Thinking comparatively about Greek mythology V, Reconstructing Hēraklēs forward in time

2019.08.22 | By Gregory Nagy

§0. Previously, in “Thinking comparatively about Greek mythology IV,” hereafter abbreviated as TC IV, I was reconstructing the mythological persona of the Greek hero Hēraklēs by tracing him backward in time, back to the earliest reconstructable phases of myths that told his story. Here in TC V, I will trace such myths forward in time, and I will start my procedure of “reconstructing forward” by concentrating on the earliest phases of myths telling of deeds performed by Hēraklēs in the service of kings. A case in point will be a mythological persona by the name of Augeias (the latinized spelling is Augeas), king of Elis. As I noted in TC IV, Hēraklēs is pictured in the act of performing a service for this king in the sculptural ensemble of Metope 12 in the Temple of Zeus at Olympia. But what is the hero actually doing here? Well, he is in the middle of shoveling an immeasurable accumulation of manure produced by countless cattle in the Stables of Augeias. I show here a line drawing and, later on in this essay, I will analyze the relevance of what is being pictured.

Metope 12: Hēraklēs clearing the Stables of Augeias, from the Temple of Zeus. Published in Curtius, Ernst, and Friedrich Adler, eds., Olympia: die Ergebnisse der von dem Deutschen Reich veranstalteten Ausgrabung. Tafelband 3: Die Bildwerke von Olympia in Stein und Thon (Berlin, 1894): plate 45.

§1. In TC IV, I started by analyzing an expanded narrative in the text of Diodorus of Sicily (4.11.3–4.26.4), dating from the first century BCE, about the āthloi or ‘Labors’ of Hēraklēs, and then I went backward from there, back in time, back to the fifth century BCE, which was the era when the Temple of Zeus at Olympia was being built. And then, in what seemed at the time to be no more than an afterthought, I was thinking back even further backward in time, centuries earlier, back to an era when Homeric poetry was emerging in the form that we still know today. Reading the Homeric text that dates all the way back to that era, I highlighted Iliad 19.95–133, where we find a compressed narrative about, again, the Labors of Hēraklēs, the word for which is, again, āthloi (aethloi), at line 133. And I argued that these two versions of stories about the Labors of the Greek hero Hēraklēs—a much newer expanded version and a much older compressed version—are comparable not only with each other but also with even older stories being told about deeds performed for a king by a persona that I described at TC IV §10 simply as a ‘Strong Man’. (My spelling Strong Man makes a distinction between this kind of man and the kind I spell strongman, which is an entirely different story, as we see if we look, but not now, at Classical Inquiries 2019.06.21.) The stories about deeds performed by the Strong Man, as I will now argue, are so old that they can be reconstructed back to “Indo-European” traditions of mythmaking—I use here, once again, the terminology of linguists who compare not only the cognate languages stemming from the so-called Indo-European language family but also the cognate myths inherited by these languages. In other words, we can reconstruct the stories about the Labors of Hēraklēs even farther back in time, over millennia, aiming to recover a prehistoric era when the Indo-European languages were as yet undifferentiated from each other. And the farther back we go in time, the more closely the myths about the Labors of Hēraklēs are linked to myths about deeds performed by Strong Man for King. But this linking becomes even more evident, as we will now see, when we reconstruct the relevant myths by tracing them forward in time. Here in TC V, then, I will take as my starting-point the various “Indo-European” myths about Strong Man and King as reconstructed by Georges Dumézil in two books I have already noted in TC IV: The Destiny of the Warrior (1970; in French, Dumézil 1969, second edition 1985) and The Stakes of the Warrior (1983b; in French, Part 1 of Dumézil 1971). Starting with such myths, I will look ahead in time to find relevant points of comparison in Greek sources. But now the evidence of such Greek sources will come mostly from narratives that deal not only with the Labors of Hēraklēs but also with what I called in TC IV the hero’s sub-Labors. In the case of the mythological figure Augeias, who will be my primary point of interest, he belongs to both sets of myths about Hēraklēs: that is to say, Augeias in his role as king of Elis figures in both the Labors and the sub-Labors of the hero.

§2. To be contrasted with Augeias, king of Elis, is, a mythological figure like Eurystheus, king of Mycenae, who has a proper place only in the myths about the Labors of Hēraklēs, not in the myths about the hero’s sub-Labors. As we will see in what follows, such a distinction between the kings Eurystheus and Augeias can be highlighted by way of reconstructing the relevant myths forward in time, not backward.

§3. But first I need to say more about the differences between forward- and backward-reconstruction. What follows is a set of relevant observations extracted from a short essay about other matters, Classical Inquiries 2017.03.23, which was in turn extracted from a much longer essay, Nagy 2011, likewise about other matters. To make cross-referencing easier, I duplicate the paragraph-numbers (preceded by “¶”) for the observations I offered in Classical Inquiries 2017.03.23, and I add the paragraph-numbers, where relevant, from Nagy 2011 (preceded by “<”).

¶1<10. For reconstructing either backward or forward in time, both diachronic and historical perspectives are needed:

A. diachronic perspectives need to be correlated with synchronic perspectives

B. these two perspectives need to be correlated in turn with historical perspectives.

¶2<11. In using the terms synchronic and diachronic, I rely on working definitions recorded in a book stemming from lectures given by the linguist Ferdinand de Saussure (1916:117). Here I paraphrase these definitions from the original French wording:

A synchronic perspective has to do with the static aspect of linguistic analysis, whereas a diachronic perspective deals with various kinds of evolution. So synchrony and diachrony refer respectively to an existing state of a language and to phases of evolution in the language. I note especially the equation here of the words diachronicand evolutionary.

¶3<12. And now I need to add that a diachronic or evolutionary perspective is not the same thing as a historical perspective.

¶4<13. The remark that I just added here about diachrony and history is based on the following formulation (Nagy 1990a 1§9 = p. 21n18):

It is a mistake to equate diachronic with historical, as is often done. Diachrony refers to the potential for evolution in a structure, whereas history is not restricted to phenomena that are structurally predictable.

¶5<15. In the same general context, with reference to synchronic as well as diachronic approaches to the study of cultural evidence (Nagy 1979/1999 §25 = p. xv with n3), I quote a relevant formulation by the anthropologist Pierre-Yves Jacopin: “Both synchrony and diachrony are abstractions extrapolated from a model of reality” (Jacopin 1988:35–36).

¶6<16. A key word in the formulation by Jacopin is model. Both synchronic and diachronic perspectives are a matter of model building. We can build synchronic models to describe and explain the workings of a structure as we see it attested in a given historical context. We can likewise build diachronic models to describe and explain how that given structure may have evolved from one of its phases into other phases. What we have built, however, is a set of models to be tested on historical realities. The models are not the same thing as the realities themselves. And the realities of history as a process are not dependent on such models. History may either confirm or upset any or all aspects of our models, since the contingencies of history do not need to follow the rules of existing structures.

¶7<17. The aim, then, in applying synchronic and diachronic perspectives is to build synchronic and diachronic models for the description of structures and for visualizing the evolution of these structures. And the building of such models may be applied not only to linguistic structures but also to the cultural structures of traditions in general (as I showed in both Nagy 1979/1999 and 1990a).

¶10. Before I continue, I need to delimit my use of the terms synchronic and diachronic, as well as my use of the term historical.

¶11<20. I offer here two different ways of further delimiting the terms synchronic and diachronic, thus bringing them into sharper focus:

A. The terms synchronic and diachronic need to be applied consistently from the objective standpoint of an outsider who is thinking about a given structure, not from the subjective standpoint of an insider who is thinking within that structure (Nagy 1990a 0§11 = p. 4). Such an objective standpoint enhances the synchronic as well as the diachronic perspectives that are needed for describing structures and for explaining how these structures evolve. This way of looking at a given structure helps avoid the pitfall of assuming that one’s own synchronic or diachronic perspectives are identical with the perspectives of those who were part of the culture in which that structure was historically anchored. Such an assumption runs the risk of misreading the historical context in which the structure is attested.

B. Whereas synchronic and diachronic perspectives are needed to describe a given structure as it exists at a given time and as it evolves through time, historical perspectives are needed to describe what actually happened to that structure. As I noted already, what happened in history can be unpredictable, since we cannot predict the contingencies of history. So, when it comes to reconstructing what happened to a given structure, it is not enough to use a purely diachronic perspective. As I have also already noted, a purely diachronic perspective is restricted to phenomena that are structurally predictable.

¶12<21. Here I come to the third of the three delimitations I am now proposing: in analyzing a given structure, synchronic and diachronic perspectives need to be applied before historical judgments or prejudgments can be made.

¶13<22. The delimitation I have just outlined is especially important in situations where we find little or no historical evidence for earlier attestations of a given structure. I am addressing here one of the biggest problems that historians face when they try to view structures over time. If they apply only a historical perspective as they reconstruct a given structure backward in time, back to the era when that structure is actually documented, they find themselves limited to the realities they find in that era. And the only way they can reconstruct further back in time is to find further documentation stemming from earlier eras.

¶14<23. By contrast, a diachronic perspective provides also for the reconstruction of realities that that are historically undocumented. And reconstruction from a diachronic perspective is not restricted to the hindsight of history. A diachronic perspective not only makes it possible to reconstruct backward in time by tracing the evolution of a given structure back to undocumented phases of that structure. It also makes it possible to reconstruct forward in time.

¶15<24. In my own work on linguistics (Nagy 1972|2008:19), I applied the concept of reconstructing backward and forward in time with reference to the term Common Greek, which refers to a diachronic model developed by linguists. I offer here a summary:

I am speaking here about the historical evidence for a chronological demarcation between pre-documented and documented eras of the Greek language. Experts used to place this demarcation somewhere around the eighth century BCE, which is the era when alphabetic writing was first being introduced into the Greek-speaking world. The Greek language as it existed in what was understood to be the pre-documented era on the farther side of this demarcation could only be reconstructed diachronically, all the way back to a hypothetical proto-language known to linguists as Common Greek. This proto-language, Common Greek, is not a historical reality but a construct, a diachronic model. But then a major shift in demarcation took place, signaled by the decipherment of Linear B, which was a system of syllabic writing that dates back to the second millennium BCE. Once the decipherment revealed that the language written in this script was an earlier form of Greek, the documented era of the Greek language needed to be pushed back into the second millennium BCE, and this newly demarcated older era could now reveal new historical facts about the language. These new facts in some ways confirmed but in other ways contradicted the reconstructions achieved by way of diachronic perspectives that had already been developed before the decipherment of Linear B (Nagy 1972|2008:33). Those previous reconstructions, which were dominated by the hindsight of later history, needed to be modified in the light of earlier history. So now a new diachronic model of Common Greek needed to be built by way of reconstructing backward in time, even farther back than before. And, now that an earlier historical phase of Greek had been discovered, this discovery required re-adjustments in how we reconstruct forward in time from that earlier phase to later phases.

¶16<25. From this example, we can see that the diachronic process of reconstructing forward as well as backward in time depends on the data provided by historical evidence. But the actual reconstruction of structures depends primarily on diachronic and synchronic perspectives and only secondarily on a historical perspective. I say this because the historical perspective works only by hindsight, whereas the diachronic perspective allows for foresight as well, so to speak, by way of the procedure I describe here as reconstructing forward in time.

¶17<26. For a prime example, I highlight here a set of findings achieved by applying another diachronic model. This model is another construct built by linguists, and this one is even bigger than the model of Common Greek. The diachronic model I have in mind here is what German-speaking linguists call Indo-Germanic and other linguists call Indo-European or Common Indo-European or proto-Indo-European. I focus here on an example of what kinds of things we can find when we reconstruct forward as well as backward in Indo-European linguistics (Nagy 1979/1999 §21 = 339–340):

The example centers on the etymology of the Greek word pontos (πόντος) ‘sea’, which is cognate with the following words in other Indo-European languages: Latin pōns ‘bridge’, Armenian hun ‘ford’, Old Church Slavonic рǫtǐ and Old Prussian pintis ‘path’, Sanskrit pánthāḥ and Avestan pantå ‘path’. When we reconstruct all these words backward in time, back to an undocumented common proto-language known to linguists as Common Indo-European or proto-Indo-European, such reconstruction backward in time does not help us fully comprehend the semantic relationship of the meaning ‘sea’ in Greek with such divergent meanings as ‘bridge’, ‘ford’, and ‘path’ in the other Indo-European languages. It is only after we reconstruct forward in time, taking into account all the comparative evidence we derive from the cognate languages that we factored into our reconstruction backward in time, that we can comprehend more fully the convergent meaning that unifies diachronically the divergent meanings of these words. This convergent meaning has to do with a crossing, over a dangerous body of water or over some other dangerous zone, that sacralizes the one who succeeds in achieving such a dangerous crossing (Nagy 1972|2008:48–49, following Benveniste 1954|1966:296–298). Only then, only after we have reconstructed forward in time, can we understand the contexts of the word pontos (πόντος) ‘sea’ in the earliest attested phases of Greek poetry, where we see expressions of dread about dangerous sea crossings and references to the sacralizing effect of such crossings. Further evidence comes from the derivative form Hellēs-pontos (Ἑλλήσ-ποντος), which is the name of a famous strait that we know as the Hellespont and which means etymologically ‘the crossing of Helle’, referring to a myth about a dangerous crossing of this strait by a girl named Helle and by her brother, who are being carried across the dangerous waters by a ram with a golden fleece: the girl falls off the ram and drowns in the Hellespont while her brother succeeds in crossing the strait and is thus sacralized.

¶18<27. This example shows that diachronic analysis, by way of reconstructing forward in time, can enhance not only historical analysis but also synchronic analysis, since a purely synchronic analysis of the attested contexts of pontos (πόντος) would yield only the meaning ‘sea’. The underlying sense of a dangerous crossing that sacralizes would be impossible to recover without applying a diachronic perspective.

§4. With these observations in place, I return to my current project, which is a forward-reconstruction of Indo-European myths about Strong Man and King. And I return to the myth about the twelfth Labor of Hēraklēs as pictured on Metope 12 of the Temple of Zeus at Olympia, a sacred site controlled by the state of Elis at the time when the Temple was built, in the fifth century BCE. We find a vivid retelling of the myth in the Library of “Apollodorus,” dating from the second century CE. According to “Apollodorus” (2.5.5; especially pp. 195 and 197 ed. Frazer 1921 I), Eurystheus the king of Mycenae ordered Hēraklēs to remove an immeasurable accumulation of manure produced by countless cattle in the Stables of Augeias the king of Elis. In terms of the myth as as we see it being retold in the visual narrative of the metopes built into the Temple of Zeus at Olympia, this deed of Hēraklēs, the removal of manure, was the twelfth and last of the hero’s Labors, represented in the twelfth and last of the twelve metopes gracing the temple. By contrast, in terms of the myth as retold by “Apollodorus” (again, 2.5.5), this deed of Hēraklēs is not Labor 12 but Labor 5, and the retelling makes it clear that this particular Labor was invalidated by Eurystheus, on the grounds that Hēraklēs had demanded that the king Augeias should pay wages to the hero for his labor. Similarly, Eurystheus had invalidated Labor 2, the hero’s killing of the Hydra (2.5.2), on the grounds that Hēraklēs had fought the beast not alone but with the help of his nephew, Iolaos. In terms of this same version of the myth, the Labors imposed by Eurystheus on Hēraklēs were originally ten, not twelve, and it was only because Labors 2 and 5 were invalidated that Labors 11 and 12 were added, centering respectively on the Apples of the Hesperides and on Cerberus the Hound of Hādēs (2.5.11 and 2.5.12).

§5. The invalidation by Eurystheus of Labor 5 in the narrative of “Apollodorus” (again,2.5.5; pp. 195 and 197 ed. Frazer 1921 I) is of particular interest. At first, it seems to make no sense to us from a purely synchronic point of view. Why should it matter, if Hēraklēs is to be paid or not paid any wages in compensation for labor performed in the service of a king? But it does matter—from a diachronic point of view. And that is because the labor that the hero is performing here is in the service of the king Eurystheus, not the king Augeias. If the hero Hēraklēs had been exclusively in the service of the king Augeias when he was shoveling out the manure from the Stables of that king, he could have demanded compensation. But he must perform this Labor gratis, not for payment, because he is really in the service of the king Eurystheus. It was Eurystheus, not Augeias, who gave the orders to our Strong Man. And, as we forward-reconstruct the model of “Strong Man in the service of King,” the Strong Man must obey the king’s orders, no matter what, since these orders are divinely authorized. If the Strong Man does not obey, it is a “sin.” In TC IV §§4–5, I have already pointed to the reconstructions of Georges Dumézil concerning various such “sins” committed by a Strong Man in the service of a King.

§6. The Labors of Hēraklēs are ordered by the king Eurystheus, yes, but the orders of this king are authorized by Zeus. These orders, these commands, can be quite arbitrary, but the authorization is purposeful. When Hēraklēs is commanded to shovel the immeasurable accumulation of manure produced by countless cattle—and that is what we see him actually doing at Metope 12 while his divine patroness Athena is sternly looking on—his choice to go ahead and shovel can count as a Labor only if he accepts the authorization of Zeus. If he accepts—and I say it again—he must perform the Labor gratis, not for payment.

§7. So, it is irrelevant, from a synchronic point of view, that Augeias, in the story as retold by “Apollodorus” (2.5.5; again, pp. 195 and 197 ed. Frazer 1921 I), refuses to pay wages to Hēraklēs after the hero’s work is done. But it is relevant for Hēraklēs from a diachronic point of view—if the deed that he performs in this case is not a Labor but a sub-Labor, which is what the Labor becomes, a sub-Labor, because it is invalidated by the king Eurystheus.

§8. This sub-Labor, on the other hand, is still a Labor for the king Augeias. When he devalues, by way of non-payment, the work of Hēraklēs, this king is acting without the authorization of Zeus. In terms of the myth as visually retold in Metope 12 of the Temple of Zeus in Olympia, the Labor performed by Hēraklēs for the king of Elis is relevant to the local traditions of Elis, where the Strong Man will take revenge on a king who has acted in bad faith and who thus becomes polluted. In the version of the myth as reported by “Apollodorus” (2.5.5; again, pp. 195 and 197 ed. Frazer 1921 I), the magnificent scale of the Labor performed by Hēraklēs for Augeias even magnifies the need for revenge against the polluter: in this version, the Strong Man had used his superhuman strength to divert the rivers Alpheus (Alpheiós) and Peneus (Pēneiós) in the direction of the Stables, thus washing away the polluting accumulation of manure. But Hēraklēs is denied remuneration, despite the support of a prince of Elis named Phyleus. And now Hēraklēs, along with Phyleus, is contemptuously sent away by Augeias. I should add that, immediately after he is sent away, Hēraklēs manages to perform a most notable sub-Labor: he goes off to a wedding feast and kills the Centaur Eurytion, who had forced the hero Dexamenos to give his daughter away to the beast. All this is happening in a very compressed retelling by “Apollodorus”  (2.5.5; again, pp. 195 and 197 ed. Frazer 1921 I).

Hércules desvía el curso del río Alfeo
“Hércules desvía el curso del río Alfeo” [Hercules diverting the course of the River Alpheus]. Francisco de Zurbarán (1598-1664). Image via Wikimedia Commons.
§9. But, eventually, Hēraklēs has his revenge: the hero raises an army and goes to war against Elis, killing Augeias and installing the prince Phyleus as the new king there. All this is happening in another very compressed retelling by “Apollodorus” (2.7.2; pp. 249 and 251 ed. Frazer 1921 I).

§10. And here is what I find most remarkable about this retelling: it is what “Apollodorus” says right after having said that Hēraklēs installed a new kingship in Elis. The very next thing that the hero does, says “Apollodorus” (again, 2.7.2; pp. 249 and 251 ed. Frazer 1921 I), is to establish the athletic competition, agōn, of the Olympics. We see here an explicit correlation of this athletic festival with the very idea of sovereignty. And it is most relevant, I think, that Elis in the fifth century BCE had gained control of Olympia, the official site of the Olympics, and that this control was eventually celebrated by way of building, in the fifth century, the Temple of Zeus in Olympia. In the twelfth metope of the Temple, the twelfth Labor of Hēraklēs could be seen as a primal event that had set off the chain of events leading to the foundation of the Olympics.

§11. I close by taking one last look, for now, at the king Augeias. Even if he is a negative model of kingship, since he is pictured as a bad king, he is nevertheless a good fit for an “Indo-European” pattern of stories about a Strong Man in the service of a King. In such stories, as we will see in TC VI, there are multiple opportunities for “sinning” to be committed by the King, not only by the Strong Man. I should add that the name of the king Augeias, like the name of the hero Hēraklēs, has an Indo-European etymology. As I have argued (Nagy 1979/1999 15§9 = pp. 174–175), the Homeric name of Augeas/Augeias, which is Augeíās (as in Iliad 11.701), is the Ionic reflex of *Augā́ās (by way of *Augḗās), derivative of a noun that survives as augḗ, referring to sunlight. And it is no coincidence, I think, that Augeias was the son of the Sun according a myth mentioned in passing by “Apollodorus” (2.5.5). Similarly, I have also argued (again, Nagy 1979/1999 15§9 = pp. 174–175) that the Homeric name of Aeneas/Aineias, which is Aineíās, can be explained as the Ionic reflex of *Ainā́ās (by way of *Ainḗās), derivative of a noun that survives as aínē ‘praise’ (as in Herodotus 3.74, 8.112), which is a by-form of aînos, again in the sense of ‘praise’. On Aeneas as a royal personality, I refer to my comments introducing Rhapsody 20 of Iliad 20 in SCIO. Here is what I say there, “By virtue of being the son of Aphrodite/Venus, Aeneas possesses a genealogical and dynastic charisma that threatens to overshadow the purely epic charisma of his Iliadic opponent Achilles.”


See the dynamic Bibliography for Comments on Comparative Mythology.