim that we should only consider something juridical if a sanction

  is attached. On the contrary, Ulpian explicitly affirms that only those laws for which no

  sanction is provided must be considered perfecta, while the presence of a sanction constitutes the law as imperfecta or minus quam perfecta (Ulpian, prol. 1–2). In the same sense, the impunity of lying in many ancient ordinances does not mean that its punishment is

  a province of the gods. If anything, it is possible that here we have to do with a sphere

  of language that stands before law and religion and that the oath represents precisely the threshold by means of which language enters into law and religio.

  In Plescia’s monograph on the oath in Greece, we read, “As a general rule, one may

  say that, until the end of the sixth century, divine punishment of perjury was still an

  effective deterrent against the misuse of the oath. From the fifth century, however, the

  individualism and relativism of the sophistic movement began to undermine the old

  notion of the oath, at least among a certain segment of the population, and fear of the

  gods, in case of perjury, began to wane” (Plescia, 86–87). These affirmations, however,

  reflect only the opinion of the author. This claim is based on the misunderstanding of

  a passage from Plato ( Laws 12.948b–d), obviously ironic, in which Radamanthys, who is credited with introducing oaths into trials, is praised for having understood “that the men of his time had a clear belief in the existence of gods—and naturally so, seeing that

  most men at that time were the offspring of gods, he himself among others, as the story

  declares.” The irony is again accentuated by the fact that Plato, firmly opposed to the use of the oath of parties in trials, adds that Radamanthys “administered an oath to the disputants regarding each matter in dispute, and thus secured a speedy and safe settlemen
t.”

  Equally ironic, and devoid of all nostalgia for a supposed ancient devotion, is the reason adduced immediately after for the exclusion of the oath of the parties: “But nowadays,

  when, as we say, a certain section of mankind totally disbelieve in gods, and others hold

  that they pay no regard to us men, while a third party, consisting of the most and worst

  of men, suppose that in return for small offerings and flatteries the gods lend them aid

  in committing large robberies, and often set them free from great penalties—under such

  conditions, for men as they now are, the device of Radamanthys would no longer be

  appropriate in actions at law.” The essential objection to the oath of the parties is actually, as is said immediately after, that making the parties in the trial swear is equivalent to compelling them to perjury: “For truly it is a horrible thing to know full well that, inasmuch as lawsuits are frequent in a State, well-nigh half the citizens are perjurers” (see, again,

  THE SACRAMENT OF LANGUAGE

  323

  in the Laws 10.887a, Plato’s irony concerning the attempt at “assuming in our legislation the existence of gods [ nomothetountes ōs ontōn theōn]”).

  13. Another institution with which the oath is closely connected is the sacra-

  tio. The ancient sources and the majority of scholars, in fact, agree in seeing in the oath a form of sacratio (or devotio, another institution with which consecration tends to be confused). In both cases a man was rendered sacer, that is,

  consecrated to the gods and excluded from the world of men (spontaneously, as

  in the devotio, or because he had committed a maleficium that rendered it licit for anyone to kill him). “One calls sacramentum (one of the two Latin terms

  for oath),” one reads in Festus (466.2), “an act that is done with the sanction

  of the oath [ iusiurandi sacratione interposta].” As Benveniste writes: “the term

  sacramentum . . . implies the notion of making ‘ sacer.’ One associates with the oath the quality of the sacred, the most formidable thing which can affect a

  man: here the ‘oath’ appears as an operation designed to make oneself sacer on

  certain conditions” (Benveniste [2], 2:168/437). And Pierre Noailles can write

  of the oath in the trial in the same way: “The litigant himself has consecrated

  himself, has rendered himself sacer through the oath” (Noailles [1], 282–83). As

  Hirzel writes of the perjurer: “his situation was no different from that of the

  Roman sacer, who has devoted himself to the Manes, and just like him can

  . . . be excluded from every religious and civil community” (Hirzel, 158). In the

  same sense the oath can be seen as a devotio: “once the oath is formulated, the

  man taking it is by anticipation a ‘devoted’ person. . . . For the oath is a kind of

  devotio: as we have seen, the Greek horkos signifies an act of self-consecration by anticipation to the power of an avenging deity if the given word is transgressed”

  (Benveniste [2], 2:243/498).

  Hence the importance, in the oath, of the curse ( ara, imprecatio), which

  constitutively accompanies its utterance. Already Plutarch, in those precious

  sources for the knowledge of Latin antiquity represented by the Questiones

  romanae, informs us that “all oaths are concluded with a curse against perjury”

  ( eis kataran teleutai tēs epiorkias, 44). Scholars in fact tend to consider the curse as the very essence of the oath and therefore to define the oath as a conditional

  curse: “The curse appears as the essential part of the oath. Since this essential

  aspect of the oath was displayed in the purest and strongest way in them, oaths

  of imprecation were held to be the most powerful. The curse is what is essential

  and originary” (Hirzel, 138–39); “To swear is first of all to curse, to curse oneself

  in the event that one says what is false or does not do what has been promised”

  (quoted in Hirzel, 141).

  324

  HOMO SACER II, 3

  Bickermann has observed that the curse can, however, be lacking (although

  the examples cited do not refer to Greek or Latin sources) and that there can be

  imprecations without an oath (Bickermann, 220/889). The opinion of Glotz, ac-

  cording to whom the curse necessarily accompanies the oath but is not identical

  to it, therefore seems more correct, and it is in this sense that one must under-

  stand the recommendation, contained in official documents, to “add the curse

  to the oath” ( tōi horkoi tan aran inēmen [Glotz, 752]). It is necessary, moreover, to specify that the oath often involves both an expression of a bad omen and a

  good one and that, in the most solemn formulas, the curse follows a blessing:

  “To those who swear loyally and remain faithful to their own, may children give

  them joy, may the earth grant its products in abundance, may their herds be

  fruitful, and may they be filled with other blessings, them and their children;

  but to perjurers may the earth not be productive nor their herds fruitful; may

  they perish terribly, them and their stock!” (Glotz, 752). The blessing can, how-

  ever, be lacking, while the curse must normally be present (Hirzel, 138). This is

  the rule in Homer, in whom the curse is accompanied by eloquent gestures and

  rites, as when, in the scene in which the Trojans and the Achaeans exchange

  oaths before the duel of Paris and Menelaus, Atreus pours wine on the ground

  from a bowl and utters the formula: “whichever host of the twain shall be first to

  work harm in defiance of the oaths, may their brains be thus poured forth upon

  the ground even as this wine” ( Il. 3.299–300).

  The oath seems, then, to result from the conjunction of three elements:

  an affirmation, the invocation of the gods as witnesses, and a curse directed

  at perjury. In the same sense, one can say that the oath is an institution that

  joins an element of the pistis type (the reciprocal trust in the words offered) and an element of the sacratio-devotio type (the curse). But, in reality, the three institutions are so closely intertwined terminologically and factually (as in the

  term sacramentum, meaning both oath and sacratio) that the scholars, although without drawing all the consequences of this proximity, tend to treat them as a

  single institution. We would do well not to forget that the series pistis- horkos-ara or fides-sacramentum refer to a single institution, certainly an archaic one, that is both juridical and religious (or prejuridical and prereligious) and whose meaning and function we must seek to understand. But this means that the oath

  seems to lose, in this perspective, its specific identity and become confused with

  the fides and the curse, two institutions whose nature—above all as regards the

  curse—is not entirely clear and, in any case, has received relatively little atten-

  tion from scholars. An analysis of the oath will thus first of all have to confront

  the problem of its relationship with the curse.

  THE SACRAMENT OF LANGUAGE

  325

  א The description of the scene of the oath in the Critias (119d–120d) shows very

  well the mutual belonging of pistis, horkos, and ara. The taking of the oath is here defined as a way of “pledging trust,” and on the other hand it is the oath itself that

  invokes ( epeuchomenos) “great curses”: “And when the kings were about to give judgment they first gave pledges one to another of the following description [ pisteis allēlois toiasde edidosan]. . . . And inscribed upon the pillar, besides the laws, was an oath which invoked mighty curses
upon them that disobeyed [ horkos ēn megalas aras epeuchomenos

  tois apeithousin]. . . . And after this they drew out from the bowl with golden ladles, and making libation over the fire swore to give judgment according to the laws upon the

  pillar and to punish whosoever had committed any previous transgression.”

  14. Once we examine more carefully the constitutive elements of the oath,

  however, we are faced with an uncertainty and confusion in terminology that

  is somewhat surprising. One of the characteristics of the oath on which all the

  authorities, both ancient and modern, from Cicero to Glotz, from Augustine to

  Benveniste, seem to be in agreement is the calling of the gods as witnesses. In his

  commentary on Aristotle’s De interpretatione (4a), Ammonius thus distinguishes

  the oath from the assertion ( apophansis) by means of “the testimony of the god”

  ( martyria tou theou). The oath, according to this endlessly repeated doctrine,

  is an affirmation to which divine testimony is added. The imperative formulas

  martys esto (Pindar Pyth. 4.166: karteros horkos martys estō Zeus, “as a mighty oath, may Zeus be our witness”; trans. altered) or istō Zeus ( Il. 7.411: horkia de Zeus istō,

  “Let Zeus be witness to this covenant”), attested in the ancient sources, do not

  seem to leave room for doubt on this matter.

  But is this really so? It has been observed that the testimony at issue here

  differs essentially from testimony in the proper sense, like that of a witness in a

  trial, because it cannot be contested or verified in any way (Hirzel, 25). And not

  only does the number of the divinities invoked tend to increase well beyond the

  number of “legal Gods” ( nominoi theoi or theoi horkioi) to the point of including eight, sixteen, and finally “all gods” (as in the Hippocratic oath), but at times

  they include rivers, trees, and even inanimate objects (the “bridal bed,” lechos

  kouridion in Il. 15.39). What is decisive in every case is that in the oath it is not in any way really a matter of a testimony in a technical sense, because unlike

  every other conceivable testimony, it coincides with the call and is accomplished

  and exhausted together with it. Things do not change if, as some sources allow