1. For important recent ethnographic works of this type, see Abu-Lughod (1986), Lindholm (1982), Lutz (1988), Myers (1986), M. Rosaldo (1980, 1984), R. Rosaldo (1984), Schieffelin (1976, 1983), White and Kirkpatrick (1985). Lutz (1988) also contains an important theoretical critique. For an extensive recent review of anthropological work on emotion, see Lutz and White (1986). Heelas (1986) offers a survey of cross-cultural differences in emotion. Harré (1986) is a recent collection on the theory of the social construction of emotion.

2. The following paragraphs up to and including that on behaviorism rely heavily on Lyons (1980:1-32).

3. Kemper (1987; see also 1981 ) is a recent attempt to reconcile a physicalist with a social constructionist perspective on emotion. He says four primary emotions—fear, anger, depression, and satisfaction—are caused by the action of neurochemicals on the body's autonomic system. All other emotions, in his opinion, are derived from the primary four by processes of socialization and social construction. But see note to below.

4. For a recent critique of this aspect of Durkheim's theory of ritual, see Kapferer (1979).

5. Spinoza's was also a cognitive theory of emotion, but he seems to have reduced it to a belief accompanied by a feeling. His theory is wanting in that it does not distinguish emotional feelings from sensations and in that it lacks an evaluative component such that emotions can be motives (cf. Lyons 1980:37-40; Leavitt 1985).

6. A basic introduction to the social construction of emotions may be found in Harré (1986). Some applications and some conflicting views on the approach in anthropology may be found in Shweder and LeVine (1984), and a lively, if somewhat idiosyncratic, discussion may be found in Solomon (1976). Lyons is a good introduction to the important work done by philosophers, but unjustly neglected by anthropologists, on emotions.

Theoretical influences on social constructionism have been many. Berger and Luckmann (1966) drawing on phenomenology have provided an important general statement on social constructionism. Wittgenstein's (1958, 1980) later theories of language and his statements on emotions in particular have also been important, as has the work of his interpreter Winch (1967). Mead (1962), Cooley (1964), and their successors in the symbolic interactionist school provided important insights on emotional socialization and on emotions as social emergents. Denzin (1984) is a statement of the extremes to which that approach may go in trying to elaborate on the emotional experience of the intentional ego.

7. Although I have drawn from many sources in identifying these characteristics, I am particularly indebted to Armon-Jones (1986), Harré (1986), Averill (1980, 1986), and Solomon (1976).

8. "Academic psychologists have begun to accumulate evidence, however, suggesting that the number of danger response elicitors present from birth is much smaller than was once thought" (Lutz 1983:257).

9. There is a Hindi verbal construction mahsus karna or mahsus hona . But this means "it seems like" or "I feel" in general; it is not used with any specific nouns for emotion, such as anger, or the like.

10. For further references and studies on the Schacter and Singer study, see Gordon (1981:573) and Kemper (1987: 272-274 ). Kemper (1987, 1981) rejects Schacter and Singer's conclusions in favor of four physiologically grounded primary emotions: fear, anger, depression, and satisfaction. Many arguments can be raised against Kemper's thesis. I mention three. First, he assumes that emotions by definition are physiological sensations. He thereby separates them from all those characteristics of emotion outlined in this essay. Second, one could as easily and as arbitrarily start with other emotions, such as humor and love which Kemper considers secondary, as paradigmatic and be led to a richer conception of emotion. Such a different beginning could include physiological correlates as well as all that is humanly important in the concept of emotion. Finally, implicit in his assumption is the idea that emotions must be measurable phenomena (cf. Lutz 1988:220). Because neurochemical correlates of some emotions are measurable, by a process of circular reasoning they are emotions. Kemper relegates to a footnote the following statement: "Autonomic differentiation of emotions does not imply that persons experiencing emotions are always aware of, or can report correctly, their underlying physiological processes" (Kemper 1987:271, n. 3). But, if the emotion is the physiological sensation, why does it not tell one what one's emotion is? Precisely because sensations are not emotions and need some social context in order to be interpreted as emotions, just as the Schacter and Singer thesis states.

11. "Intentions, unlike the behaviors they intend, are not behaviorally observable. Neither, therefore, are the emotions" (Solomon 1976:166).

12. See Bedford 1986; Solomon 1976:163-170; Lyons 1980: 17-25.

13. For anthropological studies elaborating this point, see M. Rosaldo (1980), Myers (1986), Lutz (1988), Abu-Lughod (1986).

14. This is merely to interpret emotions in terms of their activity, not in terms of functional theory in which function means contribution to the maintenance and survival of the whole. See Greenberg (1957:75-85).

15. For further information on the rasa theory, its textual bases, and interpretive variations of it, see de Bary et al. (1958:258-275).

16. My understanding of Derrida in the following paragraphs is based upon Derrida (1973, 1976), Spivak (1976), Leitch (1983), and Ryan (1982).

17. The "same... is not the identical. The same is precisely differance (with an a), as the diverted and equivocal passage from one difference to another, from one term of the opposition to the other... the other as 'differed' within the systematic order of the same" (Derrids 1973:148).

18. This is true even for the natural sciences. According to Thomas Kuhn (1970), the continuity of scientific progress is illusionary. When a scientific revolution, such as that brought about by Einstein's theories, occurs, then many old formulas expressed in signs or symbols are carried over. But their basic meaning changes within the context of the new theoretical system and its very different underlying assumptions.

19. Barnett (1977) mentions Dumont (1970) as the source of this term, although Dumont gives it a different meaning than does Barnett. Marriott and Inden (1977) develop the idea of the identity of substance and value in India and mention Barnett's dissertation. All acknowledge, in regard to this term, the stimulation of David Schneider's (1968) ideas on American kinship.

20. For an enlightening discussion and application of this insight to an ethnography of south India, see Daniel (1984).

1. Thoughts put into my head by Clifford's (1983) account of Griaule, Stocking's (1968) account of Boas, Allen's (1985) account of Mauss, Crapanzano's (1986) account of Geertz, Malinowski's (1967) account of himself, and work by Kristeva (1984) and Bakhtin (1981).

2. To them I was America, and they were India to me. They understood that I was observing them in order to learn about their way of life and write about it. It was important to them that they be represented well in the world, and so they offered a particular face to me. They would represent themselves in other ways to other people; the representation depended upon the audience or, more precisely, upon their assessment of the audience. In turn, their image of this audience would devolve partly from what face the audience presented to them. We constructed each other and ourselves with respect to each other, and, because we began as strangers with few rules in common, we were probably dancing a rather strange dance.

3. I write this in response to the idea, first articulated by Boas and still widely subscribed to by many anthropologists, that conscious explanations of cultural practices on the part of the practitioners are secondary elaborations that only obscure the true nature of the practices in question. The Marxist definition of ideology as an expression of class interest contributes to this view. Certainly descriptions of a culture coming from actors within the culture cannot be disinterested. The questions are whether a disinterested description of human life is ever possible and whether an ethnographer can in any case truthfully represent herself as being outside the cultural system she describes. I find it most reasonable to assume that an indigenous analysis of a cultural system is no more likely to be erroneous and distorted than an outsider's analysis of that same system, and it will certainly have a larger store of information as its base.

4. Here I can scarcely begin to outline the history of the idea of anpu in Tamil culture. Tamil "poetry of the interior," a large body of lyric poetry among the earliest Tamil literature, voices sentiments that are uncannily similar in some ways to modem Western romantic love. The term anpu appears to have essentially the same set of meanings in this ancient poetry as it does in modern Tamil Nadu. Tirukkural^{[ *]} and Tolkappiyam , early books of social and linguistic ideals, describe the delights that parents may find in the play and speech of their children, not as heirs, but merely as children. Modern and ancient Tamil literature idealizing mother love is extensive, as is the literature and mythology on love between siblings (e.g., Ponnar Cankar^{[ *]} Katai ), love between lovers (e.g., Cilipatikaram ), and love between spouses (e.g., Kamparamayanam ). The religion of bhakti , which originated in Tamil Nadu and still is "the religion of the masses" there (though it is a sentiment, not a creed), is based upon the premise that natural human love is the most powerful force available to human beings; directed toward a deity, this love can easily free the human spirit. So one's chosen god is adored as a mother, or as a child, or as a lover, or as a friend. A Tamil individual's relationship to a deity, if the person has a deity, is always a relationship of love. The more powerful the relationship with the deity, the more intense the emotion, although the devotion may be founded upon deep anger just as one's love for one's parents may be. Very commonly, Tamil people will say that, in worshiping a deity, the particular materials offered or rituals performed are not important but the feeling one has for the god is important. So Shaiva religion is called anpu mikunta matam (the religion filled with love). Similarly, in their relationships with each other, human beings in Tamil Nadu will often affirm verbally that anpu is all that matters, and they will break social rules in order to make this point (for ethnographic accounts see, Daniel 1984:233-278; Singer 1972:148-245).

5. This is a statement I heard only once, in the context of a brief conversation. Some individuals regarded the expenditure of money as necessary to the enactment of love ( anpu ). Others saw the exchange of money as opposed to anpu . Because money exchanges often were a sign of market relationships devoid of personal commitments, some people that I interviewed refused offers of money in exchange for interview time; others accepted money as a gift in the expectation that more such gifts would follow.

6. One reader of an earlier version of this essay has suggested that the apparent unkindnesses that took place in this family under the name of anpu were no more than outlets for suppressed tensions; in particular, mothers who mistreated their children were perhaps taking out on the children their resentment at being subordinated to men. I think that this would be an incorrect interpretation of events, for, in this family, tensions even over such matters as money and sex were not suppressed but freely ventilated. Nor were women as a class subordinated to men as a class: if a woman was angry with a man, she took it out on him directly. For reasons that I have discussed above, I think it would he misleading for readers to imagine that Tamil people who enact anpu in ways that appear paradoxical are pasting an ideological veneer over their raw aggression. Culture is not just a set of labels for things, thoughts, or feelings. It shapes all three from the bottom up. For Westerners to assume both that they know how people of another culture feel and that the Tamil accounts of feelings are mere rationalizations for behavior whose underlying motivations Westerners know better than the Tamil would be counter to the spirit of anthropology.

7. Kantirusdi^{[ *]} would be most accurately glossed in English as light from the eyes. The Tamil term itself, unlike the English term "evil eye," suggests not malice but dangerous power.

In both Indian mythology and everyday life, eyes are treated as receptacles of the most important life fluids and as emitters of powerful transformative emotional forces (Maloney 1976; Eck 1981). These forces are as substantial and material as water, fire, or blood (Babb 1981). The power in the eyes has a dangerous erotic component. A woman may lose her sakti either by looking with desire at a man or by being viewed with desire by one (Egnor 1980, 1983). Emission of light from the eyes is, in Shaiva and Buddhist mythology, parallel to emission of semen from the penis (Obeyesekere 1984; O'Flaherty 1973).

For Indians, the emotional power of the mother in any form is dangerous: it is intense, and it can easily turn into rage. The child cannot protect himself against it, and there is no mediator between the child and his mother. The mother herself must keep it under control. Therefore she does not gaze too intently at the child she bore.

8. This was in contrast to Ayya's idea that a child should be nursed for "at least three years" (he himself was nursed for five). Anni, arguing with him, had said that a child would be a burden to nurse for so long. Ayya had replied, "Is the fruit a burden to the fruit tree?" Then Anni had said, "After the tree had dropped its fruit, if you tried to tie it back onto the branches again, yes, it would be a burden." Ayya was delighted with this response and recounted the story in his lectures.

9. The custom of a wife's avoiding her husband's name in India is interpreted by some observers as a sign of respect, even subordination. Yet in Tamil Nadu, name avoidance can occur even in the absence of any other signs of respect. Such signs include the use of respectful pronominal forms ( ninkal^{[ *]} , avar, avarkal^{[ *]} ), respectful bodily postures and facial expressions (crossed arms, smiling, standing, or squatting rather than sitting or lying down), an attitude of assent and willingness to serve. Such external forms, which are complex with many nuances, are in general supposed to indicate an internal feeling of respect for the person toward whom they are directed, though dissimulation is certainly part of the game. Expressions of respect occur in face-to-face encounters between people of clearly unequal caste, economic, or political status, between people who are unequally educated, between people of widely separate ages, and between both the bride's and groom's kin at weddings. But these conventional expressions indicating acceptance of one's own subordination are noticeably absent in the behavior of many Tamil wives toward their husbands, and I have never heard of any Tamil woman explaining her avoidance of her husband's name in terms of his superiority to her or in terms of distance between them (distance and hierarchy being the two essential components of respect relationships as social scientists are prone to see them). Moreover, name avoidance between spouses in Tamil Nadu is often reciprocal, and sometimes an individual will avoid the name of a kinsperson whose rank is lower than his own. For all these reasons, I feel that a Tamil wife's avoidance of her husband's name cannot be adequately explained in terms of respect. Because Tamil women themselves explain this custom as a means of protection ( kappu ) of the husband, I have chosen to discuss it under the topic of containment. It appears related to the observance of nompu , a fast to protect the husband's life, after which the wife ties a string around her wrist to show that she has fasted for this purpose.

Whether the husband is to be protected for the sake of anpu , or for some other reason (e.g., the guardianship of one's own status as a cumankali^{[ *]} , an auspicious married woman), is not such an easy question to answer. Certainly anpu is supposed to be what binds husband and wife to each other. One standard question I asked interviewees in 1984 was, Among what pair of persons in a family should there be the most anpu ? The stock answer was that anpu should be strongest between husband and wife.

10. Sometimes such practices were explained in terms of protecting the child from the evil eye, kantirusdi^{[ *]} . The power of kantirusdi was not simply a matter of malevolence or envy on the part of onlookers, as the danger of the mother's eye illustrates. Nor was it a matter of demonic forces, for demons are attracted to flaws and impurities and to people in isolation, not (in Tamil Nadu) to those who are well and surrounded by love. Rather, the hiding of a child's beauty and of one's love for it could be seen as a special case of the strong and pervasive sentiment in India that perfection in and of itself is deadly (see Daniel 1984; Narayan 1972:52-55); perfect love, perhaps, is most deadly of all.

11. Here I follow Kapferer's (1983) definition of ritual, as an intentional patterning of the act after the idea.

12. One colleague suggests that pride turned humility into public acts meant to be interpreted as love, and dominance turned acts of servitude into acts meant to be interpreted as love. A compromise between this reader's formulation of events and my own might say that the availability of anpu as an interpretive device enabled actors to transform potentially humiliating situations into vehicles for the expression of pride and so on, and to do so in a way credible within the Tamil cultural context. Tamil Shaiva mythology is replete with paradoxical expressions of love and antihierarchical messages (see note 6 above). Ayya was simply bringing the spirit of this mythology home.

13. See Appadurai (1981) for a detailed discussion of ways in which acts of feeding and eating in Tamil Nadu become messages with negotiable interpretations about kinship, religion, and emotion.

1. For a recent and theoretically fairly rigorous framework for understanding the significance of cross-societal differences in the position of the elderly, see Foner (1984). A classic older work on the subject is Simmons (1945). See also a number of recent collections of essays on aging in crosscultural perpsective, for example: Fry (1980), Amoss and Harrell (1981), Myerhoff and Simic (1978), Sokolovsky (1983), and Brown and Kerns (1985).

2. This research was carried out from September 1974 to February 1976 in an urbanized former village in the city of New Delhi. Fieldwork was supported by NIMH Grant No. ROI MH 24220 and by the American Institute of Indian Studies. Some other publications resulting from this research are Vatuk (1975, 1980, 1982a, 1982b, 1985, 1987).

3. The classical model of the life stages or asrama assumes a male protagonist with the marriage rite marking his entry into the Householder role. The woman enters the picture as a wife who shares with him through the sexual division of labor the duties, responsibilities, and pleasures of this period of his life. She may accompany him when he becomes a Hermit, and, if she does so, should follow a similar regimen. Of course, when he becomes a Renouncer, she is prime among those with whom he must sever his ties of attachment and interdependence.

4. Mines (1981, 1988) makes a similar point when he says that family and social controls over individuals loosen as they grow older; in middle age men and women may grasp opportunities to pursue personal predelictions of lifestyle or to engage in activities previously barred to them, either because of conflicting responsibilities or family opposition. Mines does not, however, explore the issue of gender differences, perhaps because he interviewed few women.

5. A survey of some of these data may be found in Vatuk (1982a).

6. See, for example, Harlan (1964), Raj and Prasad (1971), Soodan (1975), and Marulasiddaiah (1969).

1. My thinking in this paragraph has been much influenced by Wilson (1972), Rosaldo (1984), Levy (1984), and Myers (1979, 1985).

2. For a description and analysis of this pilgrimage, see Lynch (1988).

3. This essay gives only a truncated version of the Chaubes' history of themselves. This interpretation is not uncontested by non-Chaubes. Among themselves there is more than one version of their history, as a reading of their journals, Mathur Pradip and Mathur Hitaisi^{[ *]} , would confirm. Pandit Bal Mukund Chaturvedi has published extensively in the latter journal and is acknowledged by many in Mathura to be the most learned historian of the community.

4. This myth, which Chaubes tell about themselves, is a version of that given in Bhagavata Purana 10.32. In this way Chaubes seek to validate their position in Mathura and their relationship to Krishna. I am grateful to John Hawley for pointing out this important correspondence to me.

5. This festival takes place on the bright half of the Hindu lunar month of Kartika .

6. This figure is based upon a register book for a distribution of food called daini which is given to all male Chaubes in Chaubiya Para. The distribution was done in 1978 in celebration of a young man's sacred thread ceremony. The register lists 1, 165 families with 5,674 male members. I have simply doubled that figure for a total of approximately 11,300. My reason for doubling the figure is that in the state of Uttar Pradesh the sex ratio is biased in favor of males. Yet Chaubes say that in their community today there are more females than males, although in the past it was the opposite. Because I believe there is some truth to their assertion, I have split the difference by evening out the sex ratio.

7. The karil shrub, Capparis aphylla .

8. The Chicago school under McKim Marriott has emphasized the notion of substance as both material and moral in Hinduism. It has not related this notion to the seminal ideas of Robertson Smith and Durkheim with which it has much in common. Lannoy (1971:275) notes, ''The Hindu has never divorced the physical from the spiritual; these 'ancient physiologists' ascribed an ethical significance to physiological sensitivity.''

9. Carstairs (1954, 1967:117-119) notes that Brahmans in particular take these attitudes while Rajputs drink alcohol and praise its qualities.

10. In Sanskrit man is manas . For further explication concerning the man and its location in the Ayurvedic conception of the body see Kakar (1982).

11. This way of conceptualizing emotion as externally stimulated conforms very much to the classical rasa theory and has parallels to Wilson's (1972) causal theory of emotion. Also, there is no verb in Hindi equivalent to the English verb to feel. Anger comes to one ( gussa ana ); one does love or envy with another ( usse prem karna , usse irsa karna ). Mahsus hona is not used with nouns for particular emotions.

12. For similar reasons marijuana mixed with brown sugar ( gur^{[ *]} ) and salt is fed to cows and buffaloes when they are sick and off. their fodder.

13. "The Chaubes of the district [Mathura] are famous for being hearty eaters" (Joshi 1968: 101).

14. See Ranjan (1967) for a Chaube's nostalgic description of gardens in Mathura.

15. Jivan Lal Caturvedi (1967) gives a historical list of many famous Chaube wrestlers.

16. For an interesting discussion of the cross-cultural use of hitter and sweet as metaphorical terms pairing physical and psychological properties see Asch (1958).

17. Those whom the Mathura Chaubes label as Mitha reverse the story and say that they are the true Karua Chaturvedis because by taking flight they preserved the true orthodox religion and did not become degraded through Muslim contact and work as pilgrimage priests.

1. My research assistant, Usha Bhagat (now Mrs. Mahendra Dave), and I were members of a Cornell University interdisciplinary research group studying Khalapur from 1953 to 1956, and financed by the Ford Foundation. I owe special thanks to Usha Bhagat Dave, our Chuhra friends, Director Morris Opler, Field Directors John Hitchcock and the late Edward LeClair, and the Ford Foundation. With financial support from the Smithsonian Institution, I returned to do fieldwork in Khalapur from January to June 1984.

The native Hindi terms for the two divisions within the category of servant and artisan castes is suddh and asuddh , clean and unclean. The unclean castes tan leather, make shoes, sweep streets', and clean latrines, work Hindus consider unclean. The term untouchable was first used by the Maharaja of Baroda before the Depressed Classes Mission of Bombay in 1909 (Galanter 1972:298) and became much more widely used than unclean. Discrimination against citizens considered unclean or untouchable is against Indian law and its Constitution (see Kolenda 1985, chap. 4). Old customary practices and beliefs persist, however.

2. The psychoanalyst, Alan Roland (1982), has characterized the Hindu family as thoroughly hierarchical, one in which deference and obedience from children is rewarded by affection from elders. Here in NQ1 Bhati has presented the same analysis: Kayda -keeping (deferential and solicitious) daughters-in-law win their mothers'-in-law affection, she asserts, presumably on the basis of her own experience.

3. Although Marriott's may well be the earliest account of the North Indian superiority of the bridetaker's side over the bridegiver's as a form of hypergamy (as distinct from hypergamy between grades of families), anthropologists' recent focus upon the phenomenon stems from Dumont's ( 1957, 1961, 1966, 1970) concern with marriage alliance; his inspiration came from Lévi-Strauss (1969:269) who found that "generalized exchange" (unilateral asymmetrical giving of brides) "extends over a vast area of Southern Asia."

If North India is a hypergamous milieu, there is, of course, a problem in the definition of the boundaries and territory of North India. Hierarchical relations between bridetakers and bridegivers have been found in studies done in the Indian states of Kashmir (Madan 1975), Uttar Pradesh (Marriott 1955; Dumont 1966; Khare 1975; Vatuk 1975), Delhi (Vatuk 1975), and the Punjab (Hershman 1981), and in the country bordering northern India, Nepal (Bennett 1983). But there are exceptions in the north. Fruzzetti and Ostör (1976: 109) state "that hypergamy as status difference between wife-giving and wife-taking groups on-the-ground is a distortion of the Bengali evidence." Sirkanda villagers (Berreman 1963), in the foothills in northern Dehra Dun District, Uttar Pradesh, also fall to follow this North Indian system of inequality as do the mostly brideprice-giving middle and lower Hindu castes of Rajasthan (Kolenda 1989). It is, thus, doubtful whether some parts of Bengal, the sub-Himalayas or Rajasthan, can be included in a North Indian hypergamons region.

4. Dumont (1966:108-110) addressed the question of the kinship systems of lower castes in localities in Gorakhpur District in eastern Uttar Pradesh; he concluded that, in contrast to high castes like the Sarjupari Brahmans, they lack coherent kinship systems. I did not find this to be the case (Kolenda 1982). All Hindu castes in Khalapur seem to have essentially the same system. The main difference between the Rajput and Brahman kinship systems and those of the Merchants and the middle and lower Hindu castes in Khalapur is the practice of widow remarriage, prohibited by the former two and prescribed by the latter three social categories.

The Rajputs are one large caste-complex in northern and central India famous for its ranked clans (sets of families). For a discussion of such ranking by the Rajputs of village Khalapur, see Hitchcock (1956:51:-54).

5. Throughout this essay, I use the present tense for Chuhra and Khalapur doings and conditions as they were in 1954-56. There have been major changes in Khalapur in the intervening thirty years, but the general conditions as I describe them here have not changed.

6. Describing or mentioning similar celebrations of the brother-sister relationship are Minturn and Hitchcock (1966:36), Dumont (1966:98), Pocock (1972:97), and Bennett (1983:247-252).

7. That even marital sexual activity demeans a woman is asserted by Salman Rushdie in Sham , his novel about Pakistani Muslim Punjabis. "It was believed that the mere fact of being married did not absolve a woman of the shame and dishonour that results from the knowledge that she sleeps regularly with a man" (Rushdie 1983: 76).

8. Das (I976:6) speaks of a Punjabi sister-in-law (brother's wife) bantering with her sister-in-law (husband's sister) about sexual matters.

Bennett (1983:177), writing about eastern Nepali village women, says:

There is a great deal of vigorous sexual joking carried on among the women—sometimes even within earshot of the men. Unmarried girls were told that first intercourse would burn the vagina just as chili-peppers burn the tongue. The suggestive nature of keys being fitted into locks or of the shapes or bananas and cucumbers (and which of their neighbors was thought to relish these foods the most) was a source of endless joking among the women.

Jeffery (1979:75) also implies that in conversations between Muslim women friends in Nizamuddin, Delhi, intimate sexual topics are discussed.

9. That Rajput men when among equals also discuss sex is reported by Dorschner (1983:119).

10. See my discussion of these studies and egalitarian relations within the caste system (Kolenda 1977:181-182, 1981:313-314).

11. Owen Lynch brought to my attention Briggs's (1920:75, 86) mentions of Chamar women related to the bride singing obscene songs to the men of the groom's party during weddings. Jacobson (1982:96) also mentions obscene songs sung to the bridegroom at weddings, to which samdhis (bride's fathers-in-law—that is, the father-in-law and men in his generation including his brothers, cousins, and so on) are expected to take no offence; she writes:

In these songs ( gari git ), the women name the samdhis and accuse them and their womenfolk of engaging in immoral and indecent acts. At some of these gatherings, the bride's womenfolk throw colored water on the visiting affines, tie cowbells around their necks, rub their face with wet flour, and even dress them ridiculously in women's blouses and forehead spangles. Throughout these raucous performances, the women keep their faces properly veiled. The samdhis are expected to take no offense at these songs and insults.

Hershman (1981:164-165, 187, 196-197, 206) makes much of joking between bridegroom and his sails and bride's women: "The songs sung cruelly satirise the paucity, poverty and ill looks of the members of the marriage party" (165).

12. In her treatment of cross-sex affinal joking relationships between brother's wife's brother and sister's husband's sister, husband's younger brother-older brother's wife, and sister's husband-wife's sister, Vatuk (1969:107-109) sees the applicability of Brant's (1948:160-162) generalization that cross-sex joking relationships occur between persons who are potentially marriageable, but she points out that there is no likelihood of such marriages among the high-caste populations she studied (Vatuk 1969: 109).

Widow remarriage is common among the Chuhras and other middle and lower castes in Khalapur; a woman might marry her husband's younger brother or her sister's widower-husband. These men are definitely considered as possible mates in secondary marriages for a widow (see Kolenda 1982), and joking relations do occur between these potential mates.

13. Caca is a father's younger brother. A child never calls the father bap or pita , correct terms, while his or her father is of child-bearing age; instead, the child avoids suggesting procreative activities of the parent by calling him caca , a person who may well not yet be married.

14. After his description of the relationship between Hindu brothers-in-law, Apte cites a proverb about making a maternal uncle the butt of a joke ''among the Marathi speakers on the west coast of India.'' Because Apte himself is aware of the paucity of descriptions of the relationship between wife's brother and sister's husband, he would probably be quite willing to admit that his own description of the relationship is probably not accurate for all Hindus of North India.

15. Hindus believe any bodily exuviae or product is polluting, including saliva; hence, they find biting, kissing, or licking someone polluting.

16. In discussing a joking relationship between a brother's wife and a husband's sister, Vatuk says (1969:109) that "one aspect of the humour is the suggestion that an exchange of women has or will take place." Among many North Indians, exchange marriage is unthinkable (Karve 1965:125; Dumont 1966:104-107; Kolenda 1978: 266, 1987: 2o2-2o3).

17. Kharu's making of the author, Pauline in these notes, into a father's mother ( dadi ) was probably a way of showing respect, but perhaps it was also a way of recognizing that she moved about the village freely the way only a dadi , an elderly woman among them, would do.

18. Apte (1985:43) states, "The existing ethnographic literature also indicates that institutionalized joking rarely occurs between females of the same generation." He then mentions only three instances he found in the literature. This institutionalized joking between samdhan and samdhan , the two mothers-in-law of a married couple, among the Chuhras would make a fourth instance.

Relatively few recent ethnographers mention the relationship between parents-in-law. Bennett ( 1983:160) says that the relationship between the two sets of parents-in-law is particularly "strained." She goes on to say that "all affinal relations are marked by distance-respect behavior." Jacobson (1982:95) says that a Bhopal Hindu woman veils her face in the presence of her samdhi but that, if two samdhans meet, they both should be veiled although able to speak to each other.

Hershman also makes no mention of joking between samdhi-samdhan but speaks of respect between them (Hershman 1981:203). Vatuk (1969:107-109) does not discuss joking relationships between male affinal relatives, such as that between samdhis . In fact, Vatuk (1975:183) characterizes the latter relationship as an avoidance relationship, as are relationships between co-parents-in-law involving the opposing mothers-in-law and the opposing mother-in-law and father-in-law. Such respect relationships are the opposite of the joking relationships between the co-parents-in-law among the Chuhras.

19. Jacobson (1982:92) speaks of the elder brother's wife teasing her husband's younger brother "about his love for his wife." Hershman (1981:175) says that "when she first enters her husband's household [she] jokingly demands payment in order not to veil her face from him [the husband's younger brother]," and a joking relationship between them thus begins (1981:195). Others speak of the relationship as one in which the woman does not have to veil her face (Mehta 1982:142) or like the relationships the daughter-in-law is used to in her natal home (Bennett 1983:213).

The most thorough treatment of the devar-bhavaj relationship is by Hivale (1943) who describes it among two tribal groups, the Gonds and the Pardhans of east Mandla, Madhya Pradesh. There are many aspects to the relationship, partly depending upon the difference in age between the devar and bhavaj . The joking he transcribes (Hivale 1943: 162-164) in which devar suggests that the bhavaj's brother has come in order to have sex with her (thus brother-sister incest); then the bhavaj replies that her nanand (husband's sister) is nearby to accommodate her brother's sexual impulse (thus an exchange marriage or sexual liaison). This indicates that joking about incest is not tabu, contrary to what Murdock and others might have supposed, nor is joking about exchange marriage. Another good treatment concerning mostly tribals is by Naik (1947).

20. That relations between male affinal kin are commonly friendly is also suggested by Minturn and Hitchcock (1966:33):

Easy, friendly relations are also common between a man and his wife's father and brothers. The husband usually accompanies his wife when she goes to visit her parents and returns to pick her up. On both occasions he often stays for a day or more, and most men look forward m these visits. Although the husband is treated with some deference by his wife's people, for they are culturally defined as somewhat lower than he in status—a fact which may account for some of his pleasure—the general tone of the relationship is comradely and without strain.

Jacobson (1980: 100) notes for her Bhopal Hindus:

But the ridicule and insults, offered as they are within the limits of the prescribed joking relationship, virtually never cause ill-will and instead contribute to ostensibly friendly relations between the two affinally-linked groups.

Hershman (1981:196) also suggests that friendly relations develop between affinal male kinsmen. Lévi-Strauss (1969:302-303) was the first to suggest that groups involved in a one-directional cycle of bridegiving were likely to develop hierarchical relations that they might or might not equalize through affection.

1. These sects ( sampradaya ) are known by a variety of names in Western sources. Vallabhites, or Vallabacaryis, as they have been called by Weber, refer to their sect as Pushti Mart. Chaitanyaites, referred to as Bengali Vaishnavas in a now familiar work in English by De (1961), refer to themselves as Madhva-Gauriya-Vaisnava^{[ *]} . This title refers to the fact that these Vaishnavas are most commonly found in Gauda country (Bengal) and follow the philosophy of the South Indian saint, Madhva. Given the complex comparative framework of this chapter, for the sake of convenience, I go back to an earlier scholarly convention of referring to the sects by the names of their founder-saints, Vallabha and Chaitanya.

2. Sapir (1977) explains that the contrast between metaphor and metonym represents, at the level of figurative language, a contrast between paradigmatic replacement (for metaphor) and syntagmatic continuity or combination (for metonymy). The same author also says that, with respect to the notion of a shared domain or a common ground, metonymy can be taken as the logical inverse of metaphor. Metaphor is the relationship of two terms from separate domains that share overlapping features; metonymy is the relationship of two terms that share a common ground but do not share common features. Metonymies are usually identified by their substitution of one cause for another: agent for act, cause for effect, container for contained.

3. Fabian (1983) incisively attacks the use of visual metaphor in anthropological discourse. He contends that persistent recourse to visualization has denied ethnographers understanding of more temporal aspects of the cultural Other, such as might come about through increased sensitivity to language.

4. Tyler (1984) provides examples of gustatory tropes in Standard Average European thought and language. For instance, we "ruminate," "digest thoughts," "chew the cud," and even find some thoughts "hard to swallow." Tyler also alerts us to the fact that one major verb for knowing in the Romance languages is sapere, "to savor, or taste.'' This he attributes to the well-known Latin and Gallic preference for gustatory sensation.

5. See Vaudeville (1980) for more examples of these episodes. This author hints at the deconstructionist possibilities inherent in sectarian literature, where, through a number of insertions and deletions, the sectarian view of Sri Nathji as a pampered and worldly prince is undermined by folk conceptions of him as a local ruffian. Thus, Krishna's princely and knavish sides often appear incongruously juxtaposed in the same episode of a sectarian text.

6. The Govinda Deva icon of the Chaitanyaite and the Sri Nathji icon of the Vallabhite sects were both manifest in this way. In Braj folklore, streams of milk are said to miraculously appear in Govardhan's main bathing pond, Manasi Gangs, during full moon. Folk beliefs in the auspicious qualities of milk are reflected in the pilgrimage practice of walking fourteen miles around the hill, dripping a continuous stream of milk from a hole bored in the bottom of a vessel. The vessel needs to be constantly refilled by one's priest who ambles alongside for the entire journey.

7. There is a compelling illustration of this in Govardhan lore. Govardhan hill, it is recounted, was so devoted to Krishna during his lifetime in Braj that it turned into a lump of butter. This is why imprints of his footprints, crown, mouth, and so on can still be seen today as geomorphic impressions in the hill's stones.

8. In his childhood Krishna steals milk and curds from his mother and other women in the neighborhood of Braj. Later, the comely adolescent taunts the milkmaids for their dairy products in the familiar episode known as the danghati-lila^{[ *]} . In this lila Krishna masquerades as a toll collector in order to trick the unwitting maidens as they take their wares to market. At the height of the fun, he unmasks himself, smashing all the pots and drenching the revellers in milk, butter, and curds. This festive event is reenacted from time to time at Barsans, another popular pilgrimage town in Brai reputed to be Radha's birthplace.

9. Seeing is an extrusive process for Hindus, an outward reaching process that in one way actually engages (in a flowlike manner) the object seen (Babb 1981). See Lynch (1988) on further aspects of darsana as seeing and its relationship to bhava , bhavana , and rasa in Braj pilgrimage.

10. Informants say that prasada is actually heavier than bhoga because it has added to it the weight of the lord's sweet love. Pious devotees are reputed to be able to tell the difference between potential offerings and those that have already been offered to and consumed by Krishna, through "feeling tones" in the food itself.

11. These metonymies take the form of synecdoches: the trope formed when a part is substituted for the whole. In speech, the names Krishna and Govardhan are used interchangeably to refer to both Krishna and the sacred hill. In ritual, a single stone from the hill often substitutes for the whole. Synecdoche is also characteristic of other sorts of ritual practices. For example, only a portion of food cooked each day in temples is actually offered in front of the temple image. After it is offered, prasada is then mixed with the remaining cooked food, transforming it, by contact, into prasada .

12. Ramanujan (1981) maintains that this circular motif is common in Hindu myth and ritual. He refers to it as an act of "mutual cannibalism," wherein the eater is eaten and the container contained, in a repeated metonymy.

13. Babu Lal Sharma, Das Bisa Mohalla, Govardhan, interview, October 22, 1979.

14. See Barz (1976) and Redington (1983) for differing opinions on bhava in this sect.

15. This process of deindividualization is rather strikingly reflected in Hindi kinship terminology, where the use of the mother term ( ma ) is reserved, not for the natural mother herself, but for a senior woman of the family, usually the mother-in- law. Vatuk (1982) notes that this strategic pattern of address mitigates against the possibility of a mother-child unit asserting its independence against the family as a whole.

16. Another interpretation has it that the number fifty-six symbolizes all possible food in the cosmos. By their reckoning there are fourteen worlds ( lokas ) in the universe and four basic substances—beverages, foods that do not require chewing, foods that are chewed, and those that are licked.

17. Goswami Krishnajivanji, interview at Jatipura, September 14, 1979.

18. In some representations (e.g., jugal Kisor, Syama-Syam , and Larili-Lal^{[ *]} ) Radha and Krishna are conjoined in a single icon; in others, a small icon of Radha is placed at the side of Krishna who stands in the classic tribhangi^{[ *]} posture (his body bent in three places, with his head to one side, his upper body twisted, and his right calf crossed in front of his left with the ball of his right foot resting on the ground); in still further variants, Radha's presence is signified by her name or a coronet only, placed on a cushion beside Krishna's solo image (Entwistle 1987:79).

19. A study of these two very different interpretations of the same emotion and their effect on ritual performances in each branch of the sect would make an interesting topic for future study. For example, large food offerings and displays are found in temples run by Chaitanyaite gosvamis in Brindaban, something one might never see among ascetics at Radhakund. In one such temple, Brindaban's Radharaman Temple, the largest food offering is the "Thirty-Six Delicacies" ( Chattisa-Vyanjana ), a feast described in Braj poetry inspired by the erotic sentiment. In this offering, a carved wooden figure of each gopi is displayed holding a dish in hand; thus each cowmaiden is believed to provide Krishna a unique and different amorous experience, here ex- pressed in gastronomic terms.

20. Narayana Maharaj, interview at Radhakund, January 12, 1980.

21. Compare Singer's (1984) data on food categories and the semiotic structure of meals in an ISKCON temple in metropolitan Philadelphia.

22. This point is rather nicely illustrated in the popular image of Krishna as "mountain holder" ( Govardhannath or Giridhari ), one of the first images to appear in Krishna iconography. This image derives once again from an episode in the Govardhan myth. Briefly summarized, after Annakuta was offered, the Vedic god Indra, for whom the offering was originally intended, felt insulted and pelted Braj with rain for seven days and nights. To protect the locals, Krishna held the mountain aloft on his fingertip, umbrella-style, above the entire region. The notion that the Will encompasses all emotional experience is visually reinforced in this important and widely revered image of Krishna.

23. Local pandas explain that one stone is standing Krishna, with impressions of Krishna's crown ( mukut^{[ *]} ) in it; the second, a low-lying stone, is said to be Mount Govardhan, complete with the imprint of a mouth. Reflected in this ritual image is the same bifurcated image alluded to in the Govardhan myth. The two stones are seen as one, each a devotee of the other: standing Krishna as devotee of the lower stone, Govardhan hill, and vice versa. They are dressed as mirror images each afternoon, with identical faces, costumes, and jewelry. The fact that these twin images are located in a large pond at the center of the hill further enhances this mirror effect.

24. Vatuk and Vatuk (1979:179-189) discuss the symbolic importance of sweets in the social and ritual life of North India: "The role of sweets in lubricating all types of social intercourse in this part of India, and the mental association created by this role, have been discussed here to demonstrate that frequent and heavy consumption of sweets is conceived of, in this culture, as an activity of very positive value and, in fact, as an obligatory activity in terms of the individual's successful participation in his community's social and ritual life."

25. Shrivatsa Goswami, interview at Jaisingh Ghera, Brindaban, January 30, 1979.

1. Braj is the region around the city of Mathura in the modern Indian state of Uttar Pradesh where Krishna is supposed to have spent his infancy and youth. It is a major center of pilgrimage for Vaishnavas.

2. My dissertation focuses on temple organization and worship in Pushti Marg (Bennett 1983). Barz (1976) is an excellent study of the sect. Other works in English include Bhatt (1979), Jindel (1976), Marfatia (1967), Parekh (1943), Shah (1969), and Toothi (1935).

3. See Wach (1948: 128) and Barz (1976:39-41) for definitions of the Hindu sampradaya and Burghart (1978) for an explanation of its structure.

4. Vallabhacarya is generally regarded as an incarnation of Lord Krishna, or more specifically as an Incarnation of the Mouth of Krishna ( Mukhavatara ) and an Essential Form of Agni ( Agnisvarupa ). His male descendants are also revered as incarnations of Krishna, although their spiritual standing has been ambiguous and controversial (Bennett 1983:78-128).

5. The Bhagavata-Purana^{[ *]} is a ninth-century South Indian Sanskrit epic that describes the earthly life of the cowherd god Krishna during his sojourn in Braj. The Rasapancadhyayi^{[ *]} , or the five chapters of the tenth book which cover the Rasa Lila , is a particular favorite of the Krishnaite sects and cults.

6. The appearance of the image of Sri Nathji and its early history are recorded in a popular sectarian chronicle written in the Braj language and translated as "The Account of the Manifestation of Sri Nathji" (Harirayji 1968), the first part of which has recently been translated into English by Vaudeville (1980).

7. The original deities along with their present locations are as follows: Sri Nathji (Nathdwara), Sri Navanitpriyaji (Nathdwara), Sri Mathureshji (Kota), Sri Vitthalnathji^{[ *]} (Nathdwara), Sri Dvarakanathji (Kankaroli), Sri Gokulnathji , (Gokul, Braj), Sri Gokulcandramaji (Kamavan), Sri Mukundrayji (Varanasi), Sri Balkrsnaji^{[ *]} (Surat), and Sri Madanmohanji (Kamavan). It should be noted that, although temple worship is relatively standardized, there are nevertheless significant variations in ritual style affected partly by the different regions where temples have proliferated and partly by the segmentary structure of the sampradaya . For example, initiates of the fourth house assert a far greater degree of independence than their cosectaries in other houses by revering their own preceptors (descendants of Vitthalnatha's fourth son, Gokulnathji) as the only legitimate spiritual successors to Vallabhacarya. They reinforce this distinctiveness by slight variations in ritual practice, in the sectarian mark painted on the forehead, and in the wording of the Brahma-sambandha initiation formula.

8. There are two rites. The first, which normally takes place during the candidate's infancy as a prelude to full initiation, is popularly known as Taking the Name or Taking the Necklace ( nam-lena, kanthi-lena^{[ *]} ). Positioning himself or herself crosslegged upon the floor with the guru to the right he or she repeats after the guru three times the eight-syllabled formula translated "Lord Krishna is My Refuge," after which he or she wears a necklace of wooden beads cut from thin stems of the holy basil ( tulasi ). The second rite is normally performed before marriage. The candidate fasts for a period of twenty-four hours at the end of which he or she takes a ritual bath. Then, standing before an image of Krishna and clutching a tulasi leaf in the right hand the candidate repeats after the guru the Brahma-sambandha mantra . Having uttered this dedication the initiate places the tulasi leaf at the foot of the image and assumes the status of an adhikari , or one entitled to follow the Path of Grace. English translations of the initiation formula can be found in Mulji (1865:121), Growse (1883:287), and Barz (1976:85).

9. There are interesting parallels between seva in the devotional context and in political life, particularly with regard to the ideals of humility, selfless service, and anonymity; see Mayer (1981).

10. Giriraja is worshiped in the form of a small stone from Govardhan hill dressed in a yellow smock and adorned with a tiny flower garland. The svarupa stands on a shelf in his room, and devotees approach it by performing an obeisance ( caranasparsa^{[ *]} ).

11. It is significant that the word darsana implies the subject, the "seer" ( drsta^{[ *]} ), rather than the object, that which is seen (Bhatt 1979:18). In a philosophical context it implies the realization of, or an insight into, the nature of reality.

12. Priests are known as bhitariyas because they perform seva in the inner ( bhitari ) rooms of the temple. Traditionally, they belong to one of three Brahman jatis: Audicha, Sanchora , or Girinara .

13. Srngara , meaning "adornment," also denotes the rasa of erotic love which is a principal sentiment of bhakti . The beautiful adornment of the deity is both an expression of the devotee's passionate love for Krishna and a spectacle that is capable of amusing feelings of love in the hearts of those who attend this darsana .

14. Most pada sung in temples are attributed to four disciples of Vallabhacarya, including the great bhakti poet Surdas, and four disciples of Vitthalnatha. With reference to their literary skills they are known as the Eight Seals ( Astachapa^{[ *]} ), but more significantly they are known by their divine identities as manifestations of the eight cowherd companions of Krishna Gopala ( Astasakha^{[ *]} ). Moreover, by virtue of their pure devotion, these poets were able to participate in the secret nighttime lilas as the eight intimate female companions ( Astasakhi^{[ *]} ) of Krishna (Barz 1976:12-13).

15. Arati is the waving of one or more burning cotton wicks ( batti ) dipped in a pot ( divara^{[ *]} ) of thee in a circular motion before the image. In this context atati is performed to remove the harmful effects of the evil eye ( nazar utarna ) to which beautiful babies are particularly susceptible. Yashoda is concerned to dispel any envious feelings that might have been directed toward her beautiful child while he was tending the cows.

16. Devotees often point out that their deities are only smeared with sandalwood paste during the summer and never in winter because in cold weather the cooling properties of sandalwood would cause them considerable distress. They add that such care is seldom shown in other Hindu temples where priests blindly follow ritual procedure by using sandalwood throughout the year without a thought for the deity's comfort.

17. Those ritual acts that involve the touching of the image constitute the most intimate form of worship and hence are reserved for the chief priest and his immediate assistants.

18. The Rahasya Bhavana-Nikunja Bhavana (Prabhu 1968) is an intriguing exposition of the bhavana associated with the articles of worship in which nothing is too trivial for the attribution of aesthetic significance.

19. The temple is divided into inner ( bhitari ) and outer ( bahari ) rooms; the former includes the deity's private apartments and the kitchens which only priests may enter, and the latter includes the various courtyards where devotees assemble. The word aparasa is probably derived from the Sanskrit asprsya^{[ *]} , meaning "not to be touched." Priests enter khasa or "strict" aparasa as distinct from a lesser state of purity known as sevaki aparasa , which enables lay devotees to prepare betel, milk-sweets, and flower garlands for the deity.

20. The three principal temple kitchens are Dudh Ghar (reserved for foods prepared from milk and excluding grains), Ansakhari Ghar or Balabhoga (for preparations derived from grains or vegetables which are cooked by frying in clarified butter, being less resistant to impurity than milk preparations), and Sakhari Ghar (for preparations derived from grains and vegetables, boiled in water, dry-roasted on a griddle, or fried in vegetable oil, being highly susceptible to impurity). Sakhari and ansakhari approximate to kacca and pakka , the terms popularly used in northern India to denote categories of prepared food.

21. Aspects of culinary style reinforce the maternal approach to seva . Many preparations are prepared as if for a young child and hence are known as "baby food" ( balabhoga ). Like all babies Krishna is particularly fond of milk, curd, butter, sweets, and rice pudding. Hot spicy foods are used sparingly. Savory wheatcakes ( puri ) are prepared with copious amounts of ghee so that they are soft and easy to chew. Betel nut is ground to an unusually fine consistency for the same reason.

22. Harper (1964) has argued that the priest, the offerings, and the deity's surroundings must be kept pure in order w prevent the deity sustaining impurity. But in Pushti Marg there is no sense in which it is conceived that the deity can be polluted, as I suspect is the case with Hindu deities in general; see Fuller (1979:469).

23. The sequence has an interesting secular parallel in pati-seva , the selfless devotion of a wife toward her husband, demonstrated by the wifely custom of taking meals after the husband has eaten. This is not a form of "respect pollution," serving to reinforce the inferior hierarchical status of a wife vis-à-vis her husband, as Harper (1964) understood it. Indeed, the food remains she consumes do not necessarily comprise food polluted by the husband's touch or saliva ( jutha^{[ *]} ) because traditionally they remain within the ritually pure cooking area. Rather they are "leftovers" inasmuch as the meal is prepared for her husband so that any remains become a token of his replenishment. More important, the devoted wife cooks solely for the pleasure and well-being of her husband and not selfishly as a means of satisfying her own appetite. Some devotees describe their relationship with Krishna in terms of a wifely model. One interpretation of the bhava of the sectarian mark worn on the forehead is that it is like the bindi worn by women as a sign of their happily married state ( saubhagini ).

24. Unfortunately, I do not have the space here to elucidate this area of sectarian food ritual (see Bennett 1983:227-234). It is interesting that certain foods are prohibited or discouraged in seva "because they are red—like blood." This category includes the seemingly innocuous watermelon. It was pointed out to me that the manner by which the watermelon is carved with a knife to reveal the red fleshy interior strongly suggests animal sacrifice. Thus, by disemboweling the melon it is likely that the devotee would compare his actions with the blood sacrificer. Such thoughts are repulsive and would render the watermelon unsuitable as an offering to Krishna. For the same reason, devotees engaged in the cutting of vegetables prior to cooking avoid the verb katna^{[ *]} , "to cut," owing to its associations with the carving of meat; instead they prefer the verb samvarna^{[ *]} which is free from such unpleasant associations.

25. See Harirayji (1970:93).

26. See Harirayji (1970:181).

27. Whenever the deity is offered bhoga the containers should be completely full, expressing the devotee's overflowing bhava .

28. Devotees of Vallabhacarya Sampradaya locate the "lotuslike mouth" ( mukharavind ) of Krishna at Jatipura; it is a simple cleft in the rock on the lower slopes of Govardhan hill. At Annakuta the village is packed with pilgrims who come to see the mouth soaked in libations of milk and presented with a grand feast.

29. The word pusti is derived from the Sanskrit verbal root pus^{[ *]} . which expresses the action of thriving, increasing, prospering, and of being nourished, well fed, and healthy (Barz 1976:86). Hence pusti is synonymous with divine grace which nourishes the soul. Pushti Marg is the path of spiritual nourishment through grace. Nevertheless, there have been critics of the sect for whom pusti has meant sensual nourishment, the condition of being well fed and prosperous. The reference to sectaries as the "Epicureans of India" cited at the outset of this essay typifies this view. I noted a similar play on the word among nondevotees in Ujjain when they referred derisively to devotee businessmen. For them, pusti implied the hoarding of wealth, or, with reference to shopkeepers of somewhat obese form, a condition of being sated with the sacred food. In fact I have shown that there appears to be a marked correspondence between physical and spiritual nourishment in the sampradaya 's tradition with food serving as the chief mediator. Whereas food provides for the sustenance of the body, grace provides for the sustenance of the soul. Food and grace are subtly commingled in prasada . On tasting prasada the devotee is nourished by the grace of Krishna.

1. For a brief synopsis of the history of this transformation and its causes, see the introduction in my monograph on the devadasis of Puri (Marglin 1985). By studying and performing in the latter context in India and through conducting fieldwork among the women temple dancers of a major Hindu temple, as well as studying the dance from the temple women, I became aware of the great difference between the two types of events.

2. The dance form itself does not differ significantly from the temple to the stage. The greatest change is in the segmentation of a dance sequence which in the temple may last as long as an hour. One major adaptation of the dance form to the stage was the segmentation of dance sequences into much shorter items, ranging from about five minutes to about fifteen minutes, and the creation of a particular sequence of items. A typical stage recital follows a certain sequence of pure dance items not accompanied by sung poetry, nrtta^{[ *]} , and of interpretative dance items accompained by sung poetry, nrtya^{[ *]} . The choice is based on considerations of variety and contrast to elicit maximum interest from an audience that must be entertained.

3. Because I have not witnessed repeated instances of these rituals, their indexical values are not accessible to me.

4. The ceremony for the women differs only in some small detail from that for the men (Marglin 1985:67-72).

5. This injunction against having children is not found among the devadasis of Tamilnadu studied by Amrit Srinivasan (1985).

6. The Sakta sectarian tradition is a form of tantrism whose followers worship the goddes Sakti .

7. The information that the devadasis face north, which I had when writing my book, was based on asking two devadasis to describe their orientation while they danced. Because this questioning took place in their own houses, as well as in my own house, they were disoriented, and apparently I deduced the wrong direction from their verbal explanations. Thereafter, Puma Chandra Mishra and I made diagrams of the temple including positionings of the dancer north, west, and south and showed them to the rajagurus and the devadasis . The unambiguous result from this more precise method of enquiry was that the devadasi faces south.

8. Raja from the Sanskrit root raj- , to be red, to be colored, comes to mean menstrual blood and derivatively dirt or dust since menstrual blood is—like other eliminations from the body—both polluted and dirty. This however does not prevent it from being a powerful sexual fluid, the source of life because out of menstrual blood the mother forms and then feeds her fetus. Raja in Oriya is also used to mean the colorless female sexual fluid a woman is believed to secrete during sexual intercourse. The devadasis did not perform during their menses when they were impure and did not enter the temple. I am grateful to Owen Lynch for pointing out to me the double meaning of the word raja .

9. Ron Hess in collaboration with Indian film makers has made a film largely based on my study of the rituals of the devadasis . Entitled "Given to Dance: India's Odissi Tradition," it is available through the Madison, Wisconsin, South Asia Program.

10. The devadasis , but not the stage performers, recite the drum syllables. On the stage the syllables are recited by the drum master, not the dancer herself (at least in an Odissi dance performance). Music elaborates this rhythmic pattern by giving the long beat variable lengths (measured in time units) and by creating a beat half the value of the short syllable. This elaboration and the various possible combinations of three beats (of long, short and half-short ones) have enabled the creation of more than forty different rhythmic patterns ( tala ).

11. For the historical antiquity of Indian classical dance as well as its treatment in various texts, I refer the reader to Kapila Vatsyayana's (1968) definitive study on Indian classical dance.

12. It is not necessary to go into a detailed description and recounting of these, and I refer the reader to two illustrated treatises on Odissi dance. Because I am concerned with the regional style in this paper, the reader must bear in mind that there are several other regional styles of Indian classical dance. One treatise in English is in a special issue of the magazine Marg (1960); the other is in Oriya by Dhirendranath Patnaik (1958).

13. This use of the dancer, as person, contrasts rather markedly with the use of the body in ballet or even in modern dance in which limbs and extremities (as they are called in the Western dance discourse) are primarily subordinated to line and are most often used to conjure an aerial fluidity and lightness. The limbs and extremities such as hands, feet, head, and face are primarly used to continue or break a dynamic line originating more at the center of the body. Their use is, except for intentionally dramatic moments, not based on the raw material of spontaneous communicative behavior. Thus, a Western audience is not as directly engaged in the role of a dislogical partner; instead, the dancers place the audience in the more passive position of appreciating the visual and dynamic forms created by the bodies of the dancers.

14. Besides the relevant passages in the Natyasastra , I have also used the relevant passages in the great tenth-century Kashmiri philosopher-saint Abhinavagupta's commentaries on Bharata's verses on rasa translated by Masson and Patwardhan (1970), Masson and Patwardhan's own interpretations of Abhinavagupta, as well as S. K. De's (1925) commentary on Abhinavagupta's understanding of rasa .

15. My own discussion of rasa in this essay should in no way be taken as an attempt to add to this extremely learned and scholarly literature. I am in no position to do this. Here I attempt only to use these exegeses for a semiotic interpretation of the inner transformations in the spectators which are signaled by outward visible acts and are matter-of-factly said to have taken place.

16. The following verse of a palm leaf manuscript written by a devadasi in the nineteenth century (see Marglin 1985:90) addresses this issue:

If a man desires the body of a dasi at the time of her seba [ritual service, i.e., the dance], This man, by order of the king should be heavily fined, Such a man would be a criminal in front of the great Lord.

17. The following passage from Hegel is a good example of what I refer to:

These earliest and still most uncontrolled attempts of imagination and art we meet most signally among the ancient races of India... These people... through their confused intermingling of the Finite and the Absolute,... fall into a levity of fantastic mirage which is quite as remarkable, a flightiness which dances from the most spiritual and profoundest matters to the meanest trifle of present experience, in order that it may interchange and confuse immediately the one extreme with the other. (G. W. F. Hegel 1835; quoted in Mitter [1977:213])

18. In fact, this is an image of Shiva, half-man and half-woman ( ardhanarisvara ). For a discussion of the significance of this Shaivite iconography in the midst of a Vaishnavite ritual, see Marglin (1985:chap. 7).

19. The Vaishnavite bhakti tradition has within it many sects and subtraditions, each with its own customs, favorite texts, and specific interpretations. As far as I was able to discern, the evening ritual represents a specifically local Oriyan variant of this religious movement with its own peculiar interpretations and theology. It is closely related to, but not identical with, the neighboring Gaudiya Vaishnavite tradition of Bengal. I think that the Oriya tradition has many more Sakta influences than other Vaishnavite traditions. On Oriya Vaishnavism, see Mukherjee (1940).

20. For translations of the Gita Govinda I refer the reader to Lee Siegel's (1977) literal translation accompanied by an exhaustive scholarly commentary on the text, its author, and the cultural, historical, and religious background from which they emerged. Barbara Stoler Miller's (1977) more poetic translation has a much shorter but also very incisive introduction to the text.

21. For a detailed discussion of this point, see Marglin (1985:200-203).

22. These names do not correspond to current musical classification. Because music and dance are transmitted orally, there is no way of knowing to what notes and to what beats these names corresponded.

23. For a fairly recent work discussing the special place of Radha in Vaishnavite ritual and theology, see John Hawley and Donna Wulff (1982).

1. See, for example, the New York Times , November l, 1984, p. 13.

2. Puri is from the Khatri (mercantile) rather than Jat (agriculturalist) section of the Sikh Panth (community). Khatris claim elevated status within the Panth by virtue of the fact that Nanak, the first Sikh Guru (preceptor), and his successors were all Khatris. However, although Khatris remain an influential caste in the urban areas of Punjab, Jars have come to predominate within the Panth. This is true among the Sikh diaspora as well as in the Punjab. The concern with izzat (honor), the topic of this essay, is especially a Jat Sikh concern. The fact that Purl is a Khatri, not a Jat, may have something to do with his response to the assassination, but it cannot entirely account for the moral sensitivities of his gora followers.

3. The Ghadar Party was a revolutionary organization, centered in North America, which sought the overthrow of British rule in India. Although the leadership of the party came from the Hindu, Muslim, and Sikh immigrant intelligentsia, support of the immigrant Punjabi Sikh masses was crucial to its ability to raise funds and volunteers. Hopkinson had been used elsewhere along the Pacific Coast by Canadian and American agencies interested in gathering information on potentially revolutionary activities among South Asian immigrants. For a detailed analysis of the Ghadar Party, including discussion of Hopkinson's activities, see Purl (1983). Although he does not discuss the concept directly, Puri's analysis suggests that appeals to izzat were effective in generating support from the Sikh immigrants of peasant stock, as "heaped symbols of shame and oppression were used to generate a certain auto-intoxication of disgrace" (Purl 1983:119).

4. Gurduara (or gurudvara ) comes from guru (preceptor) + dvara (door) = "the residence of the Guru." I employ the more commonly encountered anglicized spelling, gurdwara, throughout this essay.

5. I speak of a "Sikh community" in Vancouver to the extent that most self-described Sikhs in Vancouver continue to seek to influence one another's lives in the name of common membership in the Sikh Panth. At its "Desh Bhagat Temple" the local East Indian Defence Committee—an offshoot of the Indian and Canadian Communist parties (Marxist-Leninist)—celebrates Mewa Singh as one of a line of local Sikhs (including Ghadar revolutionaries, Indian National Army soldiers, and recent Naxalite terrorists) martyred through their involvement in revolutionary anti-imperialist and class struggle. The Khalsa Diwan Society and the Akali Singh Society, at their respective gurdwaras, represent Mews Singh as a local martyr who, like earlier Sikh martyrs in India, gave his life for the perpetuation of his people and his religion in the face of hostility from the dominant society. The appropriation of Mews Singh's martyrdom by such different groups suggests that he is a significant collective symbol of a ''Vancouver Sikh community."

6. I employ the conventional transliteration, izzat , which reflects the pronunciation of the term in the Majhi dialect (of the Amritsar-Lahore area) on which literary Punjabi and most Punjabi dictionaries are based. The pronunciation in the dialects of Doaba and Malwa (the areas of the Punjab south of the Sutlej River from which most of my informants originate) is more accurately rendered as ijjat .

7. A near-synonym, commonly used in Doabi Jat Sikh conversation, is man^{[ *]} (respect, prestige, pride, veneration, arrogance).

8. Malcolm Darling, the British Indian civil servant whose classic books on Punjabi village life ( The Punjab Peasant in Prosperity and Debt, Rusticus Loquitur , and Wisdom muff Waste in the Punjab Village ) constitute a regrettably ethnocentric but nonetheless valuable ethnographic record, recognized both the importance and difficulty of the concept. He notes of izzat that it is "a word for which there is no precise English equivalent, denoting objectively, social position, and subjectively, amour-propre " (1934:42, n.3). See also the discussion of the term in Pettigrew (1975:58-59).

9. See, for example, Charles Lindholm's (1982) illuminating discussion of personal honor as central to the social organization and emotional life of the Swat Pukhtun of Northern Pakistan.

10. W. H. McLeod (1976:37-58) discusses in some detail the way in which what people have come to regard as general Sikh characteristics (including the outward Khalsa form and the martial reputation) are largely Jar customs and attributes institutionalized within the Panth. McLeod argues convincingly that the living human Sikh Gurus, all of whom were Khatris, gradually came to adopt Jar practices as Jats came to predominate among their followers. Nevertheless, even today the Panth continues to include those Nanakpanthis who emphasize the more quietistic teachings of the first Guru, as well as those "orthodox" Sikhs who emphasize the militarism of the Khalsa (whose creation is credited to Guru Gobind Singh, the last human Sikh Guru). If one can speak of emulating the "dominant caste" within the Sikh Panth, one would have to recognize the remnants of contending Khatri and Jar models. Concern with izzat is not normally the intense concern for Khatris that it is for Jats, and this may partially account for the Jar stereotype that Khatris are, by nature, cowardly and spineless. It appears that Jars and Khatris are socialized with different transactional strategies (see Marriott 1976; esp. 132-133) and, consequently, different moral sensitivities.

11. In his recent, posthumously published book, Women's Seclusion and Men's Honor (1988), David Mandelbaum identifies and analyzes what he terms a "purdah-izzat complex" extending—with some regional, caste, class, and religious variation—from Pakistan, throughout most of northern India, to Bangladesh.

12. Mrs. Gandhi's assassins will, by having acted to defend the honor of the Panth, also presumably ensure the honor of their families and descendants. In fact, Beant Singh, the Sikh bodyguard and a presumed assassin, slain during the incident, is already being honored as a martyr in his village; and his shrine will likely continue to be focal point of veneration by members of his lineage.

13. Sending individuals away from the village is, of course, something quite different than migrating together as a family unit. The preference for the former type of migration, which does not involve relinquishing landholdings in the ancestral village, may distinguish landowning Jar Sikhs from other Punjabi migrants.

14. I do not believe that izzat tells the whole story about Sikh endogamy (see Dusenbery 1988, in press), but it plays a significant role. Vaughn Robinson (1980) notes how concern for izzat works against the marriage opportunities of highly educated, British-raised Sikh girls whose sexual purity and ability to submit to the demands of life in her husband's family may be questioned by exposure to Western society. My own research suggests that Jar Sikhs in Vancouver have sought to avoid certain jobs (e.g., cleaning, service, sales) that, although they may bear no stigma to non-Punjabi Canadians, are deemed beneath the dignity of a Jar Sikh and detrimental to family reputation.

15. Most of my informants estimated the percentage of Jats among the Sikh old-timers in British Columbia at 90 percent or more. The lowest estimate known to me is Adrian C. Mayer's, who writes that "the vast majority of Vancouver Sikhs are from the same caste—the Jat. Perhaps not more than one-fifth represent other castes" (1959: 13).

16. For accounts that attempt to assay the effects of the recent immigration on particular North American Sikh communities, see Buchignani, Indra, and Srivast[a]va (1985), Chadney (1985), Dusenbery (1981), La Brack (1983, 1988). For a general overview and specific case studies of Sikh migration and the experience beyond Punjab, see Barrier and Dusenbery (1989).

17. I have discussed this process in greater detail in, especially, Dusenbery (1986, 1988, in press); see also Dusenbery (1981).

18. The amrt^{[ *]} pahul or amrt^{[ *]} samskara^{[ *]} ceremony effects the incorporation of one into the Khalsa Panth, the so-called "Brotherhood of the Pure" (see Dusenbery in press). By tradition, any five worthy Khalsa Sikhs can be constituted as the panj pyare (literally, live beloved) to administer this ceremony. Puri's institutionalization of Sikh "ministers" is but one of his innovations in building an ecclesiastical hierarchy unprecedented among Sikhs.

19. I analyze the commitment mechanisms used in the early 3HO recruitment process in Dusenbery (1973). I focus on the yogic to Sikh transformation of the organization in Dusenbery (1975).

20. For an account of early Cora Sikh involvement with Punjabi Sikh gurdwaras in Los Angeles, see Fleuret (1974). For accounts of early Cora Sikh involvement with Punjabi Sikh gurdwaras in northern California, see La Brack (1974, 1979) and Bharati (1980).

21. Somewhat conflicting accounts of the factional dispute can be found in accounts commissioned by the British Columbia Police Commission (D. Singh 1975:39-44), reported in the Sikh press by the head of the local 3HO ashram (G. R. Singh 1975), produced by an outside UNESCO researcher (Scanlon 1977), and presented in a University of British Columbia master's thesis (Campbell 1977:74-102); see also Dusenbery (1981). The main factional disputants were a "businessmen" faction, so-called because it drew its strength from the successful, established immigrants who had seen to the building of the present gurdwara in the late 1960s, and a more "orthodox" faction, representing recent immigrants and numbering among its leaders some non-Jat professionals with whom the Gora Sikhs believed themselves to have some affinity. The source of most of the violence was a "communist'' cadre of Naxalites affiliated with the East Indian Defence Committee and the CCP-ML, a group that each of the two major factions felt was working in league with its opponents.

22. For Punjabi Sikhs, pulling off a man's turban is a serious challenge to his izzat . Similarly, roughing up someone in public is as much an assault on his izzat as it is on his physical person. For comparative purposes, see the detailed discussion of "the dialectic of challenge and riposte" in "the competition of honour" among the Kabyle of North Africa in Bourdieu (1966).

23. Aside from occasionally attending gurdwara functions, the only semiofficial relationship that any Cora Sikh maintained with either of the two main Vancouver gurdwaras was through a "Sunday school" class taught by the wife of the ashram head at the Akali Singh gurdwara. After I left Vancouver, members of the local ashram ran a Montessori School on Khalsa Diwan Society property, but even that relationship soon dissolved with recriminations on both sides.

24. Contributions toward construction, improvement, or maintenance of a gurdwara are meritorious gifts to the Guru. Because the amount of a donation is public knowledge (amounts of gifts are read to the congregation and subsequently published), conspicuous philanthropy is both religiously meritorious and secularly good for one's reputation. As a consequence, gurdwaras tend to be quite lavishly endowed institutions in wealthy settings. During my fieldwork, several million dollars worth of property purchases, building projects, or expansions and remodelings were scheduled at some seven present or planned gurdwaras in the greater Vancouver area.

25. I have employed the Sanskritized form, dharma , rather than the Punjabi, dharam , because that is the form used by the Gora Sikhs. In everyday conversation, Gora Sikhs use ''Sikh dharma ," "Sikh religion," and "Sikh way of life" interchangeably.

26. I would argue that, whereas they are insensitive to izzat as a moral affect, the Gora Sikhs have a highly legalistic understanding of "the dharma " as a moral imperative underlying their ideology and actions. It seems telling not only that the Gora Sikhs call their religious body, Sikh Dharma, but also that their claims to orthodoxy rest on strict adherence to the "moral duty" they identify with the Sikh rahit maryada (literally, prestigious code for conduct). A generally accepted version of the rahit was formally issued by the Shiromani Gurdwara Parbandhak Commitee in 1950. The fact that most Punjabi Sikhs are ignorant of and indifferent to the exact contents of this document—and, especially in the case of Jat Sikhs, would not feel their identity as Sikhs dependent on strict adherence to its rules of conduct—merely confirms the Cora Sikhs in their conviction that they, rather than the body of Punjabi Sikhs, are the true upholders of "the dharma ." For a further discussion of the Gora Sikhs' understanding of religion, see Dusenbery (1981, 1986, in press). For the text of and commentary on the Sikh rahit maryada , see McLeod (1984: 79-86).

27. "Blood for Blood" ran the lead headline in Vancouver's Punjabi-language Indo-Canadian Times (vol. 7, no. 23) for the week ending June 15, 1984. Elsewhere, the intense emotions Punjabi Sikhs experienced at the time are partially reflected in the following accounts: "We have suffered and suffered terribly in every respect during the last 2 years. Our prestige has gone down. Our honour has been compromised and the very source of our spiritual sustence [ sic ] has been cruelly hit" (W. Singh 1985:42). "Some of us outside Panjab had visualized the possibility of alienation which the entry of troops into gurdwaras would create among the Sikhs. But we had not reckoned with the intensity of the humiliation they have felt. Every Sikh we met was distressed by the Army action. They think that there has been an assault on their identity" (Chowdhury and Anklesaria 1984:143). "Virtually to a man, the 14 million strong [Sikh] community felt as if it had been slapped in the face. . .. The feeling of hurt and humiliation among Sikhs runs so deep that they seem to feel that they are a persecuted minority'' (K. Singh and K. Nayar 1984:37, 41). "The Sikhs feel totally alienated and isolated. Their pride and self-respect have been badly hit and the festering wound inflicted may take decades to heal, though the memory of the tragic happening will remain treasured in the Sikh psyche" (B. Singh 1984:4a).

28. See, for example, the following articles: Sardarni Premka Kaur Khalsa (1984), Shakti Parwha Kaur Khalsa (1984), Harbhajan Singh Khalsa (1984), Sikh Dharma Secretariat (1984a, 1984b). Takhat is an alternative spelling of Takht .

29. The articles in Peristiany (1966) on honor and shame in Mediterranean society, although not systematically focused on the contrast of the person in Mediterranean and modern Euroamerican societies, nevertheless contain valuable insights into the very different social selves presupposed. Peristiany and other contributors to the volume point out that honor in Mediterranean societies is the concern of equals or near equals interacting in public settings. In Punjabi society, izzat as a moral affect is similarly relevant to public relations between would-be elite persons, families, or castes. Within the private sphere of the family and in public interactions with inferiors or superiors, other moral appraisals apply.

30. Punjabi Sikhs and Gora Sikhs construe dharma in very different ways. For Cora Sikhs, Sikh dharma , as a "religious" code for conduct, constitutes a privileged source of ultimate morality applicable across all contexts. For Punjabi Sikhs, Sikh dharma , however privileged it may be, is but one of a number of moral codes for conduct impinging upon the person. See Dusenbery (1988, in press) for a more detailed discussion of the different concepts of the person presupposed by Cora Sikhs and Punjabi Sikhs.

31. Even fear of dangerous things in children may be more culturally learned than innate. Catherine Lutz reports that "academic psychologists have begun to accumulate evidence...suggesting that the number of danger response elicitors present from birth is much smaller than was once thought" (1983:257). If this were considered in a cross-cultural context, it might be nil.

1. The Bhagavata Mahatmya is found in the Uttara Khand of the Padma Purana^{[ *]} and is also included in some Indian editions of the Bhagavata Purana^{[ *]} . A glorification of the Bhagavata Purana^{[ *]} , it is generally regarded as being later in date than the Padma Purana^{[ *]} itself, representing an addendum by Vaishnava commentators. It is probably a fifteenth-century work (Gelberg 1983:212).

2. From a conversation with Sripad Baba in Brindaban, February 22, 1982 (direct quotation in English). Baba ( baba ) and sadhu ( sadhu ) are terms used in Brindaban to refer to mendicant ascetic holy men.

3. Dimock (1966: 165) comments on the word dhama :

In Vaisnava^{[ *]} thought, Krsna's^{[ *]} surroundings, both objects and persons, are considered extensions of Krsna's^{[ *]} self, and are collectively called Krsna's^{[ *]} dhaman . The dhaman , then, makes a simultaneous appearance on the earth with Krsna^{[ *]} ; like him, it participates simultaneously in both phenomenal and non-phenomenal worlds. The earthly Vrndavana^{[ *]} , the Vrndavana^{[ *]} described in the Bhagavata , is identical with the heavenly Vrndavana^{[ *]} .

4. This interpretation of Brindaban is held especially by the Bengal Vaishnavas, followers of the medieval saint Chaitanya, although it is also shared by many Brindaban residents and pilgrims regardless of sect. The word Brai ( Vraja ) refers to the cultural and linguistic area generally associated with Krishna that includes Mathura, the place of Krishna's earthly ''birth," and other sites of his youth such as Gokul and Govardhan. Brindaban, however, is often cited as the center of Braj because the most esoteric of Krishna's sports, the circle dance with the cowherd girls, rasa-lila , occurred there. In popular speech and religious thought, Brindaban and Braj are practically synonymous, Brindaban representing the "essence" of Braj, that is, the place where the circle dance, an event that is eternally transpiring in the spiritual Brindaban, took place. See Hardy (1983) for an analysis of the development of Krishna- bhakti , including the evolution of Brindaban's theological importance.

5. The word videsam does not necessarily indicate another country but can refer to other places within India. According to Rama Nath Sharma, Sanskritist and Panini scholar at the University of Hawaii, however, the interpretation given here by the ascetic is a common Indian interpretation and logical for the context in which it occurs.

6. Vaishnavas ( Vaisnava^{[ *]} ) are worshipers of Vishnu ( Visnu^{[ *]} ). Krishna is considered an incarnation of Vishnu by the general Hindu population; therefore, his devotees are also Vaishnavas. For the Bengal Vaishnavas, Krishna is the supreme God, and Vishnu is an expansion of Krishna. Sampradaya means sect or tradition.

7. Bhaktivedanta was initiated by Bhaktisiddhanta Sarasvati, founder of the Bengal Vaishnava monastic institution, Gaudiy a math ( Gauriya^{[ *]} matha ), in November 1932 (Satsvarupa Goswami 1980:297).

8. "New religious movements" is a sociological term used to describe religious phenomena that do not easily fit into traditional categories such as sect, cult, or church. Wilson (1982:17-20) has suggested that although the religions they incorporate may not actually be new, they emerge in a social context where other religious traditions already exist. He suggests that they generally fulfill the following requirements: They offer "a surer, shorter, swifter, or clearer way to salvation"; conduct an "implicit assault on spiritual elifism"; are available to a wider public than traditional paths of salvation; offer techniques that are accessible to the average man; facilitate spiritual mobility; and are therapeutic and generally life enhancing for the ordinary individual.

9. As early as 1957, Bhaktivedanta Swami had formulated the idea that if Americans and other Westerners accepted the ideas of "Vedic" culture, there would be a subsequent revitalization in India because Indians were highly motivated to emulate the West. He reasoned that if he could travel to the West and then return with devoted converts, this would initiate a revival of spiritual culture in India (see Satsvarupa Goswami 1980:212, 1983:121).

10. In most Brindaban temples the main deities are of Krishna and his divine consort Radha, but the central images in the ISKCON temple are Krishna and his brother Balaram. According to Satsvarupa Goswami, Bhaktivedanta chose these images for four reason: (1) because in most Brindaban temples Radha and Krishna were the primary images, ISKCON's temple would be unique; (2) the ISKCON land was located in an area called Raman-reti, which Bhakivedanta said was where Krishna and Balaram had played together as children; (3) the Chaitanyaites consider Balaram to be the first expansion of Krishna, Sankarsana^{[ *]} , who upholds all the universes, and Vaishnavas should therefore worship Balaram for spiritual strength; (4) Chaitanya is considered to be Krishna Himself, and Chaitanya's associate Nityananda is seen as an incarnation of Balaram, a fact that the "Krsna-Balarama^{[ *]} temple would proclaim to the world" (Satsvarupa Goswami 1983:178-179). The Krishna-Balaram deities, therefore, are worshiped on the central altar. To their right are the images of Chaitanya and Nityananda, and on their left stand the more familiar forms of Radha and Krishna.

11. During my fieldwork in Brindaban, there were 101 ISKCON devotees living in Brindaban: 39 adult members lived at the temple house, 45 students were enrolled in the ISKCON-run school ( gurukula ), and 18 devotees lived at other places in town.

12. The term "sacred complex" as originally used by Vidyarthi (1978), refers to an interrelated set of objects that together compose a sacred space, in India usually a place of pilgrimage. The sacred complex includes sacred objects and their physical locations, the performances that take place at these locations, and the people who participate in these performances. My survey data from Brindaban indicate that 98 percent of all one-day pilgrims visit the Krishna-Balaram temple. For pilgrims that stay two days or more, 94 percent visit the ISKCON complex.

13. Direct informant quotations in English.

14. See De (1961, 1963); Kane (1971); Gerow (1977); Kapoor (1977); Shrivatsa Goswami (1982) for a more complete discussion of mystical emotions in Bengal Vaishnava theology.

15. More than being an incarnation of Krishna alone, Chaitanya is accepted as the dual incarnation of both Krishna and Radha, his consort and energy of supreme pleasure ( hladini-sakti ). This perception of Chaitanya is usually traced by scholars to the Caitanya-caritamrta^{[ *]} of Krishnadas Kaviraj (De 1961:52) who wrote slightly later than the original Goswamis. Though this may be the first textual reference to Chaitanya's duality, Dimock explains that the idea existed even during Chaitanya's lifetime, and that Krishnadas in all likelihood received it from his teachers, the Goswamis (Dimock 1966:32, 150).

16. .Chaitanya is credited with writing only eight verses, called Siksastaka^{[ *]} , and these can be found in Krishnadas's Caitanya-caritamrta^{[ *]} and in Rupa Goswami's Padavali (De 1961:113).

17. The "Six Goswamis" of Brindaban were intimate disciples of Chaitanya Mahaprabhu who were sent there by him to establish the new sect's headquarters, develop the location as a pilgrimage center, and codify the religion's theology, philosophy, and ritual. By name they are: Rupa, Sanatan, Jiva, Raghunath Das, Raghunath Bhatt, and Gopal Bhatt. Goswami ( gosvami ) is interpreted to mean one who is master of his senses. The term also is a title taken by the priestly lineages of Brindaban who trace their ancestry back to the sixteenth-century founders of the town and whose duty it is to tend the temples there.

18. Chaitanya taught that this chant was especially beneficial for the present epoch ( Kali-yuga ), and chanting this mantra forms the core practice for devotees of ISKCON. Its thirty-two syllables are: Hare Krishna, Hare Krishna, Krishna Krishna, Hare Hare; Hare Rama, Hare Rama, Rama Rama, Hare Hare.

19. He wore a waistcloth, rather than a dhoti as ISKCON members do. The dhoti is a long cloth wrapped around the waist and drawn between the legs, forming a loose pantlike garment.

20. Indian scriptures are called sastra . The specific texts included in the sastra category vary from sect to sect, but Dimock points out that for the Bengal Vaishnavas "the term shastra means the Bhagavata-purana^{[ *]} " (Dimock 1966:183). Vaidhi bhakti is ritual activity based on scriptural injunction.

21. According to Anthony Wallace (1956:265), a revitalization movement is any "deliberate, organized, conscious effort by members of a society to construct a more satisfying culture." Taking his cue from general stress theory, Wallace suggests the period of revitalization, in which the movement emerges, is preceded by phases of individual stress and cultural distortion; these upset the system's homeostasis, or steady-state. The paradox is that if a revitalization movement is to be successful, it must eventually reach a steady-state condition itself, thereby setting up the conditions where the cycle may be repeated.

22. Letter from Shubananda Das, September 4, 1985. Parentheses appear in the text.

23. Letter to all ISKCON temples from Hridayananda Dasa Gosvami and Ramesvara Dasa Swami, both GBC representatives, dated June 17, 1976.

24. See Hein (1972) and Hawley (1981) for descriptions of the rasa lila ( rasa-lila ) dramatic form.

25. Quotations of sadhu translated from a combination of Hindi and English.

26. Quotations of merchant translated from a combination of Hindi and English.

27. The Radha-Vallabha sect was founded by Hit Harivamsha (Hits Harivamsa^{[ *]} ) who based its theology on the primary importance of Radha. Although in ritual and spiritual practice Krishna becomes secondary, the Radha-Vallabha temple's main image is Krishna. There is no anthropomorphic image of Radha, but rather a silver tablet with her name inscribed upon it is placed beside the Krishna image.