2. Horses, Heroes, and Sacrifice

The prominent position of the formula ὠκέες ἵπποι and related expressions attests the special role played in Greek epic by horses and the physical abilities that distinguish them. The similarity of formulas in both Indian and Iranian oral poetries makes it clear that the phrase indicates a common poetic inheritance. It is a sensible assumption that a formula about horses that is important enough to be maintained in all three cultures would accompany other poetic and cultural treatments of horses that were also commonly inherited. In fact, the similarities in the treatment of horses found in the descendant traditions do extend beyond vocabulary.

They are especially manifest in the relationship between horses and heroes, which appears to be rooted in a common predisposition toward the anthropomorphizing of horses and the concomitant hippomorphizing of humans. [1] Perhaps this tendency is to be expected given the nature of the genres in which our evidence is found, yet the particularities of this phenomenon are sufficiently unique that a common inheritance may reasonably be sought. Additionally, research into a prominent horse sacrifice ritual in the parent culture reveals several insights into the nature of this common inheritance, namely that a ritualistic shift in identity between horse and human was central to the event. A survey of our Greek, Sanskrit, and other IE evidence will demonstrate the case for common inheritance and provide a vantage point from which surprising aspects of the Greek treatment of horses become comprehensible. I will first review some idiosyncrasies of the relationship between horses and heroes in Greek epic, in order to isolate it as a phenomenon worthy of this sort of analysis. Then I will evaluate the evidence for the PIE horse sacrifice and discuss its relevance to the current issue, ultimately suggesting that a unique model for understanding the relationship between horses and humans informed that ritual and that that model, if not the ritual itself, exerted lengthy influence over Greece.

Before proceeding to cross-cultural and historical parallels, then, I will discuss some special details of horse/hero comparisons that are generally found in Greek epic. The fact that there is a close connection between the two is not surprising, but the depth and nature of this link is truly unique and not, I think, common outside of the IE world. First, the similarity between horses and heroes begins at the level of ontology. [2] Not only are both essential figures of martial excellence, but horses are seemingly unique in the animal world for sharing with humans, especially heroic humans, the capacity for divine lineage and semi-divine identity. Several Greek gods have immortal horses that are known by name, such as those of Ares, Helios, and the Dioskouroi. [3] Yet immortal horses that live among mortals are more important for the current discussion, and there are several. For example, Pausanias describes the very powerful horse, Areion, owned by Herakles and then by Adrastus, which was born from the coupling of Poseidon and Demeter when both gods had assumed the form of horses: [4] This horse is also mentioned in the Iliad in connection with other divine horses: The local Trojan horses mentioned here are the divine horses given to Laomedon by Zeus as recompense for the abduction of Ganymede. The horses of Aeneas, which feature so prominently in the Iliad due to their abduction by Diomedes and subsequent victory in the chariot race at the funeral games of Patroclus, are descendants of these horses. [5] In book XX, Aeneas recalls the story of the three thousand mares of Erichthonius, which mated with Boreas, the god of the North Wind, and produced twelve colts that could run over the top of the sea. [6]

Furthermore, part of what makes Achilles special on the battlefield is that he has three horses attached to his chariot—unlike most fighters, who have only two—and that two of his are immortal. At the time of the events of the Iliad he has only recently added the mortal horse, Pedasos, while his permanent team of Xanthos and Balios are the immortal offspring of the Harpy Podarge, an especially appropriate name for the progenitor of famous steeds. Horses, then, share with humans, especially heroic humans, an ontological position unique among living things: they may descend from the gods and maintain some divine capacity derived from them. In the same way that heroes are marked out from the rest of humanity by their divine lineages, so then are their horses. Like the demigods they are descended from immortals, equine or godly, and in some cases capable of inheriting from them immortality themselves, as Herakles did from Zeus and his horse Areion did from Poseidon and Demeter.

Other divine animals do exist in the form of the signature animals of the gods, such as the owl of Athena. Yet this does not mean that these animals are immortal, at least not that they are immortal in the same way that gods and divine horses can be, because these animals are not usually given individual identities, names, or genealogies. They are also depicted only as the companions and emblems of the gods, and not as their descendants, a status which seems reserved for humans and horses alone. They are essentially iconographic and rarely constitute actors on the mythical stage. [7] The character of divine horses seems to be fundamentally different from that of other animals, and to draw them ontologically closer to the heroes whom they accompany than to other animals.

The potential of horses to share with heroes characteristics of essential identity extends to a similarity in poetic treatment as well. Horses within the Iliad are regarded as heroes by the human characters of the poem as well as by its narrator, and occasionally treated like their owners themselves. For example in book VIII Hector urges his four horses to battle, [8] calling them by name and saying: The fact that these horses drink wine at all distinguishes them from other animals since wine is a mark of human civilization. This act essentially isolates a species boundary, a boundary that positions horses as closer to humans than to other animals. The use of the phrase ὅτε θυμὸς ἀνώγοι is also striking and is translated here as “whenever your heart bade,” and is assumed to refer to the horses. It could, however, be argued that this clause refers to Andromache’s desire rather than to the horses’, that it should be translated as “whenever her heart bade.” Yet the formula πιέειν ὅτε θυμὸς ἀνώγοι appears elsewhere twice in the Homeric corpus and in each case πιέειν complements ἀνώγοι, [10] so the heart of the figure by whom the wine is to be drunk is the subject of the verb ἀνώγοι. If that relationship holds true in this usage then the horses here display a degree of agency otherwise reserved for humans. It may be pointed out that this entire line was athetized by the Alexandrian editors, signaling their discomfort with the depiction of the horses drinking wine at all. Be that as it may, the image did appear in the ancient tradition, and did so because the offering of wine to heroes was a regular feature of Greek oral song and these horses are here treated as heroes, rather than as ordinary animals.

This attitude toward horses and heroes is also observable in the narrator of the poem himself. Near the end of the catalogue of ships he asks the goddess for special inspiration so that he may know who the very best of the warriors at Troy were: Her inspired answer is:

The narrator in fact also often compares heroes to horses, and even interweaves elements of the horses and their owners’ identities. The epic fighters in general wear horsehair-crested helmets, giving the impression that they have a horse’s mane. Horses are fitted out for battle along with warriors, who don a piece of clothing that makes them look, at least a little, like horses. Griffith expresses this very well: “The horse-hair plumes of these helmets blend valor with exuberance, horsiness with humanity.” [11] The dressing of both hero and horse works to assimilate the two figures visually. Some heroes can also be specially marked out by their identification with their horses, as a quick look at Diomedes and his horses will prove. In book V, when Aeneas is talking to Pandarus about the unknown man raging on the battlefield, Pandarus replies that he knows that man to be Diomedes, in part because of his horses. He says: It is ironic that these horses by which Diomedes is recognized are the very horses that he will forsake for the semi-divine Trojan horses that he steals from Aeneas. Yet, the story of Diomedes in the Iliad is largely the story of a lesser hero stepping temporarily into the role of Achilles, and, in a sense, his heroic ascent is paralleled by his equestrian ascent. Pandarus, on the other hand, is best known for his archery and is unique among the Trojan fighters in not having any horses at all. When Aeneas asks him to assume the role of his charioteer in book V he does so because Pandarus has just explained that he has no horses of his own because he left them all in his homeland. [12] He says in line 201 ἦ τ’ ἂν πολὺ κέρδιον ἦεν, “it would have been much better [if I had brought horses].” It is difficult to avoid thinking that the life that he leads at Troy, tricked into breaking the truce and finally killed, pierced through the tongue by Diomedes, is somehow presaged by his unusual dearth of horses, the warrior’s standard accoutrement and the very status marker in which Diomedes is excelling.

Furthermore, Diomedes upbraids the aged horseman Nestor for being out on the battlefield past his prime by criticizing his horses and by extension Nestor himself: Nestor’s old age is mirrored in the slowness of his horses while the superiority of Diomedes is reflected in the high quality of his, despite the fact that the horses have only recently been stolen from Aeneas. That the horses did not originally belong to Diomedes is not an issue: his present status as a fighter is perfectly matched by his current steeds, perhaps especially because he has proven his superiority over Aeneas by taking those horses. Diomedes’ superior stature means that the worthier horses are properly his regardless of their provenance.

This identification of horse and hero finds its most consistent expression in Achilles and his mixed triad of mortal and immortal horses. Achilles’ own horse-like qualities are signaled by his legendary speed as well as by his sharing of the epithet “swift-footed” with Iliadic horses. In turn, the human-like quality of his horses is made clear by the famous episodes in which they shed tears at the death of Patroclus [13] and in which one of the horses, Xanthos, rebukes Achilles for accusing the horses of allowing Patroclus’ death. [14] The close identity between Achilles and his horses is reflected in their relationship to immortality. As mentioned earlier, Xanthos and Balios [15] are the immortal offspring of a Harpy, and Pedasos, though mortal, is nevertheless swift enough to keep pace with the immortal horses: The mixture of mortal and immortal components in his chariot team reflects the genealogy of Achilles himself, the son of a goddess, Thetis, and a human, Peleus. [18] We hear twice in the Iliad, once from Odysseus and once from Apollo, that no one but Achilles can drive his horses properly and that this is due to a similarity in their makeup. One of these instances occurs in book X, when Dolon confesses to Odysseus that he spied on the Trojan camp because Hector had promised him the horses of Achilles in return. Odysseus replies: The team is unmanageable to others because it is partly divine but can be mastered by Achilles, who is also partly divine, the son of the goddess Thetis. Yet it is singularly important that neither Achilles nor his chariot team is entirely divine, because the narratological focus is frequently not on the two immortal horses but on the mortal one, Pedasos, just as the narratological focus of the epic is on Achilles’ mortality despite his superhuman stature.

The relevance of Pedasos’ mortality is made abundantly clear at lines XVI 466–471. While Patroclus is fighting in Achilles’ stead and making use of his horses, Sarpedon throws a javelin at Patroclus: Pedasos’ purpose, narratologically, is to die. Soon thereafter the immortal horses see Patroclus die as well, and as Zeus observes them shedding tears, he says: These horses are also closely linked to Achilles’ immortal mother, Thetis, because they were given to Peleus as a gift from the gods when Thetis was forced to marry him. The plight of these immortal horses next to a mortal companion and accompanied by mortal riders recalls the frequent lamentations of Thetis at having been forced to marry a mortal and to bear a mortal child whom she must see die. It is the plight of immortals who are close to morals. The unusual combination of mortal and immortal in Achilles’ chariot team reflects the uneasy identity of Achilles himself, more than human but ultimately not fully divine. Pedasos’ tragic demise foreshadows Achilles’ own impending death as well as that of Patroclus, and we see that the Iliad’s tendency towards comparison of heroes and horses finds its most consistent expression in Achilles, the figure that is both most horse-like and most heroic.

As we turn to Ṛgvedic evidence I should make clear that the genre of the Vedas is quite unlike that of the Greek epics. They are liturgical hymns without extended narrative and thus provide very different sorts of evidence, yet horses do feature prominently in them and their treatment is comparable. In the Vedas, for example, there is ample reference to horses’ ability to share with humans the capacity for divinity. The Ṛgvedic corpus mentions no fewer than four immortal horses by name, Dadhikrā, Tarkṣya, Paidva, and Etaśa. [19] These, like the horses of Ares and the Dioskouroi, are all heavenly horses, but there is also a handful of Ṛgvedic hymns dedicated to horses that live among humans, namely the hymns to the sacrificial horse.

In approaching these hymns we should bear in mind how unusual they are, and the significance of that oddity. They occur in the Ṛgvedic corpus, which constitutes the earliest group of Vedic hymns. There are over a thousand of these hymns, and they almost exclusively praise divinities. The divine identity of the addressee is so expected that the addressee is referred to technically as the devatā, “divinity.” The devatā of each hymn is specifically designated by commentators and constitutes one of the four categories by which Sanskrit commentaries consistently identify each hymn, others being the composer’s name, the meter, and the hymn’s ritual application. The devatā of these hymns is identified simply as aśva, “horse,” according to this principle, which the commentator Sāyaṇācārya explains as follows: aśvasya stūyamānatvādyā tenocyate sā devateti nyāyenāśvo devatā, “because it is the horse who is being praised, the horse is the devatā, according to the principle that the devatā is that which is addressed by him [scil. the ṛṣi].” [20] His explication of this principle seems motivated by the oddity of a horse serving as the subject of a hymn, since this function is usually served by gods. In consideration of the fact that the corpus of Ṛgvedic hymns does contain actual divine horses and that Ṛgvedic poetry and culture descend from an Indo-European tradition, in which horses may have been important religious symbols, it seems sensible to think that the capacity of the horse to function as the addressee of such a hymn in some way continues an Indo-European tradition of sacred and divine horses.

The content of Ṛgveda I.162 presents a tantalizing, if small, piece of evidence of this. The very first line of the hymn states: vājíno devájātasya sápteḥ pravakṣyā́mo vidáthe vīryā́ṇi; “we proclaim in assembly the virtues of the prize-winning god-born horse.” [21] Sāyaṇācārya does provide two synchronically feasible explanations for this: that horses were first born among the Gandharvas, a group of theriomorphic divine creatures, or that this has something to do with the first lines of the Bṛhadarāṇyaka Upanishad, which declare that the dawn of each day is the head of a horse and dusk its tail. Yet the fact that there were other examples of horses, or horse-like creatures, of supermortal status does not preclude the possibility that this, as well as other comparable phenomena, continues a pre-Indian tradition. Based on the presence of actual divine horses and the unusual role that horses can serve in Ṛgvedic hymns, it seems likely that the culture of the audience of the Ṛgvedic hymns accorded to horses an affinity to the immortal world that is more like that expected in humans than in other animals, much as the world of Homer does.

A further similarity of horses to heroes is also revealed by the very first word of the quoted passage, vājin, which is frequently used to describe horses in the Ṛgveda. I have translated the word as “prize-winning” to call to mind Homeric parallels, such as ἀθλοφόρος. The connotations of vājin presumably extend to horse racing, which can itself be a form of symbolic combat, but vāja, from which this adjective is derived, also indicates the loot taken from conquered enemies by a warrior. The Vedic hymn IV.38, to the immortal stallion Dadhikrā, directly compares him to a king, nṛpatiḥ, translated as “lord of men”:

Kings in the context of the Ṛgveda are often praised as warriors, as they are in Homer, so nṛpatiṃ, the accusative form of nṛpatiḥ, here is a heroic epithet. This metaphor extends to describe Dadhikrā fighting in battle and snatching war spoils from enemies. Dadhikrā is propitiated in his capacity as hero, as, in fact, are the other divine horses, who are consistently described as invincible, vanquishers of chariots, etc. One of these immortal horses, Paidva, even takes the same epithet as Indra, ahihan, “dragon-slayer.” [22] In the earliest literature of both Greece and India, then, horses and heroes are handled in notably similar ways and display a similar ontological positioning. In India too horses function as a reflection of the hero and are treated in ways that render them very much like heroes, while heroes are treated in ways that render them like horses. A fascinating manifestation of this overlap between horse and hero occurs in heroic names that involve horses. [23] While discussing these well-documented names, such as Vīśtāspa (Avestan “swift horse”), Puhvel argues convincingly that these names are not simply bahuvrīhi compounds that indicate possession, e.g. “he of whom there is a swift horse,” but that they are meant to convey a sharing of quality between the horse and man. He says that “in onomastic usage the line gets blurred, even as ‘Crazy Horse’ hardly evokes the proud owner of a deranged equine.” [24] The compounds indicate possession in a way that draws on the ability of one’s horses to function as reflections of one’s own identity with the result that the owner of swift horse possesses some of the qualities of a swift horse himself. Horse and hero, then, are equated on a poetic and onomastic level. If this tendency indeed reflects inherited PIE cultural practice, then the Greek poetic treatment of horse and hero may well be a more ancient and complex phenomenon than it first appears.

It could be argued, of course, that the evidence presented thus far does not preclude the possibility that these similarities simply reflect a coincidence of internal logic within these societies, i.e. heroes use horses in battle so the comparison of the two is simply natural and any overlap just the result of chance. Given the relationship between these cultures, in history and poetry, this would seem unlikely to me, but to bolster these conclusions one would want to see evidence of some kernel of this phenomenon in the parent culture as well. Since we have no literature from that culture, however, such evidence is inherently difficult to obtain, to say the least. In this particular instance, however, we may be in exceptionally good luck. This is so because of one particular element of IE religious practice that has to do with horses, to which I will now turn.

The ritual wherein I seek corroboration of the antiquity of the phenomena discussed above is a horse sacrifice, a ritual that appears in multiple IE cultures with such similarity that it is regularly regarded as a reflection of religious practice in the parent culture. The issues surrounding this sacrifice are something of a cause célèbre among Indo-Europeanists, yet it may seem odd that it features in my own analysis of Homeric treatment of horses, because Greece is not, in fact, one of the places where this ritual was preserved. [25] I will argue, however, not that the ritual had direct influence on Greek poetic practice itself, but rather that the ritual sheds light on how the parent culture of Greece thought about horses and that features of this ideology regarding horses, if not its actual ritual practice, were inherited in Greece and affected Greek culture and poetry.

We must look then to the evidence for this kingly ritual, in which the space between equine and human identity became permeable, an appreciation of which has practical implications for our understanding of the unique treatment of heroes and horses in early Greek epic. The reconstruction of an actual religious ritual, however, is not as straightforward as the reconstruction of a word or phrase, and may not even be possible, because the evolution of a culture’s religion does not progress by consistent or predictable tendencies. Therefore, although it is important here to review the evidence and even discuss at some length certain issues brought up in reconstruction, I will ultimately suggest that we should not try to reconstruct the ritual precisely, but instead examine what our evidence for it allows us to glimpse about the patterns of thought that shaped its formation generally. This, I think, will be more useful for the sort of analysis needed for the study of horses in Greek poetry and will help us avoid certain difficult pitfalls inherent to precise ritual reconstruction.

The most secure evidence for this sacrifice comes from ancient India and Ireland. Our evidence for a kingly horse sacrifice in India is extensive, comprising literary evidence as well as archaeological. The archaeological evidence includes an aśvamedha (horse-sacrifice) altar near Kalsi, inscriptions attesting performances of such sacrifices, [26] and even coinage. [27] Our literary evidence, however, provides the oldest and most thorough descriptions of the particulars of the ritual. There exist two Ṛgvedic hymns about the sacrifice, I.162 and I.163, and detailed information about the ritual procedure and associated mantras is also provided by other texts, especially the Śatapatha-brāhmaṇa [28] and the Śukla-yajur-veda. [29] Aśvamedhas also occur in both the Mahābhārata, in book XIV, the Aśvamedhikaparvan, and in books I and VII of the Rāmāyaṇa. Puhvel provides a succinct summary of the ritual: The ritual could in fact be much more elaborate than this, and could involve as much as a full year of preparations, yet this summary will suffice for our current purposes. [31]

The Irish evidence comes from the twelfth-century Giraldus de Barri, known generally as Giraldus Cambrensis. He is the author of Topographia Hibernica. Although his depictions of the ancient Irish people are not at all flattering or unbiased, he describes a ritual in Kenelcunnil, in northern Ireland, that is similar enough to the aśvamedha that its basic veracity is usually accepted: Regardless of any ethnic bias on the part of Giraldus, the details of this ritual seem too specific to be disregarded, and probably do reflect, at least in basic details, an actual ritual that is cognate with the aśvamedha. [32]

There is another, comparatively new piece of potential evidence that may strengthen this conclusion as well. In 1988, a large vase relief from İnandık was published by Özgüç. [33] The vase bears a series of friezes, and Watkins argues that these ought to be “read” from right to left and from bottom to top; this then would reveal a narrative of an elite man and woman led in a procession, at the end of which the couple are depicted as prepared to copulate “in more ferarum.” [34] Unlike Özgüç, who thinks that the couple represents a god and goddess, Watkins argues that they are best understood as a king and queen and that this entire ritual may represent some remnant of a Hittite horse sacrifice that is cognate with the aśvamedha. If Watkins is correct, we have evidence here of a PIE ritual involving horse sacrifice and its association with kings in Ireland, India, and perhaps also in Anatolia. There is, thus, reason to surmise that this was an important rite in PIE culture.

While the details of the later rituals are not widely documented outside of India, there is enough information that some scholars have attempted to reconstruct at least the essential elements of the PIE rite. [35] It is, however, my opinion that such reconstructions are always in jeopardy of privileging one or another of the later traditions. Since the details of the Indian horse sacrifice are so much better documented than the others, one could easily be tempted to include elements of the Indian rite in the PIE reconstruction that are not justified by other sources. Some of the specially Indian procedures may indeed be genuine remnants of PIE ritual practice that have simply been lost elsewhere, but there is no way to know which these are, and thus these Indian practices cannot securely be assumed to have descended from PIE traditions. Additionally, actual ritual practice in the culture of the Proto-Indo-Europeans should not be expected to have been any more consistent than ritual practices in other cultures, so as much room as is possible should be left in reconstructions for variations in ritual practice within PIE culture. Any such ritual is likely to have varied over time and even among different groups in the original community. Therefore, the safest methodology for hypothesizing about this particular ritual is to be minimalistic, using only the elements that are most secure and then to extrapolate from them a sense of the rationale of the sacrifice. This will undoubtedly give us an impression of the ritual that is unrealistically minimal, and we may suppose that in practice such sacrifices were much more elaborate than we can perceive. Yet it should provide us the surest details, which we will need in order to interpret early Greek poetic treatment of horses.

In such a minimalistic model many of the significant details of the later ritual will have to be omitted due to an inability to select between regional variants. Since my exclusion of some details is unusual I would like to explain the reasoning, before offering my conclusions. For instance, the Indian sacrifice involves the horse being killed by smothering before the copulation while the Irish tradition has it killed afterward in a manner that is unclear. Even more importantly, the Irish tradition has a king copulating with a horse, which should probably be assumed to be a mare, while the Indian tradition has a queen copulating with a stallion. It seems to me that the idiosyncrasies of these rituals are all so interrelated that it would be very difficult to select any particular facet of one ritual and project it all the way back to the parent culture. The method of execution in India, for example, is inextricably linked to the sequence of the ritual procedure there. It has occasionally been suggested that the explanation for this lies in the biology of asphyxia, in that death by asphyxiation gives rise to postmortem “tumescence,” which may have been necessary for the ensuing procedures. [36] It seems better to assume that this method of killing is related to another notable distinction of the Indian procedure: the fact that the killing occurs before the ritual copulation. The fiction of living participation on the part of the horse is evident and, in fact, essential in all the ritual procedures. Throughout the entire drama of the aśvamedha ritual, the participants, who include at least three wives and several priests, pretend that the horse is alive. For example the queen utters once she has lain down with the dead horse, ná mā nayati káścaná | sasastyaśvakáḥ, “no one is leading me [euphemistically], the horsey is sleeping.” [37] Furthermore there is no mantra associated with the killing of the horse, but this would, of course, have destroyed the fiction of the horse’s life. Traditional methods of sacrifice would have precluded, or at least dramatically complicated, the maintenance of this fiction by leaving the body of the animal visibly disfigured by the mortal wound. Thus the horse is presumably killed by smothering precisely because smothering leaves no marks. It is therefore necessary that the horse be killed before the sexual element of the ritual. If it were killed afterward, there would never be a need to pretend that the dead horse is alive, so no need for a method of execution that leaves no marks. This does not prove that the smothering could not have been part of the PIE ritual; it only requires that the smothering be tied to the ritual order: killing before sex. [38] To project the smothering onto the PIE ritual therefore requires that the ritual order be projected onto the PIE ritual as well. This is not, of course, impossible, but it is difficult to prove. [39] We cannot actually be sure then of how the horse was killed at all. My point is essentially that it is difficult to extricate this one or any one difference from a nexus of differences. If we attempt to create a terribly detailed reconstruction we are hard pressed not to retroject entire complexes of regional idiosyncrasies onto the parent culture.

We cannot be sure of the gender of the horse or even of the human participant. In our most useful bodies of comparanda, the Irish and the Indian, there is no agreement. The Hittite vase, if it does reflect a similar ritual, is of no help since both of these participants are human, perhaps playing the part of horses. In a lengthy treatment of the issue, Doniger O’Flaherty argues that the aśvamedha itself contains vestiges of a previous version of the ritual in which a man did indeed copulate with a mare. She points to a ready identification of the horse with the king, mediated by frequent solar symbolism which is pertinent to both figures, and focuses especially on the sexual abstinence, of both the king and the stallion, practiced during the preparation for the ceremony. [40] The stallion not only cannot have sex while the ritual is being prepared but, before his release for the year of wandering, is shown a pen of caged mares, whose sexually charged whinnying is a desirable omen. Doniger O’Flaherty sees in this caging and avoiding of female horses and women a symbolically manifested anxiety toward female sexual power. She sees these ancillary ritual elements as later, perhaps compensatory, additions to a ritual that had once had female sexuality as its principal feature. She argues that this may also explain the other principal difference between the two rituals, the fact that the horse is killed before the sexual act in the Indian version but after this act in the Irish. The switch from mare to stallion would make the horse the active, or penetrative, partner, and a live horse could not be counted on to perform this ritual function. The killing of the stallion and the subsequent fiction that the stallion is merely sleeping is a practical solution. Although this reasoning is tempting, I prefer to plot a more conservative course. I prefer to avoid most specific details entirely and to focus instead on the ritual logic of that sacrifice, the pattern of thought to which it testifies. In order to ensure the utmost stability of our analysis, I believe that no conclusion can safely be drawn as to the gender of the PIE horse without some agreement among our sources, and thus no safe conclusion can be drawn about either the ritual order of the death or the sex of the horse.

Despite these disagreements among the sources, however, some positive statements can be made about the PIE ritual. Our texts do agree that the king is understood to be a horse. In the Indian version, the horse and king are both meant to be sexually abstinent for a year, both being tempted by sex—the king by sleeping with his favorite wife and the stallion by being shown the penned-in mares. The sacrifice of the horse is also meant to renew the energy lost by the king. The horse also, of course, takes the place of the king in the sexual relationship with the queen. The Indian aśvamedha could have been performed for a variety of reasons, but in the Rāmāyaṇa it is performed because King Daśaratha cannot produce a child. [41] The queen’s interaction with the horse allows the couple to have children. The horse is even called the retodhā, “semen-giver,” and the command is given: reto dadhātu, “may it give the semen.” [42] The horse stands in place of the king and by its death transfers to the king its power.

The Irish ceremony casts the king in a similar position. At first glance it seems odd that the king would declare himself to be a beast while having sex with a horse—and this is precisely the sort of thing that might not be trusted in Giraldus’ account—yet it is what makes it most believable. For although the king is actually involved in this case and the animal is presumably a mare, the king’s declaration that he himself is transformed into a horse is consistent with the aśvamedha in that the king is a horse in the logic of the sacrifice. Puhvel has even suggested that Giraldus may have mutilated a ritual declaration here that was similar to “I am stallion, you are mare,” which may have cast this in the form of theriomorphic sacred marriage. [43] Whether or not this precise hypothesis is true, the Irish and Indian sources do agree in depicting the transformation of the king into a stallion and the queen into a mare. So long as this is secure, even if little else can be, substantial research can proceed.

The minimal, yet reliable deduction that we must work with, then, is that this evidence reflects a kingship ritual involving sex between a royal human and a horse, in which the king is identified with a stallion and the queen with a mare and in which the horse involved is killed at some point. The underlying ritual logic then is that of human/horse ontological overlap on a conceptual and ritualistic level.

There is one more piece of evidence for the Indian ritual which highlights the utility of relying only on the most minimalistic schematic regarding the actual ritual procedure. In addition to the archeological and literary evidence for the aśvamedha, there is also a non-Indian source. The eleventh-century Arab scholar Albêrûnî, who lived in India for much of his life, describes the ceremony as it apparently was practiced in his time. In his description there is no sex at all and the horse is a mare, yet the mare is still allowed to wander before the sacrifice just as the stallion was and, like the stallion, is followed by attendants. In this case, however, the attendants cry out as they follow her, “She is the king of the world. He who does not agree, let him come forward.” [44] The ritual has probably been altered but could preserve some variant of the original that is simply undocumented otherwise. If the details of the ritual vary this much from the Vedic period to the eleventh century, then how much could they vary from the PIE to the Vedic? This example certainly demonstrates the need for caution in identifying the gender of the original victim, and the importance of prioritizing that which is secure, namely, a unique coincidence of human and equine identity in what is likely to have been a very culturally significant ritual.

There is yet one more ritual that has often been suggested as a survival of the PIE horse sacrifice, that of the Roman Equus October. I have so far omitted it from this discussion because I do not think it is a direct descendant of that ritual, or least it cannot be confirmed as such. Yet it is worth discussing for two reasons. First, since other scholars have included it in their analysis, I ought to explain why I do not. Second, even though I think that it is impossible to prove a relationship to the ritual discussed above, it is nevertheless possible that practices concerning the Equus October still preserve evidence of early ideology regarding horses and humans that will be useful.

Our principal source is Festus, who says under his entry for “October Equus”: [45] He also discusses it under his entry for panibus:

Differences between this and the horse sacrifices discussed above are quite apparent. The ritual takes place annually while there is no evidence from the Indian or Irish versions that justifies understanding the PIE ritual as anything but a special occurrence rising out of particular circumstances. There is also no sex, which, it must be admitted, is a salient feature of the other rituals. There is also no real kingship involved. It seems then the Equus October is lacking in almost every requirement for inclusion in the group of IE horse sacrifices that descend from the PIE horse sacrifice. These differences have not, however, prevented scholars from seeing these rituals as related.

Most famously, Dumézil explained the majority of these differences as outgrowths of historical developments in Rome dealing with attitudes toward kingship. [47] As kingship disappeared from Roman society, the circumstances which had previously given rise to the PIE royal horse sacrifice could not occur. The ritual, he argues, may still have been preserved in an altered form due to its importance. This altered ritual could then have been made part of the regular Roman religious calendar and thus its performance assured. The fixing of the tail in the Regia, a site associated with Roman kingship from Numa onward, testifies, according to Dumézil, to the previously regal nature of the ceremony. He believed that the sexual nature of the ritual has been lost entirely, although Puhvel saw hints of it in the wish for fecundity indicated by the phrase ob frugum eventum. [48]

This is intriguing, and even tempting analysis, but it is not certain enough to warrant considering this ritual cognate with the aśvamedha and Irish rite at the moment. The substitution of the placement of a tail in the regia for sex with a king or queen is not convincing. It would be more convincing were the regia not the seat of Roman priestly authority. The fact that a ritual sacrifice has a connection to the home of the pontifex maximus is not odd enough to justify a special explanation. There is, then, neither sex nor a king, and, although there is a horse, there is no overlap of horse and human, which is the uniting theme of the Indian and Irish accounts. [49] Ultimately, the Equus October is not best understood as a descendant of the PIE horse sacrifice, at least not for our purposes.

Some utility may still be found in it, however, if we examine it for traces of inherited ideology concerning horses and sacrifice rather than for evidence relating to one horse sacrifice ritual in particular. For this I will point to an interesting and somewhat arcane feature of Indian sacrificial literature, the human sacrifice. The ritual is known as the puruṣamedha or naramedha, “human-sacrifice” and “man-sacrifice.” It is unclear whether this ritual ever actually occurred or if it is a liturgical pretense, but its description in the Śāṅkhāyana Srauta presents a sacrificial practice that is modeled on the aśvamedha. Just like the horse, the human victim is allowed to wander for a year and is sacrificed along with a goat, and the chief wife of the king then engages in sexual contact with him. [50] Whether or not the ritual was actually performed, it seems to have been thought logical that horses and men would be sacrificed in the same very special way. [51]

This is potentially relevant in Rome and to the Equus Octoberspecifically only because, although there is no evidence of human sacrifice in Rome, there was a certain execution that was modeled on the Equus October. In 46 BC Julius Caesar, once a pontifex maximus, had mutineers decapitated in the Campus Martius and their heads hung up in the Regia, an odd procedure that is reminiscent of the sacrifice. Cassius Dio tells us that this was, in fact, done precisely in the manner of a sacrifice, and the Equus Octoberseems to be the sacrifice that is implied: I do not claim that this reflects a shared methodology for human sacrifice in Rome and India at all, of course. I only point out that in each case, when a human sacrifice was imagined the logical ritual practice on which to model it was the sacrifice of a horse. The fact that horses occupied an ontological position that was closer to humans than other animals marked horse sacrifice as the most logical paradigm for human sacrifice. Even without direct relation then, the rituals may still testify to a similar ideological connection between horses and humans, an ideological connection that is a genuine inheritance, and that we have seen in our Greek poetic sources as well.

Again, it may be more useful to inspect ritual rationale than actual procedure, at least in this case and in this particular investigation. Such an approach allows us to look past the conflicting elements of our evidence. It allows us to observe a propensity toward a conceptual overlap of equine and human identity in the parent culture that was prominent in sacrifice. [52] We can see in the important kingly ritual, in fact, exactly the type of overlap of horse and hero in the parent culture that we needed to help us contextualize the overlap that we see in Greek poetic treatments of heroes and their horses, and that I have posited as a reflection of inherited ideology.

We should not conclude this discussion of horse sacrifice without addressing the realities of the practice in Greece. As in Rome, horse sacrifice was quite rare but did exist in ancient Greece, and appears in Greek literature as well as in the archaeological record. No Greek sacrifice resembles the IE kingly sacrifice, as argued above, but we do not need to restrict our discussion to confirmed ritual descendants. [53] For although there is no ritual in Greece of which I am aware that resembles the IE ritual, there are cultic and mythological vestiges of the nexus of hippomorphism, sex, and kingly power that reflect the ideology observed in the parent culture.

The fact that horse sacrifice is rare in Greece is not particularly surprising when one considers the fact that the eating of horse meat was taboo there, [54] and that sacrifices were in many ways religiously charged communal banquets. Thus, animals would not usually be sacrificed that would not then be eaten, a fact that shapes the Aeschylean description of the horror of human sacrifice in the case of the sacrifice of Iphigenia: θυσίαν ἑτέραν ἄνομόν τιν’, ἄδαιτον, “another sacrifice, unlawful, not to be eaten.” [55] This reluctance of Greeks to eat horsemeat, in apparent distinction from their PIE ancestors and Indian counterparts, even seems to have influenced the repartee between Hesiod and Homer in their legendary Certamen, in which Hesiod composes one line of verse and Homer has to compose the next. Hesiod composed lines for which it would be logically difficult to compose a companion line. In one instance, that difficulty seems to involve the taboo concerning the consumption of horse meat: Δεῖπνον ἔπειθ’ εἵλοντο βοῶν κρέα καὐχένας ἵππων, 107. This would lead the listener to expect the meat of the cows and the necks of the horses to be part of the meal. Finishing this line is a challenge because of the way that it plays on the cultural taboo of horse-eating. Homer, then, is quite deft to finish the couplet in a way that circumvents this problem with the line: ἔκλυον ἱδρώοντας, ἐπεὶ πολέμοιο κορέσθην, 108. This, then, renders the entire couplet as “they feasted on the meat of the cows and they loosed the sweating necks of the horses, since they had had their fill of war.” [56]

Homer’s need to alter the apparent semantic trajectory of the line grows partially out of the Greek disinclination to eat horse meat, but also coincides with a general unwillingness to sacrifice horses. It may, of course, seem odd that an animal that was once specially marked for important sacrifices becomes particularly unlikely as sacrificial victim. This need not be surprising, however, since it is possible that resistance to sacrificing a particular animal is paradoxically not unlike eagerness to sacrifice a particular animal. Each case reflects a designation of that animal as uniquely sacred. It is not hard to imagine the same reverence for an animal rendering it especially sacrificial in one time period and especially unfit for sacrifice in another. This very phenomenon is, in fact, evident in the transition from Vedic to post-Vedic attitudes toward the sacrifice of cows.

Horse sacrifice in Greece is not, however, completely unknown so much as uncommon. [57] There are, in fact, fascinating and potentially relevant references to horse sacrifices being performed to Poseidon Hippios in commemoration of his coupling with Demeter in the form of a horse herself, which was discussed above. Recall that she, in her wandering in search of her daughter, was pursued by Poseidon and took on the form of a horse to escape him. He chased her, in the form of a horse himself, and eventually he had sex with her. Whether these sacrifices or the myths that surround them have any relationship to the aśvamedha is certainly not secure. The Indian and Irish comparanda provide the essential elements needed for identification in the ritual itself, while in this case these elements appear partly in the ritual itself and partly in its mythological etymology. That is to say that the sex and hippomorphism are not in the ritual but in the myth that informs the ritual. It is not impossible that the PIE horse sacrifice was itself linked to a mythological etymon which is uniquely preserved in the myth associated here; Doniger O’Flaherty has, in fact, argued for just such a possibility. [58] I will not claim that this mythology ought to be treated as the continuation of cognate tradition without hesitation (and it certainly should not be used to inform our reconstruction of the PIE ritual) but it does reflect an overlap in ideology, that of hippomorphic transformation and sex.

Additionally, there is a sacrifice of a horse mentioned by Pausanias that, although not much like the Indian or Irish accounts, does contain some important links to kingship and to the Iliad. He mentions the peculiar existence in Sparta of a grave marker for a horse which was sacrificed at the gathering of the kings of Greece to compete over Helen: Although this bears no obvious relationship to the PIE ritual, there was perhaps in ancient Greece some connection between kingship and horse sacrifice. These were all kings who were gathered, and the man selected to marry Helen would become the king of Sparta. [59] Again, we see a continuity of ideology, preserved here in legend rather than in ritual.

Another Greek horse sacrifice must be mentioned due to its obvious prominence, even if it is difficult to interpret. The legendary Trojan Horse may be best understood as a horse sacrifice. The horse was offered by the Greeks to the Trojans and brought inside to be presented to the goddess at her sanctuary, as a sacrifice would have been. Burkert points out that the striking of the horse with a spear by Laocoön seems to reflect a sacrificial rite and even hints that this element may preserve, in altered form, an earlier, more explicitly sacrificial version of the story. [60] This idea seems to have been readily accepted in ancient Rome, since Festus says that the common belief at Rome was that the historic origin of the Equus October lay in the slaughter of a horse in retribution for the horse that brought about the fall of Troy, the mythical homeland of the Romans. [61] The sacrifice of the horse with a spear, like the one used by Laocoön, must have either developed from this belief, or at least given credence to it.

Finally, Odysseus, the devisor of the strategy, is himself sometimes said to have been transformed into a horse. Sextus mentions this story in the midst of a condemnation of what he saw as widespread scholarly reluctance to distinguish fact from fiction. Although Sextus mentions this story as if it was a normal topic of discussion among the educated, the obscurity of the other accounts among which it is mentioned may lead us to think otherwise. There are, however, other sources for this story. We find some corroborating data, for example, in Ptolemaios Chennos, who retells some useful pieces of information which, he claims, come from Herodotus: Finally, we learn from Servius that Odysseus is said to have died when he, still in the form of a horse, was stabbed with a spear. This particular source is especially tantalizing since the transformation is mentioned here as part of an essential, and fairly minimal, biography of the epic hero, as if this transformation was not terribly obscure to Servius. These scant bits of evidence, however, do not allow us to piece together a mythic narrative, nor even to be certain that they refer to the same story. They may, after all, simply refer to mythical variants linked through the common motif of Odysseus’ equine transformation. Even without an overarching narrative, however, it is still evident that Odysseus was said to have been the subject of equine transformation and that, at least according to Servius, he was sometimes said to have been killed while in horse form, stabbed in a way not dissimilar to Laocoön’s stabbing of the Trojan Horse.

In this last account, just as Burkert suggests, Odysseus seems to have been equated with the Trojan Horse itself, and the entire Trojan Horse episode seems to contain reflections of a horse sacrifice. I do not wish to suggest that this links the scene with the PIE ritual, of course, but it does suggest some history of horse sacrifice in the ancient Greek consciousness and, perhaps, the connection of horse sacrifices to transitions of power, an association that may have been inherent in the PIE ritual as well. It also bears further testimony to the existence of hippomorphism in Greek folklore. Other examples considered so far include the story of the hippomorphism of Demeter and Poseidon and an alternative genealogy of Achilles’ immortal horses, who are said by Diodorus Siculus to be hippomorphized Titans. [63]

It seems then that Indian and Irish horse sacrifice rituals provide sufficient evidence for assuming that speakers of Proto-Indo-European performed a kingship ritual in which the human participant, or participants, was imagined as undergoing a hippomorphic change of some sort. Our Greek evidence, additionally, provides evidence that a mythology of the hippomorphizing of humans and the connection of horses to kingship were still active in ancient Greek culture as well. [64] It may therefore be inferred that the existence of these concepts in the PIE tradition had some influence on depictions of Greek heroes, who were of course warrior-kings, compared to and ontologically linked to horses, themselves ontologically similar to humans, especially heroic humans.

Finally, as discussed above about India and potentially Rome, the depiction of human sacrifice in Greece can provide useful evidence to confirm this continuity of ideology. Our Greek evidence does not provide an obvious cognate with the Indian ritual but, again, presents a few important pieces of evidence. [65] The seventh-century burial at Lefkandi, for example, is often seen to document a human sacrifice, a satī, that may have accompanied the sacrifice of horses. [66] Greek literature provides evidence as well, in at least two prominent forms. The first comes in Achilles’ sacrifice of twelve Trojan youths at the funeral of Patroclus, at which the youths are hurled into the fire. This sacrifice is particularly interesting because it is accompanied by the slaughter of πίσυρας ἵππους ἐριαύχενας, “four strong-necked horses.” [67] Other animals were sacrificed as well, but it is the horses and the men to whom Achilles draws attention as he tells the other Greeks to remove the bones of Patroclus from the pyre. He says that Patroclus’ bones will be easy to distinguish because they are apart from the rest, which are ἐπιμὶξ ἵπποι τε καὶ ἄνδρες, “the horse and the men mixed together.” [68] Additionally, the dressing of the victims for the sacrifice is explicitly mentioned for the other animals, yet the horses and men are simply “thrown” on the pyre. [69] It is not impossible that the slaughter of the men and horses should be understood as having occurred before they are thrown on the fire, but it does seem as if they were thrown on alive. If this is true, the horses and men are treated in a uniquely similar way and truly marked out from the other victims.

It can be argued that this constitutes an actual sacrifice to the dead Patroclus, or that it simply results from an overflow of Achilles’ fury, but in either case it is contextualized by Achilles’ desecration of the corpse of Hector as part of a general pattern of grief-induced sacrilege. Although horses were, on occasion, sacrificed in some parts of ancient Greece, they were not a regular sacrificial animal, and in the context of the Iliad it seems that the slaughter of the horses somehow parallels that of the youths and that both reflect the quality of Achilles’ transgressive behavior. The sacrifice of humans is again modeled on the sacrifice of horses. [70]

Both in the actual liturgy of the puruṣamedha and in several other notable sacrifices in the PIE world there is a tendency to compare humans to horses. I believe that this preserves an inherited tendency toward the hippomorphizing of humans and anthropomorphizing of horses that is linked to a conceptual overlap of the species that is traceable to prehistoric times and the common culture of the Indo-European peoples. To return to the epic heroes, then, the similarity between horses and heroes reflected in both our Greek and our Vedic Sanskrit sources is not a fully isolated phenomenon, but one special realization of a broader tendency, one that is evident not only in our literary sources but in sacrificial practices continuous from distant antiquity.

The horse sacrifice described in Plutarch’s Life of Pelopidas may bear special testimony to the ideological nexus described above. Since the passage has never, to my knowledge, been analyzed in the light of IE horse sacrifices, I think that it is worthwhile to attempt to examine it through the information presented above. The Greek legends of the Leuktridai and of the sacrifice offered to them on the eve of the Theban battle against the Spartans in 371 BCE contain descriptions of a horse sacrifice that is very infrequently discussed in scholarly studies but incredibly interesting in light of the PIE horse sacrifice and the symbolic logic between humans and horses to which it testifies. It does not, I think, constitute a direct Greek descendant of the famous PIE ritual, but it does reflect categories of ideological inheritance outlined above.

Elements of the story are scattered across numerous sources, and few of these agree in exact detail, but a composite of the texts presents enough common features to allow construction of a basic narrative. [71] Sisters known as the Leuktridai lived somewhere near the Leuktrian plain. [72] These girls are usually two in number, and their names, when given, are Molpia and Hippo, Melitia and Hippo, or Thano and Euxippe. Despite some disagreement, one of their names is consistently hippologicial (Hippo and Euxippe). [73] These women were raped by Spartan men and subsequently died by their own hands, according to most accounts. [74] The site of their death was then prophesied to be the future setting of a Spartan military defeat. This eventually came to pass in 371, when the Spartans faced the Thebans. [75]

Pausanias and Plutarch both relate that on the eve of that battle a sacrifice was performed in honor of the girls. Pausanias offers no details about the nature of the sacrifice, but Plutarch says in two passages that it was a horse sacrifice and that it was requested by an image of the girls’ father that appeared to the Theban general in a dream. In the Amatoriae Narrationes (774D), Plutarch says that it was a white horse. He does not specify the age or gender of the horse but simply calls it a ἵππος, which may refer to either a mare or stallion. The Life of Pelopidas is much more detailed and differs markedly in the color of the horse: [76] The sacrifice was then made and the Thebans went on to victory.

The relevance of the discussion to this particular scene is presumably clear. The horse here is not just sacrificed instead of a girl but as a girl. The seer makes this explicit, and the horse and girl are the same gender, both young (it is a πῶλος), and they both have the same color hair (παρθένον ξανθήν; χρόα στίλβουσα τῆς χαίτης πυρσότατον). [77] The horse is sacrificed for girls who themselves have an equine element in their identity, at least onomastically, and the horse is offered as a substitute for a human girl whose identity it reflects. What we are seeing here is, in fact, sacrificial hippomorphism like that expressed in the Irish ritual, pervasive in the aśvamedha, and structural in the puruṣamedha. Even if this is not a real ritual it participates in the same logic that shaped the ancestral ritual. It is the same symbolic logic present in Achilles’ killing of twelve young Trojan warriors and twelve horses in Iliad book XXIII. This sacrifice does not deal with kingship, but it does deal with political power and patriarchal authority, as the aśvamedha does. There is no equine sex, but there is sexualization of different sorts, both in the initial rape and perhaps in the character of the girl and horse, or girl/horse, whose youth, gender, and general visual allure are erotically colored. Although this is clearly not a literal ritual continuation of the PIE horse sacrifice itself, it reflects the same ideological nexus that informed that ritual.

Early in this chapter I attempted to highlight a species boundary that is implicit in early Greek poetry, one that separates humans and horses from other animals. The existence of this boundary has widespread ramifications in Greek literature and informs human and equine poetic comparisons as well as mythical examples of hippomorphism. The fact that many of these phenomena are reflected in the related poetry and culture of India, as well as in other IE cultures, led me to propose that these are not purely Greek developments but instead local outgrowths of an ancient cultural ideology, inherited in Greece and common to other IE cultures. As proof of this I pointed to evidence concerning horse sacrifices in Greece and elsewhere. For in these rituals we find evidence of this phenomenon and proof that it appeared not just in daughter cultures but in the parent culture as well. Not only do we have evidence that an ontological overlap of horse and human was central to the kingly horse sacrifice, but even our human sacrifice texts seem to imagine human sacrifice as modeled on horse sacrifice specifically, in Greece as well as in related cultures. We have good reason to believe, then, that Greek mythic and poetic treatment of horses and humans developed out of inherited cultural tendencies.