After the in-depth contextual analysis of mênis provided above, it should be possible to resolve some of the lingering questions about its etymology. In this appendix, I propose to (1) point out the existence of a word ultimately related to mênis that is attested in the oldest surviving texts from ancient Iran, the Avestan Gāthās , and from ancient India, the hymns of the Rig Veda, and (2) determine which of the currently competing etymologies of mênis best concurs with its meaning.

All of the current etymologies actually do converge on one point, namely, that Greek mênis is ultimately a derivative of the Indo-European root *men- ‘activate the mind’. [1] The noun that I wish to bring into consideration is an archaic derivative of this same root in Vedic Sanskrit, manyú- ‘zeal, desire, anger’, an action noun exactly cognate with the Avestan word mainiiu- (maniiu-) ‘(good/bad) spirit’. The only archaic words parallel to it in form are vāyu ‘wind; the war god, Vāyu’ in Vedic and its exact Avestan cognate vaiiu- ‘air’, both from a root vā- ‘blow’, IE *h ₂ ueh ₁ (attested in Greek ἄησι ‘blows’); and Greek υἱύς ‘son’, Tokharian B soy, Tokharian A se ‘son’ from *suh _x -ı̯u-. In regard to manyú- , there is disagreement on the point at which its stem ends and its suffix begins, but it is clear that, like mênis, it is an ancient derivative of the root *men- . [2]

In the Gāthās , the Old Avestan hymns of Zarathustra (= Zoroaster), mainiiu- is the name for the twin conflicting spirits, one good and the other evil, who stand at the origin of the world:

In the sixth stanza of this same hymn, we learn that the daēuua , the traditional gods as opposed to the Ahura of Zoroastrianism, chose the deceptive, evil mainiiu-,and “they then rushed into fury with which they have afflicted the world and mankind.” So at the beginning of Yasna 45, Zarathustra warns:

The world was first destroyed when the daēuua chose the evil mainiiu- . Now it is time for the correct choice. The hymn continues:

The two mainiiu- represent a contrastive pair of initial beings such as we have seen in the context of Hesiodic cosmogony. [5] According to some authorities, they are the Zoroastrian replacement for the inherited Indo-Iranian primordial god, Vāyu, a cosmic starting point whose name means ‘air, wind’. [6] As Malamoud points out, the word mainiiu- by itself is so lacking in intrinsic color that it is necessary to specify whether it is spaniiå ‘good’ or angrəm ‘bad’. [7] But the concept of a cosmogonic moral force represented by this word is not irrelevant to the study of Greek mênis, as I hope to demonstrate.

In the Vedic hymns, by contrast, the cognate of mainiiu-, manyú-, offers a range and richness of significations that borders on incoherence. Surveying its uses in the Rig Veda, in the more recent Atharva Veda , and in rituals described in subsequent texts, Charles Malamoud has attempted not to dismiss the multiplicity of its meanings but to discover their underlying core. [8] His point of departure for integrating the whole history of meanings in this word is one of its latest attestations in the Rig Veda, Hymn 10.83, in which the noun manyú- has become a new, “abstract” divinity, abstract in that he lacks a biography and a developed mythology. For Malamoud, such late hymns represent not a new departure, but a recapitulation of the representation of manyú- in all the earlier parts of the Rig Veda: “for the most recent parts of the collection, the whole Veda is alluded to in each hymn.” [9] Hymn 10.83 and its shorter sequel, Hymn 10.84, were actually recited to foment the lust for battle in warriors, and they were also part of a divination ritual that preceded the battle itself. While murmuring these two hymns to Manyu, a diviner would stand between the two opposing armies and set fire (with a fire called āṅgirasa ‘burning coal’, a term associated with the fire god Agni) to a special bundle of branches. The side that drew the smoke from the burning bundle was declared the loser. [10]

These ritual uses of the hymns to Manyu make clear that the frame of reference for manyú-in them is the world of the warrior, but that is only one, albeit a fundamental, aspect of the term. In fact, Georges Dumézil has tried to show that Manyu in 10.83 is operative in all three spheres of the Indo-European ideology as he reconstructs it, namely, the priest’s domain of sovereignty and magic, the warrior’s domain of force, and the people’s domain of fertility and well-being. [11] Here are the first four stanzas of the hymn whose content he analyzes in this way:

Dumézil points out that in the second strophe, Manyu is first identified with the warrior par excellence, Indra; then with Varuṇa, a sovereign and priestly god whom Dumézil considers to be here represented with the attributes of the fire god, Agni; and lastly, the hymn tells us that the tribes of humans praise Manyu. It is problematic to associate Varuṇa, as Dumézil wishes to, with the term hotar ‘sacrificer’ and the epithet Jātavedas, both of which belong to Agni, a divinity who can bypass trifunctional sequences since he fits all three of the functions. [13] I would suggest that Agni as fire god has been inserted into the three-part sequence with good reason, because of the overwhelming association of manyú- in this and other hymns (as also in the ritual noted above) with fire. [14] In the concluding line of this same strophe, that association is explicit: “protect us, Manyu, together with the burning fire!” But there are other trifunctional features in this hymn in which the tripartite sequence remains intact; they strengthen Dumézil’s point. For instance, in the first and fourth strophes, there is a sequence of terms referring to the force gained by devotees of Manyu. In the first strophe mention is made of sáhas, which denotes force superior to that of another person, power over another; then ójas , a term for physical force that is part of the vocabulary of Indra; these two nouns are then followed by the adjective víśvam ‘complete’, a word related to and, according to Dumézil, playing upon the noun víś ‘clan’, regularly used for massed humans in references to the third function. Then, in the fourth strophe, the three terms recur: “Common to all populations (viśvácarṣaṇih), dominator endowed with superior force (sáhas), grant us formidable force (ójas) in battles.” [15]

Taking his cue from Dumézil’s demonstration that Vedic manyú- is not simply a term restricted to the warrior sphere but functional in other contexts as well, Malamoud asserts that it is also not a term for a passing emotion like anger or rage. Instead it is the name of a permanent faculty in both humans and gods whose content varies in precise ways. As he states it, manyú- is a faculty “placed on the same plane as the essential forces of cosmic life.” [16] Among humans, manyú- belongs to specific individuals who are typically hostile or wicked: strangers (1.104.2: dā́sa; 7.60.11: arí), enemies (6.25.2, 10.152.5: amítra), malicious persons of one sort or another. [17] Also attested is a more generalized manyú- in groups of people that does not escape the attention or surpass the power of Indra (7.78.6), Varuṇa (7.61.1), and Agni (8.71.2). For Malamoud, all these instances of apparently negative manyú- are in fact self-affirmations by individuals or groups that put them into active conflict or competition with others, not proof that manyú- itself is inherently negative. [18] He cites the following lines as evidence that one person’s negative manyú- can be counteracted by Indra’s positive manyú-:

Among the gods—principally Indra and those associated with him, but also (Mitra-) Varuṇa, Brahmaṇaspati, and Agni— manyú- is “the élan that makes [the god] accomplish the deeds that affirm his divinity.” [19] For instance, the manyú- of Mitra-Varuṇa correlates with their specific function as divinities:

For Agni, his manyú- is explicitly reflexive—what suits his particular desire and expectation:

Finally, Indra’s manyú-, the most commonly attested, was expressed when he slew the serpent Vṛtra with his thunderbolt as Sky and Earth trembled. [22] Here is a sample of passages:

It should be clear by now that the personages with whom Manyu is identified in the second strophe of Hymn 10.83—Indra, Varuṇa, Agni, and the tribes of humans—are in fact those to whom the noun manyú- is attributed in the earlier hymns. But Malamoud pushes his analysis of the word’s core meaning a step further to explicate the sense of two words attested as qualifiers of manyú-, a noun which rarely has any. One is the adjective satyá, the other the rare word svayaṃbhū́ (attested only at 10.83.4 in the Rig Veda). Malamoud defines satyá to mean “destined for accomplishment, sure to be realized,” such that manyú- so qualified is the “élan which makes someone bring to fulfillment their desires, translate their thoughts into deeds.” [23] The manyu– of Indra is satyá in 4.17.10 and causes universal fear; or the poet says at 10.112.8:

where satīná means ‘realized’. So in regard to the following assertion of the primacy of manyú- in Indra’s exploits, which was deprecated by Louis Renou on the grounds that it was a “contamination” caused by a rhetorical “condensation of formulas”:

Malamoud counters that such a “condensation” could not have taken place unless manyú-actually was a primordial determinant of deeds at the origins of the world, like Old Avestan mainiiu-. [24] Such a quality in manyú- is confirmed by the second qualifier explicated by Malamoud, the neologism svayaṃbhū́ ‘spontaneously generated’, a rare epithet elsewhere attested, for example, of the king in the rājasūya ritual, in which he “symbolically creates space and time, organizes the universe around him, and reproduces the cosmic process of creation and maturation.” [25]

In sum, Vedic manyú- has the following semantic and thematic attributes:

• It denotes not a transient emotion, but a permanent, primordial, conflictive, and active cosmic force among gods and men.
• It operates in all three of the domains of Indo-European society as reconstructed by Dumézil.
• It is fundamentally characteristic of Indra, in whom it is expressed by the fear-inspiring thunderbolt and fire.

Old Avestan mainiiu- shares in the first of these attributes, but not the others. Given the Zoroastrian function of mainiiu- at the zero point of the world, however, inherited features of the term may have been stripped from it. All three of these semantic and thematic attributes are present in Greek mênis and are especially strongly associated with Zeus. [26] In the Indic pantheon, however, Indra has taken over the functions of the inherited divinity represented in the Greek pantheon by Zeus, wielder of the thunderbolt and king of the gods. Accordingly, I would suggest that the semantics and thematics of manyú- listed above were associated in Indo-European with the root *men- and the lord of the pantheon. They were thence inherited, on the one hand, by Vedic manyú-, which is chiefly associated with Indra, and, on the other, by Greek mênis, which is chiefly associated in Greek epic with Zeus, his thunderbolt, and its destructive fire. Homeric mênis is also actively associated with the root noun of *men-, Greek ménos, [27] to which it is ultimately related.

The details of the connection between the root *men- and the word mênis are the subject of confusion and dispute. In order to clarify the issues, I first offer a summary of the three currently competing views on the formation of mênis. Then I will explain and justify my preference among them.

In 1931, Eduard Schwyzer proposed that the word was derived from the root *mnā– attested in the verbs mi-mnḗ-skō ‘remind’ and mn ā́ omai ‘be mindful of’. [28] With the addition of a noun suffix –ni-, Common Greek *mnā-nis became mânis (mênis in Attic-Ionic) when the first nasal consonant disappeared as a consequence of its dissimilation from the second one. Schwyzer found formal support for this hypothesis in the poetic expression mnā́mōn Mênis teknópoinos ‘mindful (from the verb mn ā́ omai) child-avenging Mênis’ attested in the opening choral ode of Aeschylus’s Agamemnon (155). But he subsequently lost confidence in the etymology, apparently because there is no parallel in Greek for the dissimilation of the nasal consonants. [29]

In 1977, however, Calvert Watkins revived the etymology that Schwyzer had himself withdrawn. Watkins tried to demonstrate that mênis is not only dangerous to provoke but also a dangerous word to utter. On his view, it is unthinkable that a speaking subject, god or Achilles, use the word to speak of his or her own mênis. Every time the occasion presents itself, persons speaking of their own mênis use a substitute, such as the words khólos or mēnithmós or ménos. [30] At the same time, Watkins asserted, mênis is only used once by any mortal subject; the one example, it turns out, occurs at a moment in the Iliad that requires an absence of reticence, when Phoenix asks Achilles to return to battle. [31] The usage restriction that Watkins noticed is significant for the etymology of the word, because if mênis is a tabu word, then, like other tabu words and phrases, it could have undergone unpredictable phonological deformation (as opposed to an otherwise unattested regular phonological change). The problem of the unparalleled loss of the nasal consonant in Schwyzer’s etymology is solved.

In addition, Watkins found more contextual evidence (like the Aeschylean phrase cited above) to confirm Schwyzer’s etymology. He showed that the word ópis is a near synonym of mênis with a straightforward etymology: it is a derivative of the root *op-, which, like the root of mn ā́ omai ‘remember, keep in mind’, denotes a form of sense perception. One can compare its verbal derivative opízomai ‘look at, pay attention to’, a word which is actually attested in epic with the noun mênis as its direct object. [32] As mênis is to ópis, so mnā́omai is to opízomai. Watkins pointed to passages in epic or lyric poetry where mênis is associated with or even defined by forms of the verb mémnēmai ‘remember’. Since the root *mnā (< *mneh ₂ -) is a “Theme II” enlargement of the root *men – (attested in Greek ménos, etc.), it must have been generated in Indo-European rather than independently in Greek and Sanskrit, even though the Theme II root is only attested in Greek and late Vedic Sanskrit. In support of this part of his hypothesis, Schwyzer had already found parallels to the suffix -ni- in Armenian and Old Church Slavonic, where it is attached to roots whose form is comparable to *mnā -. Watkins concludes that, along with the formation of its root, the suffixation of mênis is of Indo-European date. [33]

In 1985, Patrick Considine, in disagreement with Watkins and Schwyzer, revived a different neglected hypothesis on the etymology of mênis. It was first suggested in 1949 by Carl Darling Buck, [34] and it is most easily expressed as an analogy: baínō ‘go’: bā-, bē- :: maínomai ‘rage’ : mā-, mē-, whence Common Greek *mā-nis, Attic-Ionic mênis. The advantage of this hypothesis over Schwyzer’s is clear: the form of mênis can be explained in positive terms, in parallel to an attested phonological process, rather than by way of a unique dissimilation or a tabu deformation. [35] But the Buck-Considine etymology is certainly speculative. It explains one enigma by comparing another, since the alternation between baínō ‘go’ / bā- and bē- (*g ^ṷ em –/*g ^ṷ eh ₂ –), which can only be a morphophonemic phenomenon, is an isolated occurrence among the Indo-European languages. Considine tries to prove that phaínō ‘show’/phā-, phē- is a parallel to it, suggesting that Vedic bhánati ‘speaks’ presupposes a root *bhen- whose alternation with *bhā – would be the same as *g ^ṷ em –/*g ^ṷ eh ₂ -. [36] For phā- in contrast to phan- the only attested form is pephḗsetai (future of the verb phainō), a hápax legómenon that has been explained as analogous to bḗsetai ‘he will go’! [37] But an isolated form should not be the basis for analogical formations. Reduplicated future middles like memnḗsomai provide a better model, as Chantraine pointed out. [38] Considine’s whole argument is neither more nor less than a possibility. [39]

In terms of the phonological criteria of historical linguistics, it is clear, Watkins’s etymology is the strongest of the three. But the validity of his revival of Schwyzer’s etymology depends on the notion that mênis was a tabu word. Watkins supports that idea not by discussing the meaning of the word, but by citing a usage constraint on the root noun mênis in Epic diction: a speaking subject cannot utter the word in referring to his or her own mênis. But the value of this usage constraint as evidence of a language tabu has been seriously challenged. Jean-Claude Turpin has pointed out that tabu words are deformed so that they can be uttered. [40] Apparently, the deformation of mênis was a failure! Turpin also maintains that a tabu on the first-person usage of a word is unexampled and inconsistent with the structure of the relationship between the persons of the verb in general terms. In his view, if a first-person usage was forbidden, so should a second-person usage. [41] So Turpin willingly embraces Schwyzer’s original hypothesis despite its unresolved phonological difficulties. [42]

For my part, I propose to answer Turpin’s objections to Watkins’s hypothesis and to provide additional support for the notion that mênis was a tabu word: its attested meaning. First, with regard to the contradiction in maintaining that a word is at the same time tabu and deformed, since the purpose of deforming a word is to make it possible to utter it: the loss of one consonant, the second nasal of *mnā-nis, barely qualifies as a deformation, but the only way to salvage Watkins’s hypothesis is to suppose that the word, once deformed, has become tabu again. That seems doubtful at best. Yet I do know of one parallel: the English word pee was originally a children’s euphemistic form, like French pipi, for the vulgar word piss, but it is no longer perceived as a euphemism and has become mildly obscene. [43] If there is one example of such a phenomenon, there may well be others. It is plausible, it seems to me, that a deformed word can again become tabu if, for one reason or another, the word in question loses its oppositional relationship to the tabu word from which it was originally deformed. If the tabu word has disappeared completely, or if the derivational relationship between the two words is lost completely for other reasons, the original word’s functions can be taken on by the new, deformed word, which will then also acquire the same prohibitions as the old one, provided that the social conditions that motivated the deformation in the first place continue to flourish. Such a process is just a special case of the lexical renewal rule specified in Kurylowicz’s Fourth Morphological Typology, and, as Claude Sandoz has made clear, tabu words can and do disappear. [44] In the case that concerns us, it behooves us to suppose that the prohibition on the undeformed form of mênis was so strong and effective that this form has disappeared, while the reasons to prohibit the utterance of the word remained.

It is worth stressing, however, that even if such a semantic process and its consequence, phonological deformation, did not take place in the case of mênis, the word may still be tabu. The usage and semantic content of a word in its proper social context determine whether or not it is tabu, not its deformation, and for mênis, there are other plausible explanations of its form. Turpin realizes that the constraint Watkins has noticed—the absence of subjects speaking of their own mênis—requires an explanation. He has even added to Watkins’s dossier by pointing out that derived forms such as the verb mêníō are not attested in the first person. [45] According to Turpin, however, these usage restrictions can be explained by “the very nature of the idea.” Mênis is not “the anger of the gods,” but it is in itself a divine object; consequently, it is never said to reside in the mental equipment of Homeric persons as do other words meaning “anger.” The epic even takes the trouble to make explicit the emotional consequences of mênis as though they were not in fact part of its meaning, as though mênis were a “numinous object” independent of human feeling. That is why one cannot speak of one’s own mênis. [46]

Turpin’s observations on the independence of mênis are pertinent. Yet if it is not possible to attribute mênis to oneself, then why on his explanation is it possible, as he admits it is, to attribute mênis to “you” or “him” or “her”? Furthermore, his description of certain states of mind as external objects in the epic world is questionable. The strict separation between mind and body that is second nature to us cannot be projected upon the Homeric world, since the physical and external aspect of mental states that we consider immaterial and interior is a constantly recurring feature of epic language and thought. In this respect, mênis does not seem to me different from a term such as átē ‘moral blindness’, also a divine and numinous force: Agamemnon wishes to represent his átē as something distinct from himself, but in a context where he intends to dissociate himself from it and deflect blame for what he has done. [47] Mênis is independent and “almost a hypostasis,” as Turpin does well to point out, but these qualities do not require that epic personages cannot consider it to be theirs.

The last problem that Turpin raises is the credibility of a tabu on the first grammatical person that does not also fall upon the second (and for that matter, also the third). In fact, Greek epic contains at least one exact parallel to such a tabu. When epic characters are asked their identity, they do not respond with their name, only with their patronymic and place of birth; but in the second and third persons, people often refer to each other by name with or without a patronymic. [48] The prohibition against speaking one’s own name can, however, be violated in special circumstances, such as the moment when Odysseus finally identifies himself to the Phaeacians by his name and his patronymic, or when a goddess reveals herself to mortals. For Sandoz, who describes a whole range of language tabus from “systematic avoidance” to “simple reticence,” such flexibility in the rules of speech is predictable and commonplace. [49] For an example of an exclusion of the first person in contrast with the two others that is closer in context to mênis, I cite Walter Burkert’s dictum on Greek funerary customs and beliefs: “Ritual and belief are almost exclusively concerned with the death of others; one’s own death remains hidden.” [50] In this perspective, the prohibition against using the word mênis by a subject who is speaking of his own mênis can be motivated by reasons and feelings comparable to those evinced by one’s own death. With respect to the horror surrounding mênis, it suffices to consider the complete list of verbs of which it is the direct object in epic diction: hupodeídein ‘be afraid of’, aporrhí̄ptein ‘hurl away’, aleúesthai ‘shun’, opízesthai ‘watch out for’, apoeipeîn ‘renounce’. [51]

So it is possible to salvage from Turpin’s critique the hypothesis that mênis is a tabu deformation. Even so, the extent to which mênis is tabu needs to be qualified. According to Watkins, epic persons behave as if forbidden to speak of their own mênis. Instead of uttering it, they employ a substitute word, such as ménos, khôlos, mēnithmós, or ópis. But in fact this speech tabu is not very stringent. On two occasions, Achilles speaks of his own mênis with a verbal derivative of the root noun, (‘while I have / had mênis’—emeû apo mênis antos: Iliad 9.426, 19.62), and on two other occasions he uses the secondary suffixed form, mēni-thmós (16.62, 202). A gross offense to convention does not seem intended. The relaxation of a speech tabu for secondary forms of a tabu word is possible but not automatic. If the derivational link is apparent to native speakers, and the denotation of the founded form is unchanged from the founding form, it is almost difficult to credit. To cite an example of the relaxation of the tabu for a founded form, a word like “Christian” has no tabu attached to it although “Christ” is tabu in many contexts (as in ‘taking the Lord’s name in vain’), because the semantic reference and usage of the tabu term and its derivative are not the same. On the other hand, French emmerder in its transitive form is as vulgar as its founding form merde, because the semantic and the derivational links are transparent and in force; but the reflexive form of the same verb, s’emmerder, is not vulgar. When Achilles says emeû apo mênis antos in Iliad 9.426, he is patently referring to the theme of the poem expressed in its first line with a verb overtly derived from the noun. The same is true of the other instances cited just above where he uses a mênis derivative. Moreover, Watkins mistook the overall rarity of mênis in spoken discourse. [52] Epic mortals (Kalkhas, Iliad 1.75; Aeneas, 5.178; Phoenix, 9.517; Menelaos, 13.624; Telemachus, Odyssey 2.66; Nestor, 3.135; Aithon/Odysseus, 14.283) utter the root noun as frequently as the poetic narrator (six times) and more often than the gods (Hermes, Odyssey 5.146; Athena, Iliad 5.34; Thetis, 19.35; Aphrodite, Hymn to Aphrodite 290). Nor is the distribution of mênis derivatives meaningfully different from that of the root noun. It is spoken by mortal speakers eleven times, by the narrator seven times, and by gods once. On each occasion the derivative either refers to an otherwise identifiable instance of a nonderived form of mênis, or it occurs in a context comparable to one in which the root noun is also used. [53]

On balance, then, and with the qualifications stated as to the nature and strength of the speech tabu, Watkins’s revival of Schwyzer’s etymology appears to be the most consistent with the word’s meaning and usage. In fact, the absence of a stringent speech tabu against mênis supports Watkins’s overall argument better than its presence. It seems plausible to argue that what once was a tabu word has been deformed and is therefore relatively utterable. Reticence is by no means ruled out by the semantic or the phonological evidence, and reticence is in fact a diminished form of speech tabu. [54] This discussion inevitably raises a further question about the semantic content of mênis that no one has yet addressed. To put it simply, why should this word be tabu? Whence the horror surrounding it? The hypothesis that mênis is a tabu word may be sustainable on the basis of its usage and despite Turpin’s pointed objections, but it is still not clear why a word that means “solemn anger,” as Watkins defines it, should be dangerous to speak (and possibly still dangerous even after deformation). The answer to this question, however, imposes itself after a systematic review of the usage of the word in context like the one undertaken above in Chapters 1 and 2. Mênis is an emotion that acts to change the world. It is not a word for “solemn anger” but the sacred name of the ultimate sanction against tabu behavior, and epic personages invoke it to forestall people from breaking fundamental cosmic rules. So it would not be surprising if the word mênis were endowed with the same dread as the behavior that can bring it down upon a whole community. It is also economical if a language tabu explains both the usage and the form of mênis.