This chapter takes account of the tendency of Sanskritists and other scholars of Indian antiquity to disregard religious form not tied directly to texts. Similarly, it takes account of the tendency of anthropologists and other ethnographers who distrust texts and are therefore not willing to look beyond their own ethnographies for the possibility of important antecedents in classical textuality. I aim to close these gaps, at least to some degree, with respect to possession by integrating material from Sanskrit sources into a primary discussion of both the language of possession in modern India and models of possession that are intertwined with these languages.
First I must state emphatically that it is impossible to provide a complete record of the language of possession in all South Asian languages or, for that matter, in all eighteen recognized “national languages” of India or even in more than a few of them with any degree of authority or thoroughness. Comparative studies of the vocabulary and linguistic idiom of any single religious phenomenon as it is represented in all the major regional languages is, to my knowledge, uncharted territory, not to speak of similar comparative studies of nonliterary or “tribal” languages, as well as dialects and subdialects of a single major language such as Hindi or Tamil. The difficulties involved in amassing the necessary raw lexical data for such a project are, quite simply, overwhelming. Perhaps this will be possible after another quarter of a century, when optical character recognition equipment for South Asian scripts are perfected and databases with literary, lexicographic, and ethnographic material are more complete and integrated. Under current conditions, therefore, I can strive only to produce a few examples of vocabulary and idioms for possession in a few South Asian languages for which such studies have been undertaken. Unfortunately, even the best ethnographers, linguists, and lexicographers have sometimes neglected to provide the specialized vocabulary of possession in their work. It is often assumed that the phenomenon itself is all that is important, not how the group or individuals under study represent it linguistically or what the cultural or mythic comportment underlying this linguistic usage might be. This neglect has contributed to the mistaken notion that possession is just one thing and renders it easy fodder for established Western categories. While lamentable, this tendency is also correctable.
The Indian languages best served by sensitive ethnographers of possession are Hindi, Marathi, certain languages and dialects of the Himalayas (including Ladakhi and Nepali), Bengali, Tamil, Malayalam, and (interestingly) Tulu. The most intensively researched languages are Tamil, Malayalam, and the Himalayan languages, as these opposite ends of the subcontinent have become preferred territory for possession research.1 The primary sources necessary for a proper determination of possession in regional languages are, in addition to the ethnographies, the early regional literature, dictionaries of vernacular languages modeled after European lexicographies, and traditional lexicographic works (kośas).2 The primary benefit of the early vernacular literature is that it provides insight into the history of both the language used and the experiences denoted by that language. The disadvantage is that this literature was not written to present complete linguistic records, but is intrinsically fragmentary. The dictionaries benefit from their completeness, but are far too often riddled with problems. First, they are often skewed toward Sanskritic interpretations. Second, they too often passively accept Western categories. This is especially prominent when translating from one language to another, for example, from English to Tamil or Hindi to English. It must be said, however, that the problem of translating concepts and categories—when the assignment is to translate words with the briefest precision—is, in most cases, an exercise in futility. Third, a lexicographic format almost always struggles to meet the demands of context, which is to say that most dictionaries provide practically no historical accounts of word usage.3 This renders it impossible to know which words are most important and widely used and which are rare and perhaps idiosyncratic. As for the ethnographies, their great advantage is a vivid portraiture, while, as mentioned, their disadvantage is an often complete absence of deep historical and linguistic contextualization.
The following account attempts to consider images and models of possession and exorcism, for example, possession as “riding” (Hindi, Nepali, Simhala, Malayalam), as “dancing” (Tamil, Malayalam), as an “attack” (Hindi, Nepali), as a force “coming into the body” (e.g., Hindi, Marathi, Tulu, Irula), as “play” of the deity (Hindi, Marathi, Nepali), as a kind of “ecstasy” (Bengali, Marathi, Nepali), as a “weight” (Bengali), as a marker for intense emotional engagement (Sanskrit, Malayalam, Bengali, and many others), as an idiom for impersonation (Tamil), as an emblem of political oppression (Ladakhi), as a sign of debilitated life force (Ladakhi, Sanskrit), as part of a multicultural dialogic interaction (Ladakhi, Nepali), or as a symptom of a multilayered world visible “as if in a mirror” (Tibetan, Sanskrit). Similarly, exorcism has a varied set of images, for example, it can occur in an imagined legal court (Urdu) or serve as communication with a spirit for the sake of rectifying its own history of wrongdoing (Urdu, Simhala). In addition, the rich vocabulary found for exorcists and other possession mediators reflects healing and other culturally integrative roles taken on by these ritual functionaries, including conjurer, scholar, priest, herbalist and medicine man, storyteller, protector of the evil eye, and performer. The semantics of possession in regional South Asian languages reveals the important role that it has assumed in South Asian culture and therefore sets the stage for the ensuing discussions of possession in the ancient and classical literature, literature that cannot reflect such a diverse number of languages or local cultures. This should contribute to ongoing discussions of the continuities and disjunctions within South Asian cultural history, to the development of the language of emotion in South Asia, and to reflections on the relationship between the folk and the classical in South Asia, a discussion that closes this chapter.
Hindi
Variants of the Sanskrit āveśa and grahaṇa appear widely distributed in regional Indian languages. The unexpectedly late semantic conflation of √gṛh with ā√viś in Sanskrit is seen in several Middle Indo-Aryan languages contemporaneous with such usages in Sanskrit texts from the last few centuries of the first millennium C.E., which in turn set the stage for the all but interchangeable use of these verbs in New Indo-Aryan. In Hindi, āveś in the early Sanskritic senses of pervasion, absorption, and entry has been completely superseded. In addition to spirit possession, which is also a common meaning for the Hindi verbs caḍhnā (“to climb or mount,” hence its sexual connotations; “to become possessed”) and khelnā (to play) (cf. below, Marathi kheḷ,4 Panjabi kheḍnā,5 and Bengali khelā6), āveś discloses the senses of “charge, agitation, intense emotion, frenzy, wrath.”7 Erndl reports the use of the Hindi verb khelnā as illustrating the sense of possession as play, more specifically, the play of the deity with the person possessed. “When a woman is possessed, the goddess is said to take on a ‘wind form’ [pavan rūp], enter her, and ‘play’ within her.”8 Some of the terms for possession noted by K. P. Shukla, who conducted a sociological study of traditional healers in rural areas near Varanasi in 1969–71, are havā lagnā (to feel a breeze), pher (turn, curvature), kisī ne kucch kar (or de) dīyā hai (giving someone something, doing something to someone; probably indicating black magic).9 Ann Gold notes the term bhāv ānā (a feeling is coming) used in Rajasthan,10 while Beatrix Pfleiderer and Lothar Lutze note jāgar (waking [the deity]).11 Ruth and Stanley Freed note bhūt kī bīmārī (spirit illness), bhūt grasth (ghost possessed), and bhūt lagnā (attacked by a ghost).12
While not noting the vocabulary for “possession,” Ariel Glucklich, who worked on magic in Varanasi, does record the well-known colloquial term for a capricious spirit or ghost (bhūt-pret), which is attested throughout northern and central India.13 Bhūt (Skt. bhūta) is a general word for any kind of ghost, while pret (Skt. preta) more specifically designates a wandering spirit of the dead, often a deceased ancestor whose funerary rites remain incomplete or were performed improperly, if at all. Both bhūta and preta are past participles in Sanskrit; thus they may be said to represent the past, entities that have existed and gone away. We may suggest, then, that any model of spirit healing or exorcism admits the possibility that the obstruction and uncertainty that intercedes between the past, bhūta and preta, and the future can be ruptured and extinguished. In this sense, negative possession, which is to say possession by a bhūt-pret, locks an individual into a past that can be abrogated and (re)connected to the future only when the possessed personality is effaced. We explore this notion more fully in discussions on possession in the Mahābhārata below, especially in unpacking the story of the possession of Aśvatthāman by Śiva, the most powerful of bhūt-prets.
Urdu: Hanumān’s Court at Bālājī
One of the most striking models of possession and exorcism, in which bhūt-prets play a significant role, occurs in the environs of the Bālājī temple in Mehndipur, Rajasthan, as well as at several other spirit healing centers in India.14 Most of the clients at Bālājī are Hindu, from Hindi-speaking areas of India.15 Yet in a slight acknowledgment to the once-dominant Indo-Islamic culture, the clinical model is replete with Urdu terms. These terms are drawn from the British colonial legal system, which adopted the Persian-Urdu language of the Islamic legal system of the Mughals. This is described in the work of Antti Pakaslahti and Graham Dwyer,16 but a summary is apposite here. The possession/exorcism rituals at Bālājī are not standardized; instead, they bear the distinct imprimatur of different healers and exorcists, almost all of whom work outside the main temple.17 Nevertheless, all the healers situate their exorcisms (bhūt-pret utarṇā)18 in an adālat or Islamic court, drawing their inspiration from the concept utilized at the main temple. This, it is important to note, is not an actual court; rather, it is an envisioned court and background metaphor, employed for instituting their processes of possession, exorcism, and healing.19 The primary deity, Hanumān as Bālājī, is the supreme magistrate. He is assisted by two subordinates, Bhairavjī (the local manifestation of the protector divinity Bhairava) and Pretrāj (King of the Spirits of the Dead), who are Bālājī’s messengers (dūt)—as Hanumān is Rāma’s messenger (dūta). Before visiting the office—usually a small shrine room—of one of the many clinicians in the town, Bālājī is first petitioned (darkhvāst) at the main temple. Whether pilgrim or afflicted petitioner, Bālājī is proffered sweets (laddū), parts of which are offered into a sacred fire (havan) in his name, while the remainder is returned to the client as prosād. After this offering, a larger petition (arzi) of sweets is served to Bhairavjī and Pretrāj in side annexes of the temple, where their shrines are located. Through their consumption of the prasād, the power of the deities is said to be infused into the clients. One or two days later, trance commences, usually during scheduled offerings (ārati) or devotional singing (kīrtan), which may occur in either the main temple or the “office” of the exorcist. This kīrtan (at least at the locale in Bālājī that I observed) consists of deafeningly amplified recordings of devotional singing to the melodies of Hindi film songs.20 This engages the largely Hindu audience21 of spirit afflicted (bhūt-pret kī bīmārī) individuals and their families, who chant along with the music, “Jai Mā, Jai Mā” and so on, and clap rhythmically—some of the standard accoutrements of shamanic “trance” music. The music that most often sends afflicted members of the audience (or concerned relatives) into trance, at least in the adālat that I witnessed, is instrumental “snake-charmer” music from a Hindi film of the late 1980s, Naagin, which, appropriately, inspires some of the possessed to writhe like serpents.
The word used for this trance is the Urdu peśī, literally a “hearing” or “appearance in court,” in which the classic symptoms of possession occur. Initially these include rhythmic swaying (jhumnā), literally “to move as if one were intoxicated, as a drunkard,”22 but eventually violent whirling of the head and upper body, hectic chaotic movements such as somersaults, writhing, sobbing, and so on. Women’s hair is always unbraided and left loose during the proceedings.23 During the peśī, at least in my observation of it, the client or the clinician’s assistant, performing as a surrogate on behalf of the client by temporarily taking the bhūt from the client, becomes possessed by Bhairavjī or Pretrāj.24 This double possession of the invasive bhut along with Bhairavjī or Pretrāj is understood as a battle between bhut and dūt, negative and positive entities and polarities of personality.25 It is often cast as a battle; the dūt may be assisted by an army of positive spirits, with its own hierarchy, called phauj (Urdu for “army”; or dūt saphāī [sepoy]),26 constituted of former bhūts that have been transformed into this lower grade of protective dūt by the exorcistic and analytical procedures of the clinicianritualist(s).27 If the clinician or his or her assistant has taken on the bhūt, the battle transpires within his or her body, which has also become possessed by the dūt.28 S/he then acts out the violence of battle in chaotic movements, with eyes turned back, hair flying, mouth foaming, and other familiar effects of possession. If the client has retained the bhūt (and often if not), one of the regular actions of the assistant is forceful beating on the back (usually the lower part of the spine) of the victim with the fists, followed by wrenching upward movements of the fists, said to drive the bhūt(s) up the spine and out the top of the head.
This, according to the model of the adaālat, is the punishment (sazā, daṇḍ) meted out to bhūt by dūt. It forces wildly uncontrollable shaking of the head and rocking of the upper body. Often the clinician (or the assistant) and the victim are both in possession, writhing and wrestling, both sets of eyes locked or rolled back. But the clinician or assistant, with much more experience and comfort in the adālat and a seasoned eye for the dynamics of psychosocial issues, particularly the manner in which mental distress bears on social dysfunction, has marginally more control of the situation. Therefore he or she is in a much better position to control the agendas and paths of rectification and healing. This process of administering psychological justice in the peśī usually must occur many times before the healing is fully effected, which is to say before the bhūt is completely extricated. This is because, according to Pakaslahti’s informants, “the possessing spirits are held to be deceitful (cālak) and full of tricks (dhokhebāz).”29 In this way the client becomes empowered or resocialized to the point of return to the normal flow of family life.
The penultimate step in this legal proceeding is the bayan, a confession or statement by the bhūts, who identify themselves, announce their departure, their next destination, and perhaps other significant information, such as why they have taken up residence there. Communication with a spirit, or more often interrogation, even harshly worded, occurs frequently in South Asian possession. This is considered beneficial for both the victim and the spirit.30 In this case, the confessions of the bhūts are spoken out after the battle has been suspended because of exhaustion of either the client or the clinician or his or her ritual assistant who has borne the bhūts. In this way the battle lines are drawn between bhūt and dūt: counterpossession of positive entities is required to combat possession by negative ones. Usually, the bhūt reveals the required information upon the relentless interrogation of the clinician or healer.31 This part of the adālat resembles a talking cure, a local variant of psychoanalysis that deals directly with the bhūt-pret, in which the clinician attempts to address the problems of the entity and elevate it into the ranks of the phauj, thus freeing the afflicted individual of his or her madness or dysfunctional condition. This is the local variant of the widespread phenomenon in both Hinduism and Buddhism of the conversion to good causes of “criminal gods and demon devotees.”32
As it turns out, most of the afflicted, as Dwyer has discovered, believe that the afflicting bhūt-pret is a Muslim spirit (jinn/jind), a dissatisfied ancestral spirit (pitṛ devatā), or a wild sexy female spirit (chūdāl, cuṛail) sent by a tāntrika in the employ of an enemy.33 Pakaslahti notes that the Hindi terms bhūt, pret, jinn, saṇkat, upari havā, and so forth “practically never occur with such adjectival attributes as ‘evil or malevolent’ (burā, niṣidh, pāpī, duṣṭ, etc.). The only preceding adjective used was, in fact, ‘unhappy,’ dukhi, connected to their descriptions as souls stuck in the past, hovering unfulfilled in the air, unable to proceed forward.”34 The implication here is that the conception of these as “evil spirits,” with the weight of the word “evil” emphasized, is a Western construction imposed on the Indic material. In fact, the Sanskrit material agrees with Pakaslahti’s findings; I have not come across any reference to a spirit that was “evil,” per se. This reconfirms that the exorcisms are designed to alleviate misery rather than evil, an important distinction when considering the healing process.
The final procedure in the adālat is parhez, home instructions, which often includes dietary advice—usually the elimination of heavy or inertia-producing (tāmasik) foods such as alcohol, meat, and garlic (because bhūts are attracted to these) and the addition of foods categorized as “cold” in āyurvedic medicine, such as barley—and the prescription of gifts of uneaten food offerings to cows and dogs. Despite this density of Islamic terminology, the terms for possession itself, in this case spirit-induced illness, are the more standard Hindi upari havā (breeze from above), bhūt kī bīmārī (spirit sickness), and jāḍu ṭonā or caukī (sorcery),35 while the chief clinician/priest/ exorcist/psychologist is termed bhagat, in conformity with other north Indian terminology.36
Although this model of possession, exorcism, and healing is exceptional in South Asia, it is not conceptually dissimilar to important models of death, possession, and corporeality found in China. I am not arguing for its transmission from China to India (or vice versa); the evidence is much too fragmentary and temporally disparate to suggest any kind of mechanism for transmission. I mention this only to point out that the idea of a court as a venue for an ontological life and death struggle of this kind is not entirely idiosyncratic. While addressing the oxymoronic topic of “postmortem immortality” in Daoism and early Chinese Buddhism, the Sinologist Anna Seidel notes that Han dynasty (206 B.C.E.–220 C.E.) funerary texts do not assign the dead to various realms based on ethical issues of karma, retribution, and expiation, as in later Buddhism, but present a “bureaucratized netherworld,” with its “hierarchy, its tax offices, tribunals, and prisons.”37 Like the bureaucratized parallel universes of classical China, the adālat features a hierarchy of celestial officials who judge and assist the condition of the individual. But the adālat is also the locus for attending to ethical issues centered on individuals who are dysfunctional and resistant within the structure of “family values.” In both cases, the rationale appears to be to present to the afflicted (and their families) closed structures within which mastery and control over spirits may be demonstrated and, in some cases, transmitted.
In later periods in China, the task of constructing a legal model for possession and exorcism was accommodated by the convergence of government official and priest, functions rarely shared in classical India. According to Edward Davis, who has written an extensive study of possession ritual in Tang (618–907) and (especially) Song dynasty (960–1278) China, Daoist ritual masters were almost always portrayed as government officials. The Tang court, Davis writes, “continued to employ at least fifteen master spirit-mediums (shiwu 
) in the Imperial Divination Office (Taibu shu 
) of the Court of Imperial Sacrifices (Taichang Si 
) plus a number of related exorcists known as ‘Spellbinding Erudites’ (Zhoujin boshi 
, Zhoujin shi 
, or Zhoujin gong 
).”38 During the Song dynasty, a “Rite of Summoning for Investigation” (kaozhao fa 
) developed, an elaborate ceremony that included, among other elements, “the visualization of the spirit-generals and soldiers, the summoning of the demon, and the interrogation of his identity.”39 The spirit-soldiers (jiangli 
), rather like those operating in the adālat, were marshaled to locate and identify an afflicting spirit, then at the command of the ritual master, “to seize, bind, flog, pummel, yoke, burn, freeze, or crush the demon.”40 This is reminiscent of the adālat, where the rhetoric of violence is commonly employed to exorcise bhūt-prets. In both cultures, disembodied ancestral spirits seek a return to the terrestrial world. And in both the adālat and the “Rite of Summoning for Investigation” (as well as other related Chinese models of exorcism), the burden of guilt is shifted from the individual to the invading spirit. “[I]t is the demon who confesses, not the patient,” writes Davis, in words that could be applied to the adālat, “and the demon confesses because it is only the demon who has something to confess. By disrupting the link between illness and morality in the patient, the … codes, regulations, and protocols of the fashi [
]41 locate guilt solely on the side of demons.”42 In South Asia, this practice of shifting the burden of guilt to a spirit, thus freeing the individual supposedly harboring the spirit from the social opprobrium of dysfunctionality, is not simply a psychotherapeutic folk strategy employed by exorcists and village healers; we see it in the Mahābhārata, composed perhaps two thousand years ago. In the story of Nala (nalopākhyāna), which we deal with extensively (Chapter 6), Nala’s conscience is clear in his rapprochement with his wife Damayantī, whom he has recklessly betrayed, because he has understood that he was not responsible for his actions while possessed by Kali, the spirit of darkness, a spirit that is in the end bound and burned.43 Davis adds—once again with words strikingly resonant of the adālat and, indeed, classical Indian mythic and medical literature—that “if the patient has become demonized, the demon ultimately becomes humanized.”44 The spirit seeks a return, a renewed strategy for expression and communication. Finally, Seidel points out that the Chinese word for “demon, ghost, wraith 
” is a homophone of the word for “return” (kuei [gui] 
).45 And, as we have seen, the primary Sanskrit words for invasive spirit, bhūta and preta, are past participles for common verbs of becoming and forward motion.
I have presented a somewhat detailed description of this not just because I have observed it, but because it contains many of the features of possession found in classical texts. First, these texts, being classical, have as their target audience an educated elite. Pakaslahti, citing data collected by Dwyer, states that “seeking help from Bālājī is not correlated with illiteracy, lack of education, low social status, poverty or a rural background as the patients are predominantly relatively well-educated, from higher castes, middle social class, and of urban domicile.”46 Thus this contravention of norms expected from a reading of Lewis’s and other well-known studies applies not just to classical, but to contemporary possession as well. Second, many of the phenomenological features of possession at Bālājī resonate with those in antiquity. These include positive oracular possession (āveśa) and negative disease-producing possession (grahaṇa), physical symptoms that approximate textual descriptions, the development of a meaningful and meticulously described imagery on which to fix the setting of polyphonous or multitiered personalities, a devotional impetus to the process, and the engagement of possession by Hanumān and Bhairava, the two divinities directly mentioned in the only Sanskrit manuscripts that I was able to locate that deal exclusively with possession as a prescribed practice. In addition, Hanumān as lord of bhūt-prets is reminiscent of other deities who command armies or troops, notably Gaṇeśa, “Lord of Hosts,” usually of undefined character or, more appositely, of Khaḍgarāvaṇa, a minor deity often associated with possession, who is specifically called bhūtarāja (King of Spirits)47 and bhūteśaṃ (Lord of Spirits).48
Bengali
In Bengali, √bharā (to fill) is attested in the sense of possession, in addition to its expected etymological usage.49 The nominal form bhar means “weight,” hence the two meanings conflate: being filled with or bearing the weight of a spirit or deity. The extended sense of possession as intense emotional engagement is also found in Bengali: for example, karuṇā-bharala (filled with [i.e., immersed in] compassion).50
Deborah Bhattacharyya, based in Bengal, examines possession and other behavior that “specifically refers to the collapse of the boundaries between the inside and outside of the individual”51 and are treatable through folk psychiatry. She describes three paradigms of “madness” (pāgalāmi): (1) bhūt bharā (ghost possession); (2) tuk or tuktāka (black magic or sorcery); and (3) māthār golmāl (malfunctioning of the head). The most common terms she found for possession are, in addition to bhūt bharā, bhūte pāoẏā (the getting or receiving of a ghost) and the Sanskritism apadṛṣṭi (inauspicious glance, evil eye),52 which compares to the attested Sanskrit term dṛṣṭipātam.53 Although the term apadṛṣṭi is found in Bengali literature (for example, by Tagore),54 I am unable to find it used with this particular meaning in any Sanskrit text. More common in spoken Bengali for “evil eye” is bad naẓar, borrowed from Urdu.55 I suspect that apadṛṣṭi and other Sanskrit words used more frequently in modern Indo-Aryan than in Sanskrit, such as ārūḍha for possession in Sri Lanka (but unlike āveśa, which is used commonly in both Sanskrit and modern vernacular languages), are the products of local traditions of applied Sanskrit that grew up alongside the literature but are not fully reflected within it; in other words, they were clever attempts to Sanskritize the vernaculars.56 While it is reasonable that ārūḍha should denote spirit or deity possession, it is not so reasonable in the case of apadṛṣṭi. Indeed, Bhattacharyya was surprised to find it used in this sense. Before she heard it used unmistakably in this sense, she writes, “I thought that māthāra golamāḷa was a general term for madness and apadṛṣṭi one of several causal agents of madness.”57
June McDaniel notes three general terms in Bengali to indicate religious ecstasy: bhāva, mahābhāva, and bhāvāveśa.58 All of these are Sanskrit borrowings, about which I say more below. She further isolates three types of goddess possession: (1) bhar nāmā (nāmā; to descend or alight; also bhar karā, bhar hāoẏā), oracular possession, often experienced in a state of apparent pain, in which the “person often writhes on the ground, gasping and screaming”;59 (2) āveśa, which, based strictly on McDaniel’s informants in Bengal, is defined as “an emotional state arising out of love for the deity,” a “spontaneous, nonritual possession by a deity, often accompanied by visions of lights or paradises”;60 and (3) mahā-ullās (the ritual group possession of the Sakta cakra).61 The latter is voluntary and transgressive, occurring after the participants “ingest ritual substances.”62
Marathi
In Marathi āveś carries the same semantic range as it does in Hindi. J. T. Molesworth notes many compounds borrowed from or modeled on Sanskrit, probably limited to Marathi literature (they are less likely in speech, even among upper castes): kāmāveś (absorption or immersion in love), krodhāveś (anger), lobhāveś (greed), śauryāveś (heroism), piśācāveś (evil or evil spirits), harṣāveś (laughter), śokāveś (grief), īrṣyāveś (envy), kṛpāveś (a state of grace), and mohāveś (delusion).63 Unfortunately, Molesworth does not provide sources, but these appear to have been drawn from (highly Sanskritized) translations or perhaps from independent texts in Marathi on poetic theory.64 These terms bespeak the increasing use of āveś with words for emotions and states of mind that are attested by the end of the first millennium C.E. in Sanskrit as well as in Mahārāṣṭrī Prakrit. Rājaśekhara, writing in the Western Deccan around 900 C.E., has the following line in his ornate Prakrit drama Karpūramañjarī: tā vasantavaṇṇaṇeṇa siḍhilaāmi se taggadam āvesam (“Well, I must sing the praises of spring-time to him and so slacken his passion [āvesam] for her”).65
The dictionary of Old Marathi by S. G. Tulpule and Anne Feldhaus, covering texts through the mid-fourteenth century, lists three meanings under āveś/āveś: (1) force, power (jor); (2) fullness of heart (bhāvanotkatatā); and (3) wrath, anger (krodh, rāg).66 These meanings are well placed within the extended horizons of āveśa in Sanskrit literature. In spite of this Sanskritic influence in Marathi literature, the vernacular angāt yeṇe (lit. “coming to the body”) is the usual Marathi term for possession by a deity or saint.67 This is similar to the Hindi idiom for a deity “coming over” an individual: ūpar āna or ūpar honā: for example, us ke ūpar sāṃp hamesa hai (there is always a snake over him).68
John Stanley reports that the distinction between divine or desired possession (equivalent to āveśa) and forced or demonic entry (equivalent to grahaṇa or the general semantic tendency of praveśa) is maintained in contemporary Maharashtra (in other languages, such as Bengali, the distinction does not appear to be so marked). “Maharashtrian culture makes a sharp distinction between the two phenomena. Moreover, when the phenomena are analyzed as religious experience, they can clearly be seen as polar opposites.”69 Bhūt bādhā (obstruction by an evil sprit), conceptually similar to praveśa or grahaṇa, is experienced as “wrongness, chaos, disorder.”70 Conversely, “[f]or most people angāt yeṇe is immediately subjectively self-authenticating.”71 Stanley discovered that devotees of Khaṇḍobā drew a further distinction between spirit possession and infusion of power: Only the latter was regarded as a gift from the deity, and readiness for such an experience was rooted in bhakti. Indeed, bhakti-based ecstatic experience proved to be one of the symptoms of angāt yeṇe. Here we do see evidence of a nexus of ecstatic experience and possession, as Lewis presupposes; but the possession is not oracular and cannot be regarded as shamanic.
Feldhaus reports in detail on a kind of negative possession in Maharashtra of pregnant women by a group of minor divinities known as Sātī Āsarā (Skt. Sapta Apsaras) or Mālavayā (female beings who resemble ghosts [bhūts] or goblins [piśācas] more closely than they do full-fledged gods).72 In fact, they more closely resemble the skandagrahas (childsnatchers) discussed in the Mahābhārata and āyurvedic literature (see Chapters 6 and 12). In one ethnography, a woman from Pune who had experienced a number of pregnancies in which the infants died almost immediately after birth because she was “got” by these Āsarā was diagnosed by another woman from Pune in a state of oracular possession. The woman possessed by the Āsarā reported that the other woman guaranteed the birth of a son if she followed certain procedures: “Then she gave her word, in her possessed state [vāryāmadhye]. The god was possessing her [te dev vārā kheḷat hotā], you see. In that state she gave her word.”73 This is similar to the situation at Bālājī in that a possession was recognized and confirmed by another person in possession. The “cure” in this case was strictly ritual: The woman possessed by the Āsarā had to make certain offerings to them.
Possession and Its Discourse in the Himalayas
Oracular possession in the Himalayas is very common. Daniela Berti reports that in Himachal Pradesh nearly all the village temples in Kulu district have mediums who are possessed by the main temple deity.74 The situation appears to be little different elsewhere across the Himalayas, from Ladakh to eastern Nepal. This is equally the case in Tibet, as Per-Arne Berglie notes: “Even more than the local lama … the dpa’ bo [spirit medium, lit. “hero”] acted in the centre of the religious life of the village.”75
Smriti Srinivas demonstrates that possession among the Buddhists of Ladakh is an integrated activity in that its social, linguistic, psychological, and ritual components operate together. Because personhood is a reflection of these variables—it is an interdependence of social relationships, language use, culturally conditioned psychological states, inherited religious ideology and practice, and the vagaries of individual physical and emotional constitution—personal behavior is theoretically unstable, thus constantly in need of cultural sanction and legitimation. Within the Ladakhi cultural system, which includes social, economic, and political arrangements between Buddhists and Muslims, possession is a culturally sanctioned and recognized form of unusual, manifestly unstable, behavior that ironically serves to reverse, albeit temporarily, areas of perceived weakness or imbalance such as social or political oppression, psychological distress, or even physical debility.76 Although possession is understood in Ladakh to originate in weakness, it nevertheless occupies an important mediatory function. It does not fit neatly into either of Lewis’s categories of “central” or “peripheral” possession; rather, it combines elements of both. For example, it resembles peripheral possession in that the majority of mediums are females who are denied access to positions of power through the primary sociopolitical processes. It is central in that it is so much a part of public culture that it upholds the traditional power structures, in spite of its covert manipulation of certain mundane events. It is also central because it involves possession by Tibetan Buddhist deities with strong shamanic elements. Thus, although Lewis’s categories may be helpful in thinking about Ladakhi and other Himalayan possession, they are too categorical to be productive here.
Srinivas’s primary focus in her discussion of possession is a Buddhist monk called Gelong, who also served as a priest in one of the villages in Nubra, the region of Ladakh in which Srinivas conducted her fieldwork. Gelong was possessed by a local lha called Chamshing, a protector deity (dharmapāla) who, through many rebirths, during which he conquered his anger, rose to an exalted state from that of an enemy spirit. Possession occupies a high profile in Ladakh, not least, according to Gelong, because, after the Chinese occupation of Tibet and the subsequent destruction of the temple images in which the deities were embodied, the Tibetan Buddhist deities themselves “fled to Ladakh to seek their residence in people’s bodies.” According to Gelong’s divination, “seventy-five lhas would come to reside in Nubra (at the moment there are twenty-one), exiled from their former homes.”77
Ladakh is unquestionably a remote area, yet a large number of languages are spoken, including colloquial Tibetan, the Ladakhi dialect of Tibetan, Hindi, Urdu, Kashmiri, and local languages, such as Dogri. Although the language of popular religion in this area requires further elucidation, especially in the highly nuanced area of code-switching, the terminology for spirit possession appears to be derived from Tibetan and Indo-Aryan. The general term for a spirit or deity is lha, which may be benevolent or malevolent in nature. The possessed individual is known as a luyar, or vessel for a lha.78 Malevolent possession is caused by lha known as dre or temo, the latter being an evil female spirit. Srinivas identifies four types of ritual practitioners who treat this type of spirit possession: the lama, who “ensures peace and can help a man’s spirit reach heaven”; the onbo, who “removes evil and assists in securing earthly and other benefits”; the lhaba, who “sees the spirit of a man and its imaginings as if in a mirror”; and the larje, who “takes care of the body.”79 Possession is caused by debilitated life-force (la), which sustains the body (lus). In “negative” possession, vitiation or loss of la is conducive to lassitude, physical weariness, emotional distress, and illness.80 These, in turn, lead to a more immediate cause of possession, called mikha in Ladakhi (lit. “mouth” < Skt. mukha). Mikha refers to that which emanates from the mouth, specifically gossip, derogatory or slanderous speech, and idle or useless chatter, which, consistent with Buddhist doctrine, are considered counterproductive to religious or spiritual aims.81 Oracular or “positive” possession, however, is due to weakened la, which leads to a kind of madness characterized by loss of self-control and spiritual crisis. An individual subject to this type of possession undergoes an initiatory life-crisis ritual called lhapchok, in which the newfound status of oracle becomes formally recognized. The oracle, who can inherit its lha from lineages of practicing oracles, including parents and grandparents, now undergoes a period of training including exorcism and sucking out poison, all this while leading an otherwise ordinary life as a farmer or government worker.82
Among the many studies of Nepali possession and shamanism, perhaps the most comprehensive, particularly from the point of view of rescuing the terminology, is András Höfer’s transcription, translation, and lengthy discussions of oral texts of the Tamang shamans of central Nepal, which are sung in both Tamang and Nepali.83 Tamang is a Sino-Tibetan language, and, though it is related to Tibetan, its speakers are resistant to Tibetanization. The shamans (bombo < Tib. bon-po) represent what Höfer calls “accommodated heterodoxy,” a term that reflects the bombo’s rejection of the stigma of heresy, ascribed to them by the lamas of Tibetan Buddhism, who, Höfer believes, assigned to them the designation bombo because of their apparent similarity to the priesthood of the Bön.84 Höfer presents several terms for kinds of ecstatic experiences, many of them symptomatic of possession. A bombo quivers inside the body (sīṇsiṇ-khòlkhol), “something like the sensation you have when urinating,” reports one informant. The bombo experiences a shamanic trembling or quaking (chyèkpa) that indicates he has been “seized by the god” (lajye cuṇba). When this occurs, the deity, usually a goddess (mài, Nepali māi, māiju) or virgin (Nepali kanyā, kumārī), descends onto one’s back or into one’s body (Tamang: gori yùba; Nepali: sir/kum caṛhnu [mounting one’s head/shoulders]; āṅgmā caṛhnu [to mount the back]). Höfer cautions that having a “superhuman agent on one’s back or in one’s body” does not fully correspond to possession. The latter occurs when the bombo becomes a mouthpiece for the deity (vide Srinivas). In this state the deity plays (khelnu [to play]; khelāunu [to make play]) in the body of the bombo. In the state of gori yùba, the bombo attains partial identity with the deity, “an interpenetration, rather than a fusion, of identities. Informants likened this relationship between the bombo and the superhuman to that between husband and wife.”85
An excellent study of the interaction between Tibetan Buddhism and local shamanistic religion in Nepal, in which possession looms large, is Stan Royal Mumford’s Himalayan Dialogue. This volume draws its theoretical inspiration from Mikhail Bakhtin’s model of dialogic interaction, in which the temporal processes of cultural production are located in a relational “betweenness” characterized by elements of earlier layers of discourse becoming sedimented in later ones, even (or especially) in discourses that rivaled and eventually achieved hegemony over these earlier ones.86 This intersubjective model shares conceptual territory with Marriott’s notion of substance-code in that they both highlight a self that is decentered—or perhaps better is multicentered or polyphonous—and is at odds with notions of autonomous selfhood or identity, whether Hindu, Buddhist, or Judeo-Christian. This is helpful to the present study and would be even more so if the Sanskrit texts were more straightforwardly descriptive or if survivals from earlier vernaculars were more numerous.
An example of Mumford’s ethnographic application of Bakhtin’s largely literary use of dialogic interaction87 can be seen in Mumford’s analysis of competing attempts to exorcize suspected evil spirits from a dying twenty-three-year-old girl, possessed, according to the villagers, by an old woman identified as a witch, who transferred the spirits to the girl through a glass of milk.88 The local Tibetan lama was called in first, to perform an exorcism called “the evil spirit ransom” (’dre glud), in which a chant “calls down weapons like hailstones which will reduce the black sorcerer to dust.” After this failed, the government health assistant was summoned and proceeded to pronounce the girl beyond recovery. The last resort was the local paju shaman, a specialist in exorcisms, whose mantras and heavy-handed actions, including beating the victim about the head with a burning broom and a belated attempt at “red offering” or animal sacrifice, also failed. Indeed, so little confidence did the lama have in the paju’s ritual that he broke into it, whispering sections of the Tibetan Book of the Dead into the dying girl’s ear.89
Mumford comments that both lama and paju agree that “when established human rules are violated, the breach of social harmony brings demonic attack” (a view with strong resonances throughout South Asia).90 In spite of this agreement, the paju takes a dim view of the lama’s attempts to distance himself from personal responsibility in the violence of exorcism by calling in intermediate spirits for this work, as well as his feeble attempt to liberate the spirit. “I only try to kill the demons, there is no thought of liberating them,” says the paju.91 Mumford fruitfully employs Bakhtin’s notions of dialogic interaction to unpack the dialectics and historicity of ethical comportment, ritual (especially exorcistic) practice, and models of hierarchy, retribution, death, and transmutation. No reportage of such immediacy appears in Sanskrit literature, but the lesson is apposite: We must listen for traces of one community’s speech, discourse, and doctrine embedded in those of another. In this way, it may be possible to expose cracks in the surface and break through the canonizing agendas of much of Sanskritic thought.
Berglie has observed possession among Tibetan spirit mediums (dpa’-bo) in Nepal.92 His informants all explained their possession in characteristic Tibetan Buddhist idioms. First the “consciousness” (rnam shes) was sent away, meaning that the mind was emptied. It is important that the medium’s body be completely empty of his or her rnam shes; to be “half god, half man” was an untrustworthy state. This rnam shes was then placed under the care of Padmasambhava (Tib. Padma ’byung gnas) or a protective deity called mkha’ ‘gro Ye shes mtsho rgyal,93 while any one of a number of other deities protected and controlled the proceedings. The possessing deities then enter the body of the dpa’-bo through one of the main channels (Tib. rtsa, Skt. nāḍī) of the subtle body, though the informants differed as to whether the deities themselves entered their body or merely the deity’s rnam shes. During the ritual, mirrors are placed on a nearby altar, because it is in these that the deities are thought to reside during the séance (recall Srinivas’s observation that the ritual specialist called lhaba sees a person’s spirit “as if in a mirror”).94 Two experiences of Berglie’s informants are as follows.
When all the gods summoned have arrived, possession took place by the god most suited to carry out the task of the evening. dBang phyug [one of the informants] said that he then saw all the colours of the rainbow in the mirrors and that all became very bright when the gods arrived. When the god was to take possession of him, his body felt enlarged and as if filled with air. Sri gcod [another informant] said that when his rnam shes was going to leave his body, he saw a glowing fire of many colours. This fire grew bigger and bigger and finally entered his body, as it were. After that, everything became black and he remembered nothing further. At the end of the possession all this happened in the reversed order.95
Berglie records very little of the terminology of possession, though he does record ’bab pa (coming down), which he says is a Tibetan translation of the Sanskrit āveśa.96 This apparent paucity of terminology is not reflective of the actual situation in either literary or colloquial Tibetan. As discussed below, the process of possession found there is remarkably resonant with what is found in texts on classical yoga.
Another noteworthy study, albeit somewhat undisciplined from the perspective of contemporary anthropological discourse, is Angela Dietrich’s Tantric Healing in the Kathmandu Valley.97 This valuable resource extensively records the contemporary vocabulary of healing and spirit beings (unfortunately without diacritics), as well as attitudes about these spirit beings by both healers and clients alike. Dietrich vividly portrays possession as one of the primary modes of both disease production and healing and notes the ubiquity of oracular possession. Indeed, as reported by others, the spirit world is portrayed as alive and well in the Kathmandu Valley.98
Tamil and Simhala
In Tamil and Simhala,99 āvēcam (< Skt. āveśam) is perhaps the most common word for spirit or deity possession. Like other South Asian languages, Simhala betrays a linguistic nexus between possession and emotion: the substantive ävitu (possessed [by an evil spirit], mad) is related to the verb ävissenavā (to be angry).100 Obeyesekere, however, recognizes a much greater frequency in colloquial Simhala of ārūḍe (< Skt. ārūḍha [mounted]), divine possession, “in which the magnetism of the god infuses and suffuses the body.”101 This is distinguished from the equally common preta dosa (misfortune caused by an evil ancestral spirit or preta).102 The definitions of āvēcam in J. P. Fabricius’s Tamil-English dictionary are: “1. spectre, ghost; 2. the entrance of a demon into a person for the purposes of uttering oracles; 3. the fury or madness of a possessed person; 4. the grimaces of a person in ecstasy, religious frenzy.”103 The Tamil verb pukutal (to enter) is used in Tamil poetry where ā√viś is found in Sanskrit or Sanskritized Tamil. For example, Tiruvāymoḻi 9.6.3 reads, “With gentle grace he conquers, and enters my heart” (nīrmaiyāl nencam vancittu pukuntu), and 9.6.5 reads, “As if he were going to show his grace, / He entered into me” (tiru aruḻ ceypavan pōla ennuḷ pukuntu).104 On the word aruḻ in the Tamil Śaiva texts, Handel-man and Shulman note:
an unfortunate tendency to translate this critical term, in nearly every instance, as “grace,” with its heavy Christian connotations. Aruḷ can, it is true, correspond in Śaiva texts to Sanskrit anugraha, the god’s compassionate giving to his servants. More often, however, it approximates a notion of coming into being or freely becoming, present, close, alive. In village ritual contexts, aruḷ may mean simply “possession”—a presence intensified beyond bearing—or the experience of the deity’s true (nija, aghora) form.105
We examine this notion of possession as “coming into being” when discussing possession as a mode of devotion (Chapter 9).
Hiltebeitel records the Tamil substantive maruḷu (delusion, bewilderment), used for Asvatthaman’s murderous possession of the young Pañcapāṇḍavas in the folk Draupadī ritual.106 The use of this word for negatively charged possession distinguishes it from āvēcam. Hiltebeitel raises, without much discussion, the intriguing question of a connection between death, sleep, trance, and possession,107 a subject on which the Sanskrit texts are equally silent. Hancock notes a distinction (that I have not found in any other study) between āvēcam and other modes of deity or oracular possession. According to one of her informants, āvēcam refers to a general state of possession, while a “specific, temporary transformation” would be referred to with the observation “the ammaṉ has come” (ammaṉ vantatu).108
Another scholar with incisive thoughts on possession, and who has produced one of the most language-sensitive studies so far, is Eveline Meyer, who observed possession in the cult of the goddess Aṇkālaparamēcuvari in Tamilnadu.109 Meyer finds an ambiguity between possession and impersonation: Sometimes a man representing the goddess is possessed; at other times he does not achieve possession but completes the ritual by acting the part of the goddess. The latter is impersonation, asserts Meyer, because it is not considered to be genuine possession even by the participants. The possession about which she writes occurs within a non-brahman community, and it is important to note that both the culture and language in Tamilnadu are sharply divided between brahman and non-brahman.110 Like Hiltebeitel, she notes the word maruḷ, “used to describe the power and frenzy which takes hold of the pūcāri [Ta. pūjāri].”111 To be possessed is to “run frenzied” (maruḷ ōṭa).112 Nabokov, too, records the phrase maruḻ oṭutul (running in a state of confusion or frenzy), which she defines more strictly than Meyer as ceremonial flight after decapitation of a chicken that is intended to send a pēy (spirit) fleeing. Another form of possession is “dancing the god,” because of certain patterned or even choreographed dancing during deity possession.113 The Tamil term is teyyāṭṭam (god-dancing), while the Malayalam cognate is ttmyāṭṭam (festival dancing) or teyyam tuḷḷal (divine ballad/frenzy; Ta. tey, Ma. teyyam < Skt. daivam).114 Nabokov records the Tamil phrase kōriccu varutal (the request to come), corresponding to the establishment (Skt. pratiṣṭhā) of a deity in a shrine, examined below. This Tamil term refers to an “investiture ceremony” in which the spirits of dead relatives are installed “in the domestic sanctuary of the household as tutelary deities.”115
Possession of the deity Meyer studied was reported in the Indian press a few years ago. It is of interest that the Sunday Pioneer, a New Delhi newspaper, featured a front-page article headlined “THEY DIG GRAVES, CHEW HUMAN BONES & DANCE THE NIGHT” (March 4, 2001), which described “a frenzied procession to the graveyard where three chosen ‘incarnations’ of Goddess Angala Amman fished out the bones of the dead and chewed them up at midnight” on the night of Mahāśivarātrī (February 21, 2001). Leaving aside the obvious sensationalizing and orientalizing of this English-language report of a ritual among the relatively well-off Manai Telugu Chettiar community, who constitute 80 percent of the five thousand residents of Sundakkamuttur Village near Coimbatore, where the story was situated, the article reveals that there is a renascence of this sort of possession, at least in that village: “The annual event is a hit with the younger population. Surprisingly, the elders here have, of late, been trying to do away with the primitive ritual, but it’s the growing generation that has taken it up almost as a mission.” According to the reporter, R. Revathi, “A visit to the houses of a few bone chewers shows that they are literate and come from affluent families who have modern gadgets like television, refrigerators, washing machines, etc.” One of the bone chewers, eschewing the suspicion that they are tribals or cannibals, says: “We have Ministers and MLAs [Members of the Legislative Assembly] from our community.”
According to S. M. Chandrakumar, an advocate who has previously acted the role of one of the incarnations, “It’s our custom. In ancient times, an asura did tapasya to propitiate Lord Shiva. Gods, like Indra and Varuna, went to Shiva and asked him not to grant him a boon. But Shiva, touched by the asura’s devotion, granted him a boon of immortality. The asura then went on a rampage and killed many in Devaloka and threatened the munis. The sadhus complained to Parvati. But Shiva could not take back his boon. So he gave Parvati powers to eliminate the asura. Parvati came to Earth as Angala Parameswari Amman. She chased the asura who took refuge in a graveyard. To escape her fury, the asura entered a corpse. In order to kill him, Angala Parameswari Amman dug up every corpse in the graveyard. She ate up the flesh and chewed the bones and tied some around her waist and danced the ananda tandav. ‘We are the children of Amman and we’ve been doing this puja from time immemorial.’”
A comparative study of goddess worship in Tamilnadu and Orissa was conducted in the 1990s by Lynn Foulston.116 She cites the Tamil verb iṟaṇku (to descend), following Diehl, indicating the descent of a deity on a cāmyāṭi (see sami ati and cāmiyāti below) (god dancer).117 This, in its festival context, is a kind of positive oracular possession. Foulston notes several important facets of possession in the communities under study, including the honor and status accorded to those who enter into possession (mostly women, but also, on occasion, male temple officiants) and the imperative for the deity to possess the devotee. In Khurdapur, near Bhubaneswar in Orissa, “there is a strong indication that the goddess [Santoṣī Mā] must respond to the pūjāri if he approaches her correctly.” However, in Cholavandan, near Madurai in Tamilnadu, “possession was entirely dependent on the will of the goddess.”118 Among the purposes of possession, she notes, are problem solving, healing, and for the goddess to exercise her power, to demonstrate to her devotees that she “is a stable, constant force in an ever-changing and unpredictable world.”119
Elizabeth Fuller Collins, an anthropologist who studied possession at an important Hindu Tamil festival for Murugaṉ in Penang, Malaysia, called Thaipusam, also recognizes indigenous terminology. Collins observes that the Hindu Tamils of Penang recognize three categories of trance: sami ati (the possession of a pūjārī by a goddess or warrior deity); pey pirichu (lit. “having a ghost”; possession by ghosts or demonic spirits); and aruḻ (the trance of divine grace experienced by those who fulfill vows during religious festivals).120 Nabokov provides specificity to certain possessing spirits. The pēy “is the spirit of a man who took his life because he was prevented to wed, or because his marriage did not work out, and who henceforth ‘catches’ other men’s wives.” Nabokov understands this to symbolize the “obstruction and desecration” of marriage. “The pēy,” she explains, “represents a generic timeless entity, a ‘departed’ who cannot grow.”121 Although these representations may be valid in the community in which Nabokov worked in South Arcot district, it is likely that among the Tamils of Malaysia pēy is a general term for spirit, reflecting its Sanskritic origin, the word preta. Nabokov writes extensively of the pūvāṭaikkāri (woman who wears flowers).122 These are good spirits, who may be male or female, or even an indistinct collocation of deceased members of one’s lineage, despite the literal sense of the term. In a general sense the pūvāṭaikkāri is a deceased human spirit who returns as a god. He or she is invited to enter certain mediums during pūvāṭaikkāri pūjās, held to bless weddings, harvests, plantings, and other auspicious occasions.
Karin Kapadia has also studied possession of Murugaṉ in Tamil festivals.123 She focused her study on body piercing and the apparatus of ritual. Kapadia’s findings are very much in keeping with what is seen elsewhere, some of which we expand on later, when dealing with bhakti more formally. The word that Kapadia found used for possession among the largely middle-ranking non-brahman Tamil castes she studied was puḍiccikiraḍu. She arrived at two general conclusions: first, that devotees who experience possession are mostly female and that men in possession “become” female for the duration of the ritual, and, second, “that the lower castes do not share upper-caste assumptions regarding ritual purity,”124 an argument against the well-known assertion of Dumont and others.125 Most striking, though, is Kapadia’s observation that the social order is preserved in these rituals because they reinforce the notion that “only men can be ‘complete.’”126 By this, Kapadia means that men can become Ardhanarīśvara, Śiva who is half-man and half-woman, while women cannot assume this state. This “ritual androgyny of men”127 in possession empowers them to manifest female power (śakti), whereas women cannot manifest male power, which resides in stillness and inactivity, which Kapadia sees as the expressions of wisdom and knowledge inherited from brahmanical (viz., Sanskritic) discourse. Kapadia wisely does not attempt to generalize this observation to all of India. The point, however, as Hancock noted earlier (Kapadia seems to be unaware of her work),128 is that possession, transgressive and liberative though it surely is, reinforces key elements of the patriarchal social order.
It is striking that the psychological experience of possession of Murugaṉ, if not its ritual form, is similar to that expressed in the Akanāṉūṟu, a collection of Tamil poems dating from the first to third century C.E. George Hart notes that even at that time Murugaṉ was a deity popularly invoked for possession. Hart cites an early Tamil poem describing the dance of a shaman called in to treat a girl possessed by Murugaṉ: “It is the presence of Murugaṉ, hard to bear.”129
One final ethnographic study of Tamil possession deserves mention. This is a deeply moving study by Patricia Lawrence of the role of Tamil-speaking oracles in the recent Sri Lankan civil war.130 Lawrence’s study took place in the eastern part of the island, largely in Batticaloa district, perhaps the least-reported district of the war. “In eastern Sri Lanka,” she writes, “institutionalized possession is incorporated into the structure of pūjā ritual at temples for all local goddesses.“131 The civil war has brought increased public involvement in goddess cults, particularly rituals for Kālī that feature possession, blood sacrifice of goats or chickens (or ash pumpkins, their substitute), body piercing (especially of the cheeks with silver needles or triśūlas, the trident of Śiva, or the back with multiple fishhooks). At the time of Kāḷi’s annual festival in June, thousands of men, women, and children walk across burning coals, led by a charismatic “uttering oracle” (vākku solluṟatu), a possessed woman speaking with Kāḷi’s voice. All of this is performed with the ritual intention of healing the country’s ethnic wounds.
Lawrence makes an important contribution to our understanding of oracular possession, at least in the context of the torture, genocide, and horror that has been widely reported in this civil war. This is the element of pain. In addition to the bliss and heightened awareness of oracle and audience, there is an accompanying presence of pain. “Oracles describe śakti possession as a painful sacrifice of the body, reporting feelings of unbearable heat and a trembling like the ‘shivering when one suffers from fever.’ They may also, particularly if embodying the agency of Kāḷi, describe an overwhelming feeling of āvēsam,” which Lawrence renders as “uncontrollable emotion, fury, wrath.”132 The oracles, who have experienced the same torture, abductions, and genocide as practically all other Tamils in the region, employ possession as a therapeutic tool: as a special bonding strategy that embodies and renders sacrosanct their collective memory, translating their unspeakable suffering into a discourse of communal healing that undermines the discourse, and therefore the power, of the forces that have striven to strip them of their very humanity.133
Turning for a moment to classical Tamil, Māṇikkavācakar, in the ninth century, used the verb āṭkoḷ (to rule for oneself; lit. “to take to oneself in order to rule over [a person]”).134 This verb is usually translated “to enslave,” “accept as one’s slave,” “take possession of,” or “possess,” as noted by Yocum. Forms of this verb are not cited in the ethnographic studies, thus it must be assumed that at some point it fell out of favor in this context. Mention of possession and ecstatic dancing is found even earlier in Tamil, in the Caṅkam poetry of the early centuries of the common era.135 Perhaps the most richly detailed depiction of possession in early Tamil texts is in the well-known fifth-century epic drama Cilappatikāram. At the end of the drama the oracular possession of a brahman woman named Tēvantikai by the god Pācaṇṭaṉ is described. She was not alone, however, as the early Tamil texts frequently mention the possession of women. Indeed, for the realization of his own spiritual and religious instincts, it was not a problem for the male Mānikkavācakar to think of himself as a female devotee, subservient to Śiva.136 It is this sort of phenomenon that has led many scholars, such as Yocum and Friedhelm Hardy, to posit a movement of possession and ecstatic devotional madness from south to north India.137
The Yakädura of Sri Lanka
Paul Wirz, Bruce Kapferer, David Scott, and Beatrice Vogt,138 in chronological order, provide deep, vivid, and often startlingly different accounts of Sri Lankan exorcists. These are called ädurā (Wirz: edura) or kaṭṭāḍiya (Wirz, also Obeyesekere: kattadiya), though the more proper rendering is yakädurā.139 Scott defines ädurā as a “practitioner of the arts of controlling the malign figures called yakku” (Skt. yakṣa),140 avoiding what he rightly regards as colonialist terminology (e.g., the more literal “demon master”). Another ritual specialist called bandhanaya (Wirz) addresses afflictions caused by planetary influences (grahacāraya) and black magic (bandhena). The bandhanaya confines and disarms demons and other evil forces responsible for these afflictions based largely on astrological remedies.141 The baliädura (Vogt) “eliminates disturbances caused by gods (deva) by means of an offering (bali) and other acts performed together with the patient.”142 Vogt classifies healing systems in Sri Lanka into five groups: (1) Western or allopathic doctors; (2) vedarala or āyurvedic doctors; (3) yakädura, who are versed in bhūtavidyā (see Chapter 12) and employ tovil or yaktovil, a wide range of healing rituals that are captivating and artistically integrated, and that may include masked dances, drumming, pūjā, didactic songs, and theatrical scenes, some of which are comical and contain ritually regulated dialogue; (4) śāstrakāraya, a group that includes astrologers and fortune tellers; and (5) bhikkus or Buddhist monks.143
I limit my comments here to Vogt’s study, for five reasons: first, because it is the least known of them; second, because she is a trained psychologist whose eye is different from the others’; third, because it lacks both authorial inclination and disciplinary pressure to anchor her study in postcolonial, poststructural, or Freudian positionings, a freedom, it seems to me, that is a refreshing characteristic of continental European scholarship; fourth, because her findings are in substantial agreement with Kapferer’s sophisticated views of the ritual, healing, and socializing processes involved in yaktovil;144 and, fifth, because she grapples seriously with the relationship between indigenous healing practice and Buddhist doctrine and belief. Indeed the latter is the crux of her study and is thus central to the project of investigating links between modern practices and premodern epistemological patterns and between folk and classical discourse. Vogt describes her project as “ethnopsychological research (1985–1988) on a healing ritual that is traditionally performed in the Kandy highlands as a treatment for mental illness.”145 She apprenticed with a yakädura called Upasena Gurunnanse, but conducted interviews with a substantial number of healers and patients in order to understand the healing process in afflicted individuals.
Through correlating specific material in canonical and semicanonical Buddhist texts with the statements of interviewees, Vogt strives to show that the basis of the nonmedically trained healing practitioners of Sri Lanka is canonical and historically situated Buddhism. This may not seem remarkable given that Sri Lanka is largely a Buddhist country. However, it is surprising given the attempts of anthropologists to assign the background of the healing process to other factors. The latter is in step with the model of many Indic scholars who separate healing ritual from canonical Hinduism, assigning it, instead, to other indigenous forces. On the whole, as Vogt is aware, and as is also argued in the present volume, most ethnographers are untrained, largely because of confining disciplinary boundaries, in classical and ancient Indic texts, and as a result are uninformed of the possibility and dynamics of deep cultural and structural continuities.146
Eschewing attempts to dichotomize Sri Lankan culture between high and low or Buddhist and non-Buddhist, in which the latter is disparaged as “folk religion” or pre-Buddhist animistic cult practice, distinctions that Vogt sees as the relic of Christian scholarship with its own history of forging these distinctions in Euro-American culture in order to rescue a “true” Christianity, Vogt instead sees the healing process as practiced by folk healers in Sri Lanka as an integral part of the local construction of Buddhism. I argue below that the distinctions Vogt questions are as untrustworthy in India as Vogt cogently argues they are in Sri Lanka. Although I might take issue with some of Vogt’s psychological theorizing, because the Jungian and other non-Freudian notions of “psychodrama” on which she bases her study are as underdeveloped as Freudian theory is overapplied (on the latter, see Chapters 2 and 12), on the whole I agree with her conclusions.147 Vogt states: “The analysis of terms used in the psychology of tovil underline my fundamental recognition that tovil healing rituals are a form of Buddhist psychology.”148 As a psychologist looking meticulously at Buddhist psychological categories, Vogt argues that such ritual is as firmly grounded in Buddhism as Western psychotherapy is grounded in Christianity. To this end, she marshals considerable evidence from Buddhist precepts as well as from abhidharma, the Buddhist analysis of personality and mental processes. She notes, for example, that Upasena Gurunnanse describes his method as, among other things, “bringing the demons under the control of Dhamma,”149 then dealing with five components of the personhood (pañcaskandha) of the demon (yaka [yakṣa], bhūta) as well as with the kamma (karma) of both patient and invasive being. She also shows convincingly that the mythic and cosmological beliefs of the Buddhist patients who came to the yakädura with whom she studied in the Kandy highlands were consistent with those of the Mahāvaṃsa (mid-fifth century) and other “histories” that have been transmitted in both literary and oral form for nearly two thousand years.150
Vogt also carefully avoids characterizing tovil as exorcism. The yakädura are not “spirit conjurers,” a term that she associates with exorcists, but that has been applied misleadingly to yakädura based on precedents from the Catholic Church. Although it appears that she unnecessarily limits the realm of exorcists and exorcism in her laudable goal of recognizing and setting aside the relics of a Christian mindset in relevant scholarship, especially in the field of comparative religion, she nevertheless offers a perspicacious analysis of the skill (upakkrama; Skt. upakrama) of the Sri Lankan healer.151 In this analysis, the
healer is characterized by: 1) concretizing psychic problems as “demons”; 2) re-educating these demons and relegating them to their place within the psychocosmos of the patient; 3) establishing an interpersonal consensus with regard to these demons or the psychic problems personified by them; and 4) therapeutic-pedagogic guidance of the patient by means of communicating to him patterns of knowledge and action that will make possible a happier way of life.152
With respect to the second of these points, which has been briefly addressed above, Scott emphasizes that “the first thing the ädurās have to establish is whether the virtuous commands (aṇaguna) of the Buddha are accepted as valid by the yakku (and the other malign beings—prētayo, bhūtayo, and bahīravayo).”153 That this process resembles the process employed at Bālājī in Rajasthan is simultaneously striking and unsurprising.
Malayalam
Scholarship on possession in Kerala is now emerging, thanks in large part to the work of Sarah Caldwell, Rich Freeman, and Gilles Tarabout. Caldwell explores oracular possession among worshippers of Kāḷi. She writes about the performance of mudiyettu (divine offering),154 in which the goddess Kāḷi battles the demon Dārika. Caldwell notes that Kāḷi “will enter the actor’s body and consciousness as a great inner heat”155 and reports an informant saying, “We fix the picture of Bhagavati in our mind, and her supernatural power (caitanyam) enters our body.” Freeman asserts that the tradition of spirit-possessed dancers and priests in Kerala dates back two millennia, before Malayalam had separated from the linguistic and political domain of Tamil Caṅkam culture.156 Observing the teyyam complex in Kerala,157 Freeman demonstrates that ritual structure allows for spontaneity and improvisation, even one that contains stock phrases and memorized text.158 He insists that ritually circumscribed possession experience exhibits learned, context sensitive encoded gestures and behaviors that are not invariant, but permit a degree of improvisation. This alludes to the issue of “faking” possession, which has been the subject of considerable indigenous discussion and which Meyer generously labels impersonation. Tarabout has written a comprehensive study of sacrifice and gift in Malabar festival life, a study that includes major sections on oracular possession.159
Freeman also reports the Malayalam vocabulary of possession:
Whatever the type of being and circumstances may be, the onset of possession (āvēśam) is described as a “coming to” (varuka), “joining” (kūṭuka), or “mounting” (kēruka) of the body (meyyal). In this state the host is said to be “illumined” (veḷiccappeṭuka), “fixed” or “in-dwelt” (uṟayuka), to be “internally moved” (iḷakuka), and possessed of “spiritual vision” (darśanam).160
Freeman adds that the state of mind during the performance of teyyam “entails no loss of consciousness, or ‘dissociation’ in psychological terms, but rather a heightened sense of consciousness” in which “one’s consciousness has not traveled somewhere else, shaman-like, but that instead, one’s own body and mind are taken over and animated by a higher and more powerful (and I would add more concentrated) form of consciousness.”161
Caldwell explains that in Malayalam tuḷḷal “is a jumping, hopping dance undertaken in a fit of possession, fever, or madness.”162 It is difficult to ignore the confusion, ambiguity, or conflation of possession, fever, and madness in the eyes of Caldwell’s informants, a situation that haunts many ethnographers and surely skulks in the shadows in the present study. Nevertheless, her informants deploy Sanskrit terms when drawing broad categorical distinctions: divine possession of men is labeled dēva āveśam (which Caldwell renders “divine inspiration”), while spirit possession is called bādhā āveśam (demonic inspiration).163 Stuart Blackburn, too, is sensitive to the distinctions in possession that are revealed in the language. He reports that the general term in Malayalam for spirit possession is kōmaram; unwanted possession is rendered pēy piṭikkatu (caught by a malevolent spirit); and in a festival context a beneficent visitation is called cāmi mēl varavattu (the god descends). “False possession,” he says, is poy āṭṭam, while a “fierce dance” is payaṅkara (< Skt. bhayaṅkam) āṭṭam.164 Thus, Malayalam appears to have an extensive and highly developed vocabulary for possession.
Phillip Zarrilli, in his study of kaḷarippayattu, the indigenous martial art of Kerala, concurs with this observation of performed possession and extends it to the performance of martial arts. According to martial arts practitioners in Kerala, the essential elements for raising the kuṇḍalinī śakti are “self-confidence, doubtless heroism and internal fury.” This kuṇḍalinī śakti is “a state of intense concentration of energy (aveshakaram), ‘the power generated from concentration.’”165 Two points are significant here: first, possession can be a deliberately induced performative state; second, it commences with the identification of a specific emotive energy with which the practitioner is familiar, in this case, fury. Through repeated practice, the martial artist gains the ability to immerse the mind fully in this emotion. One of the qualities of this immersion is a condensation or compression of energy, rather than its dissipation in a multiplicity of thoughts or other forms of outward-directed activity. The maintenance of this immersion and subsequent compression during activity results in an attentive state of potentiality, power, excitement, and readiness that are observed features of the martial artist, the teyyam performer and, indeed, of anyone in a state of positive or oracular possession. It is thus appropriate that the Malayalam discourse of kālarippāyattu should employ the term āveśakāmm to signify this. These observations serve us well in the ensuing discussions of possession in “Sanskritic culture.”
Tulu and Irula
The recent major studies by Honko and Brückner of possession cults and epic performance in Tulunad in South Kanara District of Karnataka represent culminations in a long scholarly interest extending back to the Abbé Dubois in the early nineteenth century.166 The spirits invoked in the performance of the Siri epic are benevolent, which Claus points out “may be hard to believe” given “the incredible discomfort and agony evident in the bodily contortions and facial grimaces of the possessed young women.”167 Because these possession festivals (būta nemā/kolā) are similar to Teyyam in both Kerala and Coorg District of southern Karnataka,168 it is not surprising that in some cases the Tulu vocabulary was influenced by Malayalam. For example, Claus reports the phrase “darṣana169 had happened to so and so” (arugu darṣana ātuṇḍu). This is one of dozens of words in Tulu for spirit possession, which is not surprising given the Tulu recognition of the pervasiveness of spirits—evident from the studies of Honko, Brückner, and Claus. Claus reports the common expressions are bhūta pattuṇḍu (the spirit caught) and mayṭu battuṇḍu (came into the body),170 and Brückner reports nuḍikaṭṭu (oracular speech).171 Thus Tulu, barely a literary language, shares in the same locutions for possession found throughout South Asia.
Typically shamanic oracular possession is also found among men of the minimally Hinduized Irula of the Nilgiris in southern Karnataka. Zvelebil notes, “Jāya [induced possession] is performed when the Irulas need divine guidance and help to solve a personal or public problem—whenever they want to obtain any assistance from the deity. The process is termed joga ugādu (inviting god [to descend and possess]) or jāya. According to Zvelebil’s informants, the deity “then enters the heart [īruvatte] of the possessed person [termed jāyakāra], which trembles gïdigïdigïdigïdi in an increased heartbeat rate, and it changes [māṟi] his mind [ge ṇa] so that it can speak through the possessed person.” The languages uttered are not always intelligible, but are thought to be either Kannada or an archaic variety of Irula. The possession trance is induced by drumming, swaying, and singing a “deity-inviting song” (dua uga pāṭṭu) in which the deity is addressed as “one who comes riding a black bee or dragonfly (dumbiyānē).”172 As in Tulu, what is noteworthy is the striking similarity of rationale, ritual strategy, and verbal imagery to possession noted throughout the subcontinent.
Assam and Possession at Kamakhya
In view of its proximity to Bengal, Nepal, Tibet, and the rest of northeastern India, the presence of a large (and largely unexplored) indigenous Tantric literature, and a predominance of worshippers of exotic goddesses, Assam might be expected to evince a substantial amount of possession. I have no doubt that this is the case. However, this has been almost entirely unresearched. What little is documented, however, is intriguing and requires a brief explanation that contextualizes this absence in terms of the political and religious exigencies of colonial (or colonized) Assam.
Above I cited Ziegenbalg’s early eighteenth-century attitude toward the “heathens” of Malabar and the festival participants in possession there whom he labeled “devil dancers.” Furthermore, we saw the continuity (or reappearance) of Ziegenbalg’s attitudes nearly three centuries later in a few instances of postcolonial Protestant Orientalism. In the present instance we are able to see conflicting attitudes among the British toward festival possession in Assam. Two studies of Naga “tribes” were carried out from 1912 to 1920 by J. H. Hutton, a young Oxford-educated ethnographer and civil servant. He wrote two volumes on the Nagas that are by now in serious need of revision.173 In the foreword to one of them, Henry Balfour, at the time the director of the Pitt Rivers Museum at Oxford, wrote:
It is of the utmost importance not only to the Science of Man, but also to responsible officialdom, since a just and enlightened administration of native affairs cannot be established and pursued without an intimate knowledge of and sympathetic interest in the natives themselves, their customs and their point of view. Lack of ethnographic knowledge has been responsible for many of the misunderstandings and fatal errors which have tarnished our well-meant endeavours to control wisely and equitably the affairs of those whose culture has evolved under environments which differ widely from those of civilised peoples.174
This attitude was widespread among British administrators. Although Balfour was no doubt right that “knowledge of and sympathetic interest in the natives themselves” strengthened the colonialist enterprise of subjugation, it is also true that we would have much less information than is available to us today on topics like the prevalence of deity and spirit possession in India had it not been for the dedicated work of the British administrators who oversaw the production of local and district gazetteers, volumes from which I and many others continue to draw valuable material.
Hutton, one of these eager and highly educated administrators, reported a number of “supernaturals,” including the Sema aghau, equivalent to the Angami ropfu, “a personal familiar … perhaps it may be said that all persons are potentially possessed of aghau, though the existence of the aghau is not always important.”175 These may also be “house spirits,” thus testifying to their capacity to enter individual spaces. Study of spirits and possession was by no means Hutton’s forte, and he includes no vocabulary of possession. He was, however, a step ahead of most of his predecessors in remarking:
The missionaries in their blindness teach the Angani convert to regard all terhoma [deities] as evil, and mission-taught Nagas are in the habit of translating the generic terhoma into English or Assamese as “satas” [viz. Satans]. All of these “satrans,” as they call them, are, however, very far from having those qualities which we traditionally associate with the Devil, and the qualities of some of them are definitely benevolent.176
In spite of the efforts of the British colonial government, indigenous beliefs and religious festivals continue unabated in the twenty-first century. Because of Assam’s proximity to Bangladesh and the unstable political situation in India’s far northeast, it was almost entirely off-limits to Western scholars until recently, and local scholars did not appear to take much interest in the dramatics of oracular possession there.
However, one local chapbook, which describes festival possession at Kamakhya, the great temple complex near Gauhati, was produced in February 2004, with sections in Hindi, English, Assamese, and Bengali.177 This chapbook describes in detail a festival or “ritualistic dance” called Debaddhani or Deodhani, a local term derived from the Sanskrit devadhvani (sound or echo of the deity).178 This yearly festival is held late in the rainy season, beginning on the last day of the month of Śrãvana and continuing for three days, up to the second day of the month of Bhādrapada, coinciding with the annual worship of the Goddess Manasā, a popular snake deity of Kamakhya. The chapbook states that the term debaddhani “conveys the idea of the conceptual dances of the human vehicles chosen and possessed by the divinities to reveal their own intricate selves.” From nine to fifteen people, called deodhā, experience this possession.179 They are generally the same people every year and appear invariably to be men. Like many other sites of divinity in India—and we must view these possessed individuals as sites of divinity—the starting point is explained through a miracle. Once, it is said, a mahārāja with no offspring promised to offer his own head in sacrifice to the goddess if she would grant him a son. Pleased with his prayers, the goddess did so. But the mahārāja, in his joy, forgot about his promise. The goddess became enraged and in a vision reminded the king of his promise. The mahārāja promptly prepared a golden knife and chopped off his own head and offered it to Bhairavī The king thus attained divinity and possessed a local male devotee, thus beginning the debaddhani tradition.
As sites of divinity, the possessed individuals are viewed as loci of purification and, in an initiation ceremony conducted by the principal priest (daloī) of the Manasā temple, accept ritual paraphernalia, vow to eat a simple vegetarian diet (though they may smoke and drink tea), and are paid only in offerings of sacrificed goats, pigeons, and ducks, as well as gifts of ornaments, sweets, and fruit. Much of this is reminiscent of the Vedic initiation performed at the beginning of a soma sacrifice, though it is extremely doubtful that the influence was direct. After several rounds of dancing in the early afternoon, the performers are made up and dressed with garlands, bouquets, ritual insignia, cloth, swords, and other items reminiscent of the possessing deity. The chapbook does not list the names of all the possessing deities, but among them are the deified mahārāja; Mahādeva (Śiva), the ceremonial leader of the group; Kuvera, “the traditional store-keeper” of Mahādeva; and the goddesses Kāḷi, Kāmākhyā, and Calantā. The possessing deities bestow bīja and other meditation (dhyāna) mantras on the possessed, as well as visionary and oracular abilities. According to the chapbook, the signs of possession are “peculiar” or “abnormal” behavior, including fasting and unusually clean living.
[B]ut he demands raw flesh of black pigeons or of a he-goat and he takes these uncooked; he drinks the fresh blood of animals. His eyes turn bloodred, he yawns frequently and roams about aimlessly solely at the mercy and direction of the deity of possession.180 … The dance of the Deodhas consists of very vigorous movements interspersed with shrill shrieks. They run like mad men, as if they are unconscious of the physical environment and the thick crowd of spectators. They hold swords, batons, and animal offerings on their hands and shoulders.
The most dramatic moment occurs when the most deeply possessed among the deodhās, usually those possessed by Kāḷi and Calantā, dance with bare feet on the blade of a long razor sharp blade held by other deodhās. “It is indeed miraculous that none of the Deodhas have ever been hurt.” After the dancing is concluded, the deodhas sit, surrounding the vehicle of Mahādeva, and consume fresh blood and raw meat, in addition to juice of fresh coconuts and other refreshments. The description ends by noting that the deodhas form neither “a caste [n]or a privileged class; nor are they associated with the temple of Kamakhya according to the traditional categories of the temple’s functionaries.” They are, rather, self-selected as a result of inspiration from the particular deity.
The deodhas then, as sites of divinity, represent the Assamese version of what Shulman identifies as localized divinity: The deodhas manifest the immanence and transportability of divinity. This multilocality of divinity, from Śiva to Kālā to Calantā, is, as in Tamilnadu, a localization and transportation of the deity’s śakti, analogous to the Tamil aṇaṅku. In the Assamese case, however, the śakti is not so easily generalized. It is, rather, embodiment itself with the specifications of the individual divinity, including dress, ornamentation, diet, speech, spatial requirements, its position within the matrix of deities who have a special relationship with Mahādeva, the divinity of the center, and the peculiar way that it manifests power and danger.
In spite of the efforts of Meyer, Erndl, Bhattacharyya, Stanley, McDaniel, Srinivas, Freeman, Claus, Brückner, Macdonald, Pakaslahti, Gellner, Vogt, Wadley, Caldwell, Inglis, Wirz, and a few others to highlight the importance of local terminology, many ethnographers fail to provide not only the vocabulary for possession but also the specialized words or terms for the person possessed. Let us now briefly look at some of the vocabulary in Hindi, Urdu, Gujarati, Bengali, Marathi, Nepali, Tulu, Tamil, Malayalam, Simhala, and the language of the Maldive islands.
Wadley states: “In North India, the native terms for oracle are baki and bhagat. The native terms for exorcist include jhāḍṇewālā, ojhā, mantravidh [sic: should be mantravid], and mativah.”181 Wadley does not provide translations. They are: baki (babbler); bhagat (devotee); jhāḍnevālā (spiritualist); ojhā (exorcist, folk healer); mantravid (well-versed in mantras); and mativah (desire-bearer; opinion-bearer).182 Henry and Lambert distinguish between Tantric healers called sidh or sayana and mediums, bhopa or ojhā.183 Fuchs records that in central India the barwa cures by exorcism, while a janka does so by divinization.184 McClintock, in Pakistani Panjab, cites the Urdu āmil, more properly “agent” or “operator,” as “exorcist,” distinguishing it from the jādugar (magician) and pīr (saint).185 In Gujarati, according to the Broach District gazetteer, the term bhagat is mentioned as someone who is able to diagnose the provenance of a spirit, while a bhuva, who may be from any caste or religious community, is an exorcist.186 This is also a common term in Marathi, along with jhāḍ (medium).187 Bhattacharyya provides three words for folk healer in Bengali, all (seemingly) synonymous: ōjhā (Skt. upādhyāya), rojā, and guṇina.188 As noted above, Sontheimer records the Marathi devṛṣī, locally denoting a shaman, healer, and possession oracle. Sax notes the terms dañgārīya (little horsie) and paśvā (animal), for a person who is regularly possessed by a local deity.189 With respect to the latter, a relatively modern Tantric text, the Āgamarahasyam, also employs this image. In its description of śavasādhana, the practice (sādhana) of meditating while sitting on a corpse (śava) and in a cremation ground (śmaśāna), this text states that one should sit on the corpse as if riding a horse (aśvārohaṇam).190 Although this is not possession per se, the image is applied to a situation in which possession of the corpse by the deity is implicit.
As mentioned, the primary terms in Nepali for a religious healer are jh
ṃkri and paju, who can be either spirit mediums or practitioners of other shamanic arts,191 and the related jaki in Assamese.192 A primary Tibetan term for Vajrayāna yogis and ritualists who practice exorcism, sometimes applying wrathful means, is sngags-pa (nag-pa) (mantraist).193 They closely observe the vows of the Vajrayāna, which include observance of the bodhisattva vows, and require particular qualifications to be met for a practitioner applying “wrathful means.” The term, if not the qualifications, is similar to the Sanskrit māntrika, denoting an exorcist in relatively recent texts,194 as well as to the mantravādin, a category of ritual specialist noted in south India who employs therapeutic mantras and is distinguished locally from a vaid or vaidya, a doctor of āyurveda,195 and a pātrī (Skt. for “vessel”), a healer who diagnoses and cures through possession.196 Thus Brückner and Honko report that in Tulu pātri is used for a possessed person.197 Macdonald states that a jhṃkri is one who “after having first of all suffered possession by a spirit foreign to his everyday world, manages to control it and regulate it.”198 The jhṃkri appears to be a figure who moves from peripheral to central possession, in keeping with Lewis’s observations. The foreign spirit resides at the periphery of the social and mythic worlds of the jhṃkri, but is then domesticated and employed to present a possession that is central, if marginally intrusive, in the local cultus. Höfer prefers a more straightforward description of the jhṃkri than Macdonald’s “interpreter of the world.” Höfer states: “Roughly defined, the jhṃkri is a ‘shaman,’ i.e., a socially recognized ritual specialist whose main task is curing the sick; he is claimed to be capable of controlling and/or producing paranormal experience—visions, possession, etc.—allowing for a privileged, direct contact with occult forces and beings.”199 Gellner distinguishes the jhṃkri from the mediums recognized by the Newari as dyah-mā (godmother)200 or the more inclusive category dyāḥ waipĩ. About the latter, Gellner notes that they diagnose illness and other social problems and more often than not are low-caste women who suggest treatment in the form of the worship of certain deities.201
The nearly identical terms dyaḥ waimha in Newari and devatā āune in Nepali indicate “one to whom the gods come.”202 Also used for possession are the Newari verbs waye (to come) and dubiye (to enter into).203 Gellner also places the healers and mediums at the nexus of classical and folk Tantra and Vajrayāna Buddhism, in which the recitation of certain sūtra texts are used for apotropaic and protective purposes and in which Buddhist mantras are merged and jumbled by possessed healers with jhārphuk, exorcistic practices of “brushing and blowing.”204 As noted, the term for the healer subject to possession in Höfer’s study of Tamang shamanism in Nepal is bombo, one who treats individuals but has no role in communal rituals. The Tamang themselves regard this as a synonym of jhṃkri.205 Gaborieau records the Indo-Aryan word dhāmī and the non-IA ḍāṃgre in central Nepal.206 Berglie provides a detailed description of the training of a spirit-medium: the recruiting, apprenticeship, testing, acquisition of ritual paraphernalia, and an account of the trance experiences of the Tibetan dpaḥ-bo.207 Dietrich provides interesting vignettes of Buddhist, Hindu, and Muslim healers, Newari and Nepali, urban and rural in Nepal.208 Her study is consistent with earlier ones, though hers is more squarely situated in the indigenous healing traditions. Both Hindu and Buddhist spirit healers occasionally attribute possession among their clients to negative astrological influences, something I suspect is more widespread through the Indian subcontinent than ethnographers have recorded.
Anwarul Karim, a folklorist who studied shamanism in his native Bangladesh, is careful to clarify his terms. He explains that the Muslim ojhā belongs to the larger faqīr community of Sufis and are exorcists as well as healers of snakebite, female hysteria, childhood ailments, and cattle diseases. They are contrasted with the Hindu ojhā, who belongs to low-caste and peasant communities and are often ojhā by hereditary occupation.209 Karim distinguishes these practitioners from the kabirāj, a practitioner of normative Hindu medicine (he does not use the word Āyurveda, but this is probably intended).210 The faqīr invokes spirits or jinns to cure diseases, while the ojhā, either Muslim or Hindu, arrange and sing religious songs known as bhasan, padmapurāṇ or behulā laksmindar, which invoke the serpent deity Mānashā (as in the debaddhani festival in Assam) as well as other deities, both male and female, for healing.211 Shukla noted the words sokhā (one who absorbs or dries up) and guṇī (adept), in addition to ojhā and bhagat. He also noted that Muslim exorcists are called maulvī (a scholar of Islamic law), fakīr (a Muslim mendicant or saint), and namājī (a devout Muslim).212
Caldwell and Freeman inform us that, in Malayalam, the words for oracle are veḷicchapāṭu or kõmaram. Most are male priests associated with temples. However, some are females who reside in the hilly tribal areas near Palghat. Caldwell suspects that the male veḷicchapāṭu “in fact modeled their behavior on these female shamans, whose role was superseded in the Aryanized lowlands.”213 Freeman remarks that, regardless of their origins, “[t]hese priests are selected by a combination of a specific lineage eligibility, divine selection through spontaneous acts of possession, and subsequent ratification by astrological or other kinds of oracular readings.”214 Inglis notes the distinction in Tamilnadu between the cāmiyāṭi (god dancer) as the name for both the institution of nonshamanic oracular possession and its performers in the Madurai area, nērttikam (< Skt. nṛtya, [dance]; nārttikam [dancer]) (individuals who become possessed in the course of carrying out personal vow fulfillment) and kōṭaṅki (“possession specialists … who ply the trade of soothsaying in a market or other public place for a fee”).215 In Sri Lankan Tamil, Lawrence cites the word kaṭṭātikal for “oracle” and teyvam āṭukākkal for deity-dancer, as well as vākku solluṟatu for “uttering oracle” (discussed above).216
Finally, Clarence Maloney discusses fanḍita in the Maldive islands, which is “magic or religious ‘science’ of any sort, white or black, curative or preventative, fertility ritual or divination.”217 The word fanḍita, from Sanskrit paṇḍita, probably entered Divehi, the local language, through an earlier Tamil-Malayalam substrate population. In spite of an extensive catalogue of spirits, spirit possession is comparatively rare in the Maldives. This relative absence may be an influence of Islam on a local tradition that is still extensively practiced. Nevertheless, in Divehi a conjurer, exorcist, herbalist, astrologer, and protector of the evil eye (esfinna < Sinhala a svaha) is called a “fanḍita man.”218
As mentioned above, this review, lengthy though it may be, is not intended to be complete or definitive. I have made little effort to obtain data on possession for Telugu,219 Oriya, Assamese, and other northeastern languages and dialects and have not exerted much more effort for Gujarati.220 Scouring the district gazetteers, as I did in part for Gujarat, Kerala, and Maharashtra, will supply greater breadth, but I am not certain that our depiction of possession in South Asia will assume much greater depth for the effort. Nor are the data available in early highly positivistic studies of tribal areas likely to deepen our knowledge of possession on the subcontinent. I cannot make it my project here to take on the vast early (pre-1960) ethnographic literature, which, in any case, is hardly anthropological in the contemporary sense.221 However, I hope that what I have accumulated helps illustrate some of the conceptual, lexical, and semantic patterns of the language of possession as they emerge from a background in Sanskritic and other classical Indian possession as well as ways in which it may be independent of that background. These include the expression of possession as salvific, purificatory, exorcistic, therapeutic, celebratory, oracular, aesthetically potent, and socially binding—in part because it cuts across caste, class, and gender boundaries practically everywhere in South Asia.
From the foregoing discussion of lexemes and terms in many South Asian languages, one conclusion is unavoidable: that with a few notable exceptions the expansion of the lexicon for possession—and, we might cautiously presume, many of the ideologies behind this expansion—was to a great extent derived from Sanskrit and Sanskritic culture, after which it became diversified and adapted into regional languages. Despite this conclusion, the ethnographies highlight the non-brahmanical, non-Sanskritic dimensions of possession and the rejection of clerical authority in favor of a nonmediated transcendence or, at most, a transcendence that is sanctioned—and, therefore, in a sense mediated—by the dynamics of the local culture. Locally construed possession is, furthermore, an irruption of assigned hierarchical place that bears with it a consequent identification with nonhierarchical sacred place—indeed, of the individual as sacred place. Hence, it is an assertion that possession is largely the experiential property (and path) of people of lower social rank, women as well as men. It is regarded, in a critical mass of instances, as a socioreligious innovation of subalterns. Although this may be an overstatement, it is not without basis. After all, the social structure in South Asia hardened into a self-sustaining oppressive force long ago, giving rise to equally self-sustaining forces of reaction that drew form and idiom in part from these forces of oppression. In other words, though the preponderance of the evidence suggests that most possession is an indigenous phenomenon related to brahmanic oppression or “official” brahmanical religion only peripherally,222 in fact, a number of ethnographers (e.g., Sax, Knipe, Claus) have shown that brahmans participate in possession as much as anyone. Thus the “upward” accession of form and idiom is matched by a “downward” accession. As seen below, this is not a recent phenomenon: The evidence adduced from the Sanskrit texts testifies that possession has long been a part of unofficial (and preofficial) brahmanical religion; it was no less a part of Sanskritic culture than was officially engineered brahmanical religion.
Arguments that possession was “originally” a folk phenomenon distinguished from Sanskritic culture are, of course, correct, as far as they go. But these arguments cannot go very far, because folk culture by definition preceded Sanskritic culture and continued to flourish alongside it without operating in opposition to it. However, Sanskritic culture was never very far behind or far removed from folk culture. Indeed, one can argue that early Sanskritic, which is to say vedic, culture is nothing more than a poetic or literary redaction of folk culture. Folk culture was a part of Sanskritic culture from the outset. As seen in Chapter 5, possession (āveśa) was one of the defining experiences of the Ṛgvedic ritualists under the influence of soma, at least in the early, compositional periods of Vedic ritualism. One can conclude from this an identity of folk and classical during that period, the mid-second millennium B.C.E. At best this undermines the categories “folk” and “Sanskritic,” at least for these early periods of recorded Indian history. As for the later periods, however, as the gap widened between the literary presentation and the practice of Indian religions, we must conclude that the “origins” of possession in South Asia, whether brahmanical or non-brahmanical, Sanskritic, Dravidian, or “tribal,” are not particularly important.
Although the Ṛgveda suggests that its ritualists were familiar with possession, this does not mean that it was not also a widespread experience in non-vedic realms of early Indic folk culture. While noting the prominence of possession in folk ritual and theater, A. K. Ramanujan suggests that we do not find it in vedic or other exclusively Sanskritic ritual and theatrical forms. He writes that with respect to the “assumptions and characteristics” of folk ritual and theater, “notions of possession are never far from the audience’s mind.”223 We can probably assume that neither audience (and śrauta ritual was rarely as public as often assumed, based on its public patronage today) nor participants in vedic ritual, at least of “classical” or post–folk vedic ritual, looked for or expected possession. The emphasis on ritual exactitude far superseded expectations of immediate emotional uplift. Role playing was not expected to evolve into manifestation of the deities or sages whose mantras emerged from the mouths of the ritualists. Nevertheless, the evidence presented below suggests that Sanskritic, even vedic, possession cannot be ignored in discussions of more recent, even contemporary, cultural, religious, and linguistic parameters of possession.224
For example, the possession Hiltebeitel observed in the Draupadī cult in Tamilnadu occurred in a non-brahman population. This sort of folk possession, however, occurred as a result of long intertextual negotiations between an evolving local oral Mahābhārata and the Sanskrit epic,225 between Sanskrit renderings by professionals and folk retellings by local nonelite nonprofessionals (or even professionals), between classical dramatic performance that strived to maintain the Sanskrit plot lines and local performance that embraced non-Sanskritic folk elements. It is unlikely that the Draupadī cult’s Mahābhārata was a reaction against the classical text or against brahmans and brahmanical culture and religion. Instead, it is likely more accurate that it was a folk development in Tamilnadu that included some reconfigured and renegotiated elements of Sanskritic culture, among which possession was almost certainly one, given the substantial presence of it in the Sanskrit text (see Chapter 6). One fragment of evidence supporting this is the similarity of possession in the Tamil Draupadī cult and that which Sax finds more than a thousand miles away among brahmans as well as non-brahmans in another contemporary (but unrelated) Mahābhārata based sect in Garhwal. The linking element is clearly the classical Sanskrit epic in both its textual and performative incarnations.
Regardless of whether possession is viewed as brahmanical, it is, as we have often noted, prominently observed among brahmans. Not only does Sax find brahman possession in the Garhwal Himalayas, but it is widespread in the bhūta cult of South Kanara district of Karnataka, in East Godavari district of Andhra Pradesh, and elsewhere in the subcontinent. Typologically, brahman possession is little different from non-brahman possession: Much of it is oracular and performative, not unlike non-brahman possession.226 Although possession states are trumpeted (sometimes correctly) as being “of the people” rather than “of the elite” (sometimes conflating brahmans with “the elite” when it is unjustified), at the same time their experience is often framed in the language of that much-ballyhooed elite. As mentioned, a cautionary note should be voiced to the effect that the dictionaries of the regional languages and the training of most of the lexicographers and many of the ethnographers are perhaps skewed toward a more classical (hence elite) expression, yet the evidence currently available nevertheless suggests a healthy diffusion of the terminology of possession from Sanskrit into regional languages, supplementing the terms independent of classical or Sanskritic sources. Indeed, even the countervailing evidence, notably from classical Tamil, cannot safely be marshaled to argue against Sanskritic influence.227
It must be admitted, however, that linguistic borrowing and influence cannot safely be equated with cultural borrowing and influence; words and terms can be borrowed from one language into another to express notions or denote objects that are already present in the intellectual or material culture of the people into whose culture the borrowing takes place. Such linguistic borrowing can occur for reasons of prestige or as part of a general pattern of cultural and linguistic borrowing in which extensive sets of borrowed cultural and linguistic information might contain a smaller number of items already present in the culture into which the borrowing occurs. In the present case, deity or spirit possession might have been a part of tribal or vernacular culture in South Asia before the terminological borrowings for it from Sanskrit and other derivative Indo-Aryan languages occurred. Remnants of tribal shamanism in South Asia suggest that this might be the case, as Rex Jones demonstrates a continuum of ideas and ideologies between tribal shamanism and “ecstatic possession” of Buddhists and caste Hindus.228 Jones’s observations are augmented by Berglie, who discovered spirit meduimship in Nepal among reciters of the Tibetan national epic, the tale of Gesar of Ling.229 What is striking about this for our purposes is that in this epic, Buddhism, which condemns loss of control, eventually emerges victorious over folk religious practice through a narrative acculturation in which peaceful Indian Buddhist mahāsiddhas are given Tibetan origins and identities as heroes and warriors. This merging of Buddhism, folk religion, and shamanistic practice eventually manifested in the practices of epic recitation, in which Buddhist narrators sung the glory of local practices, often while possessed by Gesar himself. “Now,” writes Berglie, “it is more or less an integrated part of Buddhist practice [that] the activities of the spirit-medium are placed in a Buddhist scale of virtues.”230 It is often the case in South Asia that we see these three levels of discourse palimpsested, from the early performers of the vedic śrauta sacrifices to Kṛṣṇa līlā performers to kathāvācaks of Varanasi.
Thus the Nepali example helps shed light on the history of possession in South Asia. Both Buddhism and “Brahmānism” were conservative movements, favoring literate text-based priesthoods, deities, and practices over spirit mediums, healers, and local gods, in short, the codified and predictable over the (ostensibly) uncodified and unstructured. The difference, however, between Buddhist and early Sanskritic and Indo-Aryan cultures is that the latter had ancient folk and shamanistic practice embedded within it, even if it lost much of its fluidity and dynamism over time. So the latter were in a good position to provide cultural and linguistic borrowings of phenomena and terminology of possession into vernacular, folk, and even tribal cultures over a very long period of time. The situation that Berglie reconstructs for Nepal and Tibet seems analogous to what occurred in other parts of South Asia thousands of years ago. Although it is possible, we must not assume that tribal and folk cultures preserve intellectual notions and a phenomenology of greater antiquity than those of the literate and dominant or hegemonic cultures.
It is relevant here to inquire into the identity of the possessing gods. It is clear from the ethnographies that most of these gods are “folk” rather than “classical,” that they appear to be the property of “little traditions” rather than the “great tradition.” If this dichotomy is projected backward by more than a few centuries, which is possible given the evidence of early first-millennium C.E. Tamil literature, then can one not argue that possession was a folk phenomenon Sanskritized in order to confer upon it credibility, rather than assuming that it was vernacularized and localized from Sanskritic culture? Although this question may be merely “academic,” the linguistic evidence suggests that possession pervaded both Sanskritic and non-Sanskritic traditions equally, as just suggested. In any case, proving a direction of flow may be impossible to accomplish definitively, because it would require historical reconstruction beyond what the literary and inscriptional evidence allows.
As for Sanskritic and folk deities,231 Antonio Rigopoulos, subscribing to the doctrine of a clear dichotomy between folk and classical, states: “Not all Maharashtrian gods possess their devotees. For instance, Khandoba, Kal Bhairav, Dattatreya do possess but not Ganesha, Rama, Shiva, Krishna, Vithoba, Hanuman, or Vishnu. The list of Hindu gods who do possess, corresponds closely to those ancient non-Aryan local gods who were gradually assimilated into the brāhmaṇic tradition.”232 But this is not strictly accurate, as I have heard in Maharashtra (where Rigopoulos worked) first-hand claims of possession by Hanumān, Gaṇeśa, and Śiva, and, outside Maharashtra, of Kṛṣṇa (in Nepal),233 the Buddha (also in Nepal),234 Viṣṇu (in Sri Lanka and Nepal),235 and, as we have seen above, possession by Śiva (as Mahādeva) in Assam (and as shown in Chapter 11, in Andhra Pradesh and in China) as well as in classical Tamil literature.236 It is therefore possible to conclude that in folk contexts classical deities (and other figures) become folk, just as it may be shown that in classical contexts folk deities become classical.237 A good example of the former is in the Draupadī cult, where possession of the Mahābhārata figures occurred under a decidedly folk aesthetic, while Kṛṣṇa as the supreme deity was replaced by a more culturally appropriate flesh-eating goddess.238
In contemporary India, most deity possession, even of brahmans, is by folk deities.239 Thus the question arises as to whether this helps decide the matter of the direction of flow, from classical to folk or folk to classical, that was left undecided by resorting to linguistic evidence alone. In my view it does not; as I stated above it fuels an argument for dissolving, or at least rethinking, the distinction between folk and classical. As argued above, what we now regard as the folk experience of possession may well have been at one time characteristically classical, as evidence from the Ṛgveda, the Mahābhārata, the tantric literature, and elsewhere demonstrates. So, from a very early date a South Asian culture of possession must have existed that enveloped Aryan and non-Aryan, brahman and non-brahman.
Catherine Bell deploys the verb “traditionalize” to indicate the act of constructing tradition and the substantive “traditionalization” to indicate the process of such a construction.240 These terms are useful in thinking about the history of possession in South Asia. Based on the juxtaposition of textual and ethnographic data, it appears that possession became traditionalized as it assumed recognizable forms that were deemed religiously empowering and medically safe, while operating within generous margins of social acceptability. This assumes sanctioned and appropriate contexts for possession, which must have been abundant given the range of possession manifestations in South Asia. Furthermore, this traditionalization must have occurred gradually and, to a great extent, beyond the active intellectual reach of the literary traditions. Examples of models of possession that have syncretized and traditionalized “popular” (viz., the products of less-formally educated strata of society) with “learned” (whether Sanskritic or otherwise) elements include the following: the adālat, where little-educated non-brahmans preside over an imagined court and administer ontological justice to members of all social and religious communities; the “accommodated heterodoxy” of Tibetanized Nepali bombo shamans; the Sanskritized kaḷarippayattu of Kerala; the performance of the Siri epic of Tulunad officiated by brahmans; and, elsewhere, the marginally successful efforts of New Age trance channelers to mainstream and market their possession “products.”
One of the powerful objectives of this traditionalization is to disseminate normative and paradigmatic discourse, in keeping with hegemonic or Great Tradition strategies of legitimation, which privileges theory over experience. In these cases, traditionalization and institutionalization operate synchronically. This may be distinguished from the processes operating in classical India, where, with rare exception in the Sanskrit literature (the exceptions being āyurvedic texts, a few Tantras, and dharmaśāstra texts on karmic ripening [kar mavipāka] examined in Chapter 12), possession was never accorded a place in any descriptive or epistemological order. In the latter case, it appears that possession was traditionalized without being institutionalized. In general, it escaped the discursive formalization that has always been an outcome of śāstraic institutionalization, while it nevertheless remained enculturated within the broad boundaries of Sanskritic culture. And its disappearance from this sort of epistemological institutionalization is what has rendered it largely invisible to scholars.
The ethnographies show that practices of possession can be classified into a small number of categories, regardless of region, from Assam to Sindh, from the Himalayas to Sri Lanka. This accords well with the general distribution of the terms āveśa and grahaṇa. The phenomenon of linguistic borrowing, in this case of loanwords, has long been studied, with the general consensus that borrowings of terms with abstract or “cultural” import proceed from a politically and technologically dominant culture into the languages of those under political control or less advanced technologically, while the opposite movement is true for terms for flora, fauna, rivers, and other “natural” phenomena.241 The provisional conclusion, then, unless and until further lexical and other linguistic data become available from regional languages and dialects, is that regardless of the unique manifestation of the possession experience in the non- and semiliterate local cultures, its expression was to a great extent framed by the language and conceptualizations of the elite, dominant cultures, whether Sanskritic or Tamil. This speaks not only of the extent of cultural relations between dominant and local cultures but of the prominence possession must have once enjoyed in the dominant culture before it became locally prominent among nonclassical, non-Sanskritic cultures. This is not to argue, however, for a trickle-down theory, a gesunken Kulturgut, of possession phenomena from dominant Sanskritic cultures to subordinate vernacular ones; in general, such theories have long been discredited. The sociocultural and psychological affinities of baiṭhak as performed in Peshawar with, for example, the possession rituals among the Jalari of Andhra Pradesh and the bhūta festivals among the Tulu speakers in Karnataka testify to a broadly distributed cultural form with regional variations that was (and remains) inclusive of Sanskritic culture and, in many places, eventually came to adopt its primary expressions. It is important to note, too, that specific “religions” do not sequester particular domains of possession. Shail Mayaram, who studied possession in Ajmer, Rajasthan, records that a Hindu woman regularly reports possession of Imam Husayn, the grandson of Muhammad, known locally under the eclectic name Pīr Bābā. Conversely, Mayaram notes that a Muslim woman healer reported regular possession of a local mother goddess called Bayasaab Mātā.242
Let us now attempt a composite depiction of possession based on the ethnographies alone. Possession more often than not involves the feminine—either women are possessed or men are possessed by a form of the goddess. In fact, Caldwell suggests that men’s possession might have been inspired by women’s, at least in Kerala. Possession also tends to be ritualized, often in highly developed calendrically based rituals, religious festivals, or pilgrimages. It is often a part of religious practice (sādhtma)—also a feature of Sanskritic possession—and often involves an imitatio dei, as we investigate further in the discussion of āveśa in the realm of bhakti. Possession frequently begets violence, usually sacrificial though occasionally self-inflicted, such as the body-piercing Meyer observed in Tamilnadu and Lawrence in Sri Lanka. Often the individual possessed bears a visage of pain, writhing or screaming, though there is usually no memory of this. It is as if the pain were caused by trying to fit something big, frightening, wild, powerful, awesome, unfamiliar and beautiful into a small, limited, culturally conditioned enclosure. Sometimes, the awesome dominates, resulting in ecstatic wailing, but often the frightening dominates, as the deity (herself, usually) becomes wild and frenzied. Meyer has noted that the frightful possessing goddess is compared to a ghost or flesh-eating spirit (Ta. pēy, picācu < Skt. preta, piśācd). A summation of the ethnographic reports thus permits the observation that the cries of the possessed are, on the one hand, the cries of psychic pain and sheer fright at the power of the possessing agent, while, on the other hand, they are the joyful shrieks arising from the unleashing of repressed sexuality, aggressive emotions, and ecstasy paradoxically generated by the frightening deity (or spirit). The result is an empowerment centered on a wave of devotion to the possessing agent, like power generated from a great dam, whose bound waters nourish and drown. Possession is also healing; demarcated by a secret semantics of liminality, it assuages the wounds of individual and community.
A final question we must consider here is eligibility (Skt. adhikāra). Who is eligible for possession? Who is possessed? The issue of eligibility in India is as old as the middle vedic literature, when criteria were fixed for eligibility to perform vedic śrauta sacrifices. Discussions of eligibility were formalized in the Pūrvamīmāṃsāsūtras of Jaimini around the third century B.C.E. and have been argued, refined, and modified constantly since then.243 Without doubt, the issue of adhikāra has been a brahmanical, elitist issue that has served socially polarizing ideologies within Indian society as well as the brahamanical historiographical enterprise, as Pollock has shown.244 However, possession is a populist, egalitarian phenomenon that has persisted almost entirely beyond the reach of śāstra or formal brahmanical discourse. Why, then, should this question arise? First, possession has taken on many disguises in South Asian history in the past three or more millennia, and at least a few of them were determined or assisted by the acquisition of particular training or levels of initiation. This is evident in the classical period in certain Tantras and bhakti texts and, more recently, in possession festivals in Kerala, Tamilnadu, Garhwal, and elsewhere. Second, the question is of importance because of the likelihood of the influence of Sanskritic culture on popular possession.
Although the ritualization of possession as noted in ethnographies does not resemble classical Vedic ritual, the concept of divinization of the body is a theme that runs through Indian religion from the middle vedic literature to contemporary possession performance. We have already examined the phenomenon of the “localization” of divinity in southern India, the narrative act of establishing and circumscribing sacred space in a person, place, or thing. What is relevant here is that this act helps to establish eligibility, conferring a zone of safety, rather than danger, on possession. Possession is not risk-free; it may be dangerous or harmful unless the individual acquires eligibility through initiation, formal training, frequent exposure to others’ possession (which may also be a kind of training), or heredity. Thus eligibility is of great concern to the communities practicing possession. Certain individuals are authorized by the community and occupy a correspondingly higher place in the social structure, as would a formally educated priesthood. For example, Dirks notes that the cāmiyāṭi (possessed “deity dancers”) in the Aiyanar festivals in Pudukkottai District of Tamilnadu were “initially chosen for possessing special spiritual powers,” but that the position eventually became hereditary.245
Also prominent is an endemic sense of what is genuine possession and what is pure theater, especially in narrative or festival performance. Zvelebil distinguishes between “genuine possession” and “pseudo-possession,” the former characterized by hieratic trance and the latter either feigned outright or the product of passion. Genuine possession, he says, is a form of “regulated contact with the sacred,” while pseudo-possession is expressed in “poetic conventions dealing with the erotic.”246 This is a distinction that may apply “on the ground,” but, as shown in the following discussions, these are more generally regarded as different modes and discourses of possession than distinct states of genuine and pseudo-possession. This complicates the field of eligibility, in which Zvelebil’s category of “genuine possession” would bear the greatest weight of the criteria for eligibility. It will be of interest to see how Zvelebil’s categories withstand the evidence of classical literature.
In living cultures in which possession is recognized, much of it is oracular, used for divination, healing, and mediumship. It is in these forms that standards of eligibility are most important. But some possession, including (occasionally) oracular, is exhibited in a manner suspiciously akin to madness or intoxication. Although this worries family members, in most cases it is interpreted as the inspired presence of the deity. Thus a post facto form of eligibility is invoked. This exhibition of divine presence as frenzy (a culturally sanctioned dissociative state) is a respected and recognizable part of the tradition of imitatio dei. Already in the ninth century the Tamil poet Mānikkavācakar was calling Śiva a madman (pitttaṉ).247 To what extent madness, as well as contextual notions of eligibility and the localization of sacred space, are shared with possession as reckoned from the classical sources will come into sharper focus in the following discussions. But at the very least we can posit that possession appears to have undergone development with changing social and cultural environments, that it may not be what it once was, at least not completely—in spite of millennia of alleged obfuscation by various orthodoxies who made it their business to try to manage the unmanageable.
NOTES
1. This is borne out in a study of the bibliographies in an issue of Puruṣārtha devoted to possession; cf. Assayag and Tarabout 1999. The issue contains thirteen articles (excluding the prologue and epilogue), of which five are on Kerala, two each on Himachal Pradesh and Karnataka (one of these on Tulunad), and one each on Rajasthan, Tamilnadu, Nepal, and Kashmir. The brief article on Kashmir is by André Padoux and is the only one in the collection devoted to possession in the premodern period (cf. Chapter 10).
2. These are based on either synonyms or homonyms; see Vogel 1979.
3. Good examples are Sanskrit dictionaries. For example, Monier-Williams’s English-Sanskrit Dictionary (1851) is of no value because it simply has an English word followed by a large number of Sanskrit words, with no effort to contextualize any of them. By contrast, his Sanskrit-English Dictionary (1899) is much better in that he often (but, by no means, always) cites texts from which he draws a word’s meanings. Better still is the seven-volume Sanskrit-Wörterbuch by Böhtlingk and Roth (1855–75). Best is the Deccan College dictionary (Ghatage et al. 1976–), which, after more than half a century of work, including more than three decades of publishing (as of this writing twenty-six fascicles have been published), is about a third of the way through the first letter of the alphabet. Although the completion schedule has become something of a long-running joke, few scholars doubt the validity of the project or, for the most part, its results. Similarly, the Critical Pali Dictionary (Trenckner et al. 1948–90) begun in the nineteenth century is nowhere near completion. Most dictionaries of regional languages are utterly nonhistorical. To be generous, however, historically sensitive lexicography is definitely on the increase. See, for example, Tulpule and Feldhaus 1999, for Marathi, which tries to recognize chronological changes in meaning.
4. Sontheimer 1989b:142ff.
7. Caturvedi and Tiwari 1975; cf. bhūtabādhā.
8. Erndl 1996:178; also Erndl 1993:108. Wadley writes that khelnā “identifies an active and extreme possession which is manifested by the physical activity of the possessed; khelnā is also used to indicate goddess possession in Karimpur” (1976:237n6).
11. Pfleiderer and Lutze 1979:111.
12. Freed and Freed 1993:17.
14. Many of these are found in rural areas, including Bālājī. This contravenes Dumont and Pocock (1959:56–57), who estimated that healers, as distinguished from oracles, are found only in urban settings in India. See Chapter 12 and F. Smith 2004 for notes on the spirit healing center at Chottanikkara in Kerala.
15. The healing centers at Muslim dargāhs (such as at Badayun, in U.P., south of Bareilfy) and other Sufi-oriented Muslim healing places (such as Matka Pir in Delhi) have clientele that are at least half Hindu.
16. Dwyer’s ethnographic work (1998, 1999, 2003) was primarily on possession and healing in the main temple of Bālājī, while Pakaslahti has been documenting the work of a healer (bhagat) who conducts his sessions at one of the shrines near the temple. Bālājī has experienced exponential growth in the past twenty-five years, since Kakar was there. Pakaslahti estimates that the better known of the many healers there have hundreds of patients, who are often treated simultaneously. Most of these healers come to Bālājī for regular brief visits from other cities in north India (personal communication); also see Pakaslahti 2005.
17. Much of the following description is based on my own observations at the shrine that Pakaslahti has studied, augmented by Pakaslahti’s explanations; cf. Pakaslahti 1998:146ff. The tactics of exorcism at Bālājī vary so greatly that it is certain that some of what I observed does not occur at all of its exorcism and healing clinics. Conversely, some of the tactics employed at these other centers are not employed in the adālat that I observed. For a vivid depiction of one healing, see a video for which Pakaslahti was the primary research and scientific advisor: Kusum, by Jouko Aaltonen (Illume Co., 69 minutes; Helsinki, 2000).
19. The emerging work of Bellamy (2004) at Husain Tekri shows that this image has a broad and varied distribution in northern and western India.
20. This is a common genre of popular music in contemporary India.
21. A Sikh family was also present when I was there and responded similarly.
23. Cf. Obeyesekere 1981; Hiltebeitel 1988, 1991. Dwyer notes: “I was told that the possessed frequently experience a choking sensation (galā bhar ānā), a contraction of the muscles in the throat, and/or a sense of being bound up (bandha); and informants commented that these are the reasons why tight-fitting garments are loosened before, and sometimes during, the ritual, and why a woman’s hair must be untied” (1999:113).
24. See Davis 2001:99ff., and the discussion in Chapter 11, for surrogate possession in eighth- to twelfth-century China. Davis’s important study seeks to free possession (pingfu [to lean on and adhere to]) from its “shamanic substrate” and bring it to the center of Chinese religious practice, as the present effort aims to do for India. Davis, to the detriment of his otherwise tight and groundbreaking study, rarely refers to anthropological theory or other situated studies, at least in English. For example, his acceptance of Rouget’s definition of possession as merely “a trance of identification” (ibid.:1, 84) is often inadequate, even for his material. Nevertheless, he does refer to many studies in French and Japanese, lest anyone assume (as Janice Boddy has, apparently) that English is the sole authoritative language for the study of possession. Also important for Chinese religion is Teiser 1988. He makes a case for placing spirit possession within the “shamanic substrate” (to which Davis reacts). He writes that “the shaman is one who has special mastery over shen—spirits from afar, his own spirit, and spirituality in general. The shaman’s special access to shen, then, implies that his mediumship works both ways: he can be possessed by spirits, and he is an adept at soul journeys” (p. 145). In fact, Chinese elites—Buddhist, Confucian, and Daoist—placed these wu (shamanic practitioners) in a substrate. But these are disputes that I am hardly qualified to enter.
25. For a somewhat similar situation, see Werbner and Basu 1998:15, where a dargāh is thought of as a Mughal court presided over by a dead saint, potent with charisma, but whose primary minister is a sajjāda. Again, here we have examples of a debate between healer and jinn, where the person’s body is the battleground between the sajjāda and the client.
26. One of the Sanskrit manuscripts on possession on which I base my discussion in Chapter 11 below employs the compound kapisainyaprakāra, indicating that one is invoking not only Hanumān but also the army of monkeys (kapisainya) that assisted him in his invasion of Laṇkā. The phauj here is thus evocative not only of the Rāmāyaṇa but of other possession complexes in India.
27. For another example of this, in what must be regarded as an extended South Asian religious complex, see Bertrand 2004 on the conversion of “errant” Khmer spirits in Cambodia known as brāy to Buddhist pāramī, spirits who exhibit the Buddhist perfections (Skt. paramā). See also Vogt 1999, discussed below.
28. “She” because the assistant in this “clinic” was a sensitive woman named Meena who was herself a former patient who came to this work after a series of traumatic events befell her life leading to a nearly total psychological collapse. See also Trawick 1992:153, for a roughly similar instance, a woman who had experienced possession of Māriamman when she was young who then became a spirit healer only after a near fatal bout of tuberculosis.
29. Pakaslahti 1998:143. Parry states that in Varanasi, at least, a healer and assistant who work together are recognized by certain specific terms: “The ojhā lures and entraps the ghosts, while the sokhā supervises his efforts and is the real expert in identification” (p. 236, diacritics added). The situation was not quite the same at Bālājī.
30. For a similar process of questioning the spirit in more Islamic-centered possession, see Nasir 1987:165f. For a Buddhist context, see Vogt’s findings in Sri Lanka (below).
31. I am here reminded of a description of an exorcism found in Philostratus’s life of Apollonius of Tyana (Philostratus 1970). Philostratus probably finished writing this text in 220 C.E., while Apollonius, one of the earliest travelers from Greece to India, probably lived in the mid- to late first century C.E. There is no possibility of linking through any recoverable tradition the practice that Apollonius describes with practices in contemporary India. Nevertheless, the passage is of such interest that it is worth quoting in full:
In the middle of this conversation, the Wise Men were interrupted by the messenger bringing some Indians who needed cures. For instance, he brought forward a woman praying to them on her son’s behalf. He was sixteen years old, she said, but had been possessed for two years by a spirit with a shy, deceitful character. One of the wise men asked what her evidence was, and she said, “This boy of mine is rather handsome in appearance, and the spirit is in love with him. He will not allow him to be rational, or go to school or to archery-training, or to stay at home either, but carries him off to deserted places. My boy no longer has his natural voice but speaks in deep, ringing tones like a man; his eyes, too, are someone else’s than his own. All this makes me weep and tear my hair, and I scold my son as you would expect, but he does not recognize me. But when I decided on this journey, which I did a year ago, the spirit confessed who he was, using my son as a medium. He said he was the ghost of a man who formerly died in war, still very much in love with his wife; but the woman broke their marriage bond three days after his death by marrying another man, and from that time, he said, he had loathed the love of women and had transferred his affection to my son. And he promised that if I did not accuse him before you he would give many wonderful presents to the boy; and this rather made me change my mind. But he has been keeping me waiting for a long time now, acting as sole master of my house, with his wicked, deceitful ways.”
The Wise Man then asked her if the boy was with her, but she said, “No: I did everything to make him come, but that spirit threatened me with mention of ‘cliffs’ and ‘abysses,’ saying he would kill my son if I accused him here.”
“Don’t worry,” said the Wise Man; “he will not kill him when he has read this,” and he brought a letter out of his pocket and gave it to the woman; it was addressed to the spirit, and contained threats and warnings. (pp. 83–84)
This story has many of the elements of what may be found today, including a setting of family therapy, decided dysfunctionality and marked changes in behavior by the individual possessed, identification of the offending spirit, and threats and warnings by the exorcist.
32. Cf. Hiltebeitel 1989a and Mayer’s study of the phenomenon in the Nyingmapa sect of Tibetan Buddhism (1996:109–132). In the latter case, Śiva as Bhairava and Rudra/Maheśvara is tamed and converted to Buddhism.
33. For more on the cuṛail, see Chapter 12.
34. Pakaslahti, personal communication October 2001.
35. See Pakaslahti 1998:139. Dwyer (1998:11ff.) statistically analyzed the results of interviews with 734 patients or their caregivers. He discovered that 35 percent of women and 26 percent of men believed their afflictions were caused by bhūt-prets and that half of both female and male respondents believed that they were harmed by occult practitioners. The biggest differential was between married and single persons: “68 per cent of married persons as opposed to 28 per cent of single persons, in fact, accused their enemies of directly harming them by means of sorcery” (p. 12). Dwyer has much to say about this.
36. Information from Pakaslahti, confirmed by my own observations. I am indebted to him for taking me to Bālājī in April 2001. Casual visitors to Bālājī may examine the immediate surroundings of the main temple, including the pitṛsthān behind it. On a wide dirt track immediately behind the temple, the visitor may see on any day a fairly large number of people with large heavy stones either balanced on their heads or placed on several parts of their supine bodies. These are meant to press on the bhūt-prets in order to help expel them from the victims’ bodies. These stones are clearly much heavier than an ordinary person could bear. The victims might remain motionless in those positions for a few minutes or a few hours before family members struggle to remove the stones. The pitṛsthān rises on a hill behind the temple and resembles a graveyard of stone tumuli. These are, in fact, commemorative stones to deceased ancestors whose possessed and exorcised descendants have placed them there. This is also a site for ordeals. The “clinics” (a very inadequate word) such as I describe here are quite small, consisting of a room or two in houses outside the main temple.
37. Seidel 1987:228. The bureaucratization of celestial regions and netherworlds in China has long been a source of discussion; see Nickerson 1996.
39. Ibid.:95. Davis comments (101ff.) on the striking similarity of exorcism rituals with Song dynasty criminal proceedings.
40. Ibid.:2001:106, also 46. The notion of an army of spirits is not unique to India and China. In Uganda this notion has been elevated to real armies, battles for national political control, and mass murder; see Behrend 1999.
41. Davis 2001:45ff. for accounts of the exorcistic services of fashi, a somewhat amorphous class of ritual specialists who could be Daoist priests, laymen, or spirit-mediums. See also Schipper 1993:49ff. on the category of “barefoot master.” The fashi is the lowest-ranking priest in the Daoist hierarchy. He commands spirit militias filled with petty demons, hungry ghosts, and wandering spirits, and his troops “save the people, fight evil, and heal the sick” (ibid.:50).
45. Seidel 1987:229. See also Eder 1973:194ff., passim, on the division of the Chinese universe into yin (darkness) and yang (light): “A shen, a good spirit or god, comes from the beneficent half of the universe; a kuei, a demon or spectre from the opposite half. Every good in the world comes from a shen, every evil is inflicted by a kueï” (p. 195). Furthermore, “[a]fter death the soul is helplessly exposed to hordes of spirits” (p. 194). Only through the grace of bodhisattvas can these beings be saved. Note the vague resonance of this with the rationale behind the increasingly popular mass exorcisms of the living in contemporary American Christianity documented by Cuneo 2001. For complex demonologies in the Hindukush, with Indic, Chinese, and Central Asian influence, see Jettmar 1975:219ff.
47. Goudriaan 1977:145; this is in a collection of Balinese hymns written in very corrupt Sanskrit.
48. Cf. Īśānaśivagurudevapaddhati 43.68. For more on Khaḍgarāvaṇa, see Chapters 10 and 12.
49. Cf. Sen 1971:701–702. Sen gives the etymological meanings of bhar(a) as “weight, load, cargo; full” and of bhara as “to fill up, to thrust in.” These meanings are a short step to possession. Nevertheless, I am unable to find uses of the Sanskrit √bhṛ in the sense of possession; usually it has little meaning beyond the literal “excess, bearing; burden, weight, mass, bulk, fullness.” Hints, such as ṚV 1.60.5, vājaṃbhara, hardly serve as evidence. Östör 1980 notes the Bengali ānanda korā (to make joy), a joy that serves as the “the basis of devotion to and possession by Bhairab, Śiva’s deputy in the gājan” (p. 138). Their merrymaking, their ānanda, is a result of consumption of large quantities of country wine. See also Bellamy’s observation noted in Chapter 2, n. 196.
51. Bhattacharyya 1986:172.
52. Ibid.:59–75, 180–181. Alternative terms for māthār golmāl are māthā-khārāp (bad-head), māthā-garam (hot-head), and the Sanskrit borrowing unmāda (insanity) (p. 65). Maloney (1976:132) adds the Bengali mukh-lāge (evil face). See also Gonda 1969 for a detailed study of early Indian (and Indo-Iranian) notions of the nature and consequences of eye contact. On the evil eye, ághoracakṣuḥ (Ṛgveda 10.85.44), átjfhoreṇa cákṣusā (Atharvaveda 14.2.12), etc., see pp. 33ff.
53. E.g., Netratantra 19.2. As we see in Chapter 12, the curse of the evil eye is quite predominant in this text of the end of the first millennium C.E.
54. I am grateful to Amitava Bhattacharji for this information.
55. On the manifestations of naẓar and its relationship to desire and caste, see Pocock 1992. See also Kapadia 2000:182, for remarks on the evil eye (tirusti < Skt. dṛṣṭi) in Tamil.
56. Without citing sources, Maloney gives two additional Sanskrit words: kudṛṣṭi (weak sight, evil or sinister eye, heterodox opinion or doctrine) and pāpadṛṣṭi (sin-sight). These are, however, common enough in Sanskrit. Cf. Maloney 1976:128.
57. Bhattacharyya 1986:181. Note Bhattacharya’s alternation between the non-retroflex and retroflex “l.”
58. McDaniel 1990:21 and passim; see also McDaniel 1988:87–99.
59. McDaniel 1988; idem 1990:91; also Caldwell 1996:207. I have corrected the transliteration and diacritics of both Bhattacharyya and McDaniel.
60. McDaniel 1988:87, 96.
64. Molesworth states in his introduction that his sources consisted of paṇḍits in the Konkan area of coastal Maharashtra.
65. Karpūramañjarī, 37–38, 243.
66. Tulpule and Feldhaus 1999:72.
67. Cf. Stanley 1988:40ff.; also Sontheimer 1989b:78. Compare this with bhūt bādhā (harmful possession by a ghost); Stanley 1988:26ff. Sontheimer notes that one who is possessed by any of these the gods Birobā, Mhaskobā, and Khaṇḍobā is called devācī jhāḍ (god’s tree) (p. 193).
68. Wadley 1976:238. Wagle records bhūte āṅgī vāḍhu lāglī (became possessed by bhūts), āṅgāt vāḍhlā (possessed), and bhutācā asvār āṇūn (the bhūt was induced to possess) in court documents dated 1812 (1995:197, 211).
71. Ibid.:48. The district gazetteers published throughout India beginning in the 1870s and revised periodically for at least a century are also useful sources for possession. In spite of the normatively brahmanical presentation of religion in some of the volumes (e.g., the Satara District gazetteer), others try to present a more balanced view. For example, the Poona District gazetteer discusses the demigod Vetāḷa, the local demonology, and the ubiquity of belief in spirits and possession in Maharashtra; cf. Gazetteer of Bombay State: Poona District (1954:126–128).
72. Feldhaus 1995:48, cf. also pp. 11–13; and 126ff. for ethnographies.
75. Berglie 1976:86. Berglie worked with Tibetan refugee communities in Nepal and India in the early 1970s. He does not state that spirit-mediumship among the Tibetan refugees in Nepal is constituted differently from that in Tibet. It is unlikely to have accumulated significant influences from the indigenous Nepali population by that date.
76. Another study of possession and spirit belief among Hindus, Muslims, and Buddhist tribals in central Nepal is by Gaborieau (1975a), who has recorded statements allegedly uttered by members of a class of spirits of the dead, those who suffered accidental death, called bāyu. See also the excellent studies by Crook (1997, 1998) on oracular possession among Ladakhi Buddhists, as well as the shamanic narratives and Buddhist metanarratives that guide the process.
78. In western Nepal, Gaborieau (1975b: 158ff) has noted that the word avatār is normally employed for the possession of a deity by a medium. This possession is instigated by a tantric master who calls the deities into the body of the medium. Clearly, there is a greater infusion of Indo-Aryan (IA) terminology in Nepali (an IA language) than there is in Ladakhi.
79. Srinivas 1998:132. See Nagano 2000:587f., for distinctions made between kinds of possessed festival mediums among the Tibetans of northern Tibet and southern Qinghai province, China: “Tibetan people generally judge the greatness of a lha pa on the basis of whether or not he can verbalize the gods’ words. In addition to that, they distinguish a lha pa from a lha pa tshab on the basis of whether he has actually been possessed. They also distinguish between lha (rang) ’bab rgyab ‘the deity itself descends from heavens’ and lha phab/’bebs ‘to make a deity descend.’” See Paper 1995:54ff., 115, for a description of four shamanic specialists in ancient China. It is possible that there is a common ancestry for these groups.
81. See Mumford, who renders mi-kha as “envious talk” (1989:140).
82. Considerable information on possession in Nepal is in Hitchcock and Jones 1976; also Samuel 1993:194–196.
83. Höfer 1994. This exemplary work is first and foremost a linguistic study, though this inevitably leads to a “thick description” of Tamang shamanic ritual. The book also includes glossaries of Tamang, Tibetan, and Nepali words. Admirably, Höfer devotes considerable space to explaining his terms.
84. Ibid.:20. See also Gutschow 2004 for this and related shamans and oracles in Ladakh.
86. Bakhtin 1981; cf. Mumford 1989:13ff.
87. Mumford makes extensive use of Bakhtin’s observations (1968, 1984) on Rabelais and Dostoevsky.
88. Spirits are often said to be fond of milk; see the translations of the āyurvedic texts in Chapter 12. Note also in the same chapter the discussion of the Malayalam word kaiviṣa, a kind of poisoning in which spirit possession is involved.
90. Ibid.:144. See also the discussion of the Madanamahārṇava in Chapter 12.
91. Ibid.:148. Mumford adds, “The shaman regards his own exorcisms as more effective for that reason, justifying his own violent rites as the only effective way to expel demons from the community.”
92. Berglie 1982a, also see Berglie 1976, 2000. For more on the dpa’-bo, see p. 304f. For a wealth of material on shamanic healing and possession in northern Tibet, much of which is clearly imprinted on Nepali Buddhist shamanic possession, see Bellezza 2005. Unfortunately, this volume came to me too late to make full use of it.
93. She was a ḍākinī and one of the main consorts of Padmasambhava.
94. It is not without interest that the dpa’-bo call the mirror gling (“world”); cf. Berglie 1976:94. This appears to be consistent with the tenor of Buddhist philosophy. See my extended discussion of this in Chapter 11. See Bellezza 2005:23–28 for an interesting description of a number of Tibetan words for this magical mirror.
96. Berglie 2000:30. See also Pettigrew and Tamu (1994), who briefly describe possession among the Gurang of central Nepal. The term for possession in that dialect of Nepali is khhelye khhaba (god has caught, captured, covered) (p. 421n4).
97. Dietrich 1998, esp. pp. 39ff. for demonologies, pp. 160ff. for healing practices. Regrettably, the book does not include an index.
98. An important study of psychiatric treatment from the textual point of view, with a few ethnographic examples, is Clifford 1984. This is discussed extensively in Chapter 12. See also Merz 1996, which describes the goddess Hāratī in the Kathmandu Valley, whose cult has arisen since the mid-twentieth century. Hāratī confers on her devotees oracular possession as well as disease, but also heals. This is consistent with evidence found elsewhere for other deities. See also Diemberger 2005 on various terms for oracles and oracular possession in southern and central Tibet.
99. Cf. Hiltebeitel 1988, 1991, passim, esp. note 1988:275.
100. Cf. R. L. Turner 1969:65, #1437, also #1446, āvēsika. See also Obeyesekere 1970:98 and Hiltebeitel 1988:275. Although Simhala is an Indo-Aryan language, āvesa in Sri Lanka shares semantic features with its counterpart in Tamil.
101. Obeyesekere 1981:34, passim.
102. Ibid.:24, 84, 100f., passim.
103. Fabricius 1897 [1972]:61–62.
104. Cf. Carman and Narayanan 1989:168. I thank Vasudha Narayanan for the Tamil text.
105. Handelman and Shulman 2004: 40; see also Shulman 2002:134. On the Sanskrit anugratha, see Chapter 12.
106. Hiltebeitel 1991:344.
107. Ibid.:343f; see also Reiniche 1979.
108. Hancock 1995:89n19. This observation requires further corroboration. That I have not seen this distinction elsewhere raises the question of whether Hancock’s informant was sharing no more than her personal syntactic sensibility.
113. Ibid.:258ff. More important is Klimkeit 1976.
114. Klimkeit’s use of term teufelstänz is thus not an accurate translation, but reflects the usage of previous scholarship extending back to the early eighteenth century and Ziegenbalg. Clearly “devil’s dance” is pejorative and must be understood in the context of Protestant missionary activity on the Malabar coast; e.g., Burnell 1894, 1896. See also Brückner 1987 and, much more extensively, 1995.
116. Foulston 2002:140ff.
117. Cf. Diehl 1956:177; also Biardeau 1989b:149f., 153f. Foulston’s study is thin on linguistic data and records none from Oriya.
121. Nabokov 2000:126–127.
125. Cf. Feldhaus (1995), whose findings support Dumont.
128. Hancock 1995 and Chapter 2.
129. Hart 1975:28. Another example is found in the Shilappadikaram; cf. Adigal 1965:150–151.
132. Ibid.: 115. Hiltebeitel shows the same resolution of āvēsam in the Draupadī cult in northern Tamilnadu.
133. It is relevant here to cite Bertrand’s 2004 study of spirit possession and Buddhism in Cambodia, because what he finds resembles in part the situation in Sri Lanka. In one important way, however, it is dissimilar. In Cambodia many monks, in addition to laity, undergo possession in which they attempt to convert wayward spirits to Buddhism. This, observes Bertrand, “is a complete infraction of the Vinaya, the code of conduct that regulates the life of monks.… Trance is unacceptable for a monk, as it is proof of the absence of serenity and a lack of behavioural control” (p. 163). Thus many of these monks “chose to [temporarily] defrock rather than break the monastic rules” (ibid.). The situation in Cambodia is unique, however. The large number of new mediums there, both lay and monastic, is often interpreted “as the wish of the Buddha to aid mankind and also to pacify and educate all the souls of the dead killed under Pol Pot” (p. 166). The proximity of Buddhist monks to deity or spirit possession is discussed elsewhere as well; cf. DeCaroli 2004:25ff., 128ff.
134. Yocum 1982:189. It is also important to see the evidence from various first-millennium Tamil sources gathered by Hardy (1983:136–138, passim).
135. Yocum 1982:188ff. Anyone who has read this far would profit by referring to Yocum’s section on possession (pp. 187–194). On the matter of dating, see Tieken’s revisionist view attributing the Caṇkam poetry to the second half of the first millennium C.E. (2001) and Hart’s strong criticism of this view (2004). Both authors argue their cases well. I am not qualified to say which of the two is right. However, I must offer two observations: (1) the consensus remains that the Caṇkam poetry is of the first half of the first millennium C.E. (2) This is important because it is assumed that much of Indian devotional possession deriving from Sanskrit traditions and sources (e.g., from the traditions based on the Bhāgavata Parāṇa) ultimately derives from Tamil texts and religiosity. Tieken proposes to reverse this. If this were true (and I do not see a definitive solution), then the history of devotional, oracular possession would have to be revised, placing the time of its popularization a half a millennium later and its locus more squarely in Sanskritic rather than Tamil religiosity.
136. Hardy comments on the female presence in Tamil possession and ecstatic dancing: “Therefore it may not be accidental that in later bhakti religion, where these forms of worship appear again, the psychology of religious awareness is ‘female’” (1983:140).
137. The south Indian origin of what he rather flatly calls “emotional bhakti” is one of the main themes of Hardy’s Viraha-Bhakti.
139. Wirz 1954:14ff.; also Scott 1994:16, passim.
146. Exceptions to this perhaps harsh assertion may nevertheless be cited, e.g., Hiltebeitel, Sax, Korom, and certainly others.
147. Vogt’s psychological interpretations (ibid.:74ff.) are indebted to the work of J. L. Moreno, the founder of “sociometry” and “psychodrama.” Her interpretations are intriguing and often convincing. However, she does not support her otherwise good arguments for application of Moreno’s ideas with reasons why Moreno’s psychodrama is more compelling than Freudian or other allied psychoanalytic theories. Moreno, for better or worse, stands beyond the pale of Freudian and other more medicalized systems that dominate psychological perspectives of scholars in the humanities and that have, partly for this reason, achieved greater subtlety and nuancing than Moreno’s views.
149. Ibid.:48, yakunte dharmayan ana-kirima (Simhala); see also pp. 84–100.
150. See Walters 2000 for a powerful critique of the uses and misuses of the Pāli vaṃsas by the authors themselves, nationalist Sri Lankans of the nineteenth and twentieth century, and European scholars of the same period. Despite many valuable insights, Walters is much too (fashionably) zealous in stigmatizing the textualists upon whom his own work depends.
151. Scott’s postcolonial positionings are much more sophisticated than Vogt’s. The latter was clearly unfamiliar with Scott’s work. My comments above on the freedom of continental European scholarship to work outside the present trends of post-colonialism, etc., should be tempered by such admissions.
153. Scott 1994:228. In spite of this, it must be noted, Scott appears to be unaware of what these virtuous commands actually are. Although he adds that aṇaguna “conveys in fact the very force of the Buddha’s moral authority” (p. 25), he appears not to have studied the Buddhism that he admits rests at the basis of this healing process.
156. Freeman 2001:187, 384n1.
157. Cf. Freeman 1993:134, 1998:73–98; also see Nambiar 1993.
158. Cf. Albert Lord’s notion of the “formula” (1960:30ff).
162. Caldwell 1996:207; for this word see the earlier discussion of Meyer’s account of possession of Aṇkālaparamēcuvari.
164. Blackburn 1988:229–230nn22-27.
166. Brückner 1995; Honko 1998. Among the earlier studies are Abbé Dubois 1818:728–730, Prabhu 1977, and the works of Burnell cited in note 113. See also Upadhyaya and Upadhyaya 1984; and Claus 1973, 1979. For references to these and other studies, see Honko 1998:247nn95-97.
168. Brückner 1987:20;1995:66ff.
169. Tulu replaces the velar “s” with the retroflex “s.”
172. Cf. Zvelebil 1988:145–146, 2000:185, where the word is transliterated ja:ya. Zvelebil does not comment on why ja:ya is exclusively a male enterprise. He adds: “Foretelling the future, divination, and possession are obviously part of an archaic layer of Irula tribal religion” (ibid.). See also Zvelebil’s brief notes on pe:y/pe:, 2000:188.
173. Hutton 1921a, esp. 178ff; 1921b, esp. 192ff.
177. See Viswa Shanti Devi Yajña Committee, Viswa Shanti Devi Yajña (2004). A “world peace sacrifice to the goddess” (viśva-śānti-devī-yajña) was performed at that time. All quotations are from pp. 11–15 of the book. I am grateful to Loriliai Biernacki for lending this to me.
179. The chapbook says nine are possessed, but Loriliai Biernacki reports that, according to her informants, the number is not fixed, and as many as fifteen have been possessed in previous years.
180. Cf. the descriptions of possession found in the principal āyurvedic texts; see Chapter 12.
181. Wadley 1976:235. Her main sources for these words are Berreman 1963 and Planalp 1956.
182. I am in no position to comment authoritatively on the distribution of these words.
183. Henry 1981:303; Lambert 1988:356.
185. McClintock 1990:37, 47.
186. See the Gujarat State Gazetteers: Broach District 1961: 121, 124; also p. 124 for a vivid description of the rite of exorcism.
187. Wagle 1995:194, 198.
188. McClintock 1986:42–47. Bhattacharyya’s discussion is recommended for its linguistic sensitivity.
189. On Vodoun possession by loa, the local spirits, see Hurston 1938: “The person mounted does nothing of his own accord. He is the horse of the loa until the spirit departs. Under the whip and guidance of the spirit-rider, the ‘horse’ does and says many things he or she would never have uttered un-ridden” (p. 221).
190. phaḍ ity anena mantreṇa tatrāśvārohaṇa viśet | Āgamarahasyam 3899ab, 2:330.
191. Macdonald 1975:327n5, for regional variants of this word. Dietrich 1998:259 and passim, records the word dhami as a rural variant.
192. Viswa Shanti Devi Yajña Committee (2004):15.
193. Cf. Mumford 1989:148n12; also Aziz 1978:52ff. Mumford labels sngags pas as “magicians” who practice “violent” means of exorcism. In this sense, he may be misrepresenting the teachings and practices of sngags pas themselves, but correctly representing the views of local Tibetan villagers in remote parts of Nepal who, over centuries of direct contact and folklore, have come to see them as such.
194. See, for example, the Sanskrit manuscripts on possession discussed in Chapter 11.
195. Although this is correct in its Sanskritic reference, Gellner notes that in Newari the word vaidya refers to “any type of healer who is not directly possessed” (2001:97).
196. See Carstairs and Kapur 1976.
197. Brückner 1995:89ff.; Honko 1998:341. Claus is more specific, stating that this is used for a spirit medium (1979:29).
199. Höfer 1994:18. Macdonald’s designation is also too restrictive for Gellner, who, like Höfer, regards a jhṃkri as a shaman in a more general sense; cf. Gellner 1994:31.
200. Ibid.; also Merz 1996:347. Hāratī, the “goddess who steals,” is a Hindu-Buddhist crossover deity. She is worshipped by both Hindus and Buddhists, her temple ritual is performed by Newar Buddhist Vajrācāryas, and she is related to other deities of possession, including her husband Unmatta (“Drunken”) Bhairava and several of their children.
202. Gellner 1994:31 (= 2001:202).
203. Gellner 1994:32 (2001:203).
206. Gaborieau 1975a:73, 76ff.
208. Dietrich 1998:119–159.
210. The kabirāj is also discussed by Risley 1891:362–366.
211. Karim 1988:295. The word bhasan, writes Karim, “suggests one who is floated” (p. 307n15).
213. Caldwell 1996:223n23. It is of little value to discuss words found in dictionaries without comment or contextualization. Thus it suffices to merely mention that the old standard English and Tamil Dictionary, by Winslow, Spaulding, and Pillai (1888 [1842]), lists the following: under “possession by an evil spirit”–tēvarālan, veṟiyāṭṭu, āvēsam (p. 1019); under “oracle”–āvēsakkāram, tēvarālan, paṭimatton.
215. Inglis 1985:90–91; for a similar list see Diehl 1956:221ff. Inglis studied the similarities between possession dancers and the local manufacture of clay images, both serving as vessels for the divine. We have seen this image of the possessed as a vessel in Tibetan and in Tulu, via Sanskrit. Doubtless it occurs elsewhere in South Asia as well and is a common image elsewhere in the world. For example, Spanish attests vasos preferidos (chosen vessels) and cajitas (little boxes), for those who accept certain spirits for oracular and curative purposes; cf. Kay F. Turner 1992:126.
218. Ibid.:242ff.; for more on the evil eye in the Maldives, see pp. 258ff. This is not unlike a brujo in Mexico. On the evil eye in Sri Lanka, see Wirz 1954:178f.
219. On Telugu, see Knipe 2001. He observes that the Telugu word for “possession speech” is vakku (cf. the observation by Lawrence in Sri Lanka) and that “possession ritualists who have one or more neighborhood goddesses ‘come to’ or ‘come on’ them, to employ common Telugu parlance for possession [is] an event also known substantively by the word pūnakam” (p. 344). In a more recent article (2004), Knipe states, importantly, that “Godavari Brahmans are reluctant to mention āveśa, samāveśa, or related experiences, whereas pūnakam, the moment when Ammavāru or a deceased person descends on the occupant of a gaddi [a ‘seat’ for a possession ritualist, shared with the goddess], is eagerly discussed by all ritualists who claim routine divine visitation, as well as by their clients and assembled audiences who share each detail with the village and town folk at large” (p. 440). See also Nuckolls (1991a), who mentions that in Telugu a person on or in whose body a deity comes is called bhaktudù (m) or bhakturàlu (f) (< Skt. bhakta; a devotee). I suspect that these words have broader semantic fields.
220. However, see the 1961 edition of the Broach (present-day Bharuch) district gazetteer (123ff.). Spirits are referred to as bhūts, as is the case all over northern India, the word for possession is vaḷgan. Spirits are divided into family or household spirits (bharṇābhūt) and outside spirits (baharṇābhūt). If a man remarries after the death of his first wife, it is believed that the spirit of the first wife might possess the second. However, if post-mortem rites are properly performed, the spirit will be pacified. An array of spirits and goddesses of lower-caste Hindus are believed to be responsible for the possession. More fieldwork must be done on possession in Gujarat to fill out this very thin picture.
221. See, for example, Troisi 1976, which lists more than five hundred studies of this “tribe,” as it is still classified, of Bengal, Orissa, and Bihar, most of them in English, in many of which, doubtless, one can find references to spirit possession.
222. Nevertheless, I am reminded by William Sax that though “everyone gets possessed, high caste and low, upper caste people tend to deny it of themselves [at least in Garhwal], and attribute it to the subalterns” (personal communication, November 1999). This he attributes to brahmanical image-conjuring rather than suppression, though he does see an effort of Hindu nationalists to suppress possession and other kinds of popular Hinduism. This prurience is little different from that of their detested Victorian predecessors. Pakaslahti has hard evidence to back up Sax’s observations: “Seeking help from Bālājī is not related to illiteracy, low caste status or rural domicile, as the majority of the clientele has average or better than average education, belong to the higher castes and are of urban domicile” (1998:164).
224. Knipe notes that vaidika brahmans in the community he has studied in coastal Andhra Pradesh do “not enter into a heightened state of consciousness after drinking soma.” Yet they are “well aware of the emotional rhetoric of transformation, most dramatically in the large-scale sharing of soma in a śrauta sacrifice such as the vājapeya, pauṇḍarīka, sarvatomukha or the currently rare agniṣṭoma for a newly initiated member of the soma society” (2004:438).
225. For example, the presence in the Draupadī cult’s Mahābhārata of a Muslim devotee, Muttāl Rāvuttaṉ, who serves Draupadī by accepting blood sacrifices for her and receiving impure offerings, including gāñjā, opium, cigars, and arrack. At one temple, red and green chili peppers are offered to his dog. Cf. Hiltebeitel 1988:101–127. It is difficult to imagine a more non-brahmanical Mahābhārata.
226. Cf. Prabhu 1977; Honko 1998:389–421; also Knipe 1989. It is important not to commit the error of assuming that because the vocabulary of possession in northern India is in a large number of instances derived from Sanskrit, while in southern India it is less the case, possession in the north is more inclusive of upper classes. North Indian languages are Indo-Aryan while south Indian languages are not, so the idiom of possession in the north would “naturally” be more Sanskritic. Similarly, the presence of a Dravidian vocabulary for possession in the south cannot be used as evidence that the possession is somehow non-brahmanical. Correlation of language with culture is an idea that is dying all too slowly; one would have thought that this would have ceased sixty-five years ago with the publication of Frans Boas’s Race, Language and Culture (1940).
227. Note the superannuation of the term aṇaṅku in present-day Tamil (or Malayalam) by the concept (and term) śakti.
228. Cf. Jones 1968. See now Davidson 2002 for a wealth of evidence that tribal practices influenced the development of Mahāyāna Buddhism in the first millennium C.E.
229. Berglie 1996. This epic is still transmitted largely orally, through bards of virtually every stratum of Tibetan society. No standard edition of the text exists, nor is one likely to be produced anytime soon. Geoffrey Samuel has also done important work on the traditions and shamanic dimensions of the Gesar epic; see Samuel 1991, 1994, 2005.
231. I agree with Gellner (2001), following Dumont and Pocock, that “it is probably better to talk only of Great Traditions, which certainly exist, and to contrast them with ‘popular religion.’ To call the latter a Little Tradition either appears to imply that it is organized in the same way as the Great Traditions (which it is not) or that it constitutes a pool of survivals of pre-Hindu or pre-textual practices (which is doubtful)” (p. 100). Here I equate “folk” with Gellner’s “popular religion.”
232. Cf. Rigopoulos 1993:54n18. Rigopoulos reports that a devotee of Sai Baba, possessed by Khaṇḍobā, supplied information on Sai Baba’s personal history (47f., 53f.) (which, unfortunately, Rigopoulos accepts as pramāṇa). Sontheimer (1989a, 1989b, and elsewhere) also supports the folk vs. classical Hinduism dichotomy. But the dichotomy Sontheimer writes about is based only on recent historical and contemporary arrangements.
233. Cf. Gellner 2001:203. For certain spiritual practices (upāsana) that involve possession of Hanumān, see Śrīmālī n.d.
234. Gellner writes: “a Maharjan (Farmer) man from a village south of Lalitpur who appeared one day in Kwā Bāhāḥ (an important Buddhist sacred complex) and claimed to be possessed by Kwābāju, i.e. Śākyamuni Buddha, was met with polite but definite scepticism” (2001:211). For a description of a spirit medium who “dances” Śākyamuni Buddha in contemporary Singapore, see DeBernardi 1995:158.
235. See Holt 2004a, 2004b for Viṣṇu possession in Sri Lanka; and Gellner 1988 [2001:102] on the possession of the king of Nepal: “He was considered to be like Vishnu and to be possessed by him in ritual contexts.”
236. See also Dietrich 1998:236, for an account of a Nepali healer who claims possession by Brahmā, Viṣṇu, and Śiva, among others.
237. Not only is the latter apparent from much of the evidence presented in this volume, but one of the primary areas of Indological study in the twentieth century was the determination of Aryan and non-Aryan in early Sanskrit literature. See, for example, Deshpande 1995.
238. Cf. Hiltebeitel 1991.
In reading about South Asia, one often gets the impression that possession, along with related ecstatic behavior, belongs to the “little tradition” and is thus largely confined to the lower castes and the poor and uneducated in rural areas. However, in my experience, the phenomenon of possession in the cult of Śerāṇvālī is widespread throughout the population. I have witnessed Goddess possession in both village and urban settings, among low and high castes (including Brahmins), the poor and the rich, the uneducated and the educated, Sikhs as well as Hindus. (1996:179)
240. Bell argues that ritualization is not a “standardized process” of traditionalization, but that:
[r]itual can be a strategic way to “traditionalize,” that is, to construct a type of tradition, but in so doing it can also challenge and renegotiate the very basis of tradition to the point of upending much of what has been seen as fixed previously or by other groups. Examples abound, ranging from the ascetic internalization of the Vedic sacrifice and the iconoclasm of early Ch’an Buddhism to the Reformation’s challenge to papal authority through a recreation of the free outpouring of the spirit to the early church. (1992:124)
241. For an accessible general discussion of this, see Bryant 2001, ch. 5. As I write this in Iowa, I am surrounded by the states of Minnesota, Wisconsin, Illinois, Missouri, South Dakota and Nebraska, while the best-known place names in the region are those of the Mississippi River and the city of Chicago. All of these are Native American names, the total number of which by now surely exceeds the number of Native Americans in these areas who have any knowledge of their ancestral languages. Although an extreme example, it is nevertheless not wholly dissimilar to the results found elsewhere of linguistic and cultural contact.
243. See F. Smith 1987:59–63, 121–223.
245. Dirks 1992:231. Possession within epic or other narrativized ritual performance in south India yields rather similar results; cf. Hiltebeitel, Brückner, Biardeau, Caldwell, Freeman, and others.
246. Zvelebil 1981:82. See also Stewart 1995:574ff. for a modern story of a faked devotional possession during the life of Haridāsa, a devotee of Śrī Caitanya in the early sixteenth century.
247. Yocum 1982:144. There is earlier testimony for this in the Mahābhārata tale of Aśvatthāman, discussed in Chapter 6.