"The Horse and the Maiden" (Aeschines 1.182 etc.): An Urban Legend in Ancient Athens

Lowell Edmunds


With Appendices by Ruth Palmer

An Italian version of this paper was given at the University Milan (Dec. 4, 1997) and at the University of Trieste (Dec. 9, 1997).

At present (Dec. 1999), it strikes me that what I have written amounts to a small detail in the larger picture of the maiden, the young woman, in ancient Greece and in Athens in particular. This picture began to emerge in Claude Calame’s Les Choeurs de jeunes filles en Grèce archaïque, 1977), a book that was more than slightly ahead of its time. Now Hoffmann 1992, Sissa 1990 (orig. pub. 1987), Sourvinou-Inwood 1988, Visintin 1992, and Pellizer 1994, not to mention various writings of Nicole Loraux, have shown both the importance of the young woman’s social role and also the anxieties that surrounded this person. In this context I also mention Linda Roccos’ 1995 article on the kanêphoros. Through the kindness of James Whitley, I was able to read the typescript of a paper on horses, gender, and social status in Iron Age Athens that he gave at the AIA meeting in Chicago in 1997 (a longer version of the paper is listed as forthcoming in Works Cited). His approach has implications for "The Horse and the Maiden," and one of the great advantages of publication on the web is that it will be possible, in future editions, to bring his work into relation with mine.

A known weakness of this paper is the undeveloped (and perhaps, from the anthropologist’s point of view, outdated) concept of symbolism. I have not yet read Eco 1981. A known annoyance to readers is the transliteration of the Greek. A few words of Greek appear in the table of sources, but that is because the table is an image, not text.

I have received a lot of help with this paper. Claude Calame’s extensive comments are still not fully reflected in this version. I should like to thank:

Alan Dundes
Antonio Aloni
Maurizio Bettini
Carlo Brillante
Claude Calame
Carol J. Clover
Susan Edmunds
Mario Geymonat
Giulio Guidorizzi
Jennifer Kosak
Simone Marchesi
Richard Martin
Adrienne Mayor
Ruth Palmer
Athanassios Papalexandrou
Ezio Pellizer
Carlo Presotto
Linda Roccos
Roberta Sevieri
Monica Visintin
James Whitley


I will discuss an Athenian story attested in various sources beginning in the fourth century B.C.E.. This is the story of a maiden, called Leimônê, who was punished by her father, Hippomenes, for a sexual trespass. He shut her in a building with a horse and there she died, either by the violence of the horse or eaten by the horse. Most versions of the story end with the observation that there is still in Athens a place called "Of the Horse and the Maiden."

This story has no author. It is an anonymous creation that belongs to Greek folk narrative. If this category, i.e. folk narrative, is divided, as is customary, into the three genres of myth, folktale, and legend, then our story clearly belongs to the genre of legend.1 Furthermore, it belongs to the sub-genre called urban legend (German Stadtsage; Italian leggenda metropolitana; French légende urbaine). This is not the time for a lengthy discussion of terms and definitions. I am under no illusion that this term or any of the others just used corresponds to a mental category of the Greeks. The value of "urban legend" is that it indicates what I take to have been the status of "The Horse and the Maiden" in ancient Athens, which was clearly quite different from the canonical stories about the city’s early times and also from the great pan-Hellenic myths. I shall simply offer a working definition of "urban legend," and then continue with my discussion of "The Horse and the Maiden." An urban legend is a short, often mono-episodic narrative linked to a particular city or place. It is "performed" or transmitted primarily in the mode of conversation. It is presented as historically true, and expects credence.2 It contains elements of humor or horror. If there is horror, it is often punishishment of someone who defies society's conventions.3 Its lifespan and area of diffusion vary considerably from one example to another. It can, of course, be reused in some new place and reassigned to that new place. Or it may become known outside its original home but remain attached to that home, as in the case of "The Horse and the Maiden," which was used by Callimachus and by Dio Chrysostom as an Athenian legend.

Now, to continue my discussion of this story, an anthropological hermeneutics would attempt to show its relation to two contexts: one, the larger mythical tradition from which it might be presumed to emerge; two, its circumstances of enunciation or "performance." The meaning of the story for its original audience lies, I assume, at the intersection of performance and tradition, or narrative actualization and narrative potentiality. But the story poses problems for this approach, as I will explain.

First, some words of introduction concerning the sources and the state of the phenomenon to be explained.


In my reporting of the sources, I have paid special attention to the place in Athens to which the story is attached. The place in which the horse and the maiden were enclosed was called "Of the horse and the maiden."4 The place and its name are the most constant features of the story. Usually, as presented in the sources, the existence of the place confirms the truth of the story. But this is the logic of the narrator, who is using the story for his own purpose, as an illustration. The typical logic of the sources is: A. General proposition, e.g., Athenians were once more severe. B. Illustrative example: story of "The Horse and the Maiden." C. Confirmation of the truth of the story: the place named "Of the horse and the maiden" still exists in Athens.

If the place in fact existed, as Aeschines’ reference surely implies, then it is reasonable to assume, as scholars have always done, that the story was originally an aition for the place or for the placename.5 A first obstacle to anthropological interpretation is that we have no example of the story as an aition, no example of a primary context of enunciation, which is not to be expected if in fact this context was conversation. We have only secondary contexts, the reuse of the story for some ulterior purpose.

The nature of the place

What was the place to which the aition refers? Wilamowitz believed that there was once in Athens a funerary monument depicting the maiden and the horse and that this monument, which would have been destroyed in 480 B.C.E., carried the inscription Leimônês Hippomenous. The inscription carried no information about the identity of this Hippomenes; at some point he was identified with the early Athenian ruler, as in Aristotle’s Athênaiôn Polîteia.6   But, as Wilamowitz observes, there was nothing self-evident about this identification, and in Aeschines’ version of the story (346/5 B.C.E.), the father is an anonymous Athenian citizen.7

I would reconstruct the background differently from Wilamowitz. It seems to me that the discovery of a burial in Athens like the one at Toumba, Lefkandi could have inspired the story. In the anaktoron at Toumba were buried a man and a woman together with horses.8 I am well aware that no such burial has been found in Athens and that the burial of horses is rare and unattested anywhere else on mainland Greece except Mycenae and Lefkandi.9 Still, the same class of aristocrats who erected the anaktoron at Toumba existed in Athens at the same period. The burial of a hero under his dwelling was not uncommon in the Protogeometric period. One only has to imagine the burial of a princess in the same fashion, along with the sacrifice of a horse. At some later period, long after these customs have been forgotten, the Athenians discover the burial, and try to explain to themselves how a woman and a horse could have been buried together.10 (An example of the kind of burial that I am suggesting has been found at Archanes, on Crete.11)

Two versions

Such a burial is of course pure speculation, as is Wilamowitz’ funerary monument. Whatever the material inspiration for the story, it is clear that the aition required the narrative joining of two terms, horse and maiden. The aition fulfilled this requirement in two ways: there are two versions of the story. In the more fully attested one, the maiden lost her virginity before marriage. In some sources for this version, her father punished her seducer, too. Either he used the seducer as a horse, yoking him to a chariot, or he dragged him from a horse. Then the maiden was shut up with a horse, and died either by the violence of the horse or because she was eaten by the horse.

In the other version, less well attested, the maiden had sexual relations with a horse, and this was the reason for her punishment. A scholiast to Ovid’s Ibis describes her as in corio equino clausa, ut ab equo stupraretur, and adds that she was devoured by the same horse.

Lateness of source

In passing, I observe that the lateness of this scholiastic source is not a reason for discounting it. I take it as a general principle for this kind of research that what is late in the history of the sources may be early in the history of the story.

Both versions to be explained

So there are two versions of "The Horse and the Maiden." Neither can be shown to be the original.12 Both might be original, in the sense that both came into existence at the same time and both became current and thus passed into written sources. Neither has a self-evidently greater importance, despite the imbalance in the sources for the two versions. Both, then, must be explained.

Here is the place to refer to the articles on "The Horse and the Maiden" by Valerio Petrarca and by Paulette Ghiron-Bistagne, the only two extensive discussions of this story that I have found.13 One of these scholars concentrates on one version; the other on the other.

The second problem

The second problem for anthropological hermeneutics in the case of "The Horse and the Maiden" is that the usual method of comparison does not work. In this method, which was first elaborated in folklore studies, a myth or story is compared with a type established on the basis of a set of comparable stories.14 This method is especially useful, and will be useful in this paper, in "establishing identity or nonidentity of parallel versions of ancient myths."15 Comparison is also useful in filling in the lacunae in a fragmentary narrative16 or in explaining difficult elements in a complete story. A recent example is Maurizio Bettini’s "Il racconto di Alcmene e Anfitrione: Un’ analisi antropologica."17 This comparative method ought to be useful in explaining a very short and not very well attested story like "The Horse and the Maiden."

But comparanda are lacking. There are no modern examples. This story does not belong to an international type of folktale. There are no ancient examples, either. Petrarca attempted to transpose the story into the myth of Demeter and Poseidon, but, in order to do so, he had to postulate that one version of the story was the true one, and, further, that, in this one version, the sexual relations of the maiden with the horse constituted the essential story. His interpretation overlooks the fact that in both versions the punishment is the substance of the aition. His interpretation fails to explain, in particular, the motif of anthropophagy.

An alternate approach

Deprived of the comparative approach, one looks for some other. I will again begin from the notion of the story as an aition. As such, it is a response to an unusual cognitive demand. It was necessary that the juxtaposition of a horse and a maiden be explained. A story was generated for this purpose. The extemporaneous character of the story is immediately apparent. I have already commented on the ad hoc nature of the identification of the father as the early Athenian ruler Hippomenes. In order to acquire this identity the father of course had to have the name Hippomenes. The name of the father is suggested by the horse, as Samson Eitrem suggested long ago.18 The maiden’s name, Leimônê "Meadow," is not an Athenian name. It occurs for the first time in the "Horse and the Maiden" and reappears only once, in the fifth century C.E., in Aristaenetus.19 Perhaps "Leimônê" was chosen for our story because of its sexual connotation,20 or because horse suggested meadow. Whatever the reason for the name, it is another indication of the non-traditional character of the story.


It was not, however, a free invention, a fiction in our sense, or a fantasy, but followed two cultural "laws." One is a logic of personal and family relations. The other is the already fixed symbolic value of the elements that the aition must explain: maiden and horse. This fixed valence is partly but not entirely a matter of the appearance of these terms in earlier myths. As I will show, a larger symbolism is at work. The process of invention that I will now describe is distinct, then, from the Greimasian model, in which symbols are created by the conversion of empirical reality into so-called "figures." In "The Horse and the Maiden," we have the reuse of already created symbols or figures. The story may, of course, still conform to the Greimasian canonic schema but that is another question. This story is a meta-formation.

The cultural logic of personal and social relations

The presence of the horse in the same place as the maiden must be explained. The storyteller could not refer to the sacrifice of a horse at the tomb of a maiden, which was unknown to him. As I have already indicated, some burials of horses are known to us, but we have the advantage of archaeology, which the storyteller did not have.21 Neither did literature provide any examples. The sacrifice of horses, along with dogs and Trojan captives, on the pyre of Patroclus would have been well known, but was another matter. The bones of Patroclus are preserved separately in a gold jar; the bones of the others, humans and animals, are mingled. indiscriminately on the pyre.22The only literary evidence known to me that might have been known to the storyteller was the burial with Cimon son of Stesagoras of the three horses with which he had won Olympic victories. The source is Herodotus.23 But even this burial was very remote from the phenonenon that the storyteller had to explain.

So he could not refer to the horse as an honor. There would have to be some other reason for the horse’s presence. If the storyteller could not imagine the horse as an honor or perhaps as a companion of the maiden in the underworld, he could only imagine it as a punishment. As I will explain, the horse, though given by chance, i.e. by the circumstances that required the aition, is especially appropriate as the agent of punishment. But this coincidence belongs to the symbolism of the horse and, at this point, I am only imagining the logical formation of the narrative. If the horse was a punishment, it was a punishment for something. For what would a maiden be punished? The most readily available answer is: sexual trespass. She is already well-suited to play this role in the story because she has been long since defined, in, for example, tragedy and Hippocratic medicine, as a fundamentally unstable creature.24

In tragedy, the disobedience of the maiden takes extreme forms. She may even, like Antigone, prefer death to marriage. In the Hippocratic corpus, there is a late, fragmentary work called Peri partheniôn (De virginum morbis).25 The author describes the various afflictions caused by menstrual blood in the bodies of young women—fever, fear, visions, the desire to jump into wells or to hang themseves. What advice does the author give? "Je recommande aux jeunes filles ... de se marier le plus tôt possible; en effet, si elles deviennent enceintes, elles guérissent; dans le cas contraire, à l’époque même de la puberté, ou peu après, elles seront prises de cette affection, sinon d’une autre." Note the last phrase: "sinon d’une autre." The unwed woman will inevitably become ill! Hippocratic physiology dictates the same thing as social norms: marriage.26

To continue with the cultural logic of the story, the source of the maiden’s punishment would, of course, be her father.27

So the aition generates, from the presence of a horse as a term to be explained, the punishment of the maiden, the name of her father Hippomenes "Spirit of the Horse," and also the form of the punishment of the seducer. A horse is employed, in some of the sources, to punish the maiden’s seducer, too.

Cultural logic continued: punishment of a daughter by her father

The punishment of a daughter by her father was already available in Greek myth. Crotopos, an Argive king, executes his daughter, Psamathê, when he learns that she has had an illegitimate child.28 Acrisius sets Danaë adrift with her illegitimate son, Perseus.29 Such stories provide precedents. But they are not, as far as I can see, narrative models for "The Horse and the Maiden."

Neither are the myths in which a father sacrifices a daughter for the good of an army or a city, the most famous of which is Agamemnon’s sacrifice of Iphigeneia. (But note the curious detail in Aesch. Ag. 238: the binding of Iphigeneia’s mouth is compared to a bridle.) Ezio Pellizer has analyzed a group of these stories in Chapter 5 of Favole d’identità, favole di paura.30 In such myths, the father sheds the blood of the daughter.

In "The Horse and the Maiden," the father implicitly wishes to avoid the taint of blood.31 The punishment, in this respect, is like the interment of Antigone. But the cultural logic of relations does not determine the elaboration of the punishment of Leimônê. Why should she have been enclosed in the same building with a horse? Why not alone, like Antigone?


I now turn to the symbolism underlying the construction of the story, beginning with the version in which Leimônê is seduced.

Version 1

In a very widely attested image, the unmarried girl is a filly. Her education, which prepares her for marriage, is a process of taming. Further, when she marries, she is yoked by her husband. Marriage is the ultimate taming.32 The symbolism of the horse covers the whole life of the young woman from birth up to and including marriage. 33 If, like Leimônê, a young women has sexual intercourse before marriage, then she has violated the process. She has assumed the yoke but has remained a filly, untamed. She has created a contradiction in the symbolism of the horse as it applies to young women.

If Leimônê is reduced, by her crime, to the status of a horse, then her enclosure with a horse is appropriate.34 The punishment is symmetrical with the crime. If she acts like an untamed horse, then she will be punished by a horse that, if once tame, is made wild by hunger.

It is interesting to see how the same symbolism that underlies the story of "The Horse and the Maiden" also underlies the well-known myth of Hippodameia. Her father, Oinomaos, the king of Elis, set as a condition for her suitors that they carry her in their chariot from Elis to Corinth before he could overtake them. A superior chariot-driver, he overtook and killed all of them until Pelops came along. Pelops, who had been trained by Poseidon, defeated Oinomaos and married Hippodameia. Oinomaos was dragged to death by his own horses. In this way, the young woman, whose name, meaning "Taming of Horses," stands for female education, is finally freed from the negative horsemanship of her father and yoked in marriage thanks to the horsemanship of Pelops. As Claude Calame has shown, the myth links the two aspects of equine symbolism, education and marriage.35

In the story of "The Horse and the Maiden," these two aspects are disjoined. Because of the crime of the maiden, the horsemanship of her father, whose name means "Spirit of the Horse," will go to kill the improper suitor who yoked himself to the maiden before marriage. The maiden herself dies, too, again through her father’s use of a horse, and therefore, unlike Hippodameia, she never completes the process of taming that should culminate in marriage.

It is quite easy for us to analyze "The Horse and the Maiden" as a transformation of the myth of Hippodameia. The two stories share many of the same terms (father, daughter, suitor, horses), and it might seem that the Athenian story has been generated by the more famous one. But the bride-quest of Pelops belongs to a definable narrative type in which the father of the princess sets a test or series of tests for her suitors. Many fail; at last the hero succeeds. A quasi-historical example is the tests imposed upon the suitors of his daughter by Cleisthenes of Sicyon. It is one of the favorite stories in Herodotus, and is remembered because of the anti-hero Hippokleides.36 "The Horse and Maiden" is clearly remote from this type of story, in which the central feature is the test. Again, in the case of Hippolytus, it is easy for us to describe a structural resemblance. Hippolytus, suspected of a sexual trespass, is punished by his father, with the aid of Poseidon. The punishment entails the use of horses, in this case, Hippolytus’ own horses. But, in this myth, we have a well-established international story type, known as "Potiphar’s wife." It is attested in the Egyptian story called "The Two Brothers" and in the Bible, where Joseph plays the role of Hippolytus.37 "The Horse and the Maiden" does not belong to this type. Typology shows what this story is not, but not what it is.

Version 1, symbolism continued

I have discussed the equine symbolism for the education and marriage of the woman. This symbolism is part of a larger complex that includes negative aspects. The horse is ambiguous, representing both an asset to man and also a threatening, demonic force.38 As an asset to man, the horse has been domesticated by the bit, by the control of its mouth. As a demonic force, it panics, takes the bit in its teeth, and overcomes human control. The chariot crashes. In Greek myth and religion, as the structural analysis of Vernant and Detienne showed, Athena and Poseidon correspond to the two aspects of the horse.39 Athena invented the bit; Poseidon causes horses to panic.

The horse has forty teeth, correctly counted by Aristotle in Historia Animalium (576a7), and the most prominent of these are twelve incisors. The horse, from a certain point of view, looks as if it would be capable of anthropophagy, and in Greek myth there are two pairs of man-eating horses.40 One pair belongs to Diomedes the Thracian, and are the object of the eighth labor of Heracles.41 The other pair belongs to Glaucus of Potniae, the father of Bellerophon.42

Version 2

Another symbolic value of the horse is expressed in the other version of "The Horse and the Maiden," in which the maiden has had sexual intercourse with the horse which is then used as the instrument of her punishment.

The horse, said Aristotle, is the most salacious of animals after the human species (575b31). In a rare anecdote in Historia Animalium he reports: "In a stable in Opus there was a stallion that used to serve mares when forty years old [i.e. having lived twice as long as the normal horse]; his fore legs had to be lifted up for the operation" (576b26). Just as the horse, with its large incisors, would not refrain, at least in the imaginary realm, from eating a human, so, with its large phallus, it would not refrain from intercourse with a human female, given the chance.

It is not surprising that Greek obscenity includes equine imagery for human sexual activity. In its own way, this obscenity is normative. A man can proudly claim that he is the horseman who will mount a filly, as in Anacreon’s poem that begins pôle Thrêikiê.43 But if the woman "rides" the man, then she is a prostitute, or, if married, she is shameless.44 These sexual metaphors correspond to the non-obscene metaphors for female education and marriage. They show that the appropriate female role is the passive one. She should not be the rider but the horse. Leimônê, who has had sexual intercourse with a horse, has taken the obscene metaphor literally and has become, for sexual purposes, a mare. The two versions of the story prove to be homologous. In the first version of the story, Leimônê failed to complete the taming process that would culminate in marriage, and so she remained wild. In the second version, her wildness is signified by her assuming the sexual role of the mare. The second version of the story is therefore an overdetermination of the "wild" sexual activity in the first version.

To conclude my remarks on the second version, I will consider the possibility of mythical parallels for it. The sexual relations of a human female and a horse are unparalleled in Greek myth. I have already referred to Petrarca’s attempt to transpose the story of Leimônê into the myth of Demeter and Poseidon. Here I will add that this myth belongs to a series in which Poseidon is the progenitor of a powerful horse. The mother is variously an Erinys, otherwise unidentified, Medousa, or a Harpuia.45 In the version in which Poseidon mates with Demeter, she has transformed herself into a mare and he has transformed himself into a stallion (Paus. 8.25.4-7).46 What is striking about the story of Leimônê and the horse is that it is a human female and a "real" horse. Perhaps the only precedent for Leimônê’s deed is the intercourse of Pasiphaë and the bull, which was cited by Dio Chrysostom in the same passage in which he referred to Leimônê. But neither Pasiphaë nor Demeter is a typological parallel and neither of their myths explains the story of Leimônê.

The second version of "The Horse and the Maiden" brings into the present, i. e., into the day in which this legend originated, the earlier state of affairs when the centaurs were created, a period of generation that Carlo Brillante has called "equine."47 A story not discussed by Brillante, on which I have done only a bit of research concerns Hippê (sometimes Hippô), the daughter of Cheiron. In flight from her father (for what reason, I do not yet know), she bore Melanippê and was then transformed into a horse amongst the stars by Zeus. Melanippê had twins by Poseidon. The main source is, I think, the fragments and testimonia of Eur. Melanippê hê desmôtis. See Loeb Literary Papyri, p. 119.


I conclude with some reflections on methodology. Lacking typographical parallels for either version of the story, and lacking also any primary context of enunciation, I turned to the cultural logic of personal relations and to the symbolism of the horse.

From this analysis, something can be inferred about the primary context: the aition was strongly moralistic. It served as a warning to Athenian girls: no sex before marriage. This moral is, of course, not only Athenian but panhellenic, and, by itself, is not, perhaps, of great interest. It is only another example of an apparently universal function of myth and folklore. The moral of the story, deriving from societal norms, shapes the story, which, in turn, reinforces the norms from which it derives.48

In the secondary contexts of enunciation for the story, notably in Aristotle and Aeschines, the story still has a moral, but the emphasis has shifted to the father of Leimônê. Further, in the new contexts, the moral varies according to the purpose of the narrator. The one who retells the story provides, in effect, an interpretation. Aristotle says that Hippomenes, one of the Kodridai, wanted to remove the charge of softness from his genos, which had already lost the kingship of Athens. Nicolaus of Damascus says that Hippomenes was expelled from his rule because of the harshness of the punishment of his daughter, and several other sources follow this interpretation. In short, the punishment has a different function in the career of Hippomenes according to the narrator.

But the narrative is perhaps of greater interest than the moral. I mean, in particular, the narrative embodiment of equine symbolism, which determines all of the relations and all of the roles. The horse that the aition had to explain served for the name and the character of the father. Leimônê’s role is that of filly. The seducer is punished by means of a horse. Leimônê is punished by means of a horse. The narrative as it appears in this perspective is remarkable tautologous: it constantly repeats "horse."

Here I return briefly to the Aristotelian source for the story. In this source, there occurs the sentence: "When he caught a seducer on his daughter Leimone, he killed him by yoking him to the chariot [with his daughter], and he shut her up with a horse until she died." Editors have usually bracketed the phrase that I have bracketed in English, for obvious reasons. But, as Carlo Brillante pointed out to me, why not a double punishment? The father’s actions could, in fact, be regarded as his "interpretation," in terms of equine symbolism, of the young couple’s transgression. First, instead of their "yoking" in the legitimate marriage to which the daughter is no longer entitled, they are literally yoked to a chariot. There is the further bitter irony in the father’s gesture that, in the Athenian wedding, the central event was the procession in which the bride and the bridegroom were carried in a cart from the bride’s father’s house to the bridegroom’s. 49 Hippomenes has arranged a taunting, non-wedding procession in which Leimônê and her lover play the roles of the beasts. Then, one has to imagine, Leimônê is released from the yoke, while her lover is left to die. She has been spared, however, only for Hippomenes’ second punishment. In her new home, inside which is her proper place,50 she will have for a "husband" not a man but a horse, bad "filly" that she is. Again, Hippomenes shows a bitter irony in his literalizing of the horse metaphor.

I don’t know if "the tragic wedding" (maiden marries death) is relevant. For references, see Wilkins on Eur. Hcld. 591-92 (pp. 126-27).

I have deliberately used the word "symbolism" to refer to this fundamental element of the story. I have tried to imagine how a particular symbol could, if not generate, inform a narrative. In this respect, my method is close to the one employed by Scheid and Svenbro in their study of weaving.51 They showed that weaving is not simply a metaphor but a kind of common knowledge, a figure of thought shared by an entire civilization; further, that it engenders myth, and not only myth but also ritual and iconography. It is this generative aspect of a figure of thought that I have tried to show in the case of the horse.

My notion of symbolism is, however, not exactly the same as that of Scheid and Svenbro, who do not, in fact, use the term "symbol." I have followed the anthropologist Victor Turner in regarding the foremost property of the symbol as condensation. "Many things and actions are represented in a single formation," Turner said.52 Because of its power to condense, the symbol can bring together disparate and even contradictory meanings. Turner said of one of the symbols that he studied in the rituals of the Ndembu in northwestern Zambia: "We can see how the same dominant symbol, which in one kind of ritual stands for one kind of social group or for one principle of organization, in another kind of ritual stands for another kind of group or principle, and in its aggregate of meanings stands for unity and continuity of the widest Ndembu society, embracing its contradictions."53 Neither in Ndembu nor elsewhere is a symbol in this sense confined to ritual. It is pervasive. As a symbol in ancient Greek society, Athenian or other, the horse has this same capacity to embrace contradictions.

"The Horse and the Maiden" is a relatively short narrative and also a bizarre one. I have argued that it is not to be understood in relation to the few Greek myths to which it appears to have some affinity. Still less is it to be understood in relation to the supposed Indo-European horse sacrifice or Indo-European myths concerning horses.54 It is probably closer to twentieth-century horror films than it is to Indo-European mythology.55 In any case, "The Horse and the Maiden" undoubtedly served its purpose as an aition. The symbolism, well known to its original audience, made it easy to understand. For us, the situation is different. As André Jolles said in his introduction to Einfache Formen, the simple narrative forms are the easiest to read, and the hardest to understand.