Main Page



Gender Transgression in Ancient Religion


Related Materials

“The radically ‘de-oedipalized’ body of the priest of the goddess, in ways mysterious to us, is bound up with techniques of ecstasy no less historically tenacious than the weight brought to bear against it by the patriarchal Judeo-Christian tradition. The defiant presence of this figure in the midst of prevailing phallocentrism remains striking and unexpected. Today, long after the last temple of Cybele fell into ruin, we are discovering that the boundaries of gender are no less friable, and that the human body, which has been so deeply inscribed with the cultural construction of its meaning as to seem for all purposes what it is represented to be—natural and fixed—may yet be reinscribed with other meanings and other constructions."

Goddess of Catul Huyuk

The earliest known depiction of a goddess with the attributes later associated with Cybele—seated in a throne and flanked by lions. From the early Neolithic site at Catal Hüyük, 5th millennium BCE

The following is an excerpt from the introduction and conclusion of my article on priests of the goddess in the Old World, from the devotees of Inanna/Ishtar in Sumer, Assyria, and Babylonia to the followers of Cybele and Attis in Roman times and the hijra of contemporary and ancient India. The full article with notes can be found in History of Religions 35(3) (1996): 295-330.



Phyrgia, in modern day Turkey, was the homeland of Cybele and Attis, and their galli priests.

In the competition between Christian and pagan in the ancient world neither side hesitated to broadcast the most outrageous and shocking accusations against its opponents in the most inflammatory rhetoric it could muster. “In their very temples,” wrote Firmicus Maternus in the mid-fourth century, “can be seen deplorable mockery before a moaning crowd, men taking the part of women, revealing with boastful ostentation this ignominy of impure and unchaste bodies (impuri et impudici). They broadcast their crimes and confess with superlative delight the stain of their polluted bodies (contaminati corporis)” (De errore profanarum religionum 4.2). These infamous men, with their impure, unchaste, polluted bodies, were none other than the galli, priests of the gods Cybele and Attis, whose mystery religion constituted one of early Christianity’s major rivals. Time and again, Christian apologists cited the galli as representative of all they abhorred in pagan culture and religion. And of all the outrages of the galli, none horrified them more than the radical manner in which they transgressed the boundaries of gender.

“They wear effeminately nursed hair,” continued Firmicus Maternus,“and dress in soft clothes. They can barely hold their heads up on their limp necks. Then, having made themselves alien to masculinity, swept up by playing flutes, they call their Goddess to fill them with an unholy spirit so as to seemingly predict the future to idle men. What sort of monstrous and unnatural thing is this?” A century later, Saint Augustine found the galli no less shocking: “Even till yesterday, with dripping hair and painted faces, with flowing limbs and feminine walk, they passed through the streets and alleys of Carthage, exacting from merchants that by which they might shamefully live” (De civitate Dei 7.26).


Malibu Cybele

Marble statue of Cybele as she was typically depicted in the Roman era, 50-60 CE



The Mesopotamiam goddess Inanna/Ishtar. In mythological accounts, Inanna is rescued from the underworld by two beings described as "neither male nor female." Various classes of priests in Sumerian and Assyrian religion occupied alternative gender roles distinct from those of men and women.

It would be easy to dismiss the numerous references to galli in ancient literature, both Christian and pagan, as exoticisms equivalent to today’s fascination with gender transgression as evidenced by such films as M. Butterfly and The Crying Game. Unlike the modern figure of the transvestite, however, galli were part of an official Roman state religion with manifestations in every part of the Greco-Roman world and at every level of society. One finds the Roman elite worshiping Cybele with bloody animal sacrifices officiated by state-appointed archigalli; common freedman and plebians forming fraternal associations, such as the dendrophori and canophori, to perform various roles in her annual festivals; and the poor and slaves swept up by the frenzy of her rites, often to the consternation and alarm of their social superiors.

It is the widespread dispersion and great historical depth of the Cybele and Attis cult, as well as its appeal to multiple levels of ancient Mediterranean societies, that make its study fascinating on its own, not to mention its relevance to current debates concerning the social construction of sexuality and gender. The galli become even more interesting, however, when placed next to evidence of similar patterns of religious gender transgression from the Near East and south Asia, which suggests that goddess-oriented cults and priests are part of an ancient cultural legacy of the broad world-historical region Marshall Hodgson referred to as the “Oikoumene.”

In the discussion that follows, I will focus on three of the better-documented cases of goddess-centered priesthoods: the Greco-Roman galli, the priests of the goddess called Inanna in Sumeria and Ishtar in Akkad, and the hijra of contemporary India and Pakistan. The parallels between these priesthoods and the social roles and identities of their personnel are detailed and striking. Without ruling out dispersion as a factor, I will argue that these priesthoods are largely independent inventions whose shared features reflect commonalties in the social dynamics of the societies in which they arose, specifically, the agrarian city-state. The presence of goddess-centered priesthoods in the regions where the urban lifestyle first developed raises unexpected and challenging questions concerning the role of gender diversity in the origins of civilization....

Galli: tertium sexus

[See full article]

Hijra: neither man nor woman

[See full article]

Gala et al.: penis+anus

[See full article]

Social origins and social meanings of gender transgression

At the time of the birth of Christ, cults of men devoted to a goddess flourished throughout the broad region extending from the Mediterranean to south Asia. While galli were missionizing the Roman Empire, kalû, kurgarrû, and assinnu continued to carry out ancient rites in the temples of Mesopotamia, and the third-gender predecessors of the hijra were clearly evident. To complete the picture we should also mention the eunuch priests of Artemis at Ephesus; the western Semitic qedeshim, the male “temple prostitutes” known from the Hebrew Bible and Ugaritic texts of the late second millennium; and the keleb, priests of Astarte at Kition and elsewhere. Beyond India, modern ethnographic literature documents gender variant shaman-priests throughout southeast Asia, Borneo, and Sulawesi. All these roles share the traits of devotion to a goddess, gender transgression and homosexuality, ecstatic ritual techniques (for healing, in the case of galli and Mesopotamian priests, and fertility in the case of hijra), and actual (or symbolic) castration. Most, at some point in their history, were based in temples and, therefore, part of the religious-economic administration of their respective city-states.

The goddesses who stand at the head of these cults—Cybele, Bahuchara Mata, and Inanna/Ishtar—also share important traits. All three are credited with the power to inspire divine madness, which can include the transformation of gender. Their mythologies clearly place them outside the patriarchal domestic sphere: Cybele roams the mountains with her wild devotees; Inanna/Ishtar is the patron of the battlefield; and Bahuchara Mata becomes deified while on a journey between cities (see synopses of myths in table 2). Indeed, all three transgress patriarchal roles and structures just as much as their male followers: Cybele begets a child out of wedlock, which infuriates her father; Ishtar, the goddess of sexuality, is notoriously promiscuous, never marries, and, indeed, is herself a transvestite; and Bahuchara Mata, at the other extreme, cuts off her breasts in an act of asceticism to avoid unwanted heterosexual contact. The influence of these goddesses over human affairs is often as destructive as it is beneficial. “To destroy, to build up, to tear out and to settle are yours, Inanna,” reads one Sumerian text, and in the next line, “To turn a man into a woman and a woman into a man are yours, Inanna.” Despite the common reference to these goddesses as “mother” by their worshippers, there is much in their nature that exceeds and confounds our present-day connotations of the maternal.

How can we account for such a consistent pattern over such a broad area and time span? Without ruling out diffusion as a factor—the spread of Cybele and Attis was due in part to missionizing by galli themselves, and the influence of Mesopotamian religion certainly reached Syria and Anatolia-simple cultural exchange nonetheless seems the least likely explanation. A more promising approach would be to address three interrelated questions: What were the belief systems of the societies in which these priesthoods existed, in particular, beliefs concerning sex, gender, and sexuality? What was the nature of the social systems in which these roles originated? What was the source of their long-term popular appeal?

The eclectic approach implied by these questions-encompassing cultural, social and psychological analysis-is key to understanding cultural phenomena as social constructions. When we refuse to regard femininity, masculinity, heterosexuality, homosexuality, and social inequality in general as precultural givens, we necessarily make our task as historians and social theorists more complicated, for cultural facts are always multiply determined and their explication requires analysis of the social wholes in which they occur. The goal must be a unified analysis, one that integrates the synchronic viewpoint of culture afforded by anthropology with the diachronic perspective of historical study. In the case of the ancient priesthoods of the goddess described here, such an approach reveals their roles to be not accidental but, indeed, consistent features of the societies in which they flourished.

To begin with, in all three cultural regions, goddess-inspired priests were conceptualized as occupying a distinct gender category. As we have seen, hijra routinely refer to themselves as “neither men nor women,” consistent with the ancient Sanskrit designation trtiya prakrti. (The galli, as we saw, were also described as a tertium genus.) Similarly, the Sumerian myth called “The Creation of Man” (ca. 2000 B.C.E.) relates how Ninmah fashioned seven types of physically challenged persons, including “the woman who cannot give birth” and “the one who has no male organ, no female organ.” Enki finds each one an occupation and position in society-the sexless one “stands before the king,” while the barren woman is the prototype for the naditum priestesses. These proceedings are echoed in the Akkadian myth of Atrahasis (Atra-hasis) (ca. 1700 B.C.E.), where Enki instructs the Lady of Birth (Nintu) to establish a “third (category) among the people,” which includes barren women, a demon who seizes babies from their mothers, and priestesses who are barred from childbearing (3.7.1).



The worship of Cybele was originally centered in Phrygia (central Turkey), where she was known Kubaba or Kybele. The Romans formally adopted her worship in 204 BCE, when they brought a statue representing her from her main shrine in the Phrygian city of Pergamum back to Rome. This statue, from a site in Anatolia dating to the eighth or early seventh century BCE, depicts the Phrygian goddess with two youthful attendents playing a flute and harp.




Attis, the Phrygian shepherd, whose worship became part of the cult of Cybele. In varying mythological accounts, Attis is killed or is driven insane and castrates himself, as the result of jealousy and passions arising from an ill-fated love affair. Attis' fate served as the model for the galli priests, who underwent castration to become Cybele's dedicated and chaste servants.





Marble relief of a gallus, or priest of Cybele, with various ritual objects, 2nd cen



Marble relief depicting an archigallus, or head galli priest, 3rd century CE


Cybele Temple

Remains of the temple of Cybele on the Palatine Hill in Rome



Contemporary hijra in India

Clearly, the underlying conceptualization of gender implied by these taxonomies is at variance with the idea that physical sex is fixed, marked by genitalia, and binary. Recent reviews of Greek and Roman medical texts, for example, reveal a notion of gender as grounded in physiology, but the physiology involved is inherently unstable. Masculinity and femininity depend on relative levels of heat and cold in the body (and, secondarily, moisture and dryness). These factors determine the sex of developing fetuses, but even after birth an individual’s gender status was subject to fluctuations in bodily heat. If men were not at risk of literally becoming females, they were in danger of being feminized by any number of causes. A similar hydraulic construction of the body, as Wendy Doniger has termed it, is evident in Hindu belief as well.

The frequent references to priests of the goddess as “eunuchs” or “impotent” males points to another important commonality in the ancient construction of male and female genders. A little known episode in Roman legal history is especially revealing in this regard. In 77 B.C.E., a slave named Genucius, a priest of Cybele, attempted to take possession of goods left him in a will by a freedman, but this was disallowed by the authorities on the grounds that he had voluntarily mutilated himself (amputatis sui ipsius) and could not be counted among either women or men (neque virorum neque mulierum numero) (Valerius Maximus, 7.7.6). Presumably, only women and men qualified to exercise inheritance rights, and this privilege of their gender identity was, in turn, a function of their ability to reproduce. This seemingly minor case nonetheless underscores the way in which gender identity and citizenship were linked in societies of the Oikoumene region-that is, in patriarchal, agrarian city-states. Gender, to borrow Judith Butler’s terminology, was performative, or rather, to be even more specific, productive. Gender identity hinged not on the degree of one's masculinity or femininity, the direction of one’s sexual orientation, nor even one’s role in the gendered division of labor but on one’s ability to produce children, in particular males. In a patrilineal kinship system, it is the labor of male children on which the paterfamilias has the greatest claim. As anthropological research has shown, peasants around the world typically seek to improve their lot in life by having more children and thereby increasing the supply of labor for family-based production. Having male children is the central imperative of gender, as a social category, a role, and a personal identity in most patriarchal agrarian societies.

From this perspective, males or females who are unable to reproduce, who are impotent, whether for physiological or psychological reasons, or who lack or forswear heterosexual desire, including those who desire the same sex, all fail to qualify for adult male or female gender identity. Being neither, they tend instead to be categorized together as members of an alternative gender or of subdivisions of male and female genders. Like male and female, these roles are also attributed specific traits, skills, and occupations. In the same way that men’s activities are “male” and women’s are “female,” what galli, hijra or gala do comes to be seen as intrinsic to their alternative gender identities. At the same time, the distinctions we might make between impotency, asceticism, castration, homosexuality, and congenital sexual anomolies are generally overlooked or ignored in these taxonomies. The Narada Smrti, for example, an Indian law book from the fourth or fifth century C.E., defines fourteen categories of “impotent men,” including those “naturally impotent,” men who have been castrated, those cursed by a supernatural, those afflicted by jealousy, those who spill their seed, and those who are shy (12.11-19).

It should be clear by now that the concept of impotency had very different connotations in the societies discussed here than it has today. Here evidence from modern hijra is especially helpful. Hijra are habitually described (and describe themselves) as impotent, yet they are frequently and admittedly sexually active with men. As one hijra explained to Nanda, “When we look at women, we don't have any desire for them. When we see men, we like them.” This, and reports of the impotence of hijra initiates being tested by having them sleep with prostitutes, suggest that the contextual meaning of “impotence” is “absence of heterosexual desire.” (Indeed, any nonmasculine male or passive homosexual in present-day Gujurat is likely to be referred to as a hijra, and his behavior credited to the spell of the goddess Bahuchara.) But while hijra freely refer to their homosexual desires, the closest one gets to the representation of this in myths and tales is a kind of negative representation in which protagonists are portrayed as merely lacking heterosexual traits. It takes quite a different set of beliefs and construction of gender, as Foucault and others have argued, for identities based on sexual object choice without reference to gender—the modern taxonomy of heterosexual/homosexual-to emerge.

What are the circumstances in which gender categories are multiplied? Here the study of alternative genders in native North America provides a useful point of comparison. The role of the so-called berdache, a third gender status occupied by both males and females, has been well documented in tribes throughout the continent. Berdaches (or “two-spirits,” as many contemporary Indians prefer to call them) were not only associated with specific religious functions, they filled clear-cut economic specializations as well. Among the Lakota of the Northern Plains, for example, men became berdaches, or wintke, as a result of dreams or visions of Double Woman, a powerful and ambivalent goddess with the ability to inspire both madness and creativity. As a result of their contact with Double Woman, winkte acquired not only religious powers, such as the ability to foretell the future; they also became accomplished artisans, famous for their excellence in crafts normally pursued by women. This association between religious and economic specialization occurs consistently across tribes. Among the Pueblo Indians of the Southwest, who eschewed individual contact with the supernatural in favor of collective religious forms, berdaches joined the ranks of priests, carefully trained experts in mythological and ritual knowledge. In fact, in certain populous tribes of Florida and the upper Missouri River, berdaches appear to have formed a kind of nascent priesthood.


The level of production and social complexity of even the earliest Sumerian communities for which we have definite evidence of gala excedes that of the American Pueblos in the historic period. Rather, North American data suggest a model for how gender diversity might have originated in Mesopotamia and elsewhere at a point in time in prehistory prior to the appearance of full-scale city-states. Native North American communities, in most cases, lacked fixed social stratification, and their division of labor was based predominantly on gender. In social systems like these, the easiest way to accommodate a growth in specialization would be to simply multiply existing gender categories. Viewed from this perspective, North American berdaches were among the first specialists of their societies, those few individuals in a subsistence economy afforded the opportunity to focus on activities not directly related to food production.

In sum, both ancient and North American cultures shared beliefs concerning the malleability of physical sex and the possibility of gender transformation. The alternative gender roles we find in both areas differ most in their economic dimension. Instead of production (i.e., arts, crafts, domestic work), the original economic basis of ancient priesthoods of the goddess appears to have been administration. Temples throughout the Oikumene region were economic and cultural centers for their cities, even when secular institutions and leaders overshadowed (or absorbed) their political functions. The rituals and other duties carried out by the priests of these temples served to synchronize the political, religious, and economic life of both city and surrounding countryside. That an association between gender difference and productive specialization appears to be missing in the ancient world may reflect the lower status of women's work in patrilineal agrarian societies compared with the horticultural and hunting and gathering societies typical of North America. In the Oikoumene, class and caste superseded gender as organizing categories of labor and production. Consequently, at least in the period documented by written records, the religious dimension of the priests of the goddess is the dominant feature of their social role. This religious status was not lost even as the economic viability of the temples they once staffed declined (in late antiquity in the Mediterranean world; in the nineteenth century in south and southeast Asia). As long as the goddess survives, it would seem, so do her followers, who sacrifice male gender identity to become her devotees, even if begging and prostitution are the only economic niches left to them.

If the origins of religious gender transgression in the ancient world lie in a pre- or proto-urban, gender-based division of labor, these alternative genders survived long past the time when more complex forms of social stratification developed-as they did in those North American societies, in the southeast and on the northwest coast, where social ranking systems had developed. Berdaches were still present in these societies, often identified with the higher social ranks. To account for the continued existence of religious gender transgression requires the use of a different kind of data and a different mode of analysis. Longevity is in large measure a function of the emotional and aesthetic appeal of a religion, and this is best revealed in its art, its rites, and especially its mythology. For this purpose we need only posit the broad outlines of a psychological theory of myths, by recognizing their popularity as (in part) a function of the emotionally charged issues, conscious and unconscious, in the lives of their audiences that they portray.

Table 2 compares themes from etiological myths of galli and hijra, in the format developed by Lévi-Strauss for the purposes of structural analysis. Here the matrix is employed simply to reveal the overall pattern of parallels between these myths across their differing versions. As the table indicates, the common element of these stories is a central figure who is the subject of a violent attack, either physical or psychological, resulting in some form of mutilation, sometimes self-inflicted. Although the scenarios leading up to these attacks vary, they all involve a protagonist who thwarts patrilineal succession. Cybele, for example, abandoned at birth and raised by female shepherds, returns with an unauthorized pregnancy that provokes an attack from her father. Similarly, the southern Indian goddess, Yellama, who marries an ascetic against her parents’ wishes and then longs for sexual fulfillment, is attacked by her husband and sons. Both goddesses survive to become the founders of their respective religions (Type I in the table).

Attis and his Indian counterparts also disrupt patrilineal succession. The son of a king, he is born “impotent” and/or insists on following the mother goddess (Type II)-hardly appropriate behavior for a male heir. In most versions of his story he rejects heterosexual overtures (Type III). As Arnobius relates, he is safe only if he remains free of marriage (salvum quamdiu esset solutus a matrimonii foedere; Adversus nationes, 5.7). When Attis does enter heterosexual relations the results are disastrous. In all versions, he is subject to a violent attack or madness, precipitated sometimes by a paternal figure, sometimes by the jealous goddess.

The problem of possessive love is another key theme (Type III). The desiring subject, however, is most often female, and no account attributes incestuous motives to Attis, although this fact has been overlooked by almost everyone since Freud. Rather, Attis is the object of unwanted heterosexual overtures, caught between the social demands of the paterfamilias and the emotional demands of a mother figure who is herself caught up in the dynamics of patrilineal sexuality and marriage. I believe this is the aspect of these stories that most captured the popular imagination by speaking to emotional and interpersonal conflicts endemic to patriarchal cultures. Of course, Christian cultures are also patriarchal and patrilineal, but in polytheism gender ambiguity is given a different valuation, and sexual tensions can be freely projected onto female deities. It is the combination of these two factors-patrilineal social order and polytheistic religion-that creates the ground for the long-term appeal of goddess figures and their priests.

Cybele and Attis

Cybele and Attis on an altar from the Phrygianum, the sanctuary dedicated to their worship in Rome, late 3rd century CE.
The basilica of St. Peters in the Vatican was constructed on top of this site.

What do galli and hijra represent in all this? In cultures characterized by what Eva Keuls aptly terms the “reign of the phallus,” self-emasculation is a fantasy of escape from irreconcilable tensions by rendering oneself incapable of fulfilling either the social or sexual demands of patriarchal male roles. From the perspective of a protagonist who experiences these demands as unwanted, the results are truly nirvan, as the hijra would say, liberation. (No one, on the other hand, has claimed that Oedipus found his blindness illuminating.) The underlying hostility of this act underscores the transgressive nature of being nonmasculine and nonreproductive in a patriarchal culture. As we have seen, possession of a working phallus was identified with masculine agency and citizenship in agrarian city-states throughout the Oikoumene. Self-castration in these societies was a powerful statement of indifference toward the transcendental signifier of masculine status and agency, a statement repeated by every depiction of Attis with his robe flying open and by every hijra who lifts “her” skirts on the streets of Bombay to this day.

But there is a second, inner meaning here as well, hinted at in the story of Siva at the beginning of the world. Brahma and Visn.u had asked Siva to create the world. Siva agreed but plunged himself into the water for a thousand years to meditate. Impatient, Visnu gave Brahma the female power of creation, and Brahma created the gods and other beings. When Siva finally emerged, he found the universe already filled, so he broke off his phallus and threw it to the earth, saying, “There is no use for this linga except to create creatures.” This does not render him asexual, however. Rather, the circulation of his phallus through the cult of the linga extends his sexual power to the universe. In a similar way, the flow of blood during the castration of hijra is believed to empty them of both desire and gender. Because of this, they can serve as vehicles for the transmission of the sexual energy of the goddess-they become her phallus, as it were, erotic ascetics. Thus, the radically &ldquode-oedipalized” body of the priest of the goddess, in ways mysterious to us, is bound up with techniques of ecstasy no less historically tenacious than the weight brought to bear against gender diversity by patriarchal and Judeo-Christian traditions. The defiant presence of this figure in the midst of prevailing phallocentrism remains striking and unexpected. Today, long after the last temple of Cybele has fallen into ruin, we are discovering that the boundaries of gender are no less friable, and the human body, which has seemed so deeply inscribed with the cultural construction of its meaning as to be for all purposes what it is represented to be—natural and fixed—may be reinscribed with other meanings and other constructions.