ISSUE #005 - Aug 19, 2018

Ḍākinī, Yoginī, Pairikā, Strix: Adventures in Comparative Demonology

David Gordon White

A cluster of mythemes found in sacred and secular works from eighth- to eleventh- century Kashmir involve human or superhuman “witches” that prey on their male victims by extracting their vital organs or fluids from their living bodies. These are also attested down to the present day in many parts of South Asia. Nearly identical accounts, found in Buddhist tantric works from as early as the eighth century, followed the spread of Buddhist Tantra across all of East Asia in the centuries that followed. It is intriguing to note that Persian traditions of similar creatures predate these South and East Asian sources by several centuries; and that cognate Roman traditions are older still. In this article, the author proposes a hypothetical reconstruction of cultural exchanges that led to the spread of these mythemes, from Rome and Persia to Kashmir, India, China and Japan.

The Barber’s Tale

In 1989, the British director Peter Greenaway brought to the screen The Cook, the Thief, His Wife & Her Lover, a film noted for its brutality, raw sexuality, cannibalism, and scatological themes. The Cook’s story line centers on a creative act of revenge taken by the weak against the strong, which is perhaps the sole redeeming feature of an otherwise repulsive production. Some nine hundred years prior to Greenaway’s heavy- handed effort, a storyteller from Kashmir wove a similar tale with many of the same dark themes, but was able to do so with such a lightness of touch as to transform a revenge narrative into a comedy. This was Somadeva, who incorporated into his ca. 1070 CE anthology, the Kathāsaritsāgara (Ocean of Rivers of Story), the following tale of a barber (a person near the bottom of India’s ancient caste hierarchy), his wife, and his king. For reasons that will become clear, the story is told through the mouth of the barber himself:

Dṛḍhavarman, our [present] king, had a father of dissolute morals. I was that king’s slave and so carried out the duties appropriate to my status. One day, while he was going for a stroll in this neighborhood, he espied my wife. She was young and beautiful and he was attracted to her. He asked the people around him who she was, and they told him that she was the wife of a barber. Thinking to himself that a barber could do him no harm, he came into my house. Then after having had his way with my wife, that cursed monarch returned to his palace. As fate would have it, I was away that day, but when I arrived home the following morning I found my wife altered, and so I asked her what the matter was. With no small amount of pride, she related the events of the previous day just as they had transpired. And so it was that the king continued to increase my wife’s prestige by taking his pleasure with her—and me unable to do a thing about it. (When a debauched king is in rut, he can no more distinguish between licit and illicit love than a wind-driven forest fire can tell the difference between a verdant stand of trees and piece of straw.) Seeing no other way of dealing with the matter, I began to eat very little so as to emaciate my body and appear enfeebled. In that condition, pale, skeletal and out of breath, I presented myself to the king to offer him the services of my profession. Noting my weakness, he began to question me on how I had fallen into such a state, and when he wouldn’t stop, I threw myself at his feet, declaring, “Master, my wife is a witch (ḍākinī)! When I am asleep, she pulls my innards out through my arse! After that, she sucks them dry—and then she chucks them back inside! That’s why I’m all dried out! Even with a rich and steady diet, how could I possibly be plump?”

My words planted the seed of doubt in the king’s mind, and he thought, “Could she possibly be a witch? Have I fallen under her power? I who stuff myself with food, has she also been sucking on my guts? Tonight I will carry out a plan and so test her myself!” Upon which the king had me serve him his supper.

Upon returning home, I burst out in tears in front of my wife and, when she asked me the reason, I replied with the words “My darling! Listen to what I have to say, but promise you will reveal it to no one! The king has teeth, sharp as diamonds, growing out of his arse! Just today, while I was performing my duty, they broke my razor (which is, truth be told, of excellent quality). From now on, my razor is going to shatter every time I’m down there. How am I ever going to come up with a new razor every day? The reason I’m crying is that my income has fallen to zero!”

My words inspired my wife to make the most of that very night, so that once the king who would be visiting her had fallen asleep, she could have a peek at his marvelous anal dentition. She never dreamed that something that had never been seen since the dawn of time could be a subterfuge: even the craftiest of women will let herself be taken in by the words of a sweet talker. That night, the king came and enjoyed my wife, and then, recalling my words, feigned sleep as if he were exhausted. Thinking he was asleep, my wife softly stretched her fingers out to his arse to touch its teeth. No sooner had my wife’s hand brushed his buttocks than did the king jump up screaming, “A witch! A witch!” And ran away in terror.

Since that time, my wife (whom the terrified king has abandoned) has been satisfied with her fate in life and has had eyes for none but me. That is how, thanks to my resourcefulness, I was once able to free my wife from the clutches of the king. (Kathāsaritsāgara 6.6.145-171, in Sastri 1970: 143- 44)

Plainly, having a witch for a wife was viewed as a dangerous vocation in medieval Kashmir. This is a notion that has persisted down to the present day in South Asia: ḍākinīs are able to suck the life out of their husbands (and others), causing them to waste away. Somadeva’s story also provides an unexpected explanation for the barber’s low caste status in traditional Indian society. It is not simply because barbers deal with dead, and thereby polluting, bodily matter—the hair and whiskers that they cut—that their occupation is considered impure. It is also because not all the hair that they cut is located on the heads or faces of their clients. This is a point that social and cultural anthropologists of India appear to have missed.

The anthropologist Lawrence Babb has noted, however, that “the wives of barbers are also thought to be prone to witchcraft. Barbers’ wives often assist other women in dressing and with other details during weddings, and in so doing may see them in a state of partial undress. Exposure of this sort is thought to leave one especially vulnerable to ‘the glance’, i.e., the ‘evil eye’, (nazar) of witches” (Babb 1975: 205-6). Babb also recounts several modern-day accounts of witches, which resonate uncannily with the claims made by the barber in Somadeva’s story. One such story, told to him in the 1960s, involved a newlywed groom from Neora, a village in the Chattisgarh district of central India.

A newly married couple in Neora happened to walk by the house of a witch one night. The witch saw them and became angry. Knowing nothing of this, the couple went home to bed. Later that night, the witch came in the form of a cat. She climbed on the roof of the house and hung a string down through a hole so that the end touched the young man. She then began to suck his blood up through the string. The young man woke up, saw the string, and immediately realized what was happening. He brought in a pot of water and put the end of the string in it, and, as he watched, the level of the water went steadily down. The pot finally became empty. When the witch’s stomach was filled with water she returned to her human form. Then she fell down and died. (Babb 1975: 201)

Witches who fed upon the vital fluids of their victims were a commonplace of Somadeva’s medieval life world. Called yoginīs, śākinīs, śabarīs, or ḍākinīs, they were known to suck the life out of their victims as a means to both fueling their power of flight and their ability to change their outward form into that of a bird, mammal, or other creature. What were the sources of these traditions? In Somadeva’s time, Kashmir was the epicenter of Hindu Tantra, a ritual system whose imagery and ideology permeated every aspect of religious life, in much the same way that it continues to do in the modern-day Kathmandu Valley. A survey of certain of these tantric traditions is therefore in order. A description found in the circa eighth-century Kashmirian Tantrasadbhāva (16.163-64) and quoted in Kṣemarāja’s eleventh-century commentary on another Kashmirian work, the Netra Tantra (19.71, in Sastri 1939: 153), encapsulates the modus operandi of the tantric witches: “A female being who, for the purpose of shape-shifting, continually drinks the fluids of living beings after attracting (ākṛṣya) [them] by witchcraft (chalena)—and who after obtaining [their fluids] slaughters [her] victims—shall be known as a śākinī, ever delighting in dreadful places.” Two verses later, the Tantrasadbhāva describes another class of female beings, the śabarīs, who steal away (haranti) the “five nectars” (pañcāmṛtam) of men, by virtue of which they are able to “roam around the world in a moment” and change their form over and over again. The Rasārṇavam (18.102-6), an eleventh-century alchemical text likely written in Gujarat or Maharashtra, describes a similar practice on the part of śakinīs and a host of other airborne (khecarī) and land-based (bhūcarī) “goddesses” who consume the refined mercury that has accumulated in an alchemist’s seminal ducts. Here, it is said of them that they steal away the sleeping alchemist’s mercury-enhanced semen and his life’s blood, causing him to ejaculate the useless dross of his seed at the end of his sleep. The toll they take on non-alchemists is heavier still: they eat their flesh and bones while they are sleeping. For all this, their parasitism does not kill their victims, at least not immediately. Such is, after all, the strategy of a successful parasite.

In nearly all of these traditions, these witches are said to draw or suck out the vital fluids of their victims. The Sanskrit verbal form (ākarṣaṇa) that is used to denote this practice is also applied more generally to the type of sorcery known as “attraction”: male and female sorcerers, black magicians and tantric practitioners are able to draw their victims toward themselves, against their will and sometimes across great distances, by wielding this occult power (White 2003: 209). It is the same term that the barber uses when he recounts how his wife “pulled his innards out through his arse.”

When one turns to the Buddhist canon from the same period, one finds parallel traditions concerning the ḍākinīs. Here, a 727 CE Chinese- language commentary on the Sanskrit Mahāvairocana Sūtra, written by a monk named Yixing—in collaboration with the eminent Indian Buddhist master Śubharakṣasiṃha, who was the translator of the work itself—provides an additional element of medieval ḍākinī lore. According to this work, Mahākāla, the leader of the ḍākinī horde, is said to assert his power over the ḍākinīs by eating them, because they constantly eat the “hearts” of men who are destined to die in the course of the following six months. (Nobumi 1999: 51):

The ḍākinīs are capable of recognizing those men who are about to die: six months [in advance] they recognize them, and having recognized them, employ the following method: they take their heart and eat it . . . Those [ḍākinīs] who eat hearts attain the greatest supernatural powers (siddhis). They can [fly] to the Four Regions in a single day and obtain their every desire. Furthermore, they can control men in a variety of ways: they can overcome those who despise them by making them suffer the most horrific diseases. [However, the ḍākinīs] are unable to kill men in this way . . . [B]y recognizing those men who are going to die in the following six months, they steal away their hearts by means of magic. Once they have removed the heart, they have to replace it with something else, [which is why] these men do not die [immediately] . . . As a rule, these are fully realized yakṣiṇīs . . . (Nobumi 1999: 51-52)

A fourteenth-century Japanese work by the Tendai monk Chōgō reprises the themes found Yixing’s commentary, stipulating that the ḍākinīs eat human hearts in order to extract the “yellow of men,” which is said to exist in the form of “seven grains of jade in the form of dewdrops” located in the heart (Nobumi 1999:54, 99). This is a likely reference to the Hindu ojas, which the Ayurvedic Caraka Saṃhitā (1.17.74-75; 1.30.7-8, in Trikamji Acarya 1992: 103, 184-85) identifies as the essence of life itself, of which eight drops are found in the heart. An Indo-Tibetan Buddhist source, the eighth-to-tenth-century Cakrasaṃvara Tantra, transfers the powers of the ḍākinīs to those of the male practitioner, who, by repeating the “essence of a secret mantra, is empowered to “travel ten million leagues together with a ḍākinī,” change himself into mammalian and avian forms, steal away the organs of perception of others, and “draw out the blood of whomever one wishes . . .” (Gray 2007: 204-5).

The compilers of the Hindu and Buddhist Tantras were doing more than simply cataloguing this witchcraft lore. Rather, they enshrined it at the heart their practice. Here, we must recall the soteriological function that the tantric yoginīs played for the male tantric practitioner. The Hindu Tantras frequently evoke the two alternatives faced by males with regard to these beings: while most are doomed to become “food for the yoginīs,” the courageous tantric initiate known as a Virile Hero (vīra) could instead become the “darling of the yoginīs.” The Netra Tantra (20.16-21) goes so far as to etymologize the yoginīs’ name, declaring that they “yoke” (yojayanti, derived from the same verb yuj as “yoginī”) their victims to liberation from the karmic effects of their sinful acts by devouring their bodies, even as it asserts that the bodily destruction they inflict is not tantamount to killing. This path to liberation, called “supreme yoga” in the Netra Tantra, is more commonly referred to in the Hindu Tantras as “mingling” (melāpa) with these predatory female beings.

What was the attraction of becoming food for the yoginīs? On the one hand, this was a path to siddhis and self-deification. The eighth- century Brahmayāmala (99.13, in Hatley 2009: 422) evokes the violent (haṭha) type of mingling (melāpa) with the ḍākinī hordes, in which the practitioner actually dies. However, “because of [his] slaughter by means of the [esoteric] knowledge of the śaktis (?), by [their] power this person too, after giving up his body, again attains the state of belonging to the clans [of goddesses], and is reborn knowing his [past] birth.” In a tradition in which gnosis was the supreme path to self-deification, becoming food for the yoginīs as a preliminary to resurrection through the powers of the same would have been a desirable trade-off. One finds similar considerations in Buddhist traditions, most notably that of Indo- Tibetan Buddhism, in which the ravening bird- and animal-headed ḍākinīs play this precise role: by tearing his body apart, they cause the light of gnosis to dawn in the advanced practitioner’s consciousness. This is also the principle underlying the gcod practice, which, according to Tibetan sources, originated in India in the eleventh to twelfth century. Here, gcod refers to the “cutting off” of mental afflictions that block the path to enlightenment. Practiced by tantric virtuosi (siddhas) in cremation grounds, the ritual begins with the gcod pa dancing together with a coterie of women embodying the ḍākinīs. (Orofino 2000: 398, 402- 4) Following this, the gcod pa visualizes “the transformation of his consciousness into a ḍākinī, who cuts the head off his physical body and then places the corpse inside the cranium . . . The corpse is subsequently transformed in accordance with the wishes of the guests [the ḍākinīs, gods and spirits of the cremation ground] . . . and offered to them as a gift” (Orofino 2000: 405).

Elsewhere, the Keiranshūyō-shū—a Japanese exegesis of the Mahāvairocana Sūtra myth of Mahākāla and the ḍākinīs authored by the Tendai monk Kōjū—speaks of a certain Matarajin, “who is none other than the god Mahākāla and who is also Dakini.” Here, Matarajin announces that

when [men enter] into their death throes, I will go to them and eat the liver of their corpses. By virtue of that, [men] will, in their death throes, be able to reach true understanding [thereby permitting them to be reborn in the Pure Land]. If I were not to eat their livers, they would not be able to reach true understanding. (Nobumi 1999: 68)

In another commentary, Kōjū’s contemporary Chōgō, identifies the “heart-liver” with the “yellow of men,” the essence of life (Nobumi 1999: 99-100). Clearly the terms “heart” and “liver” were not understood in the way that we do today, as discrete organs with specific functions; rather, the terms denoted the inner seat of life, as well as repositories of the kleśas that were blockages to the wisdom of enlightenment.

Scriptural references to the yoginīs, ḍākinīs and kindred beings follow the arc of the Hindu and Buddhist Tantras. We cannot know when their traditions first appeared in Kashmir and on the Indian subcontinent, and whether they originated in some other part of the world. Following the decline of Tantra in South Asia, their names gradually fade from the scriptural and epigraphic records, even as cults of the Dakini-ten remained vibrant in Japan; however, other types of literary sources attest to the persistence of these traditions in South Asia down through the late medieval and early modern periods. A striking example is the account given by the fourteenth-century North African traveler Ibn Battuta of the predations of a “Joki” in a north Indian town in the vicinity of Agra:

In the surroundings of the city there are many voracious animals. One of its inhabitants related to me that a lion used to break into the city in the night although the gates were closed and that he used to molest the people, so much that he killed many . . . One night the lion broke into a house and carried away a boy from his bed . . . Curiously enough, some one told me that he who did so was not the lion but a man of the magician class called “Joki” who assumed the form of a lion . . . Some of the Jokis are such that as soon as they look at a man the latter instantly falls dead. The common people say that in such a case—of a man being killed by a mere look—if his chest were cut open one could see no heart which, they say, is eaten up. Such is, for the most part, the practice with women, and the woman who acts in this manner is called a “hyena” (kaftār). (Ibn Battuta, Rehla 4.35 in Husain 1976: 163-64)

The literal meaning of the Perso-Arabic term kaftār, which Ibn Battuta identifies with the female form of a male “joki”—presumably a yoginī— is “a hyena that digs up and devours dead bodies” (Steingass 1892: 1037), once again an indication of the shape-changing powers ascribed to witches in these traditions. Babb (1975: 204) relates a cognate tradition, once again from modern-day Chattisgarh, according to which witches disinter and reanimate the dead in order to drink their blood. Ibn Battuta’s account introduces another tantric tradition not encountered in the “Barber’s Tale” but found in Babb’s ethnography: this is the notion that witches are able to invade the bodies of their victims simply by gazing upon them. This theme, which I have discussed at length in another study (White 2012), first enters the Hindu scriptural record in the Netra Tantra (19.45-50). Here, the evil eye (dṛṣṭipāta) is said to be employed by both demons of various sorts—female seizers (grahaṇīs), mothers (mātṛs) and spirit beings (bhūtas)—and by “injurious ones” (hiṃsakas), i.e., male and female sorcerers or witches. While Ibn Battuta identifies the heart as the organ that these witches magically extract, other accounts focus on the liver, attesting once again to the ambiguity of the innards fed upon by them. So, for example, another Muslim author, the late sixteenth-century Abu’l Fazl, gives the following description of the jigar khwár (“liver eater”), a magician from the Sindh region who

by glances and incantations can abstract a man’s liver . . . [H]e renders the person senseless upon whom he looks, and then takes from him what resembles the seed of a pomegranate, which he conceals for a time in the calf of his leg. During this interval the person whose liver is stolen remains unconscious, and when thus helpless, the other throws the seed on the fire which spreads out like a plate. Of this he partakes with his fellows and the unconscious victim dies. He can convey a knowledge of his art to whomsoever he wills, by giving him a portion of this food to eat and teaching him the incantation. If he is caught in the act and his calf be cut open and the seed extracted and given to his victim, the latter will recover. The followers of this art are mostly women [my emphasis]. (Jarrett 1891: 338- 39)

Indic references to this type of witch have proliferated over the centuries, also appearing in an early eighteenth century Braj-language adaptation of the “Twenty-five Vampire Tales” (a popular anthology embedded in Somadeva’s Ocean of Rivers of Story), which speaks of “witches who were chewing the livers of boys.” The same reading, which is not found in Somadeva’s Sanskrit version, would reappear in the 1805 Hindustani “vulgate” version of the work (Barker and Eastwick 1855: v, 27; Forbes 1861: vii-viii). In the Hindustani, the term translated as “witch” is ḍāyan, a vernacular form of the Sanskrit ḍākinī and a cognate of the of modern Rajasthani ḍākaṇ, a witch who is also notorious for feeding on the livers of her victims (Mayaram 1999: 119). A late nineteenth-century Punjabi tradition depicts the jigar khor as a witch, who, when she takes out a man’s liver, “leaves it uneaten for two days. If after eating it she is put under the influence of an exorciser, she can be forced to take the liver of some animal and put it back to replace that taken from the original victim” (Crooke 1891: 14). The same patterns are noted by G. M. Carstairs in his classic 1983 ethnography, Death of a Witch, who goes on to make a comparative observation with respect to the Rajasthani village in which he had carried out fieldwork for several decades:

Indian witches are still believed to practice a magical form of cannibalism. In this part of India, whenever an adult or child succumbs to a wasting disease, the onlookers believe that a witch is somehow drinking her victim’s blood and devouring his or her liver. Not surprisingly the dakan is greatly feared and hated . . .The account of witches’ behaviour given in local folklore bears many similarities to medieval accounts of witchcraft in Europe. For example, witches can temporarily discard their human form and turn themselves into cats or ants . . . they can slip out of their house unseen in order to foregather with other witches at a burning-ghat . . . Here they resume their human form and ride naked on the backs of hyenas. In European lore all this was described, except that the witches rode on wolves. In both traditions, if a witch was engaged in “devouring the liver” of a victim she could only be appeased by the slaughter of a kid or a lamb. (Carstairs 1983: 56, 15-16)


Two etymologies have been proposed for the word ḍākinī. The first, which is translated into Tibetan as mkha’ ‘grom ma (a word which, in its feminine form, is a colloquial term for “bird”), derives from the Sanskrit root ḍī, “fly”: ḍākinīs fly through the air (Farrow and Menon 1992: 180- 81). The second, derived from the root ḍam, “sound,” refers to the fact that the ḍākinīs, like the yoginīs, are frequently described as a raucous, noisy lot (Heilijger-Seelens 1994: 126, n. 21). In a recent study, Olga Serbaeva (2013) has brought together a compendium of references to the Hindu tantric practice of yoginī-melāpa, by means of which male adepts are faced with the two alternatives, evoked earlier, of either being loved or eaten by the yoginīs. Serbaeva’s study is based on the eighth- to ninth- century Kashmirian tantric scripture entitled the Jayadrathayāmala, which describes the conditions under which male practitioners interacted with the yoginīs on cremation grounds and lonely mountaintops in the dark of the moon and the dead of night. There, delirious from the consumption of transgressive offerings and self- inflicted pain, solitary ascetics would offer themselves in sacrifice, yielding up their bodies and vital fluids to the yoginīs. In these contexts, the yoginīs were said to take the forms of “land-based” jackals and hyenas and “airborne” carrion-feeding birds, filling the night with the cries of kites (cillarāvam), jackals (śivarāvam) and other predatory creatures (Jayadrathayāmala 2.2.97-991). Repeated references to noise, clamor, din, and hubbub underscore the fact that the assault on the male practitioner was also an auditory one (even as he, in his possessed state, imitated those same bird and animal cries) (Serbaeva 2013: 200).

With this, we have a complete tableau of these nightmare creatures as found in scriptural literature from eighth- to fourteenth-century India, Kashmir, Tibet, China, and Japan: they were raucous female creatures—usually carrion-feeding or predatory birds or mammals—that preyed upon their victims by magically extracting their vital fluids or organs, sometimes by means of their gaze alone. Following Carstairs’s observation concerning parallels between Rajasthani and European witchcraft traditions, I now turn to Roman literary sources that refer to a class of powerful witches known as strix or striga (plural striges) in Latin (Scobie 1978: 75, n. 6).

Apart from a fourth-century BCE Greek account of the origin of the strix—as a human woman cursed to become a screech owl by a petulant Aphrodite (Oliphant 1913: 133-34)—the earliest mention of these creatures is found in the Roman playwright Plautus’s 191 BCE comedy, the Pseudolus (Oliphant 1913: 133-35). Here, a cook pours out his scorn on his rivals’ dishes, saying, “when these people season the meals they’re cooking, they don’t use spices for spicing, but instead striges, to eat out the entrails of the living guests. So that’s why,” he concludes, “people here live such short lives” (Smith 1991: 177; McDonough 1997: 319). A vivid account of their predations is found in Ovid’s Fasti (8 CE) which tells of screeching birdlike striges that prey on suckling babes with their beaks. These

are greedy birds, not those that cheated Phineus’s maw of its repast [i.e., the Harpies], though from those they are descended. Big is their head, goggle their eyes, their beaks are formed for rapine, their wings are blotched with grey, their claws fitted with hooks. They fly by night and attack nurseless children, and defile their bodies, snatched from their cradles. They are said to rend the flesh of sucklings with their beaks, and their throats are full of the blood which they have drunk. Screech-owl is their name (est illis strigibus nomen), but the reason of the name is that they are wont to screech (stridere) horribly by night. Whether, therefore, they are born birds, or are made such by enchantment and are nothing but beldames transformed into fowls by a Marsian spell, they came into the chambers of Proca [the future king of Alba Longa]. In the chambers Proca, a child five days old, was a fresh prey for the birds. They sucked his infant breast with greedy tongues, and the poor child squalled and craved help. Alarmed by the cry of her fosterling, the nurse ran to him and found his cheeks scored by their rigid claws. What was she to do? She went to Crane [a powerful protective nymph] and told what had befallen. Crane said . . . “I myself will heal the child.” Straightway she thrice touched the doorposts, one after the other, with arbutus leaves; thrice with arbutus leaves she marked the threshold. She sprinkled the entrance with water (and the water was drugged), and she held the raw inwards [i.e. innards] of a sow just two months old. And thus she spoke: “Ye birds of the night, spare the child’s inwards [sic]: a small victim falls for a small child. Take I pray ye, a heart for a heart, entrails for entrails. This life we give you for a better life.” When she had thus sacrificed, she set the severed inwards in the open air, and forbade those present at the sacrifice to look back at them. A rod of Janus, taken from the white-thorn, was placed where a small window gave light to the chambers. After that, it is said that the birds did not violate the cradle, and the boy recovered his former colour. (Fasti 6.131-68 in Frazer 1959: 326-31)

Apart from the specific animal substitute, this is the same preventive procedure as that evoked by Carstairs: whereas in ancient Rome a piglet was offered, in modern India, where the flesh of swine is an abomination, a kid or a lamb becomes the substitute for an infant’s flesh. As Christopher McDonough has noted, these were “inner assaults,” which the striges commonly carried out by “removing their inner substance by biting or sucking.” A striking example may be adduced from Trimalchio’s tale contained in Petronius’s first-century CE Satyricon (63.8-9) in which shrieking striges enter into a child’s death chamber during the period of mourning.

While I was yet a long-haired lad . . . our master’s minion died, a pearl, by Hercules, a paragon in every regard. While his poor mother was bewailing him and several of us were sorrowing, suddenly the striges began to scream (stridere). You would have thought a hound was on the chase of a hare. We had at that time a Cappadocian, a strapping, dare-devil fellow, so strong that he could lift an angry bull. He, carefully wrapping his mantle about his left arm, dashed fearlessly out the door, with drawn sword, and as it were in this spot . . . ran a woman through the middle! We heard a groan, but—I swear, I’m not lying—we did not see the striges themselves. Back came our blockhead and threw himself upon a bed. His entire body was as black and blue as if he had been beaten with a cat, for forsooth an evil hand had touched him. We closed the door and returned to our mourning, but when the mother would embrace her son’s body, she touched and saw only a dummy made of straw. It had neither heart nor insides, nor anything at all. The striges had in sooth already carried off the boy and substituted a stramenticum vavatnem. What think ye? They are night hags; they make everything topsy-turvy. But as for that tall lout of ours, after what happened then, he never came to his color, but died a few days later, a raving maniac. (Oliphant 1913: 144)

While stramenticum vavatnem has often been translated as “straw doll,” McDonough (1997: 320) suggests that the term possibly meant “the body of the boy himself, with the moisture of his vitality removed.” Elsewhere in the Satyricon, striges are said to prey upon the marrow of the living. This, as Oliphant notes, makes Petronius’s account eccentric with respect to general Roman striga lore, according to which the victims that they prey upon are still alive (Oliphant 1913: 145). This aligns with a passage from Horace’s 30 BCE Epodes (5.37-38), in which a witch named Canidia sucks out (exsecta) a boy’s marrow and liver in order to make a love potion (McDonough 1997: 319).

As Ovid’s account of Proca and the strix indicates, Roman traditions played on the ambiguity of the striges’ identity in the same way as we have seen for the South and East Asian ḍākinīs and yoginīs: were they “born birds,” or enchanted “beldames transformed into fowls”? So too, in Petronius’s tale of horror, the Cappadocian “blockhead,” upon hearing the screeching of the avian striges, is confronted by a woman, whom he runs through the middle with his sword, to no avail. Although McDonough (1997: 327) notes that in the Roman world “the belief in the strix as transformed old women was widespread,” what they were transformed from or into remains unclear, ranging from “night ravens” to screech owls, horned owls (Scobie 1978: 77, n. 21)—or, as Oliphant has argued, bats.

Pliny the Elder, in his 77-79 CE Natural History (11.232 in Rackham 1956: 578-79) describes the striges as “screech owls” that suckle infants, quite the opposite of a creature that would suck the life out of children’s breasts, as in the case of little Procas; however, the second-century BCE Titinius, quoted by Quintus Serenus Sammonicus in his second-century CE Liber medicinalis, speaks of “a black strix [that] attacks boys, offering

its foul-smelling breasts to their eager lips for suck” (Oliphant 1913: 136; McDonough 1997: 319, n. 16). The same theme would be taken up some fourteen hundred years later in the notorious anti-witchcraft manual, the Malleus Maleficarum, which the states that the striges present themselves as midwives or wet nurses “to kidnap our children” (Maggi 2006: 38, 40). Here, one cannot help but recall descriptions of the female seizers of ancient Indic tradition: the ayurvedic science of pediatrics (kaumarabhṛtya) was predicated on the assumption that certain childhood diseases were caused by “the empoisoned milk of the female seizers,” (Suśrūta Saṃhitā 1.1.8.[5], in Atrideva 1975:3 ) that is, of demonesses that kill infants by suckling them with their poisoned breasts. This was precisely the modus operandi of the demoness Pūtanā when, according to the circa third-century Harivaṃśa (50.20-29 in Vaidya 1969: 343-44), she flew through the night in the form of a bird of evil omen (śakuni) to the village of the infant Kṛṣṇa. There, she took the form of a woman to offer the child her poisoned breast, through which the young god sucked the life out of her. Elsewhere, the 200 BCE-400 CE Mahābhārata (3.219.36-37) describes how a number of female seizers prey upon the fetuses of pregnant women. One of this type is said to seize the fetus and go off, her victim being viewed as a mother “whose fetus has melted away,” while another eats the human embryo in the womb and replaces it with the embryo of a snake (White 2003: 46). These passages are early evidence for a perennial fixture of South Asian demonology (bhūtavidyā): as the demonic embodiments of childbirth complications and childhood diseases, female seizers like the Buddhist Hārīti and the Hindu jātahāriṇīs carry off (this is what the verb *hṛ in their names denotes) embryos and infants. Hindu works like the eighth- century Kaśyapa Saṃhitā would explain miscarriages through a combination of the workings of the evil eye and the ability of these demonesses to enter the pregnant woman’s body and prey on the embryo inside (White 2012: 150). None of these works, however, conflate these demonesses with witches and none describe them as removing the inner organs or life’s essence from living infants or male adults.

In the wake of the Roman traditions, accounts of the dread striga spread across the whole of medieval and modern Europe, with the name of these witches changing as they entered the lexicons of a variety of languages, ranging from the Gallic stria to the Russian stryga, the Slovenian štrija and the Spanish estriga. (Oliphant 1914: 49, n. 3) The fifth-century Frankish “Lex Salica” evokes stria that eat men, while a fragment attributed to the seventh-to-eighth-century St. John of Damascus identifies striges as women known to fly around houses and enter through closed doors to suffocate, eviscerate and devour infants. (Tartarotti 1751: 160). In 1010 we see the canon lawyer Burchard of Worms seemingly taking a page from Petronius when he writes in his circa 1010 CE Corrector of

women [who] believe and affirm to be true . . . that . . . you can go out by closed doors . . . and without visible weapons slay persons . . . and cook and eat their flesh and in place of their hearts put straw or wood or anything of the sort and when they are eaten make them alive again . . . (Kors and Peters 2000: 67)

The inquisitorial record from Italy with regard to strighe or strege is rich in references to these witches’ habits of extracting the flesh and blood of their infant victims while leaving skin and bones behind (Ginzburg 1984: 47, 135, 147, 163, 191, 248). As one moves forward in time, references to these witches’ avian or animal shape-changing abilities decrease, but do not disappear entirely. The thirteenth-century Dominican inquisitor Etienne de Bourbon speaks of striges that take the form of wolves to drink the blood of infants (Lecoy de la Marche 1877: 320) and the Italian record speaks of strege that morph into cats (Ginzburg 1984: 122). In medieval sources, such was far more common than the Greco-Roman transformation of striges into owls or other winged creature; medieval traditions attribute their power of flight to potions smeared on broomsticks or chairs (Scobie 1978: 81, n. 36; 86).

Like the South Asian sources in which several terms of different origins (ḍākinī, śākinī, śabarī, yoginī) lend themselves to being translated as “witch,” so too the medieval European inquisitor’s lexicon contained several near synonyms for striga. These included lamia (derived perhaps from the name of the Babylonian demoness Lamashtu: McDonough 1997: 336), geludes (from the Greek gello: Tartarotti 1751: 160; Oliphant 1913: 148) and masca, about which the thirteenth-century Gervasius of Tilbury wrote that

it is the wretched lot of some men and women to cover great distances in a swift nocturnal flight; they enter houses, torment people in their sleep, and inflict distressing dreams on them, so causing them to cry out. Apparently they also eat, and light lamps, take people’s bones apart, and sometimes, when they have dismembered them, put them back together again in the wrong order; they drink human blood and move babies from place to place. (Banks and Binns 2002: 722-25)

In his Teutonic Mythology, Jacob Grimm (1999: 1039, 1045, 1057, 1060-61) catalogues several cognate northern European traditions, medieval and modern, of noisy, shape-changing nocturnally active witches that fly, dance, devour children and so forth. One of these concerns a Servian witch (veshtitsa) that “tries to catch people to eat up, especially young children. If she finds a man asleep, she pushes a rod through his left nipple, opens his side, takes out the heart and eats it, and the breast closes up again. Some of the people thus ‘eaten out’ die directly, others live on for a time” (Grimm 1999: 1078).

Pairikās, Circe, Kuvaṇṇā

How are we to explain the parallels and convergences discussed to this point between the Mediterranean striges and the ḍākinī and yoginī traditions that informed Somadeva’s “Barber’s Tale”? If our focus is on the tantric traditions of these dire beings, then we can unequivocally state that they originated in South Asia. Ḍākinī and yoginī are Sanskrit terms that first appear in Indic texts. The word yoginī, which first appears in the Mahābhārata (1.60.15; 9.34.40-65) as the name for the aggregate of star-maidens in whose clutches the Moon god was drained of his vital fluids, mainly remains tethered to Hindu Tantra; although, an entire canon of Indo-Tibetan Buddhist scripture is classified under the heading of the “Yoginī Tantras.” South Asian ḍākinī lore was exported, together with the Mahāvairocanasūtra and Buddhist Tantra, by the Indian master Śubharakṣasiṃha to the Chinese monk Yixing and then on to Japan, through the writings of Tendai authors like Kōjū and Chōgō. These documented lines of transmission are clear and irrefutable.

The same cannot be said, however, for the more ancient Indic and Kashmirian witchcraft traditions that underlay these tantric appropriations: did they—or at least certain of their components— originate in South Asia, the Mediterranean world, or somewhere in between? Or, are they simply South Asian iterations of a polygenetic mytheme? Alex Scobie (1978) provides a long catalogue of myths and lore from the Americas, China, and the ancient Near East in which witches take the form of birds, often owls, to carry off children or lure men to their deaths. In one remarkable case, Trobriand Islander traditions of flying witches (mulukwausi) bring together nearly the entire cluster of mythemes found in the South Asian and Roman lore reviewed to this point:

The mulukwausi will eat out the eyes, the tongue, and the “insides” (lopoula) of the corpse; when they attack a living man they may simply hit him or kick him, and then he becomes more or less sick. But sometimes they get hold of an individual and treat him like a corpse and eat some of his organs, and then the man dies. It is possible to see this, for such a person would quickly fail, losing his speech, his vision, sometimes suddenly being bereft of all power of movement. It is a less dangerous method to a living man when the mulukwausi, instead of eating his “insides” on the spot, simply remove them. They hide them in a place only known to themselves, in order to have provision for a future feast. (Malinowski 1950: 242, quoted in McDonough 1997: 319, n. 15)

This analogous tradition notwithstanding, I would argue that the remarkable similarities between the witchcraft lore of Somadeva’s “Barber’s Tale” and allied Asian traditions on the one hand, and that of the Roman striges and allied European traditions on the other, cannot be adequately explained on the basis of coincidental independent innovation by cultures from across the globe. Here, a review of the multiple components of these two bodies of lore is in order. In both, “witches” are 1) shape-changers, i.e, human or superhuman female beings that transform into predatory or carrion-feeding birds or mammals 2) possessed of the power of flight. 3) Both prey on their victims by extracting or sucking out their inner organs or fluids “from the inside.” 4) Both are night creatures 5) that are able to break into the abodes of the living; and 6) both are identified by the raucous noise they make: the striges are so called because they screech (stridere), while one derivation for the word ḍākinī is *ḍam, “to sound, make noise.” 7) Both can be satiated by the sacrifice of an animal substitute, in which case they restore their human victims’ inner organs and life. In tantric adaptations of this last item, ḍākinīs and yoginīs reward their human victims’ self-sacrifice by restoring them to life, endowing them with gnosis, and admitting them into their own clans and lineages.

Chronologically, most of the elements of this cluster of motifs appear in Roman literature before they first appear in South Asian narratives. While the pre-tantric demonological traditions of mothers, female seizers and spirit beings as found in the Mahābhārata, Harivaṃśa and early ayurvedic literature are only a few centuries younger than those alluded to in the works of the second-century BCE Plautus and Titinius—and nearly coeval with the first-century CE Ovid and Petronius’s full-blown striga narratives—all of the South Asian traditions of tantric witches (ḍākinīs, yoginīs, etc.) postdate the Roman material by several centuries.

It is very likely that these various literary accounts emerged out of oral traditions; however, this can neither be proven, nor can

irrecoverable oral traditions be dated. There is also the possibility of an Indo-European link between these Greek, Latin, and Sanskrit traditions. This is improbable given the fact that one finds no cognate mythology in the Irish, Slavic, or Scandinavian record. One does, however, find a cognate Iranian tradition that displays some, but not all, of the common features of the ḍākinī-yoginī and strix narratives. Here, I am speaking of a class of powerful human witches-cum-supernatural demonesses attested to in Early and Middle Iranian literature, ranging from the ancient Avestan Yašts to Pahlavi literature dating from as late as the tenth century CE. Called pairikās (Middle Iranian parīg; modern Persian parī: the word is a cognate for the English “fairy”), their lore bears certain similarities to that of the witches of the Asian sources reviewed to this point. Like their sister witches from Hindu and Buddhist Tantra, the Iranian pairikās or parīs are shape-changers able to mask their monstrous aspect by putting on the appearance of beautiful women, birds (Christensen 1941: 14-15, 57, 90), rats, dogs, (Adhami 2010), or bats to destroy or carry off infants (Schwartz 2008: 98-99). Furthermore, as Siamak Adhami has noted, the pairikā of Avestan tradition “bore some similarities with the demoness generally called succubus and found in a number of other cultures, such as Sumerian Lillake, biblical Lilith, Slavic Rusalka, who deprive men, particularly religious men, of their bodily fluid at night” (Adhami 2010).

According to “popular Zoroastrian” traditions of the fourth- to seventh-centuries, the site at which the parīgs or parīs congregated in their greatest numbers was Mount Ilo, the central mountain of the Swat region of northwestern Pakistan. (Templeman 2002: 115-16) As Giuseppe Tucci (1977: 282) and more recently Jamal Elias (2011: 153) have noted, parī traditions remain particularly strong in the same region down to the present day. So, for example, the Kāfirs of the Hindu Kush identify parīs with “mothers,” and view them as the cause of childhood diseases, possession, and epilepsy (Adhami 2010). Writing in the thirteenth century, Urgyenpa noted that the women of the region knew “how to turn themselves by magical art into any form they want; they like flesh and blood and have the power to deprive every creature of its vitality and strength” (Templeman 2002: 122). Some five centuries earlier, early Hindu and Buddhist tantric scriptures identified precisely the same region, which they called Uḍḍiyāna or Urgyen, as the birthplace of cults of the yoginīs and malevolent ḍākinīs (Templeman 2002: 114; Sanderson 2007: 261). It is, therefore, probable that several elements of yoginī and ḍākinī lore originated in the same Kashmir- Gandhara-Bactria region that spawned certain of the Iranian pairikā traditions.

Absent from the Iranian lore, however, is the particular modus operandi of the striges of European lore and the ḍākinīs of Somadeva’s “Barber’s Tale.” Found nowhere else (with the troubling exception of the Trobriand Islands), this involves the extraction, often by sucking, of the victim’s inner organs, in which the bodily envelope is otherwise left intact. Now, it is possible to rationalize this technique (as Carstairs does) as simply a pre-modern explanation for internal pathologies, hemorrhaging, consumption and the like, conditions whose causes are not visible from the outside. But why, then, with the exception of the Trobriands, is it found nowhere outside of ancient Rome and medieval Asia? Here, I will argue that the lore of the Roman striges was carried directly to South Asia near the beginning of the Common Era, and then on into East Asia with the later spread of Buddhist Tantra. I will further argue that it was brought from the Mediterranean world to South Asia by a maritime route. Most plausibly, strix lore reached South Asia via one of the Roman trading ports on the coast of modern-day Gujarat or Maharashtra. From there, they were carried overland to both Kashmir and the Swat Valley. This was the route that brought portions of the canon of the Kubjikā Tantras to Kashmir and Nepal from coastal Koṅkaṇa in the tenth to eleventh century (Dyczkowski 2004: 224-28).

Now, it is the case that the earliest contacts and exchanges between South Asia and the Hellenic world followed directly on the Alexandrian invasion of the fourth century BCE. The enduring presence of the Bactrian Greeks (the Yavanas of Sanskrit tradition) is attested in coinage, cults, sculptural styles and many other cultural elements found in the region. Indeed, the cult of Dionysios, which flourished in the Swat Valley under the Greco-Bactrians and Kushans, may have been a source of some of the orgiastic elements of the yoginī and ḍākinī cults (Bopearachhi 2012). An important witness to these exchanges may be found in the mythology of the Mahābhārata itself. On the one hand, certain of the exotic vassal populations of the Pāṇḍava king Yudhiṣṭhira—people who used their feet as sunshades or who wrapped themselves in their giant ears (Mahābhārata 2.28.44, 47)—were reported by Megasthenes, the circa 300 BCE Greek ambassador to the Mauryan court, as ethnography, and so the perennial European traditions of the Sciapoda (“Shade-Feet”) and other “monstrous races” was born (Wittkower 1942: 164). Greek mythology also found its way into the Hindu epic. The great bird Garuḍa’s wholesale slaughter of the diminutive Niṣadas on a remote shoreline (Mahābhārata 1.24.2-14) borrows directly from Homer’s account in the Iliad (3.3, in Murray 1946: 116-17) of great “cranes [that] escape in the winter time and . . . clamorously wing their way to streaming Okeanos, bringing the Pygmy men bloodshed and destruction.”

The most remarkable evidence for the direct transmission of Greek mythology into South Asian tropes, however, is found in the realm of witchcraft and demonology. The most notorious witch of the ancient world was Circe, the enchantress of Homer’s Odyssey who transformed Odysseus’s men into swine. Then, after she has shared her bed with Odysseus, she restores the sailors to human form and shows Odysseus the way to guide his ship home to his native Ithaca (Odyssey 10.210-549 in Murray 1966: 358-85). One of the most popular myths in all of ancient and medieval Europe, the Circe episode was recounted and performed in a variety of languages and lands for some two thousand years (Paetz 1970). Largely overlooked is the fact that one of those languages was Pali and that one of those lands was the island of Ceylon, the modern Sri Lanka.

The Mahāvamsa, the fifth-century chronicle of Ceylon, contains the account of how Buddhism was brought to the island at the time of the Buddha’s death by a certain Prince Vijaya. In crafting his narrative, the Mahāvamsa’s author Mahānāma—a scholar monk from an important Buddhist monastery at Anuradhapura in the northwestern part of the island (Bullis 2012: 3)—greatly expanded upon a terse account found in the earlier Dīpavamsa, which had simply noted that Vijaya and his companions were cast ashore in Ceylon, where they founded cities and established a kingdom (Winternitz 1972: 214-15). It is in his expanded account that we see Mahānāma transparently adapting the Circe episode into an Asian context (Mahāvamsa 7.1-32, in Geiger and Bode 1912: 55- 57). The first recorded notice of this was penned by George Turnour, who, in his preface to the earliest English-language translation of the Mahāvamsa, wrote that “the fabulous tone of the narrative in which the account of Wijayo’s landing in Lanká is conveyed in the seventh chapter, bears, even in its details, so close a resemblance to the landing of Ulysses at the island of Circé, that it would have been difficult to defend Mahanámo from the imputation of plagiarism, had he lived in a country in which the works of Homer could, by possibility, be accessible to him” (Turnour 1837: xliv).

The salient points of comparison between the two narratives may be summarized as follows. In the Mahāvamsa, Odysseus is transformed into Prince Vijaya; the island-dwelling weaver witch Circe becomes the island-dwelling spinning demoness (yakkhī, the Pali form of the yakṣīs or yakṣinīs that the Chinese commentator Yixing identified with the heart-stealing ḍākinīs of the Mahāvairocana Sūtra) Kuvaṇṇā; Hermes becomes the great god Sakka (Indra); and the moly herb given to Odysseus by Hermes becomes the magic thread and water from a vessel given by Sakka to Vijaya. Both witches imprison the men of the hero’s crew; both heroes overcome their female adversaries with the magical objects they have received from their respective protector gods; and both complete their conquest of their respective enchantress on the latter’s canopied bed. In a most telling coincidence, the Marsians— whom Ovid identifies in his Fasti as the possible source of the spell that originally transformed human women into the avian striges—were renowned in Rome for their magical skills, which they themselves attributed to their direct descent from Circe and Odysseus! (Littlewood 2006: 47).

Figure 1: Two Yoginiis leading a victim to the slaughter, Chinnamasta Temple, Patan, Nepal. Photo by the author.

Figure 2 on the opposite page, which is based on the translations of Murray (1966: 360-77) for the Odyssey and Geiger and Bode (1912: 56-57) for the Mahāvamsa, provides a detailed synoptic comparison of Homer’s and Mahānāma’s accounts:

When he wrote his 1837 preface to his translation of the Mahāvamsa, Turnour was unaware of the fact that Homer’s account of the Pygmies and cranes had been adapted into the Mahābhārata, or that the Odyssey’s tale of Circe had been retold dozens of times over across the ancient and medieval world. He would also have been unaware that the circa 60 CE Periplus of the Erythraean Sea, an anonymous Greek work, catalogued over twenty ports of trade all along the (mainly) western coastline of India (Huntingford 1980:17-18). The author of the Periplus also had knowledge of certain inland cities of India, particularly of Ujjain, which was, according to several Indic traditions, the city in which the now lost anthology of tales adapted by Somadeva into his Ocean of Rivers of Story was first compiled during the first centuries of the common era (Renou 1963: 10). The first centuries of the common era saw Roman trading cartels launching ships by the thousands across the Arabian Sea, establishing fortified settlements at South Asian ports, and regularly sending ambassadors to South Asian courts. It was during the same Roman period that the spice trade made Ceylon, then called Taprobane, a prime destination for Roman ships. Roman coin hoards and other artifacts from the first centuries of the Common Era have been unearthed all along the island’s western coast, including at Mantai, which was the principal port for Mahānāma’s Anuradhapura, the political, economic, and religious capital of the island from the second century BCE to the ninth century CE. At its height, Anuradhapura’s trade network extended to China and Rome (Tomber 2008: 119, 127, 145-46; Bullis 2012: 114-15; Holt: 795). In the light of these data, Mahānāma’s “plagiarism” of Homer’s tale of Odysseus and Circe is a virtual certainty. Tales of the dread strix, recounted by Roman sailors, would also have reached South Asian shores in the same period, crossing many oceans and rivers of Indic oral traditions before being adapted into the tantric canon and Somadeva’s Oceans of Rivers of Story.

For References, download the PDF Issue.

Originally published in (SERAS) Southeast Review of Asian Studies Volume 35 (2013): 7-31


ISSUE #005

On the Divine Feminine
Image: On the Divine Feminine

Read and reference at your own pace.
Download this issue of Tarka as a PDF to access the full-length, unabridged articles.

Embodied Philosophy Forum

A Private Facebook Community