Achilles and Patroklos as Models for the Twinning of Identity

§0. This essay is a pre-edited draft of what became the definitive version, which is Essay 10 in a book edited by Kimberley C. Patton, Gemini and the Sacred:Twins and Twinship in Religion and Mythology (Bloomsbury Academic Publishers, 2022). That Essay 10 in the book is matched in the same book with Essay 11, “Achilles and Patroclus as Indo-European Twins: Homer’s Take,” by Douglas Frame. Here is the link for ordering the book: https://www.bloomsbury.com/us/gemini-and-the-sacred-9781848859319/

§1. Twinning in myth is a way to think about identity. As Douglas Frame shows in the essay cited in §0, which is a twin to this one, mythical twins share one identity, but this identity is differentiated. [1] That is, the fused identity of mythical twins is at the same time a split personality. In this essay, I will argue that the epic heroes Achilles and Patroklos are paired off in the Homeric Iliad in such a way as to resemble and even to duplicate such a model of twinning in myth.

§2. Two ancient Greek words that will figure prominently in my argument are therapōn, conventionally translated as ‘attendant’, and philos, meaning ‘friend’ as a noun and ‘near and dear’ or ‘belonging to the self’ as an adjective. The uses of these two words, as we will see later on, are interconnected in shaping the plot of the Iliad, since Achilles and Patroklos care for and about each other, and they care more for each other than for anyone else. Such caring, as we will also see, is at the root of the meaning of both words, therapōn as well as philos. To say it another way, such caring determines the identification of Patroklos as the virtual twin or body double of Achilles in Homeric mythmaking.

§3. Both the fusion and the differentiation of twins can be expressed in myth by way of picturing their conception as a dyadic affair, where the mother of the twins gets impregnated by two different fathers, one of whom is divine while the other is human. A classic example is the mythical dyad of Herakles and Iphikles. Here I offer a most abbreviated paraphrase of the standard version of the myth, as reported in the Library of “Apollodorus” (2.4.8):

§4. In the myth of Herakles, his immortality happens only after he experiences death. We get to see him die a most spectacular death before he ever becomes immortalized (a classic account is given by Diodorus of Sicily 4.38.4–5; the immortalization is narrated in 4.39.2–3). Still, once he is immortal, Herakles stops acting like a twin; and his mortal twin brother, once dead, is no longer of any consequence. [2]

§5. As the essay of Frame shows, however, immortality can be shared by twins. The most prominent example is the dyad known as the Dioskouroi, whose name means ‘sons of Zeus’. Despite this meaning, however, only one of these twins is the son of the immortal god Zeus: he is Polydeukes, while his twin brother Kastor is fathered by the mortal man Tyndareos. By contrast with the twins Herakles and Iphikles, who are separated from each other after the mortal Iphikles dies, these twins Polydeukes and Kastor stay together forever, since the immortal twin shares his immortality with his mortal twin while the mortal twin shares his death with his immortal twin; at any given moment, they can be either both dead or both alive (a classic account can be found in “Apollodorus” Library 3.11.2).

§6. In the myth of Herakles, by contrast, the twins are not only separated from each other after the mortal Iphikles dies. More than that, Herakles takes the place of Iphikles after the death of his mortal twin. This aspect of the myth is most relevant to my overall argument: just as appearances are deceiving at the moment when Herakles is conceived, since his immortal father appears to be a mortal at the moment of conception, so also the deceptiveness of appearances persists throughout the heroic life of Herakles, since he is known as either ‘son of Amphitryon’ or ‘son of Zeus’ after his twin Iphikles dies. Once Iphikles is out of the picture, then Herakles can take his place. As a monad who lives on without his twin, Herakles can now develop a split personality, sometimes appearing to be human while at other times appearing to be divine. [3]

§7. There is a striking parallel to be found in the case of the hero Nestor. I follow here the seminal book of Frame, Hippota Nestor, who reconstructs the relevant myth all the way back to its Indo-European origins and shows that there was a pastoral aspect of Nestor, linked with cattle, to be distinguished from the warlike aspect of this hero’s brother Periklymenos, linked with horses. [4] In terms of the myth, this distinction between the two brothers is fused after Periklymenos gets killed—and his killer, I must add, is none other than the hero Herakles (Hesiod F 33 MW; see also “Apollodorus” Library 1.9.9). Once Nestor’s brother Periklymenos is out of the picture, as Frame shows, Nestor can develop a split personality of his own, sometimes linked with cattle while at other times linked with horses.

§8. That said, I am ready to show that myth has even further ways of fusing identities and then differentiating them into a split personality. A case in point is the myth of Achilles and Patroklos as narrated in the Homeric Iliad. As we will see, there are patterns of fusion in myths about their identities, but there are also corresponding patterns of differentiation.

§9. To experience a fusion of identity, you don’t have to be born of the same mother. Certainly in the case of Achilles and Patroklos, they have different mothers as well as different fathers. Still, as we will see, Achilles and Patroklos can behave as twins behave in myth—and they can even look alike, just as twins are expected to look alike. [5] Like mythical twins, they can undergo the same experience at given moments of epic narration and, at such moments, they will look the same, as when Patroklos wears the armor of Achilles in Iliad XVI. Here Patroklos is leading an attack against the Trojans and, at this moment, and the warriors who are being attacked actually mistake him for Achilles. More than that, as we will see later, Patroklos in such moments of fused identity can even wear, as it were, the same epithets that are worn by Achilles.

§10. Appearances really are deceiving in this case. Patroklos is no Achilles. In the Homeric Iliad, Patroklos succeeds in saving the Achaeans for the moment, but he fails to kill Hektor. Instead, Hektor kills Patroklos in Iliad XVI. Later on, in Iliad XXII, Achilles will succeed in killing Hektor.

§11. So here is how the two halves of this dyad are differentiated in the Iliad: one hero succeeds where the other fails. But the two halves are also fused in the Iliad. We can see the fusion clearly when we take a second look at the killing of Patroklos by Hektor in Iliad XXII, since the ultimate killer here is not Hektor but the god Apollo. In the same way, beyond the Iliad, the ultimate killer of Achilles himself is not Paris but the same god Apollo, as we know directly from the plot outline of the Aithiopis, which is part of the epic Cycle. What fuses the dyad of Patroklos and Achilles is the twinning experience of death induced by the same one god, Apollo. And this death by Apollo, as we will soon see, is defined by the Homeric use of the words therapōn, usually translated as ‘attendant’, and philos, meaning ‘friend’ as a noun and ‘near and dear’ or ‘belonging to the self’ as an adjective.

§12. I start with therapōn. Here is a particularly revealing passage, where the narrative quotes the words of Achilles praying to Zeus:

Ζεῦ ἄνα Δωδωναῖε Πελασγικὲ τηλόθι ναίων
Δωδώνης μεδέων δυσχειμέρου, ἀμφὶ δὲ Σελλοὶ
235    σοὶ ναίουσ’ ὑποφῆται ἀνιπτόποδες χαμαιεῦναι,
ἠμὲν δή ποτ’ ἐμὸν ἔπος ἔκλυες εὐξαμένοιο,
τίμησας μὲν ἐμέ, μέγα δ’ ἴψαο λαὸν Ἀχαιῶν,
ἠδ’ ἔτι καὶ νῦν μοι τόδ’ ἐπικρήηνον ἐέλδωρ·
αὐτὸς μὲν γὰρ ἐγὼ μενέω νηῶν ἐν ἀγῶνι,
240    ἀλλ’ ἕταρον πέμπω πολέσιν μετὰ Μυρμιδόνεσσι
μάρνασθαι· τῷ κῦδος ἅμα πρόες εὐρύοπα Ζεῦ,
θάρσυνον δέ οἱ ἦτορ ἐνὶ φρεσίν, ὄφρα καὶ Ἕκτωρ
εἴσεται ἤ ῥα καὶ οἶος ἐπίστηται πολεμίζειν
ἡμέτερος θεράπων, ἦ οἱ τότε χεῖρες ἄαπτοι
245    μαίνονθ’, ὁππότ’ ἐγώ περ ἴω μετὰ μῶλον Ἄρηος.
αὐτὰρ ἐπεί κ’ ἀπὸ ναῦφι μάχην ἐνοπήν τε δίηται,
ἀσκηθής μοι ἔπειτα θοὰς ἐπὶ νῆας ἵκοιτο
τεύχεσί τε ξὺν πᾶσι καὶ ἀγχεμάχοις ἑτάροισιν.
Ὣς ἔφατ’ εὐχόμενος, τοῦ δ’ ἔκλυε μητίετα Ζεύς.
250    τῷ δ’ ἕτερον μὲν ἔδωκε πατήρ, ἕτερον δ’ ἀνένευσε·
νηῶν μέν οἱ ἀπώσασθαι πόλεμόν τε μάχην τε
δῶκε, σόον δ’ ἀνένευσε μάχης ἐξαπονέεσθαι.

“King Zeus,” he cried, “lord of Dodona, god of the Pelasgoi, who dwells afar,
you who hold stormy Dodona in your sway, where the Selloi,
235    your seers, dwell around you with their feet unwashed and their beds made upon the ground—
just as you heard me when I prayed to you before,
and did me honor by sending disaster on the Achaeans,
so also now grant me the fulfillment of yet a further prayer, and it is this:
I shall stay here at my assembly [agōn] of ships,
240    but I shall send my companion into battle at the head of many Myrmidons,
sending him to fight. Grant, O all-seeing Zeus, that victory may go with him;
put boldness into his heart so that Hektor
may find out whether he [Patroklos] knows how to fight alone,
[Patroklos,] my attendant [therapōn], or whether his hands can only then be so invincible
245    with their fury when I myself enter the war struggle of Ares.
Afterwards when he has chased away from the ships the attack and the cry of battle,
grant that he may return unharmed to the ships,
with his armor and his companions, fighters in close combat.”
Thus did he [Achilles] pray, and Zeus the Planner heard his prayer.
250    Part of it he did indeed grant him—but the other part he refused.
He granted that Patroklos should thrust back war and battle from the ships,
yes, he granted that. But he refused to let him come safely out of the fight.

Iliad XVI 233–248

§13. As we see at verse 244 of Iliad XVI here, Achilles himself refers to his nearest and dearest friend Patroklos as his own personal therapōn. So what does the context of this word here tell us about Patroklos and his relationship with Achilles? What we see is an affirmation about the meaning of the word therapōn in the narrative, and the future events of the epic will prove this affirmation to be true. As the narrative affirms, the wording of Achilles is mistaken when he expresses his own fond hopes for Patroklos. As the future events of the epic will show, Patroklos cannot fight alone, cannot defeat Hektor alone, and can succeed only if he fights as a pair, together with Achilles. Once Patroklos fights alone, he will die. And it is in this telling context, at Iliad XVI 244, that the wording of Achilles refers to Patroklos as his personal therapōn.

§14. So what does it mean in general, for Patroklos to be the personal therapōn of Achilles? As I will now argue, it means that Patroklos is doomed to die as the other self of Achilles.

§15. As we notice in other contexts as well, Patroklos is the personal therapōn of Achilles (for example, at Iliad XVI 165, 653; XVII 164, 388; XVIII 152). And, in each one of these contexts, therapōn is conventionally translated as ‘attendant’. So what does it mean in particular, that the hero Patroklos serves as the ‘attendant’ the hero Achilles? As we know from a variety of additional contexts where the relationship of these two heroes is described, Patroklos is the nearest and dearest companion of Achilles. Also, Patroklos is subservient to Achilles and to no one else. For example, Achilles orders Patroklos to mix and to pour wine (IX 202–204), and Patroklos complies (IX 205 ἐπεπείθετο); also, Patroklos serves the hero Achilles by preparing a meal for that hero and his guests, performing most of the tasks required for the preparation, especially the task of cooking the meat that will be served (IX 206–215). Helping Patroklos perform these tasks is another companion of Achilles, named Automedon (IX 209). This Automedon, as we will see, is an understudy of Patroklos: at a later point in the narrative, after Patroklos is already dead, Automedon will be described as a therapōn of Achilles (XXIV 573).

§16. More needs to be said about the occasion when Patroklos helps prepare a meal for Achilles and his guests. As the host on this occasion, Achilles assumes a primary role by actually slicing the meat before it is cooked (Iliad IX 209) and then distributing the sliced portions, after they are cooked, for his guests (IX 217), while Patroklos is left with the secondary role of distributing portions of bread that are placed into baskets (IX 216–217). After Patroklos is dead, Automedon takes his place in the secondary role of distributing bread in baskets on another occasion when Achilles acts as host (XXIV 625–626), while Achilles retains his primary role of distributing the meat (XXIV 626). As we will see later, this role of Automedon is relevant to his service as a therapōn of Achilles (XXIV 573).

§17. In brief, then, Patroklos as therapōn of Achilles is the nearest and dearest companion of that primary hero in the Iliad. As the personal therapōn of Achilles, Patroklos is a secondary hero, and he attends Achilles just as other therapontes who are secondary heroes will attend Achilles after Patroklos dies. A simpler way of saying it, as we will soon see, is that Patroklos takes care of Achilles. For the moment, though, I continue to use the conventional translation for therapōn as ‘attendant’. But there is more to it, much more.

§18. I now turn to the prehistory of the word therapōn, seeking to show that this word had once meant ‘ritual substitute’ and that it had been borrowed into the Greek language from Anatolian languages of Indo-European origin. The borrowing must have happened sometime in the later part of the second millenium BCE, during which period the two major Indo-European languages of Anatolia were Hittite and Luvian. The major political power in Anatolia at that time was the Hittite Empire. Accordingly, I will use the term “Hittite” as a shorthand way of referring to the relevant linguistic evidence.

§19. In Hittite ritual texts dating from roughly 1350 to 1250 BCE, we find these two relevant words: tarpanalli- (or tarpalli-) and tarpasšša-. [6] As Nadia van Brock has shown, these words were used as synonyms, and both meant ‘ritual substitute’. [7]

§20. Such a meaning, ‘ritual substitute’, must be understood in the context of a Hittite ritual of purification that expels pollution from the person to be purified and transfers it into a person or an animal or an object that serves as a ritual substitute; the act of transferring pollution into the victim serving as ritual substitute may be accomplished either by destroying or by expelling the victim, who or which is identified as another self, un autre soi-même. [8] According to the logic of this Hittite ritual of substitution, the identification of the self with the victim serving as the other self can take on a wide variety of forms: the victims range from humans to animals to figurines to ceramic vessels. [9]

§21. The mentality of identifying with your victim operates on homological principles. In the case of animal victims designated by the word tarpalli-, for example, one ritual text specifies that bulls are to be killed as ritual substitutes for men, while cows are to be killed as substitutes for women. [10] There are other examples of homologies based primarily on gender. In another ritual text involving the word tarpalli-, bulls and rams and other male animals are killed as ritual substitutes for the king, while corresponding female animals are killed for the queen. [11] And there are cases of tighter homologies. In yet another ritual text involving the word tarpalli-, for example, it is specified that the victims who are designated as ritual substitutes for the king include men as well as bulls and rams. [12] Further, there are other cases as well where humans are being designated as ritual substitutes. [13]

§22. The range of victims that are designated as ritual substitutes, extending all the way to humans, indicates that the victim of the ritual substitution, as the other self, can be identified as closely as possible with the human self—even if the ritual substitute and the human self may not be all that close to each other when they are viewed from outside the world of ritual. [14] What makes the substitute in ritual seem so intimately close to you is that he or she or it must die for you. Here I find it relevant to quote from a text of royal ritual substitution a most explicit formulation, expressed in dialogic format:

Keilschrifturkunden aus Boghazköi XXIV 5 I 15–16 [15]

§23. I draw special attention to these and other cases of ritual substitution where the person to be purified is a king. In such cases, as van Brock argues, the ritual of substitution is “périodique,” ideally annual; and it is a common idea, as we can see from a survey of myths and rituals around the world, that the king is an incarnation of the body politic, of society itself, which needs to be renewed periodically by being purified of pollution. [16]

§24. As we consider the relevant evidence from the Near East, a well-known model of periodic renewal is the Babylonian new year festival, centering on the sacrificial killing of a goat, and it is nowadays generally agreed that the Hittite rituals of substitution derive at least in part from the Babylonian rituals that marked this festival; [17] a related practice, attested in texts stemming from the neo-Assyrian empire of the first millennium BCE, is the periodic appointing and subsequent killing of substitute kings. Especially relevant are the correspondences of the kings Asarhaddon (680–669 BCE) and Assurbanipal (668–627 BCE). [18] Still another related practice is the ritual of the scapegoat described in the Hebrew Bible, Leviticus 16:8, where a designated goat (who is the tragos pempomenos of the Greek Septuagint and the caper emissarius of the Latin Vulgate) is not killed but expelled into the wasteland—hence the word scapegoat; this periodic expulsion, as is well known, figures as a climactic moment in the rituals and sacred narratives of the Jewish Day of Atonement.

§25. Since Hittite is an Indo-European language, no matter how deeply it is influenced by Near Eastern civilizations, we may also compare the relevant evidence of Hittite ritual formulations that are cognate with wording that we find in other Indo-European languages. A case in point is Latin sōns / sontis, meaning ‘guilty’, which is cognate with the Greek participle ōn / ontos (ὤν / ὄντος) of the verb meaning ‘to be’ as in esti (ἐστι) ‘is’ and with the corresponding Hittite participle ašān that likewise means ‘to be’, as in ešzi ‘is’. In the Plague Prayers of King Muršilis II, dating from the second half of the 14th century BCE, it is prescribed that the king is to utter a “confessional” formula, ašān-at, iyanun-at ‘it is true, I did it’ (Keilschrifturkunden aus Boghazköi XIV 8, in the second prayer); this formula is cognate with the formula implicit in Latin sōns, where the meaning ‘guilty’ is to be understood in the legal sense of ‘declared guilty’ or, to say it even more legalistically, ‘found guilty’. [19] So also in the “confessional” formula of the Hittite king, the guilty party must declare that he really ‘is’ the guilty one, that he really is ‘it’. [20] Similarly in the children’s game of tag, the formula ‘you’re it’ indicates by way of the verb ‘to be’ the identity of who will be ‘it’.

§26. A moment ago, I said that the ritual substitute can seem intimately close to you because he or she or it must die for you, and I gave the example of the formula used by the Hittite king for saying that the tarpalli- or ritual substitute will die for him so that he may live. But there are two sides to this formula. The intimate closeness is matched by an alienating distance, marked by pollution, separating the king from his substitute. I draw attention here to a most telling example. In one particular ritual text (Keilschrifturkunden aus Boghazköi XXIV 5 + IX 13 I lines 19–26), where the word tarpalli- applies to a prisoner, this ritual substitute is anointed with royal oil, crowned with a diadem, and dressed in the regalia of the king; then this tarpalli- is expelled from the king’s territory and sent back home to his own territory, so that he takes home with him the pollution that had been intimately associated with the king. [21] I stress here the intimacy of the actual transfer of pollution, even if the pollution itself is alienating. In another ritual text (Keilschrifttexte aus Boghazköi XVI 1 I line 10), it is specified that the king is to take off the royal clothing that he wears so that the prisoner who serves as his ritual substitute may now put on this same clothing. [22]

§27. In the two examples we have just seen, the ritual substitute is expelled and does not die for the king, but the basic fact remains: the king and the body politic get rid of the pollution by getting rid of the ritual substitute. Let me restate the fact by using the English word eliminate, derived from a most telling Latin word, ē-līmin-āre ‘take outside the boundary [līmen]’. So the king and the body politic eliminate the pollution by eliminating the ritual substitute. A remarkable parallel is the case of the goat that gets expelled into the wasteland on the Jewish Day of Atonement, instead of getting killed like the sacrificial goat of the Babylonian New Year.

§28. Rituals of elimination, that is, of expelling a polluted person or animal or thing, are in fact homologous with rituals of killing. For example, when an animal is designated as a ritual substitute for the king in Hittite texts, we may expect two alternative outcomes: in some rituals, the animal victim is killed and its body is burned, [23] but in other rituals the victim is instead expelled. [24] And I think that there existed a parallel set of two alternative outcomes when a human was designated as a ritual substitute. That is, I think we may expect that human substitutes could be not only expelled but also killed in rituals dating from the Hittite era, just as substitute kings could be killed in rituals dating from the a later era represented by neo-Assyrian texts. Granted, the testimony of the existing Hittite ritual texts is opaque concerning the actual killing of humans in contexts of ritual substitution, but the fact remains that there are clear examples of killings of humans in other Hittite ritual contexts. [25]

§29. Throughout this analysis I have refrained from using the term “human sacrifice,” since some readers will view the word “sacrifice” too narrowly by thinking only of the killing and subsequent dismembering and cooking and eating of animal victims. If we allowed, however, for a broadening of this word “sacrifice” to include the killing and subsequent burning of animal victims, which as we have seen is an option in the case of animal victims of ritual substitution, then the term “human sacrifice” could still apply in the case of human victims of ritual substitution.

§30. That said, I bring to a close my analysis of the relevant Hittite evidence by offering this summary, following the earlier formulation by van Brock: the mentality of substitution rituals requires that someone who is notionally close to the king must die or be in some other way eliminated so as to preserve the king. [26]

§31. I now turn to the corresponding evidence in Greek. The Hittite words tarpanalli-/tarpalli- and tarpasšša-, as van Brock has argued, were borrowed by the Greek language sometime in the second millennium BCE, and the corresponding Greek words were therapōn (θεράπων) and theraps (θέραψ), both of which can be translated as ‘attendant’. [27] Like the two Hittite words tarpanalli-/tarpalli- and tarpasšša-, the two Greek words therapōn and theraps were once synonyms, as is evident from the fact that the verb therapeuein is attested as a functional derivative of the noun therapōn in Homeric diction. We can see this functional derivation at work when we look at the context of Odyssey xiii 265, where this verb therapeuein means ‘be a therapōn’ even though it is formally derived not from the noun therapōn but from the noun theraps, which is absent from Homeric diction. [28] We find attestations of theraps only rarely, as in Ion of Chios F 27 ed. West; Euripides Ion 94, Suppliants 762. In the fragment from Ion of Chios, the plural form therapes refers to attendants who serve wine at a symposium; in the Ion of Euripides, the same plural form refers to the priests of Apollo at Delphi who serve as attendants of the god as they approach the streams of the spring Kastalia; and, in the Suppliants of Euripides, therapes again refers to attendants—in this case, the hero Adrastus is asking the Messenger whether therapes have removed the corpses of the fallen dead.

§32. So how do we explain the meaning of the Hittite words tarpanalli- and tarpasšša– as ‘ritual substitute’ when we compare the meaning of the borrowed Greek words therapōn (θεράπων) and theraps (θέραψ) as ‘attendant’? Here I return to my formulation summarizing the role of Patroklos as the personal attendant of Achilles:

§33. Building on this formulation, I will now explore another aspect of the service of Patroklos as the personal therapōn of Achilles in the Iliad: Patroklos serves as the personal charioteer or hēniokhos of Achilles (Iliad XXIII 280 ἡνίοχος). The role of Patroklos as the charioteer of Achilles is specially highlighted in Iliad XVII (475–478), where Automedon describes Patroklos as the best of all charioteers by virtue of driving the chariot of Achilles. Automedon, as we have already seen, is another near and dear companion of Achilles. And he too, as we will now see, is a chariot driver.

§34. The wording used by Automedon in Iliad XVII (475–478) in describing Patroklos as the best of all charioteers is most relevant to his own role as a charioteer. Here in Iliad XVII, Patroklos is of course already dead. He died in Iliad XVI, getting killed in place of Achilles. Back then in Iliad XVI, it was Automedon who had served as the charioteer of Patroklos. To appreciate this role of Automedon as charioteer of Patroklos, I now review what happened in Iliad XVI when Patroklos had died fighting Hektor.

§35. The setting for the death of Patroklos in Iliad XVI is a classic chariot fight. The fight starts when Patroklos leaps out of his chariot:

Iliad XVI 733

In a moment, Hector will leap out from his own chariot. Before that happens, however, Patroklos picks up a rock and throws it at Kebriones, the charioteer of Hector, hitting Kebriones on the forehead and smashing his skull (Iliad XVI 734–754). And then, just as Patroklos had leapt out of his chariot, Hector too leaps out of his own chariot:

Iliad XVI 755

Patroklos and Hector proceed to fight one-on-one in mortal combat on foot—a combat that is won by Hector (XVI 756–863).

§36. In this chariot fight that happened in Iliad XVI, I highlight the fact that it is Automedon who serves as chariot driver for Patroklos. And, at this moment, it is Patroklos and not Achilles who is the chariot fighter, since it is Automedon and not Patroklos who is the chariot driver. In preparation for this chariot fight between Patroklos and Hektor, it is Automedon, serving as chariot driver for Patroklos, who yokes the horses of Achilles to the chariot (XVI 145–154).

§37. Like Patroklos, as we have already noted, Automedon is described in the Iliad as a therapōn of Achilles (XVI 865). Also, in another Iliadic passage, Automedon and a companion named Alkimos are described as therapontes of Achilles (XXIV 573). That passage goes on to say that Achilles honors these two companions, Automedon and Alkimos, more than anyone else—now that Patroklos is dead (XXIV 575). And, in still further Iliadic passages, we see that one of the functions of these two honored therapontes of Achilles is the unyoking of horses or mules (at XXIV 576) as well as the yoking of horses (at XIX 392–393, where Automedon and Alkimos are yoking for Achilles his horses). So also, as we have just seen, Automedon yokes for Patroklos the horses of Achilles (at XVI 145–154).

§38. After the death of Patroklos, when Achilles finally rejoins the Achaeans in battle, his chariot is now driven by Automedon (XIX 395–399). As we have seen, however, Automedon had at an earlier point served as chariot driver for the hero Patroklos when that hero took the place of Achilles in war (XVI 145–154). And here I note a most telling detail about that earlier point in the narrative of the Iliad: after Patroklos is killed by Hektor, the chariot driver Automedon says that he now wants to become a chariot fighter, but he cannot fight the Trojans while he is still driving the chariot (XVII 463–465). So he asks another companion, Alkimedon, to take his place as a chariot driver in order that he, Automedon, may now become a chariot fighter:

Iliad XVII 479–480

And, sure enough, Alkimedon quickly leaps into the chariot, landing on the chariot platform (XVII 481 ἐπορούσας) and taking hold of the whip and the reins (XVII 482), while Automedon leaps out of the chariot, that is, he leaps off the chariot platform (XVII 483 ἀπόρουσε) and lands on the ground, where he can then start fighting. [29] So we see here a functioning dyadic relationship between Automedon as a chariot fighter and Alkimedon as a chariot driver, both of whom are secondary substitutes for the primary substitute Patroklos, the premier chariot driver who became a chariot fighter for Achilles and who thus died for him as his therapōn, as his personal ritual substitute. [30]

§39. I conclude, then, that the relationship of the chariot fighter to the chariot driver who substitutes for him is parallel to the relationship of a hero like Achilles to a hero like Patroklos, who is his therapōn. By now we see, on the basis of evidence from the narrative traditions of Homeric poetry, that Patroklos as the therapōn of Achilles does in fact serve as his substitute. In the end, the chariot driver in this case dies in place of the chariot fighter: that is, the chariot driver takes the hit, as it were, for the chariot fighter. But now that we see how Patroklos is a substitute for Achilles, the question remains: how is he not only a substitute but also a ritual substitute? We must now examine more closely how the actual concept of a ritual substitute, as attested in Hittite ritual texts, was translated into the ancient Greek song culture.

§40. For analyzing the concept of ritual substitute as attested in the Greek evidence, an ideal starting point is a climactic passage in Iliad XVI where the warrior hero Patroklos is killed in battle: at this moment, the hero is visualized as atalantos Arēi ‘equal to Ares’ (line 784). Now we will see that Patroklos in such a context is a stand-in for Achilles, or, in other words, Patroklos is a ritual substitute for the main hero of the Iliad. And just as Patroklos as ritual substitute of Achilles qualifies as ‘equal to Ares’, we can expect Achilles himself to qualify for epithets meaning ‘equal to Ares’.

§41. Here I find it most relevant to consider some basic facts about the use of the word therapōn, the plural form of which is therapontes. In the Iliad, warriors are conventionally called the therapontes of Ares as the god of war (II 110, VI 67, XV 733, XIX 78). [31] With this fact in mind, I will now make an argument that can be summarized in the following formulation:

§42. As an epic warrior, Achilles is a therapōn or ‘ritual substitute’ of Ares by virtue of becoming identical to the war god at the moment of death. In the Iliad, however, this relationship between Achilles and Ares is expressed only by way of an intermediary, who is Patroklos. This warrior is described not as the therapōn of Ares but rather as the therapōn of Achilles, and, as such, Patroklos is not only that hero’s ‘attendant’ but also his ‘ritual substitute’, since he actually dies for Achilles. So Achilles in the Iliad dies only indirectly as the therapōn of Ares through the intermediacy of Patroklos, who dies in this epic as the therapōn of Achilles.

§43. Here I come back to Iliad XVI 233–248, which was the first passage that we considered in this essay: there we saw that Patroklos qualifies as therapōn of Achilles only so long as he stays within his limits as the recessive equivalent of the dominant hero; once he is on his own, however, he becomes a therapōn of Ares and dies in place of Achilles. [33]

§44. As an epic warrior, Achilles qualifies as īsos Arēi ‘equal to Ares’, just like Patroklos. This description suits Achilles in the Iliad—though it applies to him only vicariously by way of Patroklos, who takes upon himself the role of a ritual substitute for Achilles. [34] In Iliad XVI 784, as we have already seen, Patroklos is called atalantos Arēi ‘equal to Ares’ at the exact moment when he is killed. And, as we will soon see, Patroklos is actually called īsos Arēi ‘equal to Ares’ at the exact moment when the story of his fatal impersonation of Achilles begins.

§45. Besides being equated with Ares, however, Patroklos is also being equated with Apollo. It happens in Iliad XVI 705 and 786, when Patroklos is called daimoni īsos ‘equal to a daimōn’. As we know from the contexts of these passages, the daimōn or ‘otherworldly force’ here is the god Apollo himself. [35] So in these contexts Patroklos is ‘equal’ to Apollo, though his identification with this god is not fully spelled out, since the word daimōn partly masks the identity of the god.

§46. As one who is equal to Apollo at the moment of his death, Patroklos participates in a specialized god-hero relationship. [36] By being equal to Ares at the moment of his death, Patroklos participates in a generic god-hero relationship that is typical of heroes who are warriors. [37] In identifying with both Ares and Apollo, however, Patroklos is experiencing something more special—something that will later be experienced by Achilles himself, who will also be identifying with both Ares and Apollo at the moment of his own heroic death, though his death scene is not directly pictured in the Iliad.

§47. In this way, Patroklos is a perfect stand-in for the main hero of the narrative, Achilles, whose specialized ritual antagonist is Apollo. Although Achilles, just like other warriors, can have as his generalized ritual antagonist the god Ares, he also has as his specialized ritual antagonist the god Apollo.

§48. Patroklos is drawn into the specialized relationship of Achilles with Apollo at the precise moment when he is equated with Ares and thus marked for doom:

τὸν δὲ ἰδὼν ἐνόησε ποδάρκης δῖος Ἀχιλλεύς·
600    ἑστήκει γὰρ ἐπὶ πρυμνῇ μεγακήτεϊ νηῒ
εἰσορόων πόνον αἰπὺν ἰῶκά τε δακρυόεσσαν.
αἶψα δ’ ἑταῖρον ἑὸν Πατροκλῆα προσέειπε
φθεγξάμενος παρὰ νηός· ὃ δὲ κλισίηθεν ἀκούσας
ἔκμολεν ἶσος Ἄρηϊ, κακοῦ δ’ ἄρα οἱ πέλεν ἀρχή.
605    τὸν πρότερος προσέειπε Μενοιτίου ἄλκιμος υἱός·
τίπτέ με κικλήσκεις Ἀχιλεῦ; τί δέ σε χρεὼ ἐμεῖο;

He [Nestor] was seen and noted by swift-footed radiant Achilles,
600    who was standing on the spacious stern of his ship,
watching the hard stress [ponos] and tearful struggle of the fight.
He called to his companion Patroklos,
calling from the ship, and he [Patroklos] from inside the tent heard him [Achilles],
and he [Patroklos] came out, equal [īsos] to Ares, and here indeed was the beginning of the doom that presently befell him.
605    He [Patroklos], powerful son of Menoitios, was the first to speak, and he said [to Achilles]:
“Why, Achilles, do you call me? what need do you have for me?”

Iliad XI 599–606

Here at verse 604 Homeric poetry declares explicitly that the application of the epithet ‘equal to Ares’ will doom Patroklos to death. [38]

§49. A generic warrior, as we have noted, is called a therapōn of Ares. Generically, then, heroes as warriors die for Ares. More specifically, however, a special hero will die for his special divine antagonist.

§50. Generically, Achilles would be a therapōn of Ares; specifically, however, we can say that he is a therapōn of Apollo, because it is Apollo who will directly kill him, as we know from the plot summary of the Aithiopis. And, while the therapōn of Apollo must be Achilles, the therapōn of Achilles must be, as we have seen, Patroklos.

§51. Patroklos must die for Achilles, who must die for Apollo. The death of Patroklos is caused by Ares generically but it is brought to fulfillment by Apollo personally.

§52. I return here to the moment when Patroklos dies. At that moment, as we saw, he is called daimoni īsos ‘equal to a daimōn’ (XVI 705 and 786). And, at that precise moment, he is in sacred space. Since war is ritual, the battleground is for the warrior a sacred space. Patroklos is doomed, selected for death, and that is why he is ‘equal to a daimōn’. But this sacred space is not only sacred: it is also violent and even sinister. When you are a warrior fighting in the sacred space of Ares, being ‘equal to a daimōn’ is to have martial fury. Ares is not only the god of war, he is the god of martial fury in war. [39]

§53. I return here to the ritual background of the word therapōn: it was borrowed into the Greek language, as we have seen, from Anatolian languages, sometime in the 2nd millennium BCE. The corresponding word in those Anatolian languages meant ‘ritual substitute’. Someone who is notionally close to the king, as we have also seen, may have to die in place of the king.

§54. Such a death, I argue, has the effect of healing society by way of healing the king, who is viewed as the embodiment of society, of the body politic. What I describe here for the first time as a healing is an act of purifying the king and his people from impurities, from pollution. If the king is polluted, then society is polluted. That is why the pollution of the king has to be transferred to a ritual substitute who will be eliminated in place of the king and will thus remove the royal pollution while also removing the pollution of society. This principle of purification has been described by van Brock as the transfer of evil, “le transfert du mal.” [40] Evil must be passed on, to a sacrificial victim.

§55. In Greek visual art, I must now add, the dead hero Patroklos can be represented as a sacrificial ram, who is shown with his throat slit open and with blood streaming from the gaping wound: such a picture is painted on an Attic vase executed by the “Triptolemos Painter,” dated around 480 BCE. [41] Similarly in Hittite rituals of substitution, as we have seen, rams can be sacrificed in place of kings.

§56. The meaning of the Greek word therapōn as ritual substitute and the function of such a therapōn as a healer helps explain why the related Greek word therapeuein means not only ‘be a therapōn’, as we have seen at Odyssey xiii 265, but also ‘heal, cure’; we still see such a meaning embedded in the English-language borrowings therapy and therapeutic. But before I cite some contexts where the ancient Greek word therapeuein means ‘heal, cure’, I must return once again to that passage at Odyssey xiii 265 where therapeuein means ‘be a therapōn’, since I have yet to explain the context.

§57. There are in fact three attestations of therapeuein where the word does not overtly mean ‘heal, cure’. The first attestation is at Odyssey xiii 265, the passage we are considering right now. Here the first-person narrator of a “Cretan tale” says that he was unwilling to ‘be a therapōn’, therapeuein, for the Cretan king Idomeneus, preferring instead to be the leader of his own companions. The second attestation is at Homeric Hymn to Apollo 390, where the god Apollo selects a group of Cretans to serve as his attendants, therapeuein, at his shrine in Delphi. Finally, the third attestation is at Hesiod Works and Days 135, where the prototypical humans who represent the second generation of humankind are said to be unwilling to serve as attendants, therapeuein, to the gods; and, as we read in the next verse, these sacrilegious proto-humans are likewise unwilling to sacrifice to the gods at their altars (Works and Days 136). [42]

§58. As we consider these three early attestations, the first one of the three is not decisive in establishing the overall meaning of therapeuein, since the story of the upstart Cretan who refused to serve as therapōn to the king of Crete has no attested parallels. Still, it is safe to say that the social position of the therapōn in this story cannot be too different from the social position of Patroklos himself, who is subservient to Achilles by virtue of serving as that hero’s therapōn. [43] But the second and the third attestations are in fact decisive: in these two cases, therapeuein refers to the service that needs to be rendered to gods by humans who are designated as the gods’ attendants. As we are about to see, the contexts of therapeuein in these two cases can help explain later attestations of the verb therapeuein in the sense of ‘heal, cure’.

§59. In speaking of later attestations, I have in mind evidence dating from the fifth century and thereafter. In this later era, therapeuein in the sense of ‘take care of’ can refer to the procedure of healing a body by removing some form of sickness or, more basically, to the procedure of maintaining the well-being of the body. To maintain the well-being of the body is to keep it healthy—that is, keeping it sound and immune from any sickness. [44]

§60. But there are also more specialized contexts of therapeuein in the sense of ‘take care of’, some of which are explicitly sacred. And, in such sacred contexts, the body that is being cared for and kept sound by those who are attending it is either (1) the notional body of a god or (2) the actual body of a cult hero. Such a sacred body can lend its sacredness to anything that makes contact with it, such as a temple or shrine or any other kind of sacred enclosure. In the case of gods, the sacred power of the sacred body can extend to a sacred simulacrum of the body, such as a sacred statue or picture or any other object that stands for the body of the god. There are many different attestations of therapeuein where the object of the verb is whatever such sacred thing or place is attended by the attendants who care for it. Here are three shining examples:

1. An Attic inscription dating from the fifth century BCE (Inscriptiones Graecae I³ 1–2 138.17) speaks of the need for therapeuein ‘taking care of’ the temenos ‘sacred precinct’ of the god Apollo ‘in the most beautiful way possible’ ([το̄ τε]μένο̄ς το̄ Ἀπόλλο̄νο[ς ἐπιμελέσθο̄ν, ὅπος ἂν κάλλισ]τα θεραπεύε̄ται).
2. A Cretan inscription dating from the second century BCE (Inscriptiones Creticae III:2 1.5) speaks of the need for therapeuein ‘taking care of’ archaic statues of divinities, (τὰ ἀρχαῖα [ἀ]γάλματα θαραπεύσαντες). [45]
3. In the Ion of Euripides (110–111), dating from the late fifth century BCE, the young hero Ion speaks of his service of therapeuein ‘taking care of’ the temple of Apollo at Delphi (τοὺς θρέψαντας | Φοίβου ναοὺς θεραπεύω).

§61. It is in the light of such relatively later attestations of the verb therapeuein that we can understand the earlier attestations of the noun therapōn in combination with the genitive case of names of gods like Apollo, the Muses, Ares, and so on. [46] Also, of special relevance to my ongoing argumentation about twinning is the attestation of the dual form theraponte with reference to the twin sons of Poseidon, who are Pelias and Neleus, described as attendants of the god Zeus himself (θεράποντε Odyssey xi 255).

§62. By now we have seen that therapeuein in the basic sense of ‘maintain the well-being’ and in the derivative sense of ‘heal, cure’ is in fact related to the idea of a ritual substitute who maintains the well-being of someone superior whom he serves by standing ready to die for that special someone. That is the therapeutic function, as it were, of the therapōn. Earlier on, I noted the English-language borrowings therapy and therapeutic. Now I note a semantic parallel in the use of the Greek word pharmakon, which means ‘drug used for healing’ or, more generally, ‘drug used for medication or for poisoning’, and we see the more specific meaning ‘drug used for healing’ embedded in the English-language borrowings pharmacy and pharmaceutical. The meaning of this word pharmakon as ‘drug used for healing’ helps explain the related meaning of a related Greek word, pharmakos, which designates a person who serves a very special ritual function. That person is what we call a “scapegoat,” that is, someone who takes the blame for a pollution that afflicts a whole society. [47] Here again we see at work the principle of a transfer of evil, comparable to what we saw in the case of the Hittite ritual substitutes.

§63. Having said this much about therapōn, I turn to the other of the two words that I intended to analyze in this essay. That word, as I noted at the beginning, is philos, meaning ‘friend’ as a noun and ‘near and dear’ or ‘belonging to the self’ as an adjective. By contrast with my lengthy analysis of therapōn, however, I can confine myself here to the shortest of summaries, since I have already analyzed this word philos at some length in my earlier work. [48] Here I attempt to summarize all that work in a single nested paragraph:

§64. Such an idea of Patroklos as the other self of Achilles is surely parallel to the idea of twinning, and this parallelism helps explain other features of Achilles and Patroklos that they share with the Dioskouroi, such as the power to heal. The therapeutic powers of Achilles and Patroklos are analyzed in this light by Douglas Frame in his twin essay.

§65. The time has come for me to conclude. As the other self who is ready to die for the self that is Achilles, Patroklos achieves an unsurpassed level of intimacy with the greatest hero of the Homeric Iliad. This intimacy is sacral, thus transcending even sexual intimacy. But this sacred intimacy has an uncanny other side to it, which is a kind of sacred alienation. As we saw in the case of the Hittite prisoner, about to be expelled into an alien realm, he must wear the clothing of the king, thus becoming ritually intimate with the body of the king. So too Patroklos wears the armor of Achilles when he dies, and he wears something else that is even more intimately connected with his best friend. Patroklos wears also the epic identity of Achilles, as expressed by the epithets they share. These heroic epithets, such as the one that makes them both ‘equal to Ares’, will predestine both of them to live and die the same way. And the sameness of their shared life and death can be seen as an uncanny mix of intimacy and alienation that only twins will ever truly understand.

Bibliography

Allen, T. W. 1912. ed. Homeri Opera V (Hymns, Cycle, fragments). Oxford.

Chantraine, P. 2009. Dictionnaire étymologique de la langue grecque: histoire des mots (ed. J. Taillardat, O. Masson, and J.-L. Perpillou), with a supplement Chroniques d’étymologie grecque 1–10 (ed. A. Blanc, Ch. de Lamberterie, and Jean-Louis Perpillou). Paris. Abbreviated DELG.

Davidson, O. M. 1980. “Indo-European Dimensions of Herakles in Iliad 19.95–133.” Arethusa 13:197–202.

Frame, D. 2009. Hippota Nestor. Hellenic Studies 37. Cambridge MA and Washington DC.

Griffiths, A. 1985. “Patroklos the Ram.” Bulletin of the Institute for Classical Studies 32:49–50.

Griffiths, A. 1989. “Patroklos the Ram (Again).” Bulletin of the Institute for Classical Studies 36:139.

Janko, R. 1992. The Iliad: A Commentary. Vol. 4, Books 13–16. Cambridge.

Kümmel, H. M. 1967. Ersatzrituale für den hethitischen König. Wiesbaden.

Lowenstam, S. 1981. The Death of Patroklos: A Study in Typology. Beiträge zur Klassischen Philologie 133. Königstein/Ts.

Nagy, G. 1972. Introduction, Parts I and II, and Conclusions. Greek: A Survey of Recent Work (F. W. Householder and G. Nagy) 15–72. Janua Linguarum Series Practica 211. The Hague.

Nagy, G. 1990a. Pindar’s Homer: The Lyric Possession of an Epic Past. Baltimore. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Pindars_Homer.1990.

Nagy, G.  1990b. Greek Mythology and Poetics. Ithaca, NY. Paperback edition 1992 https://continuum.fas.harvard.edu/greek-mythology-and-poetics/. Corrigenda for the printed version… On p, 90 line 4 up in the text proper,“wind” not “winds” and “XX 440” not “XX 40.” On p. 150, sunr̥tāvati not sunr̥tāvari two times, and Sūnr̥tāvatī not Sūnr̥tāvarī one time. On p. 203 between “same line)” and “specified,” insert “of the marital bed; similarly, she ‘recognizes’ (ἀναγνούσῃ xix 250) as sēmata (same line) the clothes…” (in the present printed version, the reference to the marital bed as sēmata at Odyssey xxiii 206 is distorted by a typographical omission of the wording that needs to be restored here: by haplography, the mention of the marital bed is omitted, and this omission distorts the point being made about the clothes and brooch of Odysseus as sēmata in their own right at xix 250). On p. 214n32, “Pausanias 8.44.4 not 9.44.4.”

Nagy, G. 1999. The Best of the Achaeans: Concepts of the Hero in Archaic Greek Poetry. Revised ed. of Nagy 1979, with new introduction. Baltimore. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Best_of_the_Achaeans.1999.

Nagy, G. 2005a. “The Epic Hero.” A Companion to Ancient Epic (ed. J. M. Foley) 71–89. Oxford.

Nagy, G. 2005b. “An Apobatic Moment for Achilles as Athlete at the Festival of the Panathenaia.” Imeros 5:311–317.

Nagy, G. 2006. “The Epic Hero.” Online and fuller version of Nagy 2005a, with notes, http://chs.harvard.edu/publications/.

Nagy, G. 2007a / b. “Lyric and Greek Myth” / “Homer and Greek Myth.” The Cambridge Companion to Greek Mythology (ed. R. D. Woodard) 19–51 / 52–82. Cambridge. Online editions at http://chs.harvard.edu/publications/.

Nagy, G. 2008. Greek: An Updating of a Survey of Recent Work. Cambridge MA and Washington DC. Updating of Nagy 1972, online ed. This edition tracks the page-numberings of Nagy 1972. http://chs.harvard.edu/publications/.

Nagy, G. 2008|2009. Homer the Classic. Online | printed version. http://chs.harvard.edu/publications/ | Hellenic Studies 36. Cambridge MA and Washington DC.

Nagy, G. 2009|2010. Homer the Preclassic. Online | printed version. http://chs.harvard.edu/publications/ | Berkeley and Los Angeles CA.

Nagy, G. 2009. “An Apobatic Moment for Achilles as Athlete at the Festival of the Panathenaia.” Online and fuller version of Nagy 2005b. http://chs.harvard.edu/publications/.

Nagy, G. 2013. The Ancient Greek Hero in 24 Hours. Cambridge, MA. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_NagyG.The_Ancient_Greek_Hero_in_24_Hours.2013.

Parpola, S. 1970 / 1983. Letters from Assyrian Scholars to the Kings Esarhaddon and Assurbanipal. I/II. Neukirchen-Vluyn.

Sinos, D. S. 1980. Achilles, Patroklos, and the Meaning of Philos. Innsbrucker Beiträge zur Sprachwissenschaft 29. Innsbruck.

Stähler, K. P. 1967. Grab und Psyche des Patroklos: Ein schwartzfiguriges Vasenbild. Münster.

Tarenzi, V. 2005. “Patroclo ΘΕΡΑΠΩΝ.” Quaderni Urbinati di Cultura Classica 80:25–38.

Van Brock, N. 1959. “Substitution rituelle.” Revue Hittite et Asianique 65:117–146.

Van Brock, N. 1961. Recherches sur le vocabulaire médical du grec ancien. Paris.

Watkins, C. 1967. “Latin sōns.” Studies in Historical Linguistics in honor of George Sherman Lane (ed. W. W. Arndt et al.) 186–194. Chapel Hill NC. Reprinted in Watkins, C. 1994. Selected Writings II (ed. L. Oliver) 405–413. Innsbruck.

Watkins 1995. How to Kill a Dragon: Aspects of Indo-European Poetics. New York and Oxford.

Footnotes