Use the following persistent identifier: http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Benveniste.Indo-European_Language_and_Society.1973.
Chapter 8. The Oath in Greece
The oath, a solemn declaration placed under the guarantee of a superhuman power that is charged with the punishment of perjury, has no Indo-European expression any more than the notion of “swearing” has. Different languages have coined expressions which relate to the particular forms assumed by the ordeal which the taking of an oath involves. Notably in Greek, thanks to the Homeric expression hórkon omnúnai, meaning “to swear an oath,” we can grasp its concrete origin: “to take hold of the hórkos,” an object charged with malevolent powers which will be unleashed in case the oath is broken. The old sacramental formula ístō Zeús is an appeal to the divinities as eyewitnesses and consequently as irrefutable judges (cf. Lat. iudex arbiter). Latin sacramentum ‘oath’, and perhaps Hittite lingāis (cf. Gr. élenkhos?), underline the potential malediction which specifically defines the binding declaration of the oath.
Of the religious expressions in which speech has a particular force and its own procedures none is more solemn than that of the oath and none would seem more necessary for the functioning of social life. Yet it is a remarkable fact that we look in vain for a common Indo-European expression. There is no Indo-European term of which one can say that it is found in all the ancient languages and that it properly refers to this notion. Each language has its own expression, and for the most part the terms used have no etymology. The obscurity of the terms seems to conflict with the importance and the ubiquity of the institution which they denote. On reflection one sees the reason for this discordance between the extent of the institution and the rarity of common forms. It is because the oath is not an autonomous institution; it is not an act which has its significance in itself and is self-sufficient. It is a rite which guarantees and makes sacred a declaration. The purpose of the oath is always the same in all civilizations. But the institution may appear in different guises. There are in fact two components which characterize it:
- The nature of the declaration, which assumes from this fact a special solemnity;
- The sanctifying power which receives and solemnizes the declaration.
These are the two constant and necessary elements of the oath. This may take two forms, according to circumstances: it will be an oath relating to truth, a declaratory oath, when it pertains to facts under dispute in a law case; or it will be a binding or promissory oath when it is used to support a promise.
One could define an oath as an anticipated ordeal. The one taking the oath stakes something that is essential to him, some material possession, his kin, even his own life, in order to guarantee the veracity of his affirmation. There is no necessary correspondence between the gestures and the various expressions of the oath; each time the oral or formulaic rite and the accompanying practices may differ. When we find that the oath is referred to by a specific term, this may apply to the actual procedure by which the oath is taken rather than to the oath itself. If we always knew the circumstances in which the swearing of the oath took place, we should be better placed to understand the proper sense of the term. But very often these conditions are unknown and the expression remains obscure.
In Germanic we have Got. aiþs, which has cognates in all the Germanic languages: OIcel. eiδr, OHG eid, OE āþ, Engl. oath, and this corresponds exactly to OIr. ōeth. The correspondence between Celtic and Germanic is so close that one wonders whether borrowing has not taken place—as happens so often with cultural terms between these two groups—and if so, in what direction. Got. aiþs and OIr. ōeth go back to *oito-, which can be interpreted as a form derived from the root meaning “to go”, and therefore as “the march.” The difficulty is to see what a “march” has to do with an “oath.” We might accept the view of the historian K. von Amira, who regarded this “march” as the act of “going solemnly to the oath”, cf. Lat. in ius ire. This is possible, but one can imagine other explanations, especially if we recall a rite which is known in a number of ancient civilizations. The swearing of an oath occasioned a sacrifice: an animal was cut in two and then the man or men who were swearing the oath had to walk between the two halves of the animal so sacrificed. This rite is already attested in Hittite, and survivals of it are found in Lithuania in the fourteenth century. At the conclusion of an oath sworn by the Grand Duke of Lithuania to the king of Hungary, the juror walked between the two halves of an ox which was sacrificed and he proclaimed that such would be his own fate if he did not keep his promise, sic sibi contingi si promissa non servaret. However, since this rite is not attested in the Germanic world, such an interpretation of *oito– remains hypothetical.
In Germanic, as in a number of other languages, but not everywhere, the noun and the verb are different. The verb is Got. swaran (Germ. schwören, Engl. swear) which translates Gr. omósai; ufarswaran is a calque of Gr. epi–orkeîn ‘to commit perjury, swear a false oath’. This verb has correspondents outside Germanic: in Italic, Osc. sverrunei, the dative singular of the nominal form, which means “to the orator, to the guarantor.” But sermo, which has also been brought into connection with it, must rather be related to serere. This same Germanic verb also yields the Icelandic svara ‘reply’, OHG andsvara ‘reply’ (Engl. answer); for the formation we may compare the Lat. re–spondeo, from which we might conclude that the sense of swaran is close to that of spondeo, that is “to guarantee, be responsible for something.” Thus the Germanic *swer– ‘to act as a guarantor’ is well suited to the notion of the “oath” which is expressed by the substantive which acts as the object of the verb.
In Greek too the verb ómnumi and the substantive hórkos are different. The verb by itself can mean “to swear,” but neither term is used in any other context than the swearing of an oath. Thus within Greek itself we find nothing at our disposal which would throw light on its real significance. Now the comparatist finds material for his reconstructions only if he can observe variations and here the sense is fixed and immobile. But the etymology of the Greek verb permits certain deductions. The root om– of the present stem ómnumi can be connected, as has been proposed long ago, with the Sanskrit verb am-, of the same sense, which is ancient and attested irreproachably in Vedic and Brahmanic texts. This correspondent is the only one which can throw light on the origin of ómnumi. In Vedic am– is found both as a simple verb and with the preverb sam-, just as we have in Greek sun–ómnumi along with ómnumi. We also have the imperative form in a legendary tale: a character is invited to swear that he will do what he says; the god says “r̥tam amīṣva”, “swear by the r̥ta” (that is, taking the r̥ta as guarantor); and the character in question r̥tam āmīt ‘swore by the r̥ta’. In the Śatapatha-Brāhmaņa: etad dha devāḥ … samāmire ‘and that the gods swore conjointly, they swore it to one another’; and again, samam–yate ‘he binds himself vis-à-vis another for a certain length of time’.
By virtue of the specific nature of the use we have the opportunity of delimiting the proper meaning of the term: am– properly means “to take, seize,” with or without a preverb; tam abhyamīti Varuṇaḥ is equivalent to the expression, with a different verb, taṃ gr̥hṇāti Varuṇaḥ ‘Varuna seizes him’. The man who is “taken, seized” by an attack of some illness is called abhyānta, the participle of the same verb am-. This is a particularly valuable pointer to the prehistory of the notion: our starting point must be the sense “seize.” Although no trace of this is left in Greek, this idea must find its place in the total explanation of the expression. For we can justify it indirectly. When Hypnos makes Hera swear that she will give him as a wife one of the young Graces, Pasithea, he asks her for a solemn oath: “Swear to me by the inviolable waters of Styx, touching with one hand the nurturing earth and with the other the sparkling sea, so that all the nether gods who surround Kronos may bear witness” (Il. 14, 271).
Let us now consider hórkos, the noun which usually functions as the object of the verb in the expression hórkon omósai. The sense of hórkos shows no variation. In the poetical language, from the time of Homer onwards, hórkos with ómnumi is the expression pure and simple of the “oath.” We may also cite the important derivative epíorkos ‘perjury’ and epiorkeîn ‘commit perjury’, a term which requires a separate examination.
We have no etymological connections which would help to explain hórkos. All that we have is the link with hérkos ‘fence’ which was suggested by the ancients and taken up again in more recent times. At first sight we have here an example of a familiar type of alternation: since hérkos is a neuter i-stem, we should expect the alternation to be hérkes-/hórko-. But the meaning of hérkos is exclusively “wall, fence, enclosure, etc.”; we have the familiar Homeric expression hérkos odóntōn ‘the barrier of the teeth’. We should in that case have to imagine that the form with the o-grade of the root vowel meaning “oath” had something to do with the notion of “barrier.” But however we exercise our imagination, there is nothing in Greek ideas that favors this interpretation, which in any case is far from satisfying. This is a reason for not neglecting the task of clarifying the sense as far as possible within Greek itself.
In the Homeric language hórkos designates every kind of oath: the type which gives a guarantee of what one is going to do, a pact; or else the type which supports a statement relating to the past, the so-called judiciary oath. Thus the sense of hórkos does not depend on the nature of the oath.
But it is important to note that the Homeric hórkos is not an act of speech. Let us read the formula of the “great oath” of the gods: “May Earth and the vast Sky above and the waters of the Styx which go down (to the lower world), which is the strongest oath for the blessed gods, be witnesses” (Il. 15, 36ff.).
Cf. The Homeric Hymn to Demeter 259: “May the hórkos of the gods, the implacable waters of the Styx, be witness.” Here the “hórkos of the gods” is put in apposition with húdōr ‘water’: it is the water of the Styx which is the hórkos.
Hesiod, in fact, in the Theogony (l. 400) makes the Styx into a nymph whom Zeus wished to honor by making her “the great hórkos of the gods.” This is why Zeus, when he wants to find out which of the gods has lied (l. 784f.), sends Iris far away to bring back the “great hórkos of the gods” in an ewer. This is the famous water which flows cold from a steep and precipitous rock, the water of Styx. We see, then, that the water of Styx by itself constitutes the hórkos of the gods, being a material invested with baneful powers.
There are other types of hórkos: Achilles desires to give to Agamemnon a solemn promise; he gives him his scepter, which guarantees the thémistes of Zeus. He adds: “This scepter will be for you a mégas hórkos” (Il. 1, 239).
This is not merely a turn of phrase: the literal interpretation leads to the identification of the hórkos with a material object: some sacred substance, the wand of authority, what is essential is always the object itself and not the act of affirmation. We can now see a possibility of harmonizing, in their primary significance, the verb and the substantive: just as ómnumi goes back to a prehistoric meaning “grasp firmly”, so hórkos, in Greek itself, suggests some material substance, hence the expression “to grasp the hórkos.” Whether it is an object or some substance, this hórkos is a sanctifying object, one which has a potency which punishes every breach of the pledged word.
This is presumably how the Greeks imagined the personification of hórkos; it is sinister. Let us quote Hesiod again: “Hórkos is the worst of the scourges for every terrestrial man who knowingly shall have violated his oath” (Theog. 231-32); cf. Works 804, where it is said that hórkos was created only to be the scourge of perjured men. “He keeps pace with crooked judgments” (ibid. 219).
The mythical imagination has done no more than personalize the notion implicit in the sense of the word, by representing hórkos as a destructive force which is unleashed in case of breach of oath, for the substantive hórkos designates a substance charged with bane, a divine, autonomous power which punishes perjury.
Behind this concept we can guess at the idea present in other terms for the oath. In Latin we have apart from ius iurandum, studied above, the term sacramentum (from which French serment ‘oath’ comes); this 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. We recall that a man who is declared sacer may be killed by anyone whatsoever.
This “consecration” recurs also in the same term of Sanskrit śapatha ‘swearing’, derived from śap– ‘to curse’ and also in Slavic in the OSl. klęti ‘to curse’, whereas klęti sę means “to swear,” just like Russian kljast’ ‘to curse’ and kljast’ sja ‘to swear’. The expression reveals the phenomenology of the oath. The person taking the oath vows himself to malediction if he commits perjury, and he solemnizes his act by touching the object or substance invested with this terrible potency.
We must now test the validity of this interpretation for the compound of hórkos which designates the “perjurer”; this is epíorkos, a term so fraught with difficulty in spite of its transparent formation that it remains the subject of discussion among scholars.
The word enters into two different constructions, the oldest having the attribute in the nominative: epíorkos omnúnai ‘to swear in such a way as to be epíorkos’; the other has an object in the accusative: epíorkon omnúnai. The first construction is found in Hesiod Works, 804, the second in Homer, e.g. Il. 3, 279.
The literal sense of this compound term has been discussed a number of times. A recent interpretation is that by Schwyzer.  To explain why epi + hórkos means “to swear a false oath” Schwyzer starts from a verse of Archilochus (Diehl, Anthol. Lyr. I, 265): “He who was once a companion has trampled on the oath” làx ébē eph’ horkíois.
This would be the literal explanation of the compound, from the fact that epí ‘on’ figures in an expression which formulates the notion analytically. This would imply that we must understand epíorkos as ho epì hórkōi < bás >, i.e. “he < who walks > on the hórkos.” But the flaw in the explanation is obvious: the essential term is lacking, for it is precisely the verb baínō which is missing from the compound. We certainly have the nominal construction of epí, but without the idea of “walking on”, trampling on.” This is why we must reject the explanation of Schwyzer.
The explanation of epíorkos ‘perjurer’ and of the verb epiorkeîn ‘perjure oneself’ must start with the observation that the form epíorkos cannot be ancient: if it were we should expect *ephorkos. It must therefore be an adjective (or a verb, according to whether we posit one or the other as the primary term) which was based on an expression in which both epí and hórkos occurred together. This expression is attested and we find it in Hesiod (Works 194) in a description of the Age of Iron. In this age, he says, no one will care about good and evil, traditional conventions will no longer be respected: “the base man will do mischief to the better, speaking in crooked words, and he will add an oath, epì d’ hórkon omeîtai.” We find here, still as distinct elements, the members of the compound epí–orkos, and we can see how they add up to the idea of perjury: there is an implicit connection between the oath which is taken and the lie (the crooked words) which it supports. The idea, therefore, is the “addition” (epì) of an oath (hórkon) to a statement or a promise which one knows is false. This is confirmed by a second example from Hesiod (Works 282): “the man who deliberately bears false witness by swearing a false oath, hòs dé ke marturíēisi hekṑn epíorkos omóssas pseúsetai …” In the Homeric Hymn to Hermes, Hermes himself gives the example of the great oath offered in support of an entirely false statement (ll. 274 and 383). Thus the fact of “adding a hórkos” (epí–orkos) always supposes, whether explicitly or not, that the person swearing will not keep his word, that he will be epíorkos. It is by implicit reference to the use of a false oath which must have become habitual (and proverbial) that the expression “to add (to one’s statement) an oath” soon came to signify “to swear a false oath,” “to perjure oneself.” Thus the term epíorkos throws light on a fact of morals; it shows that all too lightly support was given by an oath to a promise which one had no intention of keeping or a statement which one knew to be false. The evidence of language finds support, curiously enough, in a historian, the first of the Greek historians, Herodotus. He tells a story about an episode in the struggle between the Medes and the Greeks. The Lacedaemonians having warned Cyrus not to do harm to any Greek city because they would not tolerate it, the latter replied to the herald who brought this message: “I have no fear of these men who have at the center of their city a place where they assemble to deceive one another by (false) oaths” (I, 153: allḗlous omnúntes exāpatō̂si). The expression which is literally “deceive one another by oaths” evidently implies that the oaths are false. Here we see clearly how the intention to deceive turns the oath into a stratagem. Herodotus relates many other examples of this. Glaucus calmly goes to ask the oracle whether he can use an oath to gain possession of a deposit entrusted to him which he does not wish to give back. The Pythia makes this crushing reply: “There is certainly profit in thus winning by an oath and in acquiring riches. Swear, then, if you will, since death also awaits the man who keeps his word. But there is a son of the oath, nameless and without hands or feet. Yet swiftly he pursues (the perjurer) until he seizes him and destroys all his progeny and his whole house; whereas the descendants of the man who keeps his word will have the better fate hereafter” (VI, 86). Elsewhere we read how Etearchus made his guest swear to agree to all his demands and profited by it to make him kill his own daughter: the other outraged by the “deceit of the oath” (tē̂i hapátēi toû hórkou) ingeniously gets out of his obligations (IV, 154). It was also by the device of false oaths (tō̂i hórkōi kaì tē̂i hapátēi) that Ariston procured the wife of a friend (VI, 62).
The analysis of the compound epíorkos thus links up with the description of morals: in the expression which was coined at an early date for “perjury” we can find confirmation of the deceitful use of the oath in the social life of the Greeks. The only curious thing is to find that this feature is so old, since epíorkos and epiorkeîn are already in use in the Iliad. 
We have now explored, etymologically and conceptually, the interpretation of the notions connected with hórkos and ómnumi. We can now turn to the Hittite term for “to swear”: ling– ‘swear’ with the substantive lingāi– (genitive –iyas) ‘oath’, and the denominative verb linganu– ‘cause to swear an oath, bind by an oath’, notably for the taking of a military oath imposed by a chief on his troops. Sturtevant was of the opinion that the Hittite ling– corresponded to Greek élegkhos. Now élegkhos means “inculpation, a proof of guilt,” whence in the vocabulary of philosophy it came to mean “refutation.” Consequently it would follow that “to swear” in Hittite meant “to inculpate,” which would correspond fairly well with Greek and Roman ideas. The person swearing inculpates himself in advance and conditionally, and this inculpation takes effect in case of perjury.
This is an idea which recurs in the Latin expression sacramentum, and this poses a problem of law rather than one of etymology or philology. We know different senses of sacramentum: the legis actio sacramenti is a particular form of proceeding bound up with archaic practices in making a claim before the pontifex. If the proof should not have been established in the regular way, a poena would afflict the one who instituted the action. Another formula defines the soldier’s oath, which is of a special kind: consulibus sacramento dicere, ‘to bind oneself to the consuls by the sacramentum’.
Sacramentum is a derivative, not of sacer, but of sacrare ‘to declare sacer, to pronounce anathema’, the man who commits a certain offense. The sacramentum is properly the action or object by which one anathematizes one’s own person in advance (the military sacramentum) or the pledge deposited (in the judiciary oath). Once the words are spoken in the set forms, one is potentially in the state of being sacer. This state becomes effective and invites divine vengeance if the undertaking is transgressed. In all circumstances the process of engagement is ordered in the same way, and to some extent this is apparent in the terms themselves.
We now consider the formulas and the particular ways in which the oath is sworn. There is one aspect which seems to us particularly striking but which usually escapes comment. This is the formula which in Homer recurs every time the text of the oath is reproduced. Appeal is made to Zeus and to a series of gods: “ístō nûn Zeùs prō̂ta … Gē̂ te kaì Ēélios (Il. 19, 258f.) “May Zeus, the Earth and the Sun know it …” The purpose is not merely to acquaint the gods with the text of the undertaking by which one binds oneself. We must give to ístō its full etymological force: not simply “May he know”, but more accurately “May he see.” The root *wid– in this use preserves its original meaning. It calls upon the gods to be eyewitnesses of the oath. The witness at an early period is a witness insofar as he “knows” but primarily in virtue of what he has seen.
This interpretation is not a simple etymological conjecture. When the other Indo-European languages offer ancient and explicit evidence for the sense of *weid-, they agree with Greek. Thus vettar in Sanskrit, which has the same sense of “witness” is, apart from the vocalic grade of the root, the form corresponding to the Greek ístōr ‘witness’ and certainly means “the one who sees”; Got. weitwōþs, the perfect participle (cf. Skt. vidvas-, viduṣ-) is “he who knows because he has seen”; similarly the Irish fíadu (< *weidōn) ‘witness’. The Greek ístōr takes its place in the same series and the proper meaning of this root *wid– is illuminated by the rule enunciated in the Śatapatha Brāhmaṇa: yad idānīm dvau vivadamānām eyātām aham adarśam aham aśrauṣam iti ya eva brāyād aham adarśam iti tasmā eva śraddadhyāmā ‘If now two men dispute (have a law suit), one of them saying, “I have seen,” and the other “I have heard,” the one who says “I have seen” is the one whom we must believe’.
As between the one who has seen and the one who has heard, it is always to the one who has seen that we should give credence. The fundamental value of eyewitness emerges clearly from the name of the witness—ístōr. This is why the gods are taken to witness by inviting them to see. The evidence of sight is irrefutable: it stands alone.
In Latin, too, the oath is accompanied by an appeal to the gods, but the formula is different. We read it in “the first covenant” (thus Livy I, 24, 7), that between Rome and Alba. After the conclusion of the pact, the fetial pronounces the words: “Audi… Juppiter audi, pater patrate populi Albani; audi, tu populus Albanus.” Thus Jupiter, the pater patratus, and the Alban people are requested to hear. It is necessary to hear to be a witness of the oath at Rome. For the Roman, who attaches such importance to the pronouncement of solemn formulas, to see is less important than to hear.
There remains some obscurity, however, about a particular (Homeric) use of ístōr in an important passage (Il. 18, 498ff.), which we have studied from a different point of view  —does ístōr here mean “witness” or “judge”? In a scene which figures on the shield of Achilles two men appear in a dispute, which concerns the poinḗ to be paid for the killing of a man. Both of them resort to an ístōr for a decision in the case (501).
It is difficult to see how he can be a witness, since his presence would have obviated the debate; he must be an “arbiter.” For us the judge is not a witness; this variation prejudices the analysis of the passage. But it is precisely because the ístōr is the eyewitness, the only one who can settle the dispute, that made it possible for ístōr to acquire the sense of “one who decides by a final judgment on a question of good faith.”
At the same time we also grasp the proper meaning of the Latin term arbiter, which is the source of our own term. As has been expounded above,  arbiter in fact designates two functions: (1) first the “witness” (the older sense), this being the sole sense in Plautus, and even in the classical period remotis arbitris means “without witnesses.” And later (2) “arbiter.” As a matter of fact, this sense is explained by the proper function of the iudex arbiter. As we have seen, arbiter is etymologically “the one who supervenes,” as a third person, in an action in which he is a witness without having been seen, and consequently the one whose evidence settles the dispute. In virtue of the law, the iudex arbiter has the power of deciding as though he were the arbiter-witness, as though he had been present at the scene.
All this is evoked by the oath formula in Homer. Why were the gods invoked? This is because the punishment of perjury is not a human concern. No ancient Indo-European code provides a sanction for the perjurer. The punishment is regarded as coming from the gods since they are guarantors of the oath. Perjury is an offence against the gods. To bind oneself by an oath always means devoting oneself in advance to divine vengeance, since the gods are implored to “see” or to “hear”, to be present in every case at the action which binds and commits.
[ back ] 1. Indogermanische Forschungen, 45, 1927, 255ff.
[ back ] 2. In an earlier article on the expression of the oath in ancient Greece (Rev. Hist. Relig. 1947-48, pp. 81-94) we gave a different explanation of the term epíorkos. The interpretation offered here is close to that given by M. Leumann, Homerische Wörter, 1950, p. 79. The term hórkos has been the subject of articles by J. Bollack, Rev. ét gr., 1958, 1ff., and by R. Hiersche, ibid. 35ff. Other studies are cited in the etymological dictionary of Frisk, under epíorkos and hórkos.
[ back ] 3. Cf. Book Five, Chapter Two.
[ back ] 4. On arbiter, see Book Five, Chapter Three.