En poursuivant votre navigation sur ce site, vous acceptez l'utilisation de cookies. Ces derniers assurent le bon fonctionnement de nos services. En savoir plus.

  • Nick Allen and Roger D. Woodard - Hermes and Gandharvas

    Hermes and Gandharvas


    Nick Allen* and Roger D. Woodard**


    *Oxford, **Buffalo



    NB: Les caractères grecs et les signes diacritiques étant mal gérés par notre plate-forme, nous prions nos lecteurs de bien vouloir consulter cet article en téléchargeant la version pdf.


    Abstract: The dossier of the Greek god Hermes is compared with that of the Indian demigods called Gandharvas (who sometimes appear as a singular being). In both regions much cultural material bypasses the oldest sources to surface in later ones, so that in India the comparison can draw on the epics no less than the Vedas. The points of comparison are organised into five themes that link Hermes with Gandharvas in general, and twelve themes that link him with individual Gandharvas (or associated figures) – Viśvāvasu, Citrasena and Citraratha, Purūravas, Nārada, Dhṛtarāṣṭra, and Kubera. In the light of the comparisons it is proposed that the Greek and Sanskrit figures derive from an early Indo-European common origin, which itself was rooted in the Dumézilian third function.

    Keywords: Greek mythology; Sanskrit mythology; Indo-European cultural comparison; Pan; Kubera; herms; invention of the lyre; Dumézil’s third function

    Résumé: Le dossier du dieu grec Hermès est comparé avec celui de demi-dieux indiens nommés Gandharvas (qui parfois apparaissent au singulier). Dans les deux domaines, bien des matériaux culturels, négligés des sources les plus anciennes, ont fait surface dans les plus récentes, ce qui oblige a appuyer la comparaison plus sur les épopées que sur les Védas. Les points de comparaison sont organisés selon cinq thèmes liant Hermès aux Gandharvas en général, et douze thèmes qui le lient avec des Gandharvas individuels (ou à des figures associées) : Viśvāvasu, Citrasena and Citraratha, Purūravas, Nārada, Dhṛtarāṣṭra, et Kubera. A la lumière de ces comparaisons, nous proposons que les figures grecques et sanscrites soient issues d'une origine indo-européenne commune en rapport avec la troisième fonction dumézilienne..

    Mots clés : Mythologie grecque, mythologie sanscrite, comparaison culturelle indo-européenne, Pan, Kubera, hermès, invention de la lyre, troisième fonction dumézilienne

    Télécharger l'article en pdf / download in pdf: Allen et Woordard.pdf

    Comparative mythology can employ a variety of methods, singly or in combination, but no doubt one of the most cogent is the comparison of narratives. Maximally simplified, and pruned of all its normal complexities (locality, timing, agents, narrator, tropes…), a narrative consists of a sequence of events. So to compare two narratives is like comparing structure a-b-c-d-e with structure A-B-C-D-E. Rapprochements are sought between events a and A, b and B, etc., but also between relations, whether sequential (as a-b and A-B) or non-sequential (a-d and A-D). The cogency of the comparison turns on the number, quality and variety of the rapprochements (cf. Allen 2010).


    Allen Fig 1.JPG

    Figure 1. Model of the comparison of two narratives.

    The circles represent events or episodes while the vertical lines

    represent individual rapprochements.


    The present paper draws on a number of narratives, but it does not study them as wholes. Instead it extracts from them particular motifs or facts in order to compare two dossiers (collections of facts about agents). Relationships may exist between these facts but the analyst who looks for them must do so without the helpful lineality and finiteness of a narrative. It is as if a…e and A…E are jumbled together in their respective dossiers, which offer no obvious starting point, end-point, articulation or boundaries. A dossier can hardly aspire to completeness, since this would imply inclusion of the agent’s associates and associations – a domain of potentially indefinite extent. Moreover, to assess the significance of any individual rapprochement (say b-B), and hence the cogency of the whole comparison, is harder when each entity is drawn from its own amorphous cloud-like mass than when each has its place within an organized whole. Finally, the comparativist cannot expect that in practice each agent will have one and only one parallel in the other tradition. On the contrary, agent p in one tradition may resemble, not only agent P in the other, but also Q and R, while P may resemble not only p but also agents s and t – and not resemble q or r. One-to-one correspondence between dossiers, as in Figure 2, is probably the exception rather than the rule. To base a comparative article on dossiers rather than on narratives is therefore to take on a harder challenge. But although we are aware of the methodological difficulties, we hope that the method gives persuasive results in this case1.


    Allen Fig 2.JPG

    Figure 2. Model of the comparison of two dossiers

    To think about the delimitation of dossiers is to raise the question of the dates at which facts are attested. Both Indologists and classicists are well aware that their earliest sources do not present a full picture of the cultures in which their texts reached their current form, but they almost inevitably think in terms of pre-classical leading on to classical. Cultural comparativists (more so than their specialist linguistic colleagues) are at ease with a different model: Indo-European heritage contributes to the early sources but also bypasses them, so as to surface in later ones (Figure 3)2. This second model makes it legitimate to combine classical and preclassical in each of our two dossiers. Thus although the composition of the Vedic hymns is often dated to a millennium earlier than the Mahābhārata, much of our material on the Gandharvas comes from the latter. Similarly, we suppose that some traditions about Hermes bypassed the earlier sources (including the ‘Homeric’ Hymn to that god).


    Allen Fig 3.JPG

    Figure 3. Model of the field of study of this paper.

    The bold arrows emphasize the bypass, but are not intended to have quantitative implications.


    The Gandharvas are one among the many categories of supernatural beings recognized in the Indian religious tradition. They are usually situated in the middle ranks of the spiritual world, being seen less as gods or demons than as demigods. They are conceptually particularly close to groups of indefinitely numerous spirits such as Yakṣas or Kiṃnaras, with whom their dossier tends to overlap. They are best known as celestial musicians who sing to the lyre (gāndharva means ‘music’), and are often associated with Apsarases (‘nymphs’), who are celestial dancers.

    The Rig Veda stands somewhat apart, since it does not link Gandharvas with music and usually speaks not of a group but of an individual. Since the Avesta too recognizes only a single Gandarǝβa, it is often thought that the pluralization is historically secondary. At first sight this view would favour our comparison with the single figure of Hermes, but we leave the issue undecided.

    A dossier on Vedic Gandharvas is provided by Macdonell (1981:136-8, under ‘Lower Deities’), and one on epic Gandharvas by Hopkins (1986:152-9, under ‘Hosts of Spirits’)3. Monier Williams (1974) has a good summary. Somewhat arbitrarily, we have not used sources later than the epic. Hermes being much more widely known, the literature on him is copious, but it is readily accessible via handbooks and encyclopaedias (e.g. Burkert 1985:156-9, Jost 2012). Naturally we shall draw heavily on the Homeric Hymn to Hermes (hereafter HH Hermes)4, as well as referring to specialized studies.

    Earlier comparativists often linked Gandharvas with Centaurs, relying fairly heavily on the similarity of the names. The linkage is best represented by Dumézil (1929), which brings together within a common-origin framework four main dossiers: Central European carnivals, the cognate Iranian and Indian mythic beings, Centaurs, and the complex surrounding the Latin word februum. Nowadays the etymological link between Gandharva and Centaur is rejected by specialists, and Dumézil himself came to disparage his 1929 book, along with most of what he wrote before his breakthrough recognition of trifunctionalism in 1939. Moreover, Keith (1937:39) argued that the similarities between the two types of supernatural were ‘wholly overstated by Hopkins (157f)’. In particular, epic Gandharvas are never presented with the mixed horse-man form so characteristic of Centaurs. Similarly, Panchamukhi (1951:49) writes:

    The Gandharvas as a class are not known to possess a horse-head either from the literature or sculptures, though in lexicons, the word conveys among several other meanings, the sense of a horse. It is only the Kinnara that is definitely described with horse features.’


    Even so, Doniger and Smith (1991:7n37) still systematically translate Gandharva as ‘Centaur’, and the dossier of the individual Centaur Chiron presents several interesting comparisons with Gandharvas (Vielle 1996:134-6, Sterckx 2002:34-5; more detail in Vielle 2005).

    Comparativists have naturally made other suggestions about both our comparands. By way of illustration, with no claim to completeness, here are a few examples. Referring to his long-standing belief in an original connection between Rudra-Śiva and Dionysus, Schroeder (1908:19) says that the thiasos or cult-group of the latter, viz. the Satyrs, Sileni and Nymphs, have in India ‘their closest and most immediate relatives in the host of Gandharvas and Apsarases.’ Following a suggestion by Oldenberg (1993:170n352), Oberlies (2000:380) argues for a Pūṣan–Hermes comparison (his title relates to the concluding sentences of his paper). Hocart (1970:16-22) ventured a brief Agni–Hermes comparison, and van Berg (2002), while justly criticizing Hocart’s attempt, has pursued the same idea5. However, it seems that the Gandharva–Hermes comparison has not previously been seriously envisaged. We shall focus first on the Gandharvas as a category, then on some of its individual members or near-members.


    Gandharvas in general


    1. Main Wife

    In many contexts Arjuna, the central hero of the Mahābhārata, is cognate with Odysseus, the central hero of the Odyssey, and the same applies to their respective wives, Draupadī and Penelope. Both heroes marry or have sex with other females, but these two are their main wives. Draupadī’s marriage is polyandrous – she marries all five Pāṇḍava brothers; but it was Arjuna’s archery that made the union possible. Draupadī is totally faithful to the brothers, but her virtue is several times tested.

    The relevant instance occurs during the thirteenth year of the Pāṇḍavas’ exile, which they spend in disguise in the realm of King Virāṭa. Draupadī disguises herself as a lady’s maid in the service of the queen, and claims that she will be protected by her five Gandharva husbands. Each of her husbands has his own individual disguise, and she does not publicly identify them as Gandharvas, nor, it seems, do the inhabitants of the kingdom recognize what is obvious to the listener or reader. Presumably the tradition fixed on this particular category of supernatural because Arjuna was disguised as a musician, singer and dancer. In any case, when the queen’s brother Kīcaka attempts seduction, he is killed in private by Bhīma (the second Pāṇḍava brother), and she ascribes the killing to her (apparently invisible) Gandharva husbands; similarly, the next day, Bhīma kills 105 followers of Kīcaka, and again the massacre is attributed to a Gandharva (4,13-22). We can say that, within Book 4, the ‘real’ Pāṇḍavas are identified with the ‘fictional’ Gandharvas.

    In Homer Penelope too is totally faithful in her monogamous marriage to Odysseus, but not all sources agree. In Book 2 (section 145) of his History, Herodotus observes that the Greeks (in opposition to the Egyptians) identify Heracles, Dionysus and Pan as the youngest of the gods. He then goes on to place each of the three in a relative chronology in which Pan holds the most recent position. The god was fathered after the Trojan War by Hermes, when he impregnated Penelope: ἐκ ταύτης γὰρ καὶ Ἑρμέω λέγεται γενέσθαι ὑπὸ Ἑλλήνων ὁ Πάν, ‘for from her and from Hermes, say the Greeks, came Pan’. This is the earliest attestation of such a union, though perhaps not of the maternity of Pan that Herodotus endorses: Hecataeus of Miletus, as well as Pindar (FGrH 1 371), are reported to have identified Pan as the offspring of Apollo and Penelope (though see Brown 1981:64).

    Herodotus is not alone in reporting this tradition. Cicero (De natura deorum 3.56) likewise holds Pan to be the child of Hermes and Penelope, as do the mythographers Pseudo-Apollodorus and Hyginus. The latter states only this much (Fabulae 224). The former, more expansively (Bibliotheca Epitome 7.38-39), writes that, according to some, the suitor Antinous seduced Penelope, whom Odysseus consequently sent away to her father. She then travelled on to Mantinea in Arcadia (cf. Pausanias 8.12.5-6, who saw what was locally claimed to be Penelope’s tomb); it was there that she gave birth to Pan, having been impregnated by Hermes (presumably in Arcadia). Pseudo-Apollodorus also reports the alternative tradition that it was the suitor Amphinomus with whom Penelope was unfaithful, and that for her infidelity Odysseus killed her. Lucian constructs his Dialogues of the Gods 22, Πανὸς καὶ Ἑρμοῦ, around the tradition that Hermes and Penelope were the parents of Pan.

    More striking is the account, earliest attested by Servius (Aeneid 2.44), that Penelope had intercourse with all of her suitors and that from these manifold copulations there emerged a single issue – Pan: sicut ipsum nomen Pan videtur declarare, ‘just as the name Pan itself appears to declare’. The Byzantine scholar Johannes Tzetzes, citing the historian Duris of Samos (fourth/third century BC), likewise records the tradition that Pan was fathered on Penelope by all of her suitors (Scholia in Lycophronem 772). A scholion on Theocritus (1.3) states the same, making the etymological connection to which Servius alludes – that the god Πάν takes his name from the Greek adjective πᾶς, πᾶσα, πᾶν ‘all’ because he was fathered by all the suitors6. It is of course a folk etymology, and false: the god’s name was earlier Πάον (dative Πάονι in the sixth-century Arcadian inscription IG 5 2,556). A common origin with the Sanskrit divine name Pūṣan has been proposed but is far from universally endorsed (see Chantraine 1968:855).

    Of the two main wives, Draupadī is in a foreign court and in disguise, while Penelope is in her own court and her identity is known to all. Nevertheless the situation of the two women is comparable. Both are under pressure from would-be seducers or suitors, and both are in the vicinity of their husbands but cannot or do not recognise them. The killing first of Kīcaka, then of his followers, parallels the killing of Antinous, then of the other suitors (Allen 2002). But if the two main wives are cognate, what of their children?

    For each of her husbands, Draupadī gives birth to one son, the whole set being known as the Draupadeyas. These five, her only children, all of them killed in Book 10, are somewhat pallid characters, sharing little but their generation with Telemachus, the single son of Penelope and Odysseus. However, Penelope also bears the god or demi-god Pan, who is occasionally pluralized; and the Draupadeyas–Pan comparison is more interesting. Two accounts are given of the ontology of the pentad. At the start of the epic they are said to incarnate the groups of Viśvedevas or ‘All-Gods’ (1,61.88), but at the end they are clearly stated to be ‘highly blessed and energetic Gandharvas’ (18.4.11). The Viśvedevas (Macdonell 1981:130, Hopkins 1986:173-5) are a curious category of deity, whose name transparently contains the common adjective viśva meaning ‘all’, thereby paralleling the folk-etymology of Pan. The actual derivation of the Greek name does not detract from the rapprochement, which can be stated as follows: both main wives have sons who were associated by tradition with the notion of totality. This remains true however the notion is expressed, whether within the name of the incarnating gods, in the plurality of human suitors regarded as Pan’s fathers, or even in the whole set of Pāṇḍava fathers (who themselves incarnate deities). But if the sons of the two main wives are comparable, perhaps the same applies to their partners, and in particular to the Gandharvas and Hermes.

    The complex consisting of main wife, partners and sons is differently organised in the two cases, and the comparison between them is not among the most straightforward of our rapprochements. However, it is placed first since it was the starting point for this study (cf. Allen 1997:150-151).


    2. God-human margin

    As the term ‘demigod’ suggests, Gandharvas hover on the margin between men and celestial gods. Systematic surveys of Vedic or Hindu mythology tend to place Gandharvas at the end of the section on gods proper, just before mythic priests and heroes. One Upanishad (Tait. Up. 2.8) offers an ascending ranking of types of bliss: bliss associated with humans, with human (manuṣya-) Gandharvas, with celestial (deva-) Gandharvas, with pitṛs, gods… This explicit split within the category is not particularly common but, as Hopkins notes, the term covers human minstrels as well as divine ones. Nārada, as we shall see, is both a semi-divine Gandharva and a mythical human sage. Banerjea (1956:335 ff.) treats Gandharvas and other types of demigod under the heading of vyantara devatas ‘gods of intermediate position’, using an expression borrowed from the Jains. Pali tradition places the Gandabbā lowest among the devas, associating them with Asuras and Nāgas (Malalasekera 1960).

    As for Hermes, investigators have long drawn attention to what they perceive as his subservient status among the deities: “. . . Hermes resterà in tutta la tradizione greca, in una posizione subordinata tra gli dei7. When Zeus commands that Hermes lead Apollo to his stolen cattle, the HH Hermes (395-6) describes the young god’s response in these words: ἐπεπείθετο δ’ ἀγλαὸς Ἑρμῆς· | ῥηϊδίως γὰρ ἔπειθε Διὸς νόος αἰγιόχοιο, ‘And shining Hermes was persuaded; | for the mind of Aegis-bearing Zeus was quickly persuading’. But it is not clear that the rapidity of Hermes’ response to Zeus’s ‘nodding’ (νεύω) would be uncharacteristic among Olympians (compare Il. 1.528). The version preserved in the Bibliotheca of Pseudo-Apollodorus (3.10.2), says that Hermes’ response to Zeus is to deny that he stole the cattle.

    Performance of various tasks, potentially or actually menial, is attributed to the god. In Aeschylus’ Prometheus Vinctus, Prometheus refers to Hermes derisively as ὁ Διὸς τρόχις, ‘the courier of Zeus’ (line 941); as ὁ τοῦ τυράννου τοῦ νέου διάκονος, ‘the servant of the new tyrant’ (line 942); as θεῶν ὑπηρέτης, ‘underling of the gods’ (line 954; cf. 983). Sappho knows Hermes as οἰνοχόος ‘wine-pourer’ for the gods (fr. 141), as does Alcaeus (fr. 141); and in the twenty-ninth Homeric Hymn, Hermes and Hestia are praised in tandem, Hermes being invoked to assist (ἐπαρήγω) together with Hestia (lines 10-11)8. Aristophanes presents Hermes as having been left behind when the gods vacated heaven (Pax 201-2) – left to take care of (τηρέω) the things they have abandoned: χυτρίδια καὶ σανίδια κἀμφορείδια ‘ little pots and boards and jars’. At Ion 4, Euripides makes Hermes identify himself with the phrase δαιμόνων λάτρις, ‘servant of the gods’. In his dialogue between Maia and Hermes (Dialogi deorum 4:1), Lucian has Hermes ask the nymph: ἔστι γάρ τις, ὦ μῆτερ, ἐν οὐρανῷ θεὸς ἀθλιώτερος ἐμοῦ; “Is there any god on Olympus more wretched than I am, O mother?” Hermes then goes on to complain about his many tasks (πράγματα), such as σαίρειν τὸ συμπόσιον, ‘to clean the sympotic space’.

    A lowly position among the gods in itself implies a degree of closeness to mortals; and when Zeus sends Hermes to help Priam recover Hector’s body, he explains his request on the grounds that the godling particularly enjoys being the companion or friend of human beings (Il. 24.334 -5). Presenting himself as a possible servant for the suitors, Odysseus claims that it is by favour of Hermes that he is good at menial tasks (Od. 15.319). Aristophanes has a chorus address Hermes as φιλανθρωπότατε καὶ μεγαλοδωρότατε δαιμόνων, ‘the most philanthropic and bountiful of divinities’ (Pax 390 ff.).


    3. Wings

    Though gods can ordinarily move easily around the cosmos, only a few are described or depicted as winged. Gandharvas can be found in various places, in royal or divine courts, in waters, in forests or trees, but the Vedas link them prominently with heaven and the mid-air, and Oldenberg (1993:125) suggests that this was their original location. If so, it is not surprising that they should fly, and in iconography (so in the post-Vedic period), Gandharvas tend to have their upper half human (but with wings attached to their shoulders), their lower half bird-like (Banerjea 1956:281, 351-3). In the epic they are sometimes referred to as khecara or khacara, ‘sky-rangers’.

    In Homeric epic Hermes moves from place to place by flight. Thus, at Il. 24:339-346, in describing Hermes’ journey to Troy (to guide Priam), the poet sings that, with staff in hand, πέτετο κρὰτυς Ἀργεϊφόντης, ‘the strong Slayer-of-Argus flew’ (line 345). He does so after having bound beneath his feet his καλὰ πέδιλα | ἀμβρόσια χρύσεια, ‘beautiful sandals – immortal, golden’ (lines 340-41). Both phrases recur in the Odyssey (5.49, 44-5) as Hermes sets off for the island of Calypso, to communicate Zeus’ command that Odysseus be permitted to depart. Given the poet’s formulaic specification that the flight was accomplished after sandals were tied on the god’s feet, one might suspect that the imagined footwear would be the winged shoes with which Hermes is typically associated9.

    In Greek art Hermes is depicted with winged attributes as early as the end of the seventh century BC10. Dated to this period is a Melian amphora (National Museum in Athens) bearing an archaic image of Hermes who is shown wearing endromides; from these extend posteriorly large red wings (Yalouris 1953:295, with Fig.211). From only slightly later (ca. 600-590 BC) survives an olpe bearing an image of Hermes standing between two sphinxes12 (National Museum in Athens; LIMC 5.2, 230): the god wears shoes from which smaller wings project forward. Archaic representations of the god, however, occasionally present him as having a winged body. On a kylix of ca. 540 BC he is depicted not only as wearing winged endromides but as having two large wings, attached at the midline of his chest, with each wing extending back across a shoulder; and a painted image of ca. 530 BC is similar except in that the god’s wings are attached to his back (Chittenden 1947:101, with Pl. XXI, a & b)13.

    4. Wealth and livestock

    The single Gandharva typical of the Rig Veda is sometimes accompanied by the epithet viśvāvasu ‘possessing all goods’ (vasu- means ‘wealth, goods, riches, property’); and in the often cited ‘Marriage Hymn’ (RV 10.85), ‘the epithet is used by itself to designate Gandharva’ (Macdonell). Several of the topics raised by Macdonell will recur later, notably Viśvavasu as an individual, and wealth personified in the god Kubera, but the present section focuses on just one form of wealth. In the pastoral economy of the early Indo-Iranians, livestock must have been particularly important.

    Gandharvas do not look like horses but are linked with them in many ways. Twenty-seven Gandharvas first yoked the steed and placed swiftness in it (Tait. Saṃh., cf. ŚB The feminine form of their name is significant here, for Surabhi (‘Fragrant’)14 had two daughters: Rohiṇī, who gave birth to cows, and ‘the famous Gandharvī’, who gave birth to horses (Mbh. 1,60.65). The animals tend to be presented to epic heroes by Gandharvas. Citraratha promises to give to each Pāṇḍava brother 100 horses of the kind bred and ridden by Gandharvas, horses that are divinely fragrant, as speedy as the wind, and possess magical powers (1,158.45-6). At Yudhiṣṭhira’s rājasūya the king receives many gifts, but only two from Gandharvas: Citraratha (again) gives him 400 horses and Tumburu gives him 100 (2,48.22-23). According to the Vulgate Śikhaṇḍin too receives war horses from Tumburu (7,22.13). The chariot of Kubera, god of wealth, is yoked to Gāndharva horses (3,158.23).

    In Greece the second Homeric Hymn to Hermes (Hymn 18) opens as follows:

    Ἑρμῆν ἀείδω Κυλλήνιον Ἀργειφόντην

    Κυλλήνης μεδέοντα καὶ Ἀρκαδίης πολυμήλου


    Hermes, I will sing, the Cyllenian Slayer-of-Argus,

    Guardian of Cyllene and Arcadia rich-in-sheep


    At the end of this short hymn (in line 12), Hermes is addressed as δῶτορ – poetic form for δοτήρ – ἐάων ‘giver of good things’. The phrase recurs as a characterization of Hermes at Homeric Hymn to Hestia 8 as well as at Odyssey 8.335, with the plural nominative occurring in the same formula ten lines earlier, where the gods (θεοί) are called δωτῆρες ἐάων ‘givers of good things’. Hesiod (Theogony 46, 111, 633, 664) uses the same phrase to distinguish the Olympian deities from the Titans. Both Durante and Schmitt15 have drawn attention to the great antiquity of the formula, comparing Sanskrit dātā vasūnām, a Vedic epithet applied to Indra (see RV 8.51.5; compare 6.23.3, 10.55.6). As the comparison suggests, Greek ἐΰς ‘good’ is almost certainly cognate with Sanskrit vasu-, a term denoting ‘good’ and ‘goods’, as we have just seen, and, hence, applied as a categorical name to the deities of the realm of fertility and wealth, the third-function gods.

    Related themes are found elsewhere. By Polumele, a young woman whom Hermes encountered as she danced in the chorus of Artemis, he fathers a son who is named Eudorus (Εὔδωρος) ‘Good-gift’ (Il. 16.179-186). In his Works & Days (lines 69-82), Hesiod tells of the creation of the first mortal woman. Among the things that Hermes gives her is the name Pandora (Πανδώρη) ‘All Gifts’16, an appellation that can be used both to name a chthonic goddess and as an epithet of Earth17; and this δῶτορ ἐάων ‘giver of good things’ then gives her, Pandora, as a gift to Epimetheus, brother of the trickster Prometheus. She will be a bane to mortals – this is Hesiod’s sole verdict – but the only means by which progeny, human fertility, is realized (Theogony 603-12). Hermes is not so much the ‘giver of good things’ in the Pandora tradition as Hesiod weaves it into his epics (without using the identifying phrase δῶτορ ἐάων), but the ‘giver of goods’ – the stuff of fecundity, the fundamental elements that are the fruits of toil.

    The archaic application of the epithet ἐάων to Hermes specifically is completely consistent with his role as deity of fertility and abundant flocks and herds; and the ambiguity that results from a broader Greek usage of the formula appears to be equally consistent with more primitive Indo-European language of cult. Below we will draw attention to Hesiod and his conjoining of Hermes with Hecate as deities that bring increase for the herder (Theogony 444-7). When the swineherd Eumaeus entertains the disguised Odysseus in Odyssey 14, he slaughters a boar, makes offerings of bits of raw flesh wrapped in fat to all the gods and, after roasting the dressed hog, sets aside cooked meat as an offering to the nymphs and to Hermes, ἐπευξάμενος ‘after praying’ (lines 435-6). A scholiast on line 435 draws attention to Semonides (fr. 20 [West 1972]), who writes of shepherds sacrificing to nymphs and Hermes, οὗτοι γὰρ ἀνδρῶν αἷμ’ ἔχουσι ποιμένων, ‘for they are kindred with shepherd men’. The compounding phrase ποιμένες ἄνδρες ‘shepherd men’ (that is, ‘shepherds – men’) is an interesting one that recurs in Sappho fr. 105c1 and Alcman fr. 56, in which latter it receives contextualization by reference to an offering of a cheese made from lion’s milk that is dedicated to Hermes. Aristophanes has the chorus of women celebrating the Thesmophoria call upon Ἑρμῆς νόμιος ‘Hermes of shepherds’, together with the nymphs and Pan, to be pleased with their dancing (Thesmophoriazusae 977-81)18. Compare Homeric Hymn to Pan 28-47, where Pan and the nymphs sing of Hermes and of how he came (lines 30-31) ἐς Ἀρκαδίην . . . μητέρα μήλων, ‘into Arcadia . . . mother of flocks’. This was the location of the temenos of Cyllenian Hermes, where, despite being a god (θεὸς ὤν), ψαφαρότριχα μῆλ’ ἐνόμευεν, ‘he herded coarse-haired flocks’ (line 32)19.

    The Sanskrit emphasis on horses contrasts with the Greek emphasis on flocks and herds. However, as we shall see, Hermes is also given a μάστιξ or whip, which is notably used for driving horses (so Chantraine 1968:670).


    5. Drugs, crops and wealth

    Both our comparands relate to the vegetable world as well as to livestock. Vedic Gandharva guards the plant soma, and Macdonell suggests that this may help to explain his appearance in a charm for the treatment of impotence (AV 4.4.1).

    Thou plant that Gandharva dug for Varuṇa when he had lost his virility, we dig thee up as a medicine (oṣadhī) that excites the penis and causes erection (śepaharṣaṇīm, from √hṛṣ).


    When Odysseus is on his way to his first encounter with Circe, Hermes comes to meet him, promising to protect him against the witches’ wiles (Odyssey 10.287-306). He foretells that she will attempt to drug him and turn him into a pig, repeating what she has already done to those of his crew who had entered her palace earlier on. Odysseus is to threaten Circe with his sword and force her to take an oath not to harm him when they make love; otherwise, when she has his clothes off, she may render him feeble. But the hero can only resist the witches’ attack because Hermes gives him a potent herb, a φάρμακον ἐσθλόν, which the gods call molu (μῶλυ), which may share a common origin with Sanskrit mūlam ‘root’ (Heubeck and Hoekstra 1989:60). The god draws it from the ground and explains its properties. Its root is black, its flower milky white. It is difficult for mortals to dig it up, but gods can do anything.

    What the Gandharva does for Varuṇa and what Hermes does for Odysseus differ in all sorts of ways. A ritual with its foundation myth contrasts with a one-off epic event. The Sanskrit recipient of help (‘the patient’ – Varuṇa or his human counterpart) has already suffered, while Odysseus is only at risk of suffering; the drug is used respectively for therapy and for prophylaxis. Human helpers – the ‘we’ in the Sanskrit – are absent from the Greek. To obtain the drug, the Sanskrit helpers, whether divine or human, have to dig (khan-) whereas, even if Greek mortals have to dig (ὀρύσσω), Hermes can simply pull it up (ἐρύω). The root of the unnamed Sanskrit plant is only implicit (i.e. in the necessity to dig), while the colour at the root (ῥίζη) of μῶλυ is explicit. Varuṇa’s impotence is probably made explicit in his epithet mṛtabhraja, where mṛta- means ‘dead’; though the second element is etymologically obscure, its thrust is made clear by the result of the treatment. In contrast, Hermes’ reference to impotence is oblique: without the oath Circe will make the hero κακός καὶ ανήνωρ, roughly ‘feeble and unmanly’ (cf. ἀνήρ ‘man’). The females implicit in the concept of impotence go unmentioned in the Sanskrit, whereas Circe is presented at some length.

    Despite the many differences, both passages refer to a divine helper (Gandharva // Hermes), who supplies a patient (Varuṇa or human counterpart // Odysseus) with a treatment for impotence (the opposite of fertility), a treatment consisting in a plant whose root is important. Moreover, a further passage from the same Veda confirms the attribution of botanical knowledge to Gandharvas. The whole hymn extols the excellence of medicinal plants, and is used in therapeutic ritual.

    The boar knows the plant; the mongoose knows the remedial (plant). The plants that serpents and Gandharvas know, I call on them to aid this man (AV 8.7.23).


    Referring to these two passages, Dumézil (1929:150) remarks that ‘the Atharva Veda does indeed present the Gandharvas as doctors par excellence’. However their association with the plant world goes beyond drugs. They and their partners can live in certain species of tree that have religious associations (Gonda 1962:124, citing Tait. Saṃh.; and three of them have names relating to cereals. The Gandharvas who comment on the correctness of certain offerings to Agni are ‘Yavamān (rich in barley), the winnowing basket; Uddālavān (rich in paspalum frumentaceum), husbandry; and Antarvān (the pregnant), grain’ (ŚB, with Eggeling’s comment in his translation). Despite the obscurity of this passage it provides a parallel to the fact that Hermes’ powers over wealth and fertility are relevant not only to stockmen but also to cultivators.

    In HH Hermes 529-30 the golden staff that Apollo presents to Hermes is described as ὄλβου καὶ πλούτου . . . ῥάβδος, ‘a staff . . . of plenty and wealth’. Earlier in the hymn (following the theft of Apollo’s cattle), when Maia scolds Hermes for being a μεγάλη μέριμνα ‘great concern’, his retort entails an enumeration of advantages he wants to acquire: πλούσιος, ἀφνειός, πολυλήιος (line 171). While the three terms, encoding notions of ‘plenty’, show near synonymy (Richardson 2010:181), the poet’s choice of the set may possibly hold some significance within the sphere of horticultural abundance. Without providing textual citations, Versnel (2011:325) claims that the first two ‘have strong associations with corn in archaic poetry’ and adds that the last means ‘with rich cornfields’ – making reference (note 50) to the Homeric Hymn to Demeter. The derivation of πολυλήιος from λήιον ‘standing crop; grain-field’ appears probable (Chantraine 1968:636); Versnel’s reference to Demeter’s hymn seemingly alludes to line 489, in which the respectively related lexemes Πλοῦτος and ἄφενος co-occur: Demeter and Persephone are said to send to those whom they love the deity Πλοῦτος, who gives ἄφενος to mortal humans. In his description of Troezen, Pausanias (2.31.10) writes of an ἄγαλμα of Hermes Polygius (Πολύγιος) and of how Heracles once propped a club of olive wood against the image, whereupon the club took root and grew into a tree. Pausanias (1.27.1) also describes a wooden image of Hermes in the temple of Athena Polias in Athens (said to have been set up by Cecrops) that is obscured by branches of myrtle20. In an epigram from Lesbos (Kaibel 1878:330-31 [no. 812]; possibly second century AD), Hermes is invoked to bring fecundity to a vineyard. Hermes Κεδρίτης ‘of the cedar’ was worshipped, together with Aphrodite, at the sanctuary of Kato Symi Viannou in southern Crete. On this site have been found ninety-five bronze plaques, dating from as early as the seventh century BC, votives that were seemingly suspended from cedar trees at the sanctuary; one of these depicts Hermes seated in a tree (ca. 650 BC)21.

    Let us now focus on the cult image of Hermes, the δῶτορ ἐάων, that takes the form of the boundary marker called the herm. Herms were typically constructed of a worked-stone rectangular pillar surmounted by a bearded male head, with stubby arm-like blocks projecting laterally a distance below, and outfitted frontally, at about mid height, with a phallus. Prior to the advent of stone herms, similar cult objects were likely crafted of wood and perhaps of unworked stone22. Herodotus judges the herm to have been introduced to the Athenians by the indigenous Pelasgians, and the Athenians to have then passed the custom of its use to other Greeks (Pausanias [1.24.3; 4.33.3] agrees on the Athenian primacy of herms among Greeks; Thucydides [6.27.1] writes of an epichoric tradition).

    The name that the god carries, Ἑρμῆς, is almost certainly etymologically bound up with the god’s columnar cult object, also denoted by the phonic string ἑρμῆς, though the envisaged connection is not without its detractors (see Chantraine 1968:373-4). The Greek noun ἕρμα denotes a ‘prop, support’. In Homeric epic, as also at Homeric Hymn to Apollo 507, it is used in the plural of the props that are placed beneath boats to keep them secure and steady when they have been pulled ashore: thus, Il. 1.486; 2.15423. At Il. 16.549 the poet uses it metaphorically of the slain Sarpedon, who had been the ‘support’, the ‘pillar’, of his city – as it is used of the many slain suitors at Od. 23.121. The use of the term at Il. 4.117 has been considered “puzzling” (Kirk 1985:342), being found in a line that was (not unrelatedly) athetized by Aristarchus: here ἕρμα is used of the arrow that Pandarus lets fly against Menelaus.

    In addition, Greek ἕρμα is used to name a structure that consists not simply of a column but of another sort of vertical extension, one made of stones piled one upon the other. Such a cairn can itself form a base out of which a typical stylized herm extends24. According to Cornutus (Theologia Graeca 24), passers-by would pick up stones from the roads as they walked and pile them against herms as they met them: the author suggests several motivations for the practice, most being fundamentally acts of homage to the god of the herm (the traveller has no other offering to present, etc.). Such a mound is said to be a ἑρμαῖος λόφος ‘hill of Hermes’ (Scholia in Odysseam [scholia vetera] 16.471); ἕρμαιον denotes both abstractly the acquisition of goods or good fortune and concretely ‘cairn’ (beside ἕρμαξ in the latter sense); while a ἑρμαῖα δόσις is a ‘gift of prosperity’, as at Aeschylus Eumenides 947-8.

    That the herm naturally lends itself to pastoral themes, developed with images of earthly sustenance, can be seen in various epigrams of book nine of the Palatine Anthology, which provide insight into the sorts of day-to-day activities which centered around the herm. An epigram attributed to the Arcadian poet Anyte (9.314) mentions a herm that stands at a τρίοδος, the triple-fork crossroads, by a row of trees, at which travellers rest and refresh themselves with waters of an associated spring. In another, attributed to Leonidas of Tarentum (9.316), a herm is depicted that stands by a path (ἀτραπός) that traverses space between fields and city – a herm of the sort that has the head of Hermes on one of its sides and the head of Heracles on the opposite, figures styled as ὅρων φύλακες ‘guardians of the boundaries’ (line 8): here passers-by deposit offerings of fruit for the (εὐάκοος ‘inclined to listen’) god. Leonidas again calls attention to herms in 9.335 – two images (ἄγαλμα) set up by a poor wood-carrier, Miccalion, for passers-by to see. Elsewhere (9.318) the same poet frames Hermes, that is, the herm, within pasture land covered with abundant fennel and chervil: in return for being ‘gentle’ (προσηνής), the god can expect to receive both λάχανα ‘cultivated herbs’ and γλάγος ‘milk’ from the cultivator/herder. Compare the list of offerings to Hermes ἐνόδιος ‘of the roadside’ which appear in epigram 6.299 (attributed to Phanias): part of a grape cluster; bits of an oven-baked cake; a black fig, an olive, cheese slices, Cretan grain, and wine.

    The boundary marker that is associated with fertility and the acquisition of goods is a well-known Indo-European cult artifact25. Conspicuous realizations of the ancestral Indo-European object appear in the form of the terminus of archaic Italy and the especially well-attested yūpa of Vedic worship. In the cult tradition of primitive Indo-European transhumant pastoralists such a marker appears to have been erected at the distal boundary of temporarily installed sacred spaces. A ritual conducted within the space was conceptualized as a journey that advanced toward the boundary marker; attaining the marker, the worshipper accrued blessings: ‘chief among those blessings which the sacrificer obtains from the yūpa are cattle, sustenance, and prosperity’ (Woodard 2006:81). These are advantages no less conspicuously associated with Hermes, and one suspects that one element in the origin of the god was the primitive Indo-European boundary marker of blessing. Ἑρμῆς would be the ἕρμα animated and deified, and in that process he would fully mirror Roman Terminus. In Vedic cult the sacrificer and his wife ascend the yūpa on a ladder and by so doing are said to gain the world of the gods: the yūpa is intermediary no less than Hermes.




    In discussing individual Gandharvas we start with Viśvāvasu largely because of his prominence in the Rigveda (§4 above); he has been called the Gandharva par excellence (Dumézil 1929:139). He also seems to enjoy a certain priority or seniority in the epic. When Arjuna arrives in heaven during his twelve-year exile, he is eulogised by the Gandharvas ‘led by’ or ‘starting with Viśvāvasu’ (V.-prabhṛtibhir 3,44.18). In the Vulgate Viśvāvasu comes first in the list of nine Gandharva chiefs present in Kubera’s assembly hall (after 2,10.22, in Appx 1.3 line 2). He is referred to as ‘the Indra among the Gandharvas’ (12,306.36), and is the father of Citrasena (3,89.13; 165.54), whom we shall meet later. He is an impressive musician: at the lavish sacrifices of King Dilīpa, amid six thousand Gandharvas, his lyre-playing made each hearer think that the music was for him alone (12,29.64-9).


    6. Sexuality and marriage

    Gandharvas are linked with sexuality and marriage in a number of ways. Having presented the archetypal wedding of Soma and Suryā, the Rigvedic ‘Marriage Hymn’ turns to human marriage and says of and to the bride:

    Soma obtained her first; next Gandharva obtained her; Agni was your third husband; and your fourth was human-born. Soma gave her to Gandharva, and Gandharva gave her to Agni. Agni has given me wealth and sons, and now this wife (RV 10.85.40-41).


    The hymn has already implored the second husband, under the name Viśvāvasu, to leave the human bride and seek another partner (10.85.20-21), but later sources imply that this may not happen straightaway. During the first three nights after the wedding the newly-weds are not supposed to make love. According to Āpastamba Gṛhyasūtra (3.8.9), they are separated by a wooden staff anointed with perfumes and wrapped with a garment or thread (daṇḍo gandhalipto vāsasā sūtreṇa vā parivītas); this object is addressed as Viśvāvasu and urged to depart, being explained as a symbol of the Gandharva (Oldenberg 1993:125-6, Oberlies 2005). Other connections with sex are equally vivid. When touching the female’s genitals, a male lover may address them as the mouth (mukham) of the Gandharva Viśvāvasu (Śāṅkhāyaṇa Gṛhyasūtra 1.19.2). At a Horse Sacrifice Gandharvas receive the victim’s penis and Apsarases receive its testicles (Gonda 1962:126). According to a Buddhist doctrine, the soul of a deceased individual takes the form of a Gandharva and enters the mother’s womb at conception (Wijesekera 1994: esp. 193-202).

    Less vividly, the paired categories of spirits ‘preside over fertility and are prayed to by those who desire offspring’ (Macdonell 1981:137). It is not surprising that they sing and dance at the wedding of Rama and his brothers (Rām. 1,72.25). Moreover, Hindu law (Manu 3.20 ff.) recognises eight modes of marriage, named by adjectival forms of the names of supernaturals, ranging from Brahmā at the top to the demonic Piśācas at the bottom (cf. Allen 1996:14 ff.). The gāndharva mode, in fifth position, is based on mutual consent of the partners and allows them maximum independence.

    As Hopkins puts it, Gandharvas are ‘lovers par excellence’ and ‘sharp in love’ (kāmatīkṣna). When exercising their musical skills, they are regularly accompanied by Apsaras, and they apparently spend most of their remaining time sporting with the same group – little is said of their marrying or maintaining long-term relationships. Female Gandharvīs, in addition to the Ancestress of horses, are mentioned here and there, for instance Kumbīnasī, who saved the life of her husband, the Gandharva Citraratha, when he was defeated by Arjuna (1,158.32); but they are less prominent than Apsarases. The latter can sometimes become partners of mortal men, and Gandharvas can sometimes take an interest in, or possess, mortal women. Contrasting the Atharvavedic picture of Gandharvas with the somewhat heterogeneous and imprecise Rigvedic one, Schroeder describes them as follows (I translate):

    priapic impudent fellows, phallic demons, who are not satisfied with dancing, swinging, playing and making love in the company of their beautiful sweethearts, the Apsarases or Indian Nymphs; in addition they pester the wives of men, whether awake or asleep, dancing around human dwellings in the evening in all sorts of forms, sometimes horrific, but sometimes friendly and familiar – for they know how to transform themselves. They are particularly dangerous to women in childbirth, to the foetus and the new-born infant. So one has to use powerful spells and effective herbs to exorcize them and keep them at a distance (1908:61).


    Though they can be horrific, Gandharvas are more often strikingly handsome. Handsome youths (yuvānaḥ śobhanā) who are present at a Horse Sacrifice are referred to as Gandharvas (ŚB When Nala is first seen in the flesh by Damayantī and her maids, they wonder whether someone so beautiful is a god, a Gandharva or a Yakṣa (3,52.16). When Rāma comes to the palace to be appointed Prince Regent – surely looking his best – he appears the very image of a Gandharva-king, a gandharva-rāja-pratima (Rām. 2,3.11).

    Despite the early bearded images of Hermes, he is typically beardless and young, his cult being associated with children. To deal with mortals he can take the guise of a young man or prince with the first down on his upper lip, ‘one in whom the charm of youth (ἥβη) is fairest’ (Il. 24.347-8, Od. 10.277-9). The ithyphallic herms have already been mentioned, and emphasize his generally promiscuous reputation. At the port of Kyllene in Elis he was worshipped in the shape of a phallus (Burkert 1985:158, with references). The association of the cult of Hermes with the gymnasium is well known26.

    Like the Gandharvas, Hermes has a very active sex life and no generally recognized wife. According to the above-mentioned Homeric Hymn 19, he is the father of Pan by the daughter of Dryops; her joining in marriage with Hermes is described, if formulaically, as θαλερός ‘blossoming, teeming’ (lines 33-4). In the Homeric Hymn to Aphrodite 262-3 the poet sings of Hermes and Sileni having intercourse with nymphs in the depths of caves. Nonnus (Dionysiaca 14.113-15) records the tradition that Hermes is father of the Satyrs by Iphthime, a daughter of Dorus (eponymous ancestor of the Dorians). The name of Iphthime recurs in the Catalogue of Women, attributed to Hesiod (see fr. 10a), in conjunction with reference to the origin of the Satyrs and (seemingly) nymphs; compare Hesiod fr. 10b in which the nymphs are said to originate from the daughters of Dorus27. Of Hermes’ children, Pan, Eros, Hermaphroditus and Priapus all have amatory or sexual connotations. The fourth day of the month is both his day of birth and the day recommended for bringing home a wife (Works & Days 800, West 1978:352).


    7. Religious and magical knowledge

    Viśvāvasu cannot be reduced to his erotic and marital aspects. He is also distinctly learned and has a role in the transmission of knowledge. We have already met his knowledge of herbal medicine (§5); and drawing on his lengthy studies of Rigvedic religion, Oberlies (2005:98) writes of him as follows:

    As a guardian at the border of this world and the beyond and as the ‘god of transfer’ the Gandharva knows the true nature of things – ‘their innermost name(s) (RV 10.123.4) –, which he reveals to Indra (RV 10.139.6).

    The second reference is to the one Rigvedic hymn of which he is the reputed author. It is surely relevant that, as we shall see, Viśvāvasu exchanges soma for the naked Vāc (‘Voice’), who is not only female but also goddess of speech and language (cf. e.g. Ait. Br. 1.27). When the Gandharvas compete with the gods in trying to attract Vāc, they boast of their knowledge of the Vedas (ŚB

    The epic (12,306.27) presents Viśvāvasu as learned in Vedantic teachings (vedānta-jñāna-kovidaḥ), and he is able to put to the sage Yajñavalkya twenty-five questions about the Vedas. He has already heard discourses on the soul from numerous sages and supernaturals, and he transmits what he has learned on this occasion to beings elsewhere in the cosmos (12,306.82). He also possesses the magical power of vision, cākṣuṣī vidyā,which was transmitted as follows: From Manu to Soma, then in succession to Viśvāvasu, Citraratha and Arjuna (1,158.40). Though it is not clear that Arjuna used it, the power would have enabled the hero to see whatever he wanted, across the three worlds of the cosmos.

    Hermes’ religious knowledge and verbal skills are treated under §15, but much of what he does on his first day of life implies knowledge that is more miraculous or magical than naturalistic. Many of his creative actions, such as sacrifice, are subsequently followed by humanity in general, which implies that he transmitted them.


    Citraratha and Citrasena


    These two Gandharvas are juxtaposed here because, as Hopkins notes, they interact with Arjuna in comparable ways (cf. already §10), and have comparable names. The adjective citra means ‘variegated’ and occurs also in the name of the Gandharva Citrāṅgada, who kills the homonymous mortal king born in Arjuna’s grandparental generation (1,95). Since ratha means ‘chariot’ and senā means ‘army’, both names have martial connotations; and although aṅgada means ‘bracelet’, the homonymous pair fight for three years. Arjuna interacts with Citraratha in Book 1 and with Citrasena in book 3, but we start with the latter. This is not to imply that Citraratha is the lesser figure. In one section of the Bhagavad Gītā Krishna explains his own significance by relating himself to a long list of types of being: ‘Among all trees I am the Aśvattha, among divine ṛṣis Nārada, among Gandharvas Citraratha, among Siddhas the ascetic Kapila (6,32.26).’


    8. Messenger, go-between, intermediary

    Gandharvas are not only sexually active themselves, they also bring together married couples and partners (§6). Viśvāvasu’s son provides an instance where the role of go-between is combined with that of messenger from higher gods to mortals – we are dealing with what Oberlies (cited in §7) called ‘the god of transfer’.

    When Arjuna goes to heaven to visit his father Indra, the god lays on a festive welcome. During the celebrations Arjuna is observed staring at the dancing Apsaras Urvaśī. To test his son, Indra plans a tryst between hero and nymph, sending Citrasena to arrange it. It is not the Gandharva’s fault that, when the nymph comes to him, Arjuna refuses to make love to her and is consequently cursed to the eunuchism that he experiences in year 13 (3, 1.6*, after 3,45.9)28. In some respects (e.g. Allen 1996:13), Urvaśī in heaven parallels Nausicaa in Scheria: in both cases the union between the nubile female and the central hero is mooted but not realised. But Odysseus is only able to reach Scheria because Zeus has sent Hermes to release him from Ogygia – in other words, Indra’s messenger visiting Urvaśī parallels Zeus’s messenger visiting Calypso. Admittedly, the gap between Hermes’ trip and Odysseus’s meeting with Nausicaa is too long to allow interpretation of the messenger as a go-between, in the sense of one who facilitates amatory encounters. He is more naturally seen as breaking up the hero’s relationship with Calypso.

    However, Hermes as go-between is attested elsewhere in Greek tradition. Almost at the very end of the Iliad, as we have it, the epic poet gives a nod to the story in which Hermes, the διάκτορος ‘conductor’29, is conspicuously portrayed as an intermediary whose actions will eventuate in a sexual liaison. Hermes conveys a triad of goddesses into the presence of the Trojan prince Paris so that he may judge which of the three is the most beautiful; and his judgment leads to his liaison with Helen. Whatever conclusion one may reach about the authenticity of the lines in the Iliad30, the tradition is undeniably an archaic one. It is well attested in Greek art in the seventh century BC, as on the Chigi Vase of ca. 630 BC (Rome; Villa Giulia 22679), and formed part of the Cypria of the Homeric Cycle. Dumézil (1995:608-14) argued cogently that the bribes offered to Paris by the three goddesses constitute a significant Greek preservation of primitive Indo-European tripartite ideology.

    Hermes perhaps performs a similar function in scenes painted on various black-figure vases studied by Hedreen (1992). If so, in this instance the god serves inversely, as an intermediary who brings, not goddesses to a mortal, but a mortal woman to a god – namely, Ariadne to Dionysus. Consider, to take but one example, the case of an amphora housed in the Royal Ontario Museum, Toronto (304 [919.5.141] = ABV 259,21) that depicts Dionysus and Hermes side-by-side, both faced by a woman to whom Hermes gestures ‘emphatically’ (Hedreen 1992:41, see pp. 40-42 for discussion with additional examples, including red-figure). Drawing attention to Hermes’ role as psychopomp, Hedreen (p. 42) surmises that the god’s presence as intermediary is occasioned by ‘a transgression of the human and divine realms’ – realms between which Hermes, the messenger, can pass.

    Hermes can, however, play the opposite role with regard to such mortal-immortal ‘transgressions.’ A late fifth-century Attic relief, attested by various copies, depicts Orpheus, his mortal wife Eurydice, and Hermes, who appears to be about to lead Eurydice away from Orpheus. In a well-known tradition, Eurydice died, having been bitten by a snake; Orpheus subsequently descended to the realm of Hades and charmed its denizens into allowing him to return with her to the world of the living – only to lose her when he turned to look at her during the ascent, as had been forbidden to him by the infernal sovereign (Pseudo-Apollodorus Bibliotheca 1.3.2; Virgil Georgics 4.454-503; Ovid Metamorphoses 10.1-73). The relief perhaps encodes this tradition, or possibly another in which Eurydice returns to the world of the living, but only temporarily (see Bowra 1952:121-122): either way, Hermes appears to separate the couple. Hedreen (1992:42) points out that a similar scene is depicted on an amphora in London (B 257 [=LIMC 3. 391]), except here Hermes leads away not Eurydice, who had tasted death, from Orpheus, who had not, but mortal Ariadne from immortal Dionysus. Such cases recall Hermes’ role in the separation of mortal Odysseus from divine Calypso.


    9. Music Teacher

    Like his father, Citrasena must be a particularly fine musician since he is chosen by Indra as Arjuna’s music teacher. When Arjuna visits heaven and has completed his military education, Indra tells him to acquire from Citrasena the skills in dancing, singing and instrumental music that he will need in year 13 (3,45.6-7; 164.54). Indra wants the two to be friends, and Arjuna greatly enjoys Citrasena’s company.

    Hermes too is a good and innovative musician, who passes on his musical knowledge to his half-brother Apollo. This story is first reported in HH Hermes.

    On the day he is born, Hermes’ first action when he leaves the cave of his mother, the nymph Maia, is to invent the lyre (lines 39-51, §14). He then steals the cattle of Apollo (lines 68-104, §12), sacrifices two of them (lines 105-141, §13), is traced by Apollo, and tried before Zeus (§10). Though Hermes leads his brother to the surviving cattle, Apollo remains angry; but he is pacified when Hermes produces his newly created lyre and plays on it as he sings to Apollo (lines 414-34, §15). Hermes then gifts Apollo with the lyre and instructs him in its use, thus giving the latter his characteristic skill on that instrument (lines 435-502).

    Following the account of Hermes’ presentation of the lyre to Apollo, the poet of HH Hermes sings succinctly and metonymically of yet another musical invention attributed to the god of his hymn. In line 512 one reads: συρίγγων ἐνοπὴν ποιήσατο τηλόθ’ ἀκουστήν ‘he made for himself the distantly audible sound of the panpipe’. It is the instrument typically associated with the Greek herdsman and the deities of the herdsman (see West 1992:110). The Hellenistic poet Euphorion of Chalcis also knows Hermes as inventor of the panpipe, as Athenaeus points out (Deipnosophistae 4.184A) – although, he continues, the crafting of the instrument in one or another of its forms is also credited to others, including Silenus and Marsyas the satyr. In the Homeric Hymn to Pan, that son of Hermes to whom the hymn is dedicated is described as playing the panpipe (δόνακες, line 15), as is typical; and some know him as the inventor of the instrument (thus Pliny HN 7.204). Pseudo-Apollodorus (Bibliotheca 3.10.2) also presents Hermes as the inventor of the panpipe (σῦριγξ), but departs from the account of the Homeric Hymn by depicting Apollo as wanting to acquire the pipes in addition to the lyre. Here Hermes exchanges the second instrument for a golden staff (ῥάβδος) and for (τέχνη) μαντική ‘mantic skill’ (on both of which, see below): καὶ δοὺς διδάσκεται τὴν διὰ τῶν ψήφων μαντικήν ‘and when he had given [the pipe] he learned divination by pebbles’.

    When Arjuna is in heaven he acquires not only musical skills but also, from the gods collectively, a conch (śaṅkha) named Devadatta, ‘God-given’ (3,165.21-22). The conch serves as a war-trumpet and has a mighty sound (for instance, it is mahārava in 3,171.5). So, despite obvious differences, comparison with the ‘distantly audible’ panpipes is not impossible.


    10. Enemy becomes Friend

    During their initial exile Arjuna leads the Pāṇḍavas into a beautiful wood beside the Ganges. The wood is occupied by Gandharvas, whose leader, a friend of Kubera, is usually called Citraratha (he has other names). The demigod reacts angrily to the trespassers and attacks Arjuna. Responding with his Fire Weapon, Arjuna burns the colorful chariot of the Gandharva, whose life is only saved by his wife’s intercession. The males now exchange gifts. Citraratha gives magic sight (§7) and promises horses (§10); Arjuna gives not only ‘the gift of life’ but also his Fire Weapon. The alliance (saṃyoga)or friendship(sakhya)is to last for ever (1,158).

    In 3,229-34, after Arjuna returns from heaven, the Pāṇḍavas are near Dvaitavana Lake. The Kauravas arrive, with a view to gloating over their exiled rivals, and attack the Gandharvas for trying to block the intrusion; but the demigods, led by Citrasena, capture Duryodhana and others. Yudhiṣṭhira sends the Pāṇḍavas to rescue their cousins, and Citrasena is attacking Arjuna when, abruptly, the two recognize each other as old friends (they had met in heaven – see §8, 9). The humiliated Duryodhana is set free and the Gandharvas depart to heaven.

    The second story is more complex than the first since it involves three sets of combatants, not two, and two different offences – trespass by the Kauravas, and imprisonment of Kauravas, including Kaurava womenfolk, by the Gandharvas. However, both stories include a combat between the central Pāṇḍava hero and a Gandharva leader who has been angered by trespassers; and in both the initial hostility gives way to friendship. Hopkins sees the second conflict as ‘imitating’ the first, but (as we shall see) this is not the only possible interpretation of the dualism.

    The reconciliation of former enemies – that is, of the half-brothers Hermes and Apollo, is fundamental to HH Hermes. The offence causing the enmity is Hermes’ theft of Apollo’s cattle. When Apollo has tracked Hermes to the cave of Maia, he confronts the thief, threatening to imprison him in the ‘gloomy darkness’ (ζόφος) of Tartarus, where he will be ὀλίγοισι μετ’ ἀνδράσιν ἡγεμονεύων, ‘leader among feeble men’ (lines 256-9); and, after the thief has denied his act, Apollo addresses him as μελαίνης νυκτὸς ἑταῖρε ‘companion of dark night’ (line 290). Following a contentious back-and-forth of recriminations and denials, the poet defines the pair as ἀμφὶς θυμὸν ἔχοντες ‘having a divided heart’ (line 315)31. The turning of this enmity to friendship is a process instigated by Zeus, the divine judge and their father, in a scene of arbitration set on Olympus, with a dialogue that is constructed with phrasing of a legal tone. When the complaint and defense have been heard, Zeus commands that both Apollo and Hermes ὁμόφρονα θυμὸν ἔχοντας ζητεύειν ‘make a search, having a single-purposed heart’ (lines 391-2). The phrase ὁμόφρονα θυμὸν ἔχοντας recurs at Iliad 22.26332; as the enemies Achilles and Hector prepare to duel, a raging Achilles rejects Hector’s proposal of a ‘covenant’ (συνημοσύναι) that the survivor will see to it that the body of the slain is returned to his comrades, retorting (263-4):

    . . . οὐδὲ λύκοι τε καὶ ἄρνες ὁμόφρον θυμὸν ἔχουσιν,

    ἀλλὰ κακὰ φρονέουσι διαμπερὲς ἀλλήλοισιν

    . . . and wolves and lambs do not have a single-purposed heart, but instead they continuously purpose evil against each other


    For two enemies to take hold of a reciprocal ὁμόφρων θυμός is for them each to not purpose evil against the other.

    Reconciliation is not, however, realized during the arbitration, but comes only after the search is made and after Apollo’s anger has once again been roused against his cattle-thieving kinsman. Seeing the hides of the two butchered cows, Apollo begins to plait strong bonds out of withes probably with a view to binding Hermes, but the latter causes the withes to grow into the ground, to spread, and to ensnare the stolen cattle instead. It is at this point that Hermes produces his lyre, performs the Theogony, and thereby pacifies the anger of Apollo (lines 409-35). He now addresses Apollo as φίλε ‘near and dear one’ (line 469), though the term of affection is preparing the way for a request – that he, Hermes, be endowed with care of cattle (lines 490-95). Apollo, having received the lyre, assents and demonstrates as much by turning over to Hermes the μάστιξ φαεινή ‘clear-sounding whip’: Apollo βουκολίας ἐπέτελλεν ‘commanded [to him] cattle herds’ (or ‘care of cattle’; lines 497-8). Compare Hesiod who, as he praises Hecate at Theogony 445-7, sings that the goddess can increase (and diminish) the βουκολίας ‘cattle herds’ and other sorts of domesticated animals: the lines are an elaboration of 444, where one learns of the goddess that she is ἐσθλὴ δ’ ἐν σταθμοῖσι σὺν Ἑρμῇ ληίδ’ ἀέξειν, ‘good in the stables, together with Hermes, at increasing livestock’.

    But Apollo still fears that the thief Hermes may steal away the lyre from him. More than this – and intriguingly – Apollo is ‘afraid’ (perfect of δείδω) that his former enemy will steal away his καμπύλα τόξα ‘curved bow’ – the weapon of Apollo the destroyer. The line (515) is not without a broader context: a tradition of Hermes’ theft of weaponry from Apollo is well attested. Alcaeus composed a hymn to Hermes (fr. 308, together with P. Oxy. 2734 fr. 133) in which he included the tale of how Hermes stole Apollo’s quiver of arrows. The tradition is relayed by several later authors. In his hymn to Mercury (Odes 1.10), Horace alludes to the disappearance of the pharetra ‘quiver’ of Apollo, in tandem with a mention of the theft of his cattle. Lucian, in his Dialogi deorum 11.1, has Apollo declare to Hephaestus regarding Hermes: ἀφώπλισε τοῦ τόξου καὶ τῶν βελῶν, ‘he disarmed [me] of bow and arrows’. Philostratus (Imagines 1.26) reports that Hermes stealthily λύει τὰ τόξα ‘unbinds the bow/arrows’. The earlier mentioned scholiast on Il. 15.256 writes: ἀπειλουμένου δὲ τοῦ Ἀπόλλωνος, ἔκλεψεν αὐτοῦ καὶ τὰ ἐπὶ τῶν ὤμων τόξα ‘and with Apollo threatening him, [Hermes] even stole the bow/arrows upon his shoulder’.

    The episode as presented by Alcaeus is known only from the testimonia, with a specific contextualizing of the event within Alcaeus’ hymn being lacking. In HH Hermes, the only other archaic attestation of Hermes’ (here potential) theft of the characteristic weapon of Ἑκηβόλος Ἀπόλλων ‘Far-shooting Apollo’, mention of such a taking away of warrior efficacy is made only subsequent to overtures of friendship and some sanctioning action taken by Zeus in that regard: χάρη δ῾ ἄρα μητίετα Ζεύς | ἄμφω δ’ ἐς φιλότητα συνήγαγε, ‘and then All-Wise Zeus was glad | and made a covenant for mutual fraternal-allegiance’ (lines 506-7). The language is certainly that of formal pact making.34 There are σήματα of this covenant (line 509). What next follows is (1) Apollo’s expression of concern over potential theft of his weaponry – his warrior efficacy – and then (2) still another episode of making a σύμβολον ‘compact’ of fraternal alliance (lines 521-8), on which see further below.

    The tradition of former adversaries making a pact or agreement of mutual nonaggression, after which one of the pair behaves treacherously so as to rob the other of warrior prowess, is a well-attested and well-studied Indo-European trope. Indra made of Namuci a sakhā and Tullus Hostilius made of Mettius Fuffetius a socius. Both Namuci and Mettius then took measures that resulted in the weakening of that warrior who had exchanged animosity for a friendly, fraternal relationship. In each instance the weakened party, Indra and Tullus, is rescued from a loss of warrior prowess by the intervention of deities belonging to the realm of fertility: the Aśvins and Sarasvatī in India: Quirinus, Saturnus, and Ops in Rome.35 Surely strains of this primitive, inherited tradition are to be heard in the hymning of Hermes. A specter of the loss of warrior prowess remains, without embodiment, in conjunction with the making of a pact of mutual nonaggression; but the poet, or some predecessor poet, has reworked the Indo-European tradition by localizing its expression within a trough created by a doubling of the formal declaration of fraternal alliance. And this doubling is perhaps a necessary consequence of the interweaving of the primitive Indo-European tradition into a μῦθος in which a fraternal relationship will be preserved intact.

    But while there is no rupture of the pact in this Greek tradition, the combat prowess of the warrior still is vouchsafed by a figure of fertility: Hermes affirms he will not take away whatever Apollo has acquired. In other words, Hermes is not only a participant in the pact, but in the process of forging a fraternal bond with his former adversary, the δολομῆτης ‘trickster’ himself is, as it were, transformed into the deity whose domain is the fertility of animals. Notice also that just as the formal making of a covenant of non-aggression is mentioned twice, so is the endowing of Hermes with this pastoral role. We have seen just above that prior to the first oath episode (lines 506-7), Apollo entrusts Hermes with the care of herds (lines 497-8). Subsequent to the second oath episode (lines 521-8), as Apollo concludes his ensuing speech, the role assignment is repeated (lines 567-71): Hermes will have not only cattle, but horses, mules, lions, boars, dogs, sheep, and all flocks.

    Despite the complexities we mentioned, both Sanskrit stories tell of Gandharvas and Arjuna moving from enmity to friendship, while in the Greek Hermes and Apollo make the same move, but twice over. So both traditions show a certain dualism. Moreover, if one conflates the two Sanskrit stories, the combination contains many of the motifs present in the single Greek story.

    • Indra wants Citrasena to be Arjuna’s friend; Zeus wants Hermes to be Apollo’s friend.

    • In both traditions the friendship is sealed by exchanges of skills or property – in particular, it involves transfers relating to domesticates. Arjuna will receive horses; Hermes receives horses, cattle and other animals.

    • One of the friends experiences uneasy mixed feelings. Arjuna enjoys his studies in heaven but broods on wrongs he has suffered and misses his family (3,192*, after 3,45.8); Apollo is worried that Hermes may resume hostile behaviour.

    • Just before the end of the hostile phase, one of the future friends exhibits miraculous powers. Arjuna the bowman envelops the Gandharvas on all sides with an ‘arrow-net’ (śarajāla) to stop them escaping into the sky, confining them like birds in a cage (3,234.12-13, & 19). Hermes, who can foster plant growth (§5), causes Apollo’s withes to interlace so as to cover the cattle.

    • Duryodhana was not only captured by Citrasena but also bound (√bandh, 3,235.6; 238.6). Apollo plaits the withy bonds (δεσμά) possibly to bind Hermes.


    Although the allocation of these motifs to individuals does not always support our Gandharva–Hermes rapprochement, their existence supports the theory that the traditions are related. However the main point to note is the enemy-to-friend theme, which will turn out to be notably pervasive.




    Urvaśī is an Apsaras, but her lover Purūravas is born a mortal, and only becomes a Gandharva when their relationship is interrupted. The story is known from the Rigveda onwards and has a copious secondary literature.


    11. Ancestors

    We encountered Urvaśī previously in Indra’s heaven (§8), where Arjuna rejected her advances. He did so on the grounds that she was a distant ancestor. In fact, the epic gives two accounts of Arjuna’s patriline, which coincide only in part (Brodbeck 2009:21-30); but if the line is followed back for 26 or 38 generations it reaches Āyus and then his parents, Purūravas and Urvaśī (Apsarases do not age!).

    If Arjuna corresponds to Odysseus, the natural question is whether the Greek hero has any comparable forebears. In fact, his mother Anticleia is the daughter of Autolycus, and Autolycus has close relations with Hermes. The qualities they share are treated in §12, and the point here is that the author of the Catalogue of Women (Hesiod fr. 64) knows Autolycus to be no less than the son of Hermes, borne by Philonis. The same tradition is attested in many post-Homeric sources: the fifth-century historian Pherecydes (fr. 63b); Ovid (Metamorphoses 11.301-315); Pseudo-Apollodorus (Bibliotheca 1.9.16), Hyginus (Fabulae 200–201), Lucian (De astrologia 20), Polyaenus (Strategemata 1.Pro.6), Proclus (In Platonis Alcibiadem 1.216); Eustathius (Commentarii ad Homeri Iliadem 3.65; Commentarii ad Homeri Odysseam 2.246); Joannes Tzetzes (Chiliades 8.202); and various Homeric scholia. In other words, Purūravas and Hermes are both presented as ancestors of the central hero within their respective epic traditions.


    12. Theft

    In the Vedas Gandharvas sometimes guard soma, sometimes steal it (or him). At one time, soma was in the sky and the gods on earth wanted it for use in sacrifice; but when they dispatched Gāyatrī to carry it off (ā-hṛ-), the Gandharva Viśvāvasu stole it from her (parimuṣ-, ŚB The Gandharvas are ‘fond of women’ and hope to exchange their soma for the goddess Vāc; but the gods create the lute (vīṇāṃ sṛj-) and attract her with their song. She prefers this empty pleasure to the Gandharvas’ Vedic recitation (cf. §7; the text adds that women still prefer such pleasures).

    In the Purūravas story the Gandharvas again appear as thieves (ŚB – there are variants elsewhere). Urvaśī has married him on condition that he does not let her see him naked. Missing their Apsaras companion, the Gandharvas carry off (pramath-, hṛ-) two pet lambs (her ‘children’, putrān), who are attached to her bed, and when the naked husband runs in pursuit, they send a flash of lightning. On seeing Purūravas in the nude, Urvaśī disappears. The story continues, as we shall see in §13, but we now have two instances of theft by the Gandharvas, one associated with the creation of a stringed instrument, the other with the nocturnal removal of domesticated animals from someone’s bedroom or private space. In addition, epic Purūravas, a human king ‘surrounded by supernaturals’, wages war on brahmins, whose valuables he carries off or steals (hṛ-, 1,70.17-18).

    After Hermes invents the lyre, he makes his way north to Pieria in Thessaly (not far from Mt. Olympus), where he encounters θεῶν … βόες ‘cattle … of the gods’; he steals and drives away fifty belonging to Apollo (lines 68-104), slaughtering and roasting two of them (lines 105-141). This episode of the theft stands at the core of the hymn, which is in fact pervaded by references to cunning, deception, tricks; to robbery and theft, often by night; and to house-breaking, cattle-rustling and plunder. The cattle raid is a well attested primitive Indo-European theme (on which see, for example, West 2007:451-2 with bibliography), one which is encountered, for example, in Book 4 of the Mahābhārata, where the Kauravas attempt to steal cattle from King Virāṭa (cf. §1). Hermes’ theft of Apollo’s cattle must surely be viewed as a particular archaic Greek instantiation of that tradition36. The centrality of the theft within the hymn supports our view that the work preserves other matrix features of an ancestral Indo-European poetic performative tradition37.

    The next day Apollo succeeds in following his stolen cattle and confronts the infant thief in Maia’s cave. Hermes staunchly denies the accusation and offers to swear an oath of innocence by Zeus’s head (lines 274-7). The clever equivocation displayed by Hermes in his denials brings to mind Homer’s characterization of Autolycus at the naming of the infant Odysseus: the hero’s maternal grandfather surpassed humankind in ‘both stealing and oaths’ (κλεπτοσύνῃ θ’ ὅρκῳ τε, Od. 19:396). Furthermore, it was Hermes who endowed Autolycus with this exceptionality – in reward for sacrificial devotion (19.396-8). As many have noted, Odysseus has a share in the nature of his grandfather Autolycus, ‘the wolf himself’: he is “the prototype of Odysseus’ personality seen in its most negative aspects38.” Autolycus claims as much when he gives the name (19:407-8):

    πολλοῖσιν γὰρ ἐγώ γε ὀδυσσάμενος τόδ’ ἱκάνω

    ἀνδράσιν ἠδὲ γυναιξὶν ἀνὰ χθόνα βωτιάνειραν.

    For I myself have come here having inflicted pain

    on many men and women upon the man-nurturing earth.


    Like Hermes, Autolycus steals cattle. Pseudo-Apollodorus (Bibliotheca 2.6.2) writes that Autolycus stole cattle from Euboea – those animals that Eurytus sent Iphitus to find. Hesiod (fr. 67b) is said to have claimed that whatever Autolycus grasped with his hands, he made that thing ‘not visible’ (ἀείδελον); the source of the fragment is the Etymologicum Magnum (A 317), which adds that he stole horses and changed their appearance. According to Joannes Tzetzes (Chiliades 2.36), Autolycus stole horses from Eurytus; and Tzetzes (Scholia in Lycophronem 344) also knows Autolycus as a thief of horse, cattle, and sheep who would change their mark of ownership (cf. Hyginus Fabulae 201).

    At Iliad 10.266-7 the poet relates a different sort of theft committed by Autolycus. In lines that precede, one reads of the arming of Odysseus by Meriones: he gives to Odysseus bow, quiver, sword, and κυνέη ‘helmet’ of hide, decorated with boar tusks. Homer specifies regarding this helmet:

    τήν ῥά ποτ’ ἐξ Ἐλεῶνος Ἀμύντορος Ὀρμενίδαο

    ἐξέλετ’ Αὐτόλυκος πυκινὸν δόμον ἀντιτορήσας . . .

    This Autolycus removed once out of Eleon when he bored through [the walls of the] solid house of Amyntor, son of Ormeneus . . .


    The concatenation πυκινὸς δόμος ‘solid house’ seen in line 267 recurs in HH Hermes 523. While the phrase is formulaic in Homeric epic (found, in addition to line 267, at Il. 12.301 and Od. 6.134; 7.81, 88)39, its occurrence at Il. 10.266 in conjunction with Autolycus’ theft of weaponry must bear significantly on our understanding of its use at HH Hermes 523. The latter line is drawn from the episode in which Hermes gives signs of a compact of friendship following Apollo’s declaration of fear for his warrior prowess (lines 521-3) – that is, his being afraid that Hermes may steal his καμπύλα τόξα ‘curved bow’ (line 515). The compact of non-aggression into which Hermes then immediately enters entails two elements.

    First, Hermes will never steal away (ἀποκλέπτω) ὅσ’ Ἑκηβόλος ἐκτεάτισται ‘whatever Far-shooting [Apollo] might gain’ (line 522). The verb here used, κτεατίζω, occurs five times in Homeric epic – in three instances in the line αἴ κεν ἄτερ σπείρου κεῖται πολλὰ κτεατίσσας ‘if he, having gained many things, were to lie without a shroud’ (Od. 2.102; 19.147; 24.137), spoken by Penelope to the suitors concerning Laertes and the shroud she craftily weaves and reweaves to keep them at bay. A fourth occurrence of the verb in the Odyssey is found at 24.207, used of the farmland (ἀγρός) that Laertes had himself gained: ‘the exact sense is disputed’ (Russo et al 1992:382). The one remaining occurrence is at Il. 16.57, where the exact sense is quite clear: it denotes the gain of property by the exercise of physical force – that property being the woman Briseis whom Achilles gained δουρί ‘by [his] spear’ (she is δουρικτητή ‘spear-gained’ at 9.343) πόλιν εὐτείχεα πέρσας ‘having ravaged a well-walled city’.

    Second – and this is where the matter of πυκινὸς δόμος comes in – Hermes will never ἐμπελάσειν πυκινῷ δόμῳ ‘come near to [Apollo’s] solid house’ (line 523). The intent of the affirmation is clearly that Hermes will not rob the πυκινὸς δόμος of Far-shooting Apollo. With this compare a line from early in the hymn – line 178 – in which the verb ἀντιτορέω governs δόμον, as in Il. 10.267, though the modifying adjective has changed: here Hermes tells Maia that if Apollo should confront him, then he, Hermes, will go to Pytho μέγαν δόμον ἀντιτορήσων ‘to bore through his great house’, robbing it of tripods, lebetes, gold, iron and raiment – that is, to rob Apollo’s temple – μέγας δόμος – at Delphi. Within the synchronic system of the fourth Homeric Hymn as we have it, the threatened action against Apollo’s μέγαν δόμον of line 178 is most likely that one which Hermes vows not to perform against Apollo’s πυκινὸς δόμος in his compact of line 523. Along the diachronic axis, however, we must surely read the promise of line 523 in conjunction with the first element of the vow (line 522), and both of them against the background of the language of Il. 10.266-7 (Autolycus’ theft of weaponry) as a response to Apollo’s fear of being robbed of his warrior prowess. The language of Hermes’ compact of lines 522-3 of the fourth Homeric Hymn points again to the primitive Indo-European trope of the warrior weakened by a former enemy, the warrior restored by a figure of fertility.

    The cattle theft leads to a trial (lines 313-96) in which Zeus judges between Apollo (plaintiff) and Hermes (defendant). A verdict40 is delivered: Hermes and Apollo are to be of one accord, and Hermes must lead Apollo to the stolen cattle. This he does (lines 397-408). The pair travel to the river Alpheius in the Peloponnese, and the thief leads the forty-eight survivors out of the cave in which they were hidden. The release of stolen cattle from a cavernous space is a familiar one in Indo-European tradition, seen notably in the Vedic account of Indra slaying the cattle-thieving Vr̥tra and the cognate Italic tale of Hercules/Semo Sancus destroying the monster Cacus (see Woodard 2006:189-191, 196-8).

    In summary, Gandharvas steal not only soma but also lambs – and thereby Purūravas’ wife; and Hermes is the Prince of Robbers (HH Hermes 292, cf. 175), who transmits his expertise to Autolycus – apparently both his son and his human counterpart41.


    13. Sacrificial Fire

    Abandoned by Urvaśī, Purūravas wonders miserably across Kurukṣetra until he finds her, in the form of a swan. They agree to make love again in a year’s time, after Āyus is born. On Urvaśī’s advice he then asks the Gandharvas to admit him to their number. This will necessitate a sacrifice, and they instruct him how to make the fire-drill that is needed when kindling fires in certain Vedic rituals (ŚB According to the epic (1,70.21), it was when he lived with Urvaśī among the Gandharvas that Purūravas brought to earth the three fires required for almost all non-domestic Vedic sacrifices. In both cases he is effectively introducing sacrificial fire among humankind. We cannot here explore the fact that his father, or mother, was the sex-changing Ilā (1,70.16; 1,90.7), who, written in the related form Iḍā/iḍā, is involved in other stories of the origin of sacrifice.

    Let us now go back to Hermes’ original visit to the river Alpheius. He grazes and shelters the stolen cattle, but as they feed, his thoughts are directed elsewhere: πυρὸς δ’ ἐπεμαίετο τέχνην, ‘he sought after the τέχνη [‘skill, art’] of fire’ (HH Hermes 108). After succinctly describing (line[s] missing?) how Hermes prepared the materials for a fire drill (lines 108-110), the poet proclaims (line 111): Ἑρμῆς τοι πρώτιστα πυρήϊα πῦρ τ’ ἀνέδωκε ‘Hermes first brought forth firesticks and fire’42. His attention to the τέχνη of fire and the associated invention of firesticks have concrete results: when the fire has grown strong, Hermes slaughters two of the cattle, roasts flesh of the animals, and makes of it an offering in twelve portions (lines 112-137). Immediately thereafter, the poet – for the first time (line 138) – refers to Hermes as divine, denoting him as δαίμων. Commentators have noted the conspicuous contrast with the earlier characterization of Hermes (though he is ἀθάνατος ‘deathless’) as hungering for the cooked flesh of the cattle (lines 130-33). The gods do not ingest the offerings made to them but savour the aromas: ‘Hermes’ reaction here verges on the human’ (Vergados 2013:343)43. In this regard, as in others, Hermes is a liminal figure.

    The provision of fire to humankind is associated with another figure of deception, Prometheus. This fundamental similarity to Hermes as fire-initiator can be elaborated: in the tradition preserved by Hesiod in his Theogony, Prometheus’ theft of fire (lines 565-7; cf. Works & Days 47-52) is no less compounded with an account of sacrificial ritual, and one that involves, as with Hermes, illicit and deceitful behavior on the part of the sacrificer (lines 535-57). Here there is no suggestion of the theft of a sacrificial victim, but Prometheus contrives to dupe Zeus into choosing fat-wrapped bones rather than flesh as his sacrificial portion. As early as the work of Kuhn (1886:17-18) the claim was advanced that Greek ‘Prometheus’ and Sanskrit pramantha- ‘firestick’ share a common origin, but the hypothesis is no longer considered credible. Instead, the Greek term appears to be related to Sanskrit pra-math-, having the sense ‘to rob, snatch away’ (Narten 1960). If so, Prometheus is the ‘[Fire-] Snatcher’, and the account of the primeval snatching of fire in Greek and Indic tradition (identified with Mātariśvan) is of primitive Indo-European origin44. But we can now see why full studies of the topic need to include Purūravas and Hermes as well.



    Nārada is no doubt best known as a sage (ṛṣi), as which he already appears in the Atharva Veda; but the name is also that of a fairly prominent Gandharva. In the two lists of the Gandharvas mentioned above, Nārada comes explicitly in the sixteenth (and last) place (1,59,43; 114.46); and in the account of music at the court of King Nahuṣa (son of Āyus), Viśvāvasu and Nārada are the only names given (5,11.12). Sörensen (1904) has separate entries for the divine Devarṣi, son of Parameṣṭhin, and the Devagandharva, son of Muni, but suggests that the two were originally identical; similarly, Hopkins suggests a tendency for Gandharvas ‘to become earthly seers and act like saints’. Mani (2002:526, 529) is aware of the distinction but ascribes it to different births of the same individual sage.

    Nārada has many roles apart from that of musician. He is a messenger – for instance, he brings from heaven Pāṇḍu’s advice that Yudhiṣṭhira hold a rājasūya ritual (2,11.66). He acts as mediator, notably in the conflicts of Bhīṣma and Rāma Jāmadagnya (5,186.2-4) and of Arjuna and Aśvatthāman (10,14.12). He teaches Sāṃkhya philosophy to a thousand sons of Dakṣa (1,70.6), and by interrogating Yudhiṣṭhira instructs him in statecraft (2,5). More generally, being a great traveller, he distributes news, warnings, advice and prophecies. He also has an unsavory reputation as one who provokes quarrels and enjoys them (9,53.18), though elsewhere an account is given of his virtues (12,223). He is prominent in the Nārāyaṇīya (12,321-39). But we shall focus on just two aspects of his activity.


    14. Lyre and staff

    The relation between Gandharvas and music is treated usefully by Wiersma-Te Nijenhuis (1970: 62-71) in her commentary on the text and translation of the Dattilam. According to the second shloka of her text: ‘In the very beginning music (gāndharva) [was given] by the Self-existing one (Svayambhū) to Nārada and the other [gandharvas]. Then it was duly taken down to earth by Nārada’: she also mentions the myth of origin of Sanskrit theatre, given at the start of the Naṭyaśāstra. The primal performance was devised in heaven by the Creator Brahmā, who entrusted to Nārada the songs that were part of the event.

    In addition the same author adduces a valuable passage from the epic. During the Great War, Balarāma goes on pilgrimage. Having bathed in the Yamunā River, he is sitting with sages and Siddhas when the holy ṛṣi Nārada arrives. The sage is described as follows (9,53.15-18):

    With his mop of matted hair, and wearing golden rags, the great ascetic holds a golden staff (hemadaṇa) and waterpot [or gourd]. He also has with him that delightful lyre (vīṇām), made of tortoise shell45, with its pleasing sound (kacchapīṃ sukhaśabdām) – being, as he was, skilled in dance and song and honored by gods and brahmins.


    The sage brings news of the climactic duel about to take place between Bhīma and Duryodhana, which Balarāma hurries off to watch. The hairstyle and (apart from the gold) the clothing, staff and vessel are typical of ascetics; but the main point here is that the figure who introduced theatre and song to humanity carries on his travels a lyre linked to the tortoise. Presumably he already used the instrument to accompany his singing in heaven, and it could well have had the seven strings that are already attested in the Brāhmaṇas (Caland 1919:143-4). Moreover, ‘in later mythology he [Nārada] is said to be a friend of Krishna and is regarded as the inventor of the Vīṇā or lute’ (Monier-Williams 1974:537, no doubt alluding to the Purāṇas)46.

    In Greece, Orpheus was famed for his musical talents – his song and his lyre, but the creation of the instrument with which he charmed the shades was assigned to Hermes. The HH Hermes tells of the neonate leaving the cave of his mother, only to be distracted at its threshold when he comes across a tortoise. With this creature he begins to experiment: he strips flesh from shell, over which he stretches hide, adding framing and seven strings of sheep gut to craft the first φόρμιγξ ‘lyre’. The tortoise-shell lyre as described in the HH Hermes is first depicted in Greek art in the late eighth century BC (Maas and Snyder 1989:36-37).

    While the Sanskrit mentions Nārada’s staff and lyre in adjacent shlokas, the Greek separates the two objects and gives them separate origins. Hermes gives Apollo the lyre he invented at their initial reconciliation (line 496), and does not receive Apollo’s promise of the magical golden staff until the definitive reconciliation (529). Nevertheless, both belong to the same context (the Hermes–Apollo interaction), and both are at some point owned by Hermes.


    15. Myth-Teller

    Within the epic Nārada frequently narrates. Among his stories are those of Sunda and Upasunda (1,200-204); Hariścandra (2.11.52 ff.); Gālava (5,104-121); the Origin of Death, together with the Sixteen Kings who lost their sons (7, 1.8*, 12,30-31, 248-50). In response to a question by Vālmīki, he outlines the story of Rāma (Rām. 1,1). Citraratha says that he has heard ‘Nārada and other divine seers’ tell the history of the Kuru dynasty, and then himself goes on to narrate at some length (1,159-173). Moreover, no sharp division can be made between Nārada’s mythic narratives and other sorts of discourse, such as his account of the Assembly Halls of the Lokapālas (2,7-11) or his eulogy of food (13,62).

    However, Nārada does more than narrate stories that are included in the epic; both the start and the finish of the great work affirm that he narrates the epic itself. If the first narration is usually attributed largely to Vaiśaṃpāyana, that is because we are humans. Nārada recited it to the gods, Asita Devala to the ancestors, and Śuka to Rākṣasas, Yakṣas and Gandharvas (1,1.64; 18,5.42).

    Twice in the fourth Homeric hymn, Hermes is presented as engaging in the performance of μῦθοι. The first instance is specified to be an improvisational (ἐξ αὐτοσχεδίης) performance that follows immediately upon his creation of the lyre (lines 52-61): he sings of the intercourse of his parents, of the fame of his ancestry, of the splendour of Maia’s home – a Homeric hymn to Hermes within the Homeric Hymn to Hermes47. In his second performance Hermes intones a Theogony (lines 427-33) as he calms Apollo, who has been angered by finding that two of his cattle have been slaughtered. Hermes now sings authoritatively (κραίνω [used again in line 559 of the oracular bee maidens that Apollo assigns to Hermes]; see Nagy 1990:59) of Gaea and of the birth, portion, and station of the gods; the lines are reminiscent of Hesiod’s Theogony, though they begin with praises for Mnemosyne, the patron of Hermes (see line 430), rather than for her daughters the Muses. Hermes’ role as theogonic poet at this moment in this Homeric hymn cannot be separated from Apollo’s forthcoming gift to him of the staff and marks the tradition here recorded as one that reaches back to an archaic time when the Indo-European poet is also μάντις ‘seer’ no less than κῆρυξ ‘herald’ (see Nagy 1990:59-60).

    However, the comparison of Hermes with Nārada suggests that seer and herald/messenger were only two aspects of a complex role that also included at least musicianship.


    Dhṛtarāṣṭra and Kubera


    Neither of these two figures is commonly discussed in connection with Gandharvas, but we shall see that both merit their place in this essay.


    16. Human king and demigod

    Dhṛtarāṣṭra, a king, is best known as an important figure in the plot of Great Epic. He is the father of Duryodhana, the arch-enemy of the Pāṇḍavas, and he is the half-brother of Pāṇḍu. He does not appear in the long introductory list of partial incarnations in 1,61, but after the Great War Vyāsa explains to Dhṛtarāṣṭra’s wife the significance of the horrific loss of life: the heroes descended to earth to achieve the purposes of the gods. Turning from generalities to individuals, Vyāsa continues: ‘It is said that the wise Gandharva-king called Dhṛtarāṣṭra became in the human world your husband Dhṛtarāṣṭra’ (15,39.8), much as Pāṇḍu incarnated the Maruts, and Vidura (the third and last half-brother in the royal court) incarnated the god Dharma. Similarly, when Yudhiṣṭhira reaches heaven, Indra directs his attention to the intelligent Gandharva-king Dhṛtarāṣṭra, the elder brother of his father (18,4.12).

    A Dhṛtarāṣṭra also appears in sixth position in two somewhat overlapping lists of the names of Gandharvas. The first list appears before the main story starts, in a section that recalls Hesiod’s Theogony and enumerates the sons of Muni, seventh daughter of the demiurge Dakṣa (1,59.41). The second describes the gathering to celebrate the birth of Arjuna (1,114.44). The same Gandharva recurs in Book 14 (9.25, 10.1-8), and only here. To help Yudhiṣṭhira with preparations for the Horse sacrifice, Vyāsa tells the story of Marutta’s sacrifice. Initially, Indra and his priest Bṛhaspati in heaven are hostile to the earthly figures of King Marutta and his priest Saṃvarta, Bṛhaspati’s younger brother (Scheuer 1982:168-180); but later Indra supports his priest’s desire to officiate at Marutta’s ritual. Indra sends two messengers to the king, Agni with an offer of immortality, and Dhṛtarāṣṭra the Gandharva with a threat; but the earthlings stand firm. Indra approaches with a thunderous roar, but Saṃvarta promises to protect his patron. In fact Indra not only participates in the ritual but appears to do so with pleasure (prīto 14,10.27d).

    At first sight the two Dhṛtarāṣṭras are unconnected figures who just happen to share a name, and synchronically the feeble and blind brother of Pāṇḍu cannot possibly be Indra’s emissary. But viewed diachronically the picture recalls that of the two Nāradas. The Gandharva and his human incarnation must once have been very close – arguably the shared name makes them even closer than (say) Indra and his human incarnation Arjuna. So it is interesting that, on the one hand, the only story clearly featuring Dhṛtarāṣṭra the Gandharva exemplifies the theme ‘enemy becomes friend’ (§10); and on the other hand, that the most striking feature in the biography of Dhṛtarāṣṭra the mortal is its bisection. Until the end of the Great War he is officially leader of the Kauravas, the ‘Baddies’ and enemies of the Pāṇḍavas; but thereafter he becomes their friend – loved and respected especially by king Yudhiṣṭhira. Admittedly it is Indra, not his messenger, who explicitly changes from enemy of Marutta to his friend, but (without being named individually) Gandharvas and Apsarases participate in the sacrifice (14,10.26).

    So far this section has alluded to three initially conflictual relationships, which we can label as follows: Indra–Marutta or Bṛhaspati–Saṃvarta (which Vyāsa in 14.5.3 compares with Devas–Asuras); Pāṇḍava–Kaurava – in other words, the main plot of the epic; and, via §10, Apollo–Hermes. However, a fourth relationship, the climax of the main plot of the Iliad, provides a more direct rapprochement. The initial hostility is clear, in that Achilles–Priam epitomizes Greeks–Trojans; but in Book 24 the two men end up in Achilles’ tent, if not exactly as friends, at least sympathizing with each other. But more is involved than the ‘enemies become friends’ theme. As noted by Vielle (1996:122n, 155), Priam, father of the main Trojans (with his fifty sons, corresponds rather clearly to the human Dhṛtarāṣṭra, father of the main Kauravas (all one hundred of them). Furthermore, as we saw in §2, Zeus sends Hermes to help Priam in his undertaking, and the two nocturnal travellers appear almost intimate to each other: the disguised Hermes treats the old man as his father, and is addressed reciprocally as son (24.362, 371, 373). Hermes’ mission parallels that of Dhṛtarāṣṭra the Gandharva in Book 14, and his closeness to the Trojan mortal parallels the Gandharva’s incarnation in the Kaurava mortal48.


    17. Soporific implement

    In Buddhism the regents of the four cardinal points are called the Cātummahārājikā (e.g. Malalasekera 1960 s.v, Banerjea 1956:590-629). The list typically starts in the east with Dhataraṭṭha, the Pali for Dhṛtarāṣṭra, who is presented as lord of the Gandhabbas (= Gandharvas), e.g. in Dīrghanikaya 3.197 (in the Āṭānāṭiya Sutta). In the Hindu tradition the equivalent list, that of the Lokapālas, ‘the Guardians of the Worlds’, includes Kubera, and although the god is usually associated with the north, he can also occupy the east, where he may have replaced Agni (Hopkins 1986:149)49. It may or may not be significant that, from a Greek point of view, Priam’s kingdom is in the east.

    Though Kubera is not a Gandharva, he is usually defined as the God of Wealth – his various names or titles include Dhanapati (Wealth-lord) and Dhanada (Wealth-giver). This already makes him typologically close to Gandharvas (§4, 5), and he is linked to them in several other ways. On Mount Kailāsa he was consecrated to sovereignty (ādhipatya) over Rākṣasas, Yakṣas and Gandharvas (5,109.8), and on Mount Mandara (also located in the north) he, together with his chief attendant, the Yakṣa king Maṇibhadra, is served by 88,000 swift Gandharvas (3,140.4-6; 61.123). Yakṣas yoke his chariot with Gāndharva horses (3,158.23). Each of the four Lokapālas, plus Brahmā (representing the centre), has a divine assembly hall (sabhā), at which crowds of Gandharvas and Apsaras perform their music and dancing (2,7-11), but it is only the account of Kubera’s hall that lists their names. Kubera’s hall is never without these entertainers, who also amuse themselves in his pleasure garden (2,10.13; 3,152.4, 158.37). Though it is sometimes said to belong to Indra, his garden or park, called Caitraratha, was made for Kubera by the Gandharva Citraratha (Hopkins), who claims to be a close friend of Kubera’s, his priyaḥ sakhā (1,158.13).

    The closeness of Kubera and Gandharvas is reinforced by certain mediating concepts. For instance, Kubera can be called Kāmeśvara ‘Love-Lord’, and Kāma ‘Love’ is close to the Gandharvas (Hopkins 1986:164) – compare §6 above. Another of Kubera’s titles (e.g. 2,45.34) is Guhyakādhipati, Lord of the Guhyakas, who carry his hall in 2,10.3. Guhyakas are often aligned or juxtaposed with Gandharvas in lists of spiritual beings; they disappear in air like the fata morgana known as ‘Gandharva-cities’. While recognizing that Guhyakas are sometimes a distinct category, Hopkins considers it probable that theirs ‘was a general name for all the spirits of concealment’ (1986:144). The word belongs to the family related to the root guh- ‘conceal, keep secret,’ along with guhā ‘cave’ and guhyam ‘secret, mystery’. Compare too the Atharva Veda passage (8.10.28) which links Kubera and his son to the ‘milking’ of concealment (tapodhā) from the primal figure of Virāj. All of this recalls §12.

    During the great battle against Rāvana, Rāma and Lakṣmaṇa are felled by the invisible Indrajit, but when they are resuscitated Kubera sends them an eyewash enabling them to see creatures that would otherwise be invisible (3,273.10). This recalls Citraratha gifting magical vision to Arjuna (§7), but also the medical skill of Viśvāvasu (§5).

    Kubera provides yet another instance of the enemy-to-friend theme. As we saw in §10, the Pāṇḍavas’ intrusion into his grove made Citraratha angry (cukrodha 1,158.5), much as Duryodhana’s intrusion at Dvaitavana Lake affected Citrasena (who became kruddha ‘angry’, 3,230.21). And when, for the second time, Bhīma intrudes violently on Kubera’s pleasure ground, the Lord of All the Yakṣas is initially angry (cukrodha again – 3,158.22) and mounts his war chariot; but when he reaches the Pāṇḍavas he abruptly becomes friendly.

    Thus the number and variety of links between Kubera and Gandharvas amply justify including the god in the Sanskrit side of our comparison. So let us turn to Mbh. 3, where Arjuna travels to heaven to stay with his divine father Indra. Having reached the Himalayas, the hero receives weapons from the set of gods who visit him – first Śiva, then the Lokapālas (listed clockwise, as usual). Yama from the south presents Arjuna with his staff, and Varuṇa from the west gives his nooses. Kubera from Mount Kailāsa (in the north) gives something more surprising, while Indra (who has to represent the remaining quarter) defers his gift until Arjuna reaches heaven. Kubera’s gift (3,42.33) is ‘a weapon of disappearance of which he is fond’ (astram antardhānaṃ priyaṃ), and it has two further properties: it dissipates energy, vigour and splendour (ojas-tejo-dyuti-haraṃ), and it puts (the foe) to sleep (prasvāpanam, from svap-‘sleep’, cognate with ὕπνος and Latin somnus)50.

    A very similar weapon, named prasvāpa,is mentionedin the duel between Bhīṣma and the brahmin Rāma Jāmadagnya. In a dream Bhīṣma is told to use this favourite weapon (astraṃ sudayitam 5,184.11), which he will remember from a previous existence. It will give him victory by putting Rāma to sleep, but, using his beloved ‘awakening weapon’ (astrena dayitena…saṃbodhanena), Bhīṣma should then resuscitate him. The next day he is about to use ‘Sleepmaker’ when, following advice from Nārada, he withdraws it (5,186.7).

    Like Kubera, Hermes is not a particularly belligerent deity (cf. Il. 21.497-501). What he typically carries is a ῥάβδος (later the caduceus), which ‘seems to have combined the functions of a shepherd’s staff, a herald’s sceptre and a magic wand’ (Richardson 2000:309). When Hermes is told by Zeus to help Priam on his secretive nocturnal visit to Achilles, he puts on his sandals (which enable him to travel swiftly) and takes his wand, ‘with which he lulls to sleep the eyes of whom he will, while others again he wakes from their slumber’ (24.343-4). Though the text does not say so, he surely uses the wand both to put to sleep the Greek guards and, well before dawn, to waken Priam.

    If we focus first on the weapons or implements, Kubera’s simply induces sleep. Bhīṣma may well have two weapons with contrasting functions, but the wording does not exclude two uses of a single weapon. As for the users, the question arises why Kubera and Bhīṣma should share a somewhat unusual weapon, but we cannot attempt to answer it – Bhīṣma is mentioned here primarily as helping to bridge the gap between the implements of Kubera and Hermes. Our main point is that, although a wand is hardly a weapon, the gap is small enough to provide an additional rapprochement between the two gods51.


    Concluding remarks


    Clearly this paper could have been organized very differently. The decision to prioritize the Sanskrit and base the headings on Gandharvas has meant breaking up HH Hermes into component episodes and shuffling their sequence – ignoring alternative presentational options. To distinguish seventeen themes has served as a convenience, but some of them could have been run together, or subdivided. Both dossiers are so rich that they contain not only details that we have failed to study but whole themes – for instance, exchange as such; gambling and luck; the complex that links secrecy, night, sleep and death. We noted at the start the problem of the delimitation of dossiers, and we have now met many instances. To study Gandharvas is to find oneself willy-nilly involved with other types of supernatural such as Apsarases, Guhyakas, Yakṣas; with deities such as Vāc and Kubera; with more or less human sages, and with heroes who incarnate gods. Similarly, Hermes leads one on to nymphs, Satyrs and the Titan Prometheus; to gods like Pan and Apollo; to mortals such as Autolycus and Odysseus.

    Our comparison between the dossiers of Hermes and the Gandharvas implies an early Indo-European proto-dossier, which may of course have overlapped with other proto-dossiers. An obvious question is what if anything held this one together, or at least gave it some degree of coherence. One hypothesis might focus on the notion of ‘betweenness’ or mediation, and could cite the God-man margin (§2), the roles of messenger, go-between and intermediary (§8), the trickster-like ambiguity implicit in the enemy-to-friend theme (§10), and the transmission of knowledge, music, myths and rituals. But a more promising approach seems to us to build on the Dumézilian concept of ideology. In the discussion of wealth and fertility in §4 reference was made to third-function gods, and the same reference could have been made in §5 (medicines and plant fertility), in §6 (sexuality), in §12 (theft of property, especially of livestock), and in §17 (large number – 88,000 Gandharvas serve the God of Wealth). Again and again the third-functional interpretation of the Aśvins comes to mind. The Aśvins are gods but relatively close to men, whom they help; they are physicians; they are young, handsome and interested in sex (they seek to seduce Sukanyā, wife of Cyavana). Following a conflict (cf. §10), they are promoted to the rank of soma-drinkers, somewhat as Purūravas is promoted to join the Gandharvas, or Kubera to join the Lokapālas (Hopkins).

    The fact that the Aśvins are closely paired – being twins – suggests another line of thought, which can only be mentioned here. Citraratha and Citrasena are the protagonists of closely matched stories. Citrāṅgada the Gandharva and Citrāṅgada the half-brother of Bhīṣma fight each other for three years. Nārada is very often paired with Parvata (who is in fact his sister’s son). Kubera God of Wealth is often mentioned alongside the Yakṣa-king Maṇibhadra, who is ‘the tutelary deity of travellers and caravans (Sörensen 1904:464; 3,61.123 and 1.112* lines 14-15). One might even speculate about the pairing of Dhṛtarāṣṭra and Pāṇḍu, half-brothers born of the sisters Ambikā and Ambālikā respectively (Pāṇḍu’s death is due to his sexuality). Of course pairing by itself is too imprecise a phenomenon to be of much use for analysis or comparison, but it is interesting that HH Hermes so emphatically pairs the Greek god with his half-brother Apollo.

    A more important issue is whether either of our comparands represents the third function in contexts where the other functions are equally clearly represented. In Maṇipura or Maṇalūra, in the eastern quarter, Citrāṅgadā, daughter of King Citravāhana, represents Arjuna’s third-functional wife (Allen 1996). Dhṛtarāṣṭra fills the third-function slot in Arjuna’s parental generation (Allen 2012), as well as among the Buddhist Regents of the four quarters, where he corresponds to Hindu Kubera. But here is one further argument, taken from 1,158.45-51.

    When presenting his horses to the Pāṇḍavas, Citraratha refers to the thunderbolt of Indra which shattered into a hundred pieces over the head of Vṛtra. The fragments were shared out among the gods, who worship them; so, in this world, something that ensures success (sādhanaṃ kiṃcit) is spoken of as embodying the thunderbolt. The point is exemplified by reference to the varṇas. The text could be more secure and less ambiguous: for instance the bolt of the warrior is his chariot or his horses, that of the serf is his work or his obedience; however, that of the Vaiśya is his gifts (dāna). The southern manuscripts replace dāna with sīra ‘plough’, but the reading of the Critical Edition is better since the gift of horses provides the context, as is clear in the final half-shloka of the speech. Thus the passage suggests an affinity between the Gandharvas as gift-givers and the varṇa that represents the third function.

    If the Gandharvas relate to the third function and Hermes is cognate with the Gandharvas, Hermes too should have his roots in the third function. We avoid claiming that either comparand ‘belongs to’ this function since, in the course of their histories, they may very well have incorporated features from elsewhere, and we have not found a clearly defined context (such as a list, story or ritual) in which Hermes is juxtaposed with obvious representatives of other functions. But Indo-European cultural comparativism is still at a fairly early stage.



    ABV J. D. Beazley (1956), Attic Black-figure Vase-painters

    AV Atharva Veda

    FGrH F. Jacoby (1923–1958), Die Fragmente der griechischen Historiker

    Il. Iliad

    LIMC Lexicon Iconographicum Mythologiae Classicae

    Mbh. Mahābhārata (Critical Edition)

    Od. Odyssey

    Rām. Rāmāyaṇa (Critical Edition)

    RV Rig Veda

    ŚB Śatapatha Brāhmaṇa



    Alexandridou, Alexandra. 2011. The Early Black-Figured Pottery of Attika in Context (c. 630-570 BCE). Leiden: Brill.

    Allen, N.J. 1996. “The Hero’s Five Relationships: a Proto-Indo-European Story.” In Julia Leslie (ed.) Myth and Myth-making: Continuous Evolution in Indian Tradition, 1-20. London: Curzon.

    ________. 1997. “Why did Odysseus Become a Horse?” JASO 26 [‘1995’]:143-154.

    ________. 2002. « Pénélope et Draupadî: la validité de la comparaison. » In A. Hurst and F. Létoublon (eds) La mythologie et l’Odyssée. Hommage à Gabriel Germain, 305-312. Geneva: Droz.

    ________. 2003 (2000). “The Indra-Tullus Comparison.” In Joseph C. Salmons and Bridget Drinka (eds), Indo-European Language and Culture: Essays in Memory of Edgar C. Polomé, part I. General Linguistics 40:148-171.

    ________. 2010. “Hephaestus and Agni: Gods and Men on the Battlefield in Greek and Sanskrit Epics.” In W. van Binsbergen & E. Venbrux (eds) New Perspectives on Myth, 357-372. Haarlem: Shikanda.

    ________. 2012. “Bhārata Genealogy: the Close Parental-Generation Males.” In John Brockington (ed.) Battles, Bards and Brahmans, 39-50.Delhi: Motilal.

    Austin, Colin and S. Douglas Olson. 2004. Aristophanes: Thesmophoriazusae. Oxford: Oxford University Press.

    Banerjea, Jitendra Nath. 1956. The Development of Hindu Iconography. Calcutta: Calcutta University Press.

    Bowra, C. M. 1952. “Orpheus and Eurydice.” Classical Quarterly 2:113-126.

    Brelich, Angelo. 1958. Gli eroi greci: Una problema storico-religioso. Rome: Edizioni dell’Ateneo.

    Brodbeck, Simon P. 2009. The Mahabharata Patriline: Gender, Culture and the Royal Hereditary. Farnham: Ashgate.

    Brown, Edwin L. 1981. “The Lycidas of Theocritus’ Idyll7.” Harvard Studies in Classical Philology. 85:59-100.

    Burkert, Walter. 1985. Greek Religion. Cambridge, MA: Harvard University Press.

    Cairns, Francis. 1983. “Alcaeus’ Hymn to Hermes, P. Oxy. 2743. Fr. 1 and Horace Odes 1, 10.” Quaderni urbinati di cultura classica. N.S. 13:20-35.

    Caland, W. 1919. Das Jaiminīya Brāhmaṇa in Auswahl. Amsterdam: Johannes Müller.

    Chantraine, Pierre. 1968. Dictionnaire étymologique de la langue grecque. Paris: Éditions Klincksieck.

    Chittenden, Jacqueline. 1947. “The Master of Animals.” Hesperia 16:89-114.

    Clay, Jenny Strauss. 2006. The Politics of Olympus. Second edition. London: Bristol Classical Press.

    Detienne, Marcel and Jean-Pierre Vernant. 1989. The Cuisine of Sacrifice among the Greeks. Translated by Paula Wissing. Chicago: University of Chicago Press.

    Doniger, Wendy and Brian F. Smith (introduction, notes, translation). 1991. The Laws of Manu. Harmondsworth: Penguin.

    Dumézil, Georges. 1929. Le Problème des Centaures: Étude de mythologie comparée indo-européenne. Paris: Guethner.

    ________. 1965. Le livre des héros: Légendes sur les Nartes. Paris: Gallimard.

    ________. 1970. The Destiny of the Warrior. Translated by A. Hiltebeitel. Chicago: University of Chicago Press.

    ________. 1985. L’oubli de l’homme et l’honneur des dieux. Paris: Gallimard.

    ________. 1986. Loki. Paris: Flammarion.

    ________. 1995. Mythe et épopée I. II. III. Paris: Gallimard.

    Durante, Marcello. 1962. “Ricerche sulla preistoria della lingua poetica greca: l’epiteto.” Rendiconti delle sedute dell’Accademia Nazionale dei Lincei 17:25-43.

    Eitrem, S. 1912. “Hermes.” In Paulys Real-Encyclopädie der Classischen Altertumswissenschaft. Volume 8.1, pages 738-792.

    Furley, William D. 1996. Andokides and the Herms: A Study of Crisis in Fifth-Century Athenian Religion. London: Institute of Classical Studies.

    Gonda, Jan. 1962 [1956]. Les religions de l’Inde, I : Védisme et Hindouisme ancien. Translated by L. Jospin. Paris: Payot.

    ________. 1985 [1965]. Change and continuity in Indian Religion. New Delhi: Munshiram Manoharlal.

    Haft, A. 1996. “‘The Mercurial Significance of Raiding’: Baby Hermes and Animal Theft in Contemporary Crete.” Arion 4.2:27-48.

    Hedreen, Guy Michael. 1992. Silens in Attic Black-Figure Vase-Painting. Ann Arbor: University of Michigan Press.

    Heubeck, Alfred and Arie Hoekstra. 1989. A Commentary on Homer’s Odyssey. Volume 2. Oxford: Clarendon Press.

    Heubeck, Alfred, Stephanie West, and J. B. Hainsworth. 1988. A Commentary on Homer’s Odyssey. Volume 1. Oxford: Clarendon Press.

    Hocart, A.M. 1970 [1936]. Kings and Councillors: an Essay in the Comparative Anatomy of Human Society. Chicago, London: University of Chicago Press.

    Hopkins, E. Washburn. 1986 [1915]. Epic Mythology. Delhi: Motilal Banarsidass.

    Janko, Richard. 1978. “A Note on the Etymologies of διάκτορος and χρυσάορος.” Glotta 56:192-195.

    Johnston, Sarah Iles. 2002. “Myth, Festival, and Poet: The Homeric Hymn to Hermes and its Performative Context.” Classical Philology 97:109-132.

    Jost, Madeleine. 2012. “Hermes.” In S. Hornblower and A. Spawforth, with E. Eidenow (eds) Oxford Classical Dictionary. Fourth edition. Oxford: University Press.

    Kaibel, Georg. 1878. Epigrammata Graeca. Berlin: G. Reimer.

    Karavites, Peter. 1992. Promise-Giving and Treaty-Making: Homer and the Near East. With the collaboration of Thomas Wren. Leiden: E. J. Brill.

    Keith, A. Berriedale. 1937. “A new explanation of the Gandharvas.” Journal of the Indian Society of Oriental Art 5:32-37.

    Kirk, G. S. 1985. The Iliad: A Commentary. Volume 1. Cambridge: Cambridge University Press.

    Kuhn, Adalbert. 1886. Mythologische Studien. Volume 1. Die Herabkunft des Feuers und des Göttertranks. Gütersloh: C. Bertelsmann.

    Lambrinoudakis, Vassilis (in collaboration with Zoe Sgouleta and Spyros Petrounakos). 2004. “Consecration, Foundation Rites.” Addendum to Thesaurus Cultus et Rituum Antiquorum. Volume 3. Los Angeles: J. Paul Getty Museum.

    Lubotsky, A. 2007. “The Indo-Iranian Substratum.” In Christian Carpelan, Asko Parpola, and Petteri Koskikallio (eds), Early Contacts Between Uralic and Indo-European: Linguistic and Archaeological Considerations, 301-317. Second edition. Helsinki: Suomalais-Ugrilainen Seura.

    Maas, Martha and Jane McIntosh Snyder. 1989. Stringed Instruments of Ancient Greece. New Haven: Yale University Press.

    Macdonell, A. A. 1974 [1898]. Vedic Mythology. Delhi: Motilal Banarsidass.

    Malalasekera, G.P. 1960. Dictionary of Pali Proper Names. London: Luzac.

    Malamoud, Charles. 1989. Cuire le monde : rite et pensée dans l’Inde ancienne. Paris: Découverte.

    Mani, Vettam. 2002 [1975]. Purāṇic Encyclopaedia: a Comprehensive Work with Special Reference to Epics and Purānic Literature. Delhi: Motilal Banarsidass.

    Mitchell-Boyask, Robin. 2008. Plague and the Athenian Imagination: Drama, History and the Cult of Asclepius. Cambridge: Cambridge University Press.

    Monier-Williams, M. 1974 [1899]. A Sanksrit–English Dictionary. Oxford: Clarendon Press.

    Nagy, Gregory. 1990. Greek Mythology and Poetics. Ithaca: Cornell University Press.

    Narten, Johanna. 1960. “Das vedische Verbum math.” Indo-Iranian Journal 4:121-135.

    Nilsson, Martin P. 1967. Geschichte der griechischen Religion. Volume 1. Munich: C. H. Beck’sche Verlagsbuchhandlung.

    Oberlies, Thomas. 2000. “Pūṣans Zahnlücken und Hermes Vorliebe für Backwerk: Ererbte Strukturen des Pantheons der ṛgvedischen Religion.” In Bernhard Forssman and Robert Plath (eds) Indoarisch, Iranisch und die Indogermanistik, 369-389. Wiesbaden: Reichert.

    ________. 2005. “Der Gandharva und die drei Tage währende ‘Quarantäne’.” Indo-Iranian Journal 48:97-109.

    Oldenberg, Hermann. 1993 [1916]. The Religion of the Veda. Translated by S. B. Shrotri. Delhi: Motilal Banarsidass.

    Panchamukhi, R.S. 1951. Gandharvas and Kinnaras in Indian Iconography. Dharwar: Kannada Research Institute.

    Parker, Robert. 2011. On Greek Religion. Ithaca, NY: Cornell University Press.

    Penglase, Charles. 1994. Greek Myths and Mesopotamia: Parallels and Influence in the Homeric Hymns and Hesiod. London: Routledge.

    Pratt, Louise H. 1993. Lying and Poetry from Homer to Pindar: Falsehood and Deception in Archaic Greek Poetics. Ann Arbor, MI: University of Michigan Press.

    Radermacher, Ludvig. 1931. Der homerische Hermeshymnus. Vienna: Hölder-Pichler-Tempsky.

    Richardson, Nicholas. 2000 [1993]. The Iliad: a Commentary. Volume 6. Cambridge: Cambridge University Press.

    ________. 2007. “The Homeric Hymn to Hermes.” In P. J. Finglass, C. Collard, and N. J. Richardson (eds.), Hesperos: Studies in Ancient Greek Poetry Presented to M. L. West on his Seventieth Birthday, 83-91. Oxford: Oxford University Press.

    ________. 2010. Three Homeric Hymns. Cambridge: Cambridge University Press.

    Russo, Joseph, Manuel Fernández-Galiano and Alfred Heubeck. 1992. A Commentary on Homer’s Odyssey. Volume 3. Oxford: Clarendon Press.

    Scanlon, Thomas F. 2002. Eros and Greek Athletics. Oxford: Oxford University Press.

    Scheuer, Jacques. 1982. Śiva dans le Mahābhārata . Paris: PUF.

    Schmitt, Rüdiger. 1967. Indogermanische Dichtersprache. Darmstadt: Wissenschaftliche Buchgesellschaft.

    Schroeder, Leopold von. 1908. Mysterium und Mimus im Rigveda. Leipzig: Haessel.

    Shelmerdine, Susan C. 1986. “Odyssean Allusions in the Fourth Homeric Hymn.” Transactions of the American Philological Association 116:49-63.

    Siebert, Gérard. 1990. “Hermes.” In Lexicon Iconographicum Mythologiae Classicae V.1, 285-387. Zurich and Munich: Artemis Verlag.

    Sörensen, S. 1904. Index to the Names in the Mahābhārata. London: Williams & Norgate.

    Stengel, P. 1910. Opferbräuche der Griechen. Leipzig: Teubner.

    Sterckx, Claude. 2002. Le fils parfait et ses frères animaux : Lugus, Pan et les Draupadeya. Brussels: Société belge d’études celtiques.

    van Berg, P.-L. 2002. “Hermès et Agni : Etude comparative.” In Françoise Labrique (ed.) Religions méditerranéennes et orientales de l’antiquité, 185-216. Cairo: Institut français d’archéologie orientale.

    Vasilkov, Ya.V. 1989-90. “Draupadī in the Assembly-Hall, Gandharva Husbands and the Origin of the Gaṇikās.” Indologica Taurinensia 15-16:387-398.

    Vergados, Athanassios. 2013. The Homeric Hymn to Hermes. Berlin: Walter de Gruyter.

    Vernant, Jean Pierre. 2006. Myth and Thought among the Greeks. Translated by Janet Lloyd with Jeff Fort. New York: Zone Books.

    Vielle, Christophe. 1996. Le mytho-cycle héroïque dans l’aire indo-européenne : correspondances et transformations helléno-aryennes. Louvain-la-Neuve: Institut Orientaliste.

    ________. 2005.  “Correspondances et transformations en mythologie comparée indo-européenne.” In S. Vanséveren & F. Doyen(eds) Modèles linguistiques et idéologies : “Indo-européen” III : Les Indo-Européens et le modèle comparatif indo-européen dans les sciences humaines, 329-361. Bruxelles : Ousia.

    Walcot, P. 1979. “Cattle Raiding, Heroic Tradition, and Ritual: the Greek Evidence.” History of Religions 18:326-351.

    Watkins, Calvert. 1995. How to Kill a Dragon: Aspects of Indo-European Poetics. Oxford: Oxford University Press.

    West, M. L. 1972. Iambi et Elegi Graeci. Volume 2. Oxford: Clarendon Press.

    ________. 1978. Hesiod: Works and Days. Oxford: Clarendon Press.

    ________. 1982. Homer, Hesiod, and the Hymns: Diachronic Development in Epic Diction. Cambridge: Cambridge University Press.

    ________. 1992. Ancient Greek Music. Oxford: Clarendon Press.

    ________. 1997. The East Face of Helicon. Oxford: Oxford University Press.

    ________. 2007. Indo-European Poetry and Myth. Oxford: Oxford University Press.

    Wiersma-Te Nijenhuis, E. (introduction, translation, commentary). 1970. Dattilam: a Compendium of Ancient Indian Music. Leiden: Brill.

    Wijesekera, O. H. De A. 1994. “Vedic Gandharva and Pali Gandhabba.” In Buddhist and Vedic Studies: a Miscellany, 175-212. Delhi: Motilal Banarsidass.

    Woodard, Roger D. 2006. Indo-European Sacred Space. Urbana and Chicago: University of Chicago Press.

    ________. 2013. Myth, Ritual, and the Warrior in Roman and Indo-European Antiquity. Cambridge: Cambridge University Press.

    Yalouris, Nikolaos. 1953. “ΠΤΕΡΟΕΝΤΑ ΠΕΔΙΛΑ.” Bulletin de correspondance hellénique 77:293-321.


    1This paper derives from a presentation given by one of us (NA) at St Antony’s College, Oxford, in November 1994. Awareness of the ‘dossier problem’ was one factor in the delay in publication.

    2Similar points were already made by Dumézil 1929:97-104.

    3In the absence of page numbers, references to these authors refer to these passages. References to epic are to the Critical Edition of the Mahābhārata, if not otherwise indicated.

    4The HH Hermes has long been recognized as a somewhat particularistic Homeric hymn. The work’s distinctiveness among archaic (oral) poetic compositions has been attributed to what some have identified as a fundamentally comic nature (see, inter alia, Janko 1982:149; Richardson 2007; 2010:19-20; and especially Vergados 2013, with bibliography). The characterization may be in need of reconsideration. Penglase has observed that the Hymn to Hermes is distinct ‘from the other three long Homeric hymns,’ (1994:183) in that in the latter set the motif of ‘the journey,’ and affiliated motifs, is depicted in a way that suggests Mesopotamian influences; such influence is absent in the Hymn to Hermes: ‘the other hymns appear to be a result of conscious creation relying on Mesopotamian ideas and material’ (1994:185).

    5Gandharvas have also been compared to youths in the ‘men’s houses’ (ghotul) of the Muria Gonds, who are Dravidian-speaking ‘tribals’ of middle India (Vasilkov 1989-90).

    6The suitors number 108 in Homer, 136 in Pseudo-Apollodorus (though he only gives 129 names).

    7Brelich 1958:357. See Brelich’s discussion on pages 357-60; see also, inter alia, Eitrem 1912:779-80; Heubeck and Hoekstra 1989:253; Siebert 1990:286.

    8On the duo Hestia and Hermes interpreted as expressions of space and movement, see Vernant 2006:157-196, who writes (page 161):

    Hestia appears capable of ‘centering’ space while Hermes can ‘mobilize’ it because, as divine powers, they are the patrons of a series of activities dealing with the organization of earth and space and even constituting, in terms of praxis, the framework within which, for the ancient Greeks, the experience of spatiality took place . . . .

    The significance of the observation for the Indo-European ancestry of Hermes will become apparent later.

    9Though not all agree; for the positive view, and comments on the negative one, see Richardson 2000:308.

    10The earliest-known images of Hermes (between the late eighth and first quarter of the seventh centuries BC) occur on the bronze plaques from Symi Viannou in Crete, from the sanctuary of Hermes and Aphrodite (mentioned below). ‘Hermes is shown as a beardless, naked, young male holding a staff’ (Alexandridou 2011:61).

    11According to Yalouris (1953:295), the earliest representation of winged footwear is provided by an image of Perseus on a terracotta metope from Thermos, ca. 625 BC.

    12Alexandridou (2011:62) judges regarding early Attic black-figure vases that ‘on almost all the vases from funerary contexts he appears between sphinxes, and on those from sanctuaries mostly between sirens,’ until the middle of the sixth century.

    13This is not to suggest that the god is always depicted with winged attributes; this is certainly not the case. Siebert (1990:384) provides an inventory, according to sphere of activity, of ratios of images depicting Hermes (1) with winged shoes as opposed to (2) with wingless footwear / no footwear. Impressionistically, the most significant variation appears to be in the case of ‘Hermès et la musique’: in this realm of action the god is represented with winged shoes seven times more often than not (21:3). There are three areas for which winged-shoe representations are in the minority, though the differences are marginal in two of the three: ‘Hermès psychopompe et chthonien’ (13:17); ‘Hermès et l’amour’ (9:19); and ‘scènes cultuelles’ (7:10). When Hermes is depicted involved in combat the ratio shows no significant difference (15:13); however, Siebert makes an interesting observation: ‘L’absence de chaussures aillés est concomitante avec le port de l’épée.’ Siebert would see this ‘héroïque’ representation of Hermes as the inspiration for the depiction of the god without winged shoes.

    14The name Gandharva is often linked with gandha ‘smell, odour’. Thus, in the domain of Gandharvas and Apsarases Virāj provides sweet (puṇya) odour when milked by Vasuruci, his brother Citraratha serving as calf (AV 8.10.27). Conceivably the link relates to that between perfume and sex (cf. Mbh. 1,155.34-5 and §6). Hopkins raises the idea of confusion between Gandharvas and Gāndhāra in the north-west of the sub-continent.

    15See Durante 1962:28 and Schmitt 1967:142-9, both with bibliography of earlier work. Also see Heubeck, West and Hainsworth 1988:369.

    16So named, according to Hesiod (Works & Days 81-2) because all the gods had given her a gift. West (1978:164) accurately observes that ‘the reason given is not sufficient to account for her having this name, any more than Pan was really so called ὅτι φρένα πᾶσιν ἔτερψεν’, ‘because he made all [the gods] happy’ (Homeric Hymn to Pan 47).

    17For an inventory of such usages, and associated discussion, see West 1978:164-6.

    18The adjective νόμιος is more typically used of Apollo; see Austin and Olson 2004:304.

    19On the affiliations of this hymn, and particularly lines 28 ff., with the HH Hermes, see Janko 1982:184-5.

    20On this and the several other wooden images of Hermes mentioned by Pausanias, see Siebert 1990:295. On Athena Polias and Hermes, see, inter alia, Mitchell-Boyask 2008:157-9, with references.

    21See, inter alia, Siebert 1990:315; Parker 2011:234-5, with bibliography, especially the work of Lebessi on the sanctuary and its votive artefacts. For the plaque depicting the god within a tree, see Lambrinoudakis 2005:316 (plate 59).

    22See Siebert 1990:289, 294-5, with associated images and bibliography. See also the remarks of Furley 1996:17.

    23A homophone (from εἴρω) is used to denote ‘earrings’ at Il. 14.182 and Od. 18.297.

    24See Nilsson 1967:1:Tafel 33.1 and compare the comments of Burkert 1985:156.

    25For detailed discussions, see Woodard 2006, passim, but especially those of Chapter 3.

    26On affiliation of Hermes with athletic eroticism, see Scanlon 2002:250–255

    27This is the first literary mention of the Satyrs; the earliest such reference to the Sileni is in the Homeric Hymn to Pan (see Richardson 2010:252 with bibliography).

    28Starred references refer to appendixes (which begin 1.) or to footnotes that contain text rejected by the Critical Edition.

    29For discussion of the sense of the term, see, inter alia, Janko 1978.

    30On which matter, see, inter alia, Richardson 2000:276-8.

    31On the formula θυμὸν ἔχοντες, with an adjective specifying the nature of θυμός, see Vergados 2013:450.

    32As also at Homeric Hymn to Demeter 434 and Theognis 81.

    33See the discussion of Cairns 1983.

    34For discussion see Vergados 2013:549, with bibliography. On the Homeric notion of φιλότης, see Karavites 1992:48–58.

    35For detailed expositions, see Dumézil 1970:29-32; 1995:1:279-80; Allen 2003; Woodard 2013:242-3, 253-4.

    36See Johnston 2002:111–15 (and cf. Walcott 1979; Haft 1996), though the emphasis on the raid as a rite de passage is misplaced within an Indo-European context (and see Vergados 2013:285–6).

    37For the relation between the Hymn and composition in performance, see Vergados 2013:73-5.

    38Russo et al 1992:96. On similarities between Odysseus and Hermes, see, inter alia, Pratt 1993:55-67; Vergados 2013:665-7; and, especially, Shelmerdine 1986, with references to earlier work.

    39Also Homeric Hymn to Demeter 280.

    40On the legal quality of the phrasing here, see Richardson 2010:202.

    41The similarities between the infant Hermes and the lovable, playful, childish Krishna of later Vaishnavism lie outside our scope.

    42On the interpretation of the sense of ‘first fire’, in line 111, see Richardson 2010:173.

    43See the discussion of Clay 2006:122-138.

    44In addition to Narten, see Watkins 1995:256n3; West 2007:273-4. On Prometheus and the technology of fire, see Vernant 2006:263-73.

    45This is to follow the translations of Ganguli and Meiland. Wiersma-Te Nijenhuis (1970:82) thinks kacchapīṃ may refer to the shape of the instrument, not its material. In any case kacchapa means ‘tortoise’ or ‘turtle’.

    46As was seen by Dumézil (1985: 226-7, cf. 1965:161-3, 1986:149), comparison is possible with one of the Nart traditions from Ossetia, in which the invention of a twelve-stringed instrument is ascribed to the ambiguous and trickster-like figure of Syrdon. Syrdon, whose family lives underground, one night he steals a cow, whose owner tracks it down and kills the family. Finding them bubbling in a cauldron, Syrdon makes the fændyr using the hand of his eldest son and, as strings, the blood vessels of his other children. When he plays in the village square,the Narts are delighted and treat him as a brother, taking the instrument.

    47See the discussions in Richardson 2010:163; Vergados 2013:271.

    48For the return journey from Achilles’ tent Hermes in person yokes the horses and mules – compare Rigvedic Gandharva in §4. A fuller version of this paragraph could take into account the myths and rituals associated with soma. Soma is needed for the three-fire solemn sacrifices, and its acquisition involves both theft and purchase (e.g., Gonda 1985:68-9). Hector is needed for cremation, and acquiring his body involves the skills of Hermes as well as the rich ransom provided by Priam. As Malamoud notes (1989:56, citing ŚB, the ritual purchase of somaest le prototype de tous les achats.’

    49Reasons why ‘the East is called the first quarter’ are given in Mbh. 5,106.

    50Elsewhere, during the Khāṇḍava forest fire, Kubera’s weapon is a śibikā, which normally means ‘palanquin’ (1,218.31). The Vulgate reads gadā ‘a spiked club’, and elsewhere a brief simile presents the god as using the same weapon (3, 1.4*, line 235). Gandharvas may be shown holding clubs as well as lyres (Banerjea 1956: 352).

    51Possibly Hermes’ wand should also be compared with the stick in the bridal bed representing Viśvāvasu (§6), as well as with Nārada’s staff (§14).