|Durham Anthropology Journal
Volume 12(1) 2004: 37-50. Copyright © 2004, Stephen M. Lyon
Download a PDF version of this article (640 KB)
This paper examines the social impact of the rice distribution ritual called a dég, in Punjab, Pakistan. This ritual exposes modes of dominance expressed or asserted by `symbolically' violent means. I focus on one dég which was significant by its excess, providing a clear example of the roles of the `giver', the `receiver' and the rivalries between collateral groups. Bourdieu's notions of symbolic violence provide one analytical social model of a mode of domination. This paper is not a refutation of Bourdieu's analysis so much as a conceptual and geographic extension. The violence in this case study deviates from Bourdieu's pattern in that the direction of symbolic violence, or assertion of control through non physically violent means, is not directed against the receiver. I suggest that the analytical model must be expanded to include indirect symbolic violence in which the receiver is incidental to the intended direction of domination.
dég (2): n. a public feeding. An individual (usually) arranges a large quantity of rice to be prepared and distributed to anyone who comes. This may be done to give thanks to Allah for a particular incident (like the birth of a child) or to honour an important visitor.
South Asian gift giving has justifiably received considerable attention by anthropologists. Gift exchange has of course been a mainstay of anthropology since well before Mauss produced his seminal thesis on the subject [notably Malinowski 1922 on the kula exchange in the Trobriand Islands]. Mauss'  contribution to the subject was to generalise from cultural specificity to cultural universal and lay the groundwork for understanding how social groups are possible at all. Social debt, Mauss argued, is incurred through the exchange of the gift between moral persons. Parry's  thoughtful reconsideration of Mauss' essay on the gift, both expands and diminishes the general relevance of this thesis. Parry suggests that Mauss was correct in his analysis of the gift as the `general principle... [of] an absence of any absolute disjunction between persons and things' [1986: 457]. Things as commodities, on the other hand, are defined in their separation from giver and receiver (persons). Instead, commodities are represented their autonomous exchange value. Gifts are inalienably associated with both giver and receiver. The presence of commodity markets do not necessarily lead to the wholesale abandonment of gift economies, however [cf. Gregory 1980; Parry 1986; Werbner 1990]; commodities may be transformed into gifts by imbuing them with a `transcendent quality which relates to the cultural value ascribed to them' [Werbner 1990: 282].
Gift exchange, commoditisation and power in South Asia pose particular problems for anthropologists. Hierarchical gift giving appears to constitute a critical mechanism in the expression and performance of relationships between individuals and groups. Transformation of commodities into gifts appears common in North India, Pakistan and among British South Asians. This occurs despite the presence of capitalist market infrastructure with its corresponding economic rationality. Market rationality, rather than supplanting the gift in South Asia, has provided a new range of legitimate gift items (cash, electronic appliances, cars etc.). These items, to be sure, may lend themselves to gift inflation but there is no reason to suppose that they lead to lessening importance of the gift. Werbner  argued that in order to understand `hierarchical gift giving' the focus must be placed on exchange between groups. She identifies two gift exchange arenas which reinforce notions of gift exchange prevalent in anthropology: marriage and sacrifice. In both arenas exchange is subject to inflationary and hierarchical pressures. Eglar  described the competitive nature of gift exchanges linked, in particular, to marriage and birth ceremonies. Incremental increases in wedding gifts are important expressions both of social solidarity between groups and Maussian gift competition. Werbner's [1990; 1998] analyses of sacrifice in the context of sufi saints ( pir ), describes similar patterns in the manner in which gifts are defined and transformed into something else in order to escape the hierarchies associated with marriage gifts (and indeed gifts in other social contexts in South Asia).
Orenstein , comparing traditional Western feudal society to traditional Melanesian big man society, argues that institutionalised asymmetrical reciprocity can take two forms. The giver may endebt the receivers through excessive giving, or conversely, the giver may place themselves further in a position of endebtment in relation to the receivers. The form of endebtment in which the giver becomes indebted, is characteristic of egalitarian societies. Thus, societies with models of asymmetrical reciprocity similar to potlatch societies (of the North West Coast of North America) would be expected to exhibit weaker leadership lacking in strong and institutionalised authority. This is clearly not the case in Punjab, Pakistan. Nor would one expect this to be the case, since Punjabi landlords do not rely on giving as their primary strategy for gaining prestige and authority. Yet there remain interesting parallels with potlatch and big man
cultures and the culture of patronage evident in northern Pakistan [see Goodell 1985 for a more general discussion of the comparison of patronage and potlatch]. What must be explained, therefore, is why a hierarchical society, such as that of the landlord dominated areas of Pakistan, practices asymmetrical giving reminiscent of more egalitarian societies in Melanesia and North America. This paper examines this question in the context of Mauss' theory of the gift and Bourdieu's theory of violence and dominance expressed and asserted through the excessive gift. The various ways of expressing or asserting dominance between equals is important for two reasons. First, they underscore the importance of hierarchy in Pakistani culture. This goes some way to demonstrating arguments I make elsewhere that Pakistanis actively try to impose asymmetrical relations in all social contexts, even where ideally there are no apparent hierarchical differences between the individuals involved [Lyon 2002; forthcoming]. Second, they illustrate some of the ways that patron/client roles may be initially established. Gift giving is a way of asserting asymmetry between giver and receiver which can then be transformed into roles of patronage [cf. Eglar 1960; Werbner 1998]. The rice giving ritual called the dég, provides another mechanism for either challenging an existing asymmetrical relationship or introducing asymmetry in a relation of equality which can not be disrupted through normal gift giving.
The first dég I attended was a complete surprise. After eating breakfast of fried parat'ha and honey (collected by Afghan migrants), eaten with some of my hosts and their servants, I moved to one of the déra for some `structured' interviews and audio recordings. On my way to the tea shop however, I was called over to the tallest of the three hills on the edge of the village. A large number of people had gathered at the base of the hill and were sitting in the shade eating. As I approached I saw that three or four men sat huddled together around a single large plate filled with rice. They all ate the rice directly with their hands and chatted and laughed while doing so. I sat down next to them and asked why they were eating rice on the side of a hill. They explained to me that this was a dég.
Dég rituals are organised in fairly uniform ways in the area. Some household decides to organize the ritual, usually on rather short notice. It is a household enterprise, though the head of household will usually be given most of the credit. Specialist cooks are called, either from within the village or from nearby towns. In either case, the cooks are often known for this activity and one way of measuring the prestige of a dég is by seeing who is doing the cooking. For small dégs (4 cauldrons or less), the cooks are usually villagers. Larger dégs may require more help. In all cases I viewed men doing the cooking. Cooking is done in public view so women do not participate in either cooking or preparation. The dég begins with a prayer, usually by a group of men including a prayer leader, who may or may not be a maulvi but is probably considered to be well versed in the Qur'an. General thanks to Allah are given and recitation of passages of the Qur'an are performed. People begin gathering long before the rice is ready. Men and children linger in the area and chat and play while the food is being prepared. Depending on their relationship to the host of the ritual, they may help in some way in the preparation. When the rice is ready it is distributed in plates to those nearby. People share plates. The custom in the Punjab is that rice and bread are not eaten together, so the rice is eaten with the fingers. Very few ladies eat at the site of the dég. When they come then they bring a sack or some other plate and have it filled to be taken away. Very often women send their children to the site with sacks or plates and the children receive a portion. The danger of this is that the person distributing the rice may not give a child as much as he would give an adult woman. Children are occasionally told not to be greedy and sent away with only half a sack of rice. The very poor children use the front of their kamiz (long shirt) as a container. They hold out their shirt and the rice is spooned in.
People sit in clumps and tend to stay together. They chat and joke while they eat and in every sense the occasion is treated as a festive break from the normal routine. Rice is expensive for poor people since it is not grown in northern Punjab. It must come from regions further south. As a result, for many poor villagers, these ritual dégs are the most frequent occasions for them to eat rice. Rice is also associated with weddings which further marks the festive aspect of the ritual. Anyone passing by is expected to participate, including complete strangers, like truck drivers or passing goat or sheep herders.
Status hierarchy is marked in various ways in a dég. High status guests are never asked to sit directly on the ground. If the ritual is performed in someone's courtyard or guest house, high status guests are given a charpai (South Asian cot or sofa) or chair to sit on. If the dég is performed in a field then a chaddar (shawl) will be laid on the ground for the important guest to sit on. High status people eat first, except for the host, who eats very little and usually after the first wave of guests have started eating. The host may or may not sit close to his guests but in every case I witnessed, the host is clearly identifiable. He sits on a charpai and is conspicuous by the absence of food or the indifference he shows to the food. Men around the host laughing and joking deferentially with him. He will, from time to time, order the cooks to do something
different. The order itself often seems irrelevant but the act of giving orders and expressing superiority seems to be of extreme importance. High status guests are expected to greet the host personally and sit with him while they eat.
Dég s end less formally than they begin. As the rice runs out people drift away. By the time the rice is finished, the only people left are the host, his entourage and the cooks who may or may not be responsible for cleaning up.
The reasons a household may organize a dég are varied: to honour Haj pilgrims both prior to departure and upon their return, for the birth of a child, to mark the arrival of a new bride to a household (usually after the official wedding ceremonies of nikkah, mehndi and walima), to thank Allah for saving the life of an ill relative or perhaps to thank Allah for providing a relative with a good job or some windfall profit. The reason is always stated as being linked to some benevolence on the part of Allah and generally involves thanks for some event which was not entirely in the host's control.
In such a universe, there are only two ways of getting and keeping a lasting hold over someone: debts and gifts, the physically economic attachments created and maintained by the generous gift... [Bourdieu 1990: 126]
An important thread in this argument is linking the market with non-market determined decision making. The economy of rural Punjab is clearly not pre-capitalist, however, there are aspects of rural exchange relations which are not market driven, and are derived from non-capitalist economic conditions. Like the suq in Algeria, the economy of rural Pakistan is neither purely `traditional' nor `rational' (in a capitalist economic sense). Bourdieu argues that the suq in Algeria supplies both `traditional' and `rational' information [Bourdieu 1990: 124]. In these conditions, the economy does not provide the `conditions necessary for an indirect, impersonal domination secured quasi-automatically by the logic of the labour market.' [Bourdieu 1990: 123]. The opportunities for physical violence and other means of physically ensuring the subordination of individuals is limited within Punjabi villages (despite popular press reports to the contrary). The mechanisms in place for restricting powerful individuals' ability to physically dominate the less powerful would tend to make that avenue of domination less efficient than the more economical assertion of dominance through symbolic violence.
Physical violence in Pakistan carries risks which extend beyond the individuals involved or particular events. Among Pukhtuns, Barth [1959a], Ahmed [1976; 1980] and Lindholm [1982; 1996] state that physical violence is an accepted part of the cultural discourse, even between close patrilineal kin, yet the consequences of blood feud are severe. The ideal prescribes violence in more often than actually occurrence of violence. Punjabis share a system of values which praise violence as a response to certain actions, but they also tend to avoid it when possible. Individual disputants may indeed wish to commit physically violent acts, but are discouraged by their patrons and kin members. The incidence of physical violence in Punjabi
despite the impression generally created by ethnographies of the Pukhtun (including my own), violence in the valley is quite rare and everyday life is certainly more courteous and safer than ordinary life in many supposedly less violent societies. [1996: 50].
So while there is clearly a rhetoric of aggression and violence, similar to the situation described by Gilsenan  in northern Lebanon, where violence would appear to be the modus operandi of the rural areas, this is not reflected in the incidence of actual physical violence. The most frequent mode of domination between collaterals, therefore, is symbolic.
When physical violence is directed against status and material subordinates the risk of blood feud is reduced but there are, nevertheless, serious consequences which must be addressed. Landlords have the economic resources to commit acts of physical violence against villagers with little fear of effective physically violent reprisal, but must contend with other forms of revenge. Villagers, who constitute the labour force in agricultural villages, may respond by increasing complaints and lack of cooperation. Villagers can and do evade calls for corvée labour. The families of affected individuals may rally behind their relative and collectively interfere with the daily operations of running a farm. Gossip composed of insults and backbiting is common within a closed circle of people, but it may become an effective tool for checking the power of others, even the very high status, if spread to outsiders [see Gluckman 1963 on the power of gossip to check behaviour]. Villagers may insult landlords in village tea shops in the area. News of discontent spreads quickly between villages and rival landlords pay close attention to potential weaknesses of each other. In a society which places such a high premium on izzat2 (both individual and collective), claims that a man or his family are stingy or unjust or cruel, are taken seriously. Every landlord family has legends about the generosity and indulgence of its members, in order to demonstrate that they are benevolent.
Some of my informants, both landlord and non-landlord, reported that the incidence of physical acts of violence committed by landlords against villagers has decreased as a result of the democratization processes under Ayub Khan and Z. A. Bhutto. They also report that increased economic opportunity for villagers in cities has afforded them greater protection from rural landlords. Neither increased democratization nor economic autonomy, however, fully explain the reluctance on the part of powerful men, resident in the village, to engage in physical violence. It is, rather the contrary which would be surprising. If past landlords really did commit gross acts of violence against villagers then that is what would require explanation. I suggest rather that landlords have never been free to employ violence as a primary means of controlling villagers.
Landlord reluctance to commit violence stems, in part, from their desire to be perceived as `just' and benevolent. They want to be respected as well as feared, and while they may speak about their `peasants'
loving them, heads of household landlords rarely believe that villagers love them. The authority of landlords rests in large part upon the acceptance of their personal and family authority, which `can only be lastingly maintained through actions that reassert it practically through their compliance with the values recognized by the group' [Bourdieu 1990: 129].
The separation of power and violence [Arendt 1970: 46-56; Lukes 1974: 28-30] provides additional explanation for this hesitancy towards physical violence. If a landlord commits frequent acts of physical violence it implies, paradoxically, a lack of power. A powerful landlord should be able to assert his will without recourse to physical violence. It is precisely situations like this to which Bourdieu refers. Where the opportunity for physical violence is, for whatever reason, curtailed, individuals employ other means to assert their `personal authority' and domination.
I have some misgivings about applying Bourdieu's terminology to Pakistani giving in the case of the dég ritual. On the one hand, gift giving is a highly elaborated process in Pakistan. Individuals keep a careful record of the amounts of every gift and gifts may indeed be perceived as expressions of hostility, or somehow interpreted as aggressive. Gift `aggression' is observable in the manner in which a gift may be received or rejected [Eglar 1960: 173-174]. Gifts must be reciprocated and when this is not possible the nature of the relationship shifts. The relation is no longer that of co-equal but of super- and sub-ordinate. It may then develop into a clear patron/client relationship. There can be little dispute that dominance, or at least hierarchy, may be asserted or expressed in gift giving. When dealing with highly formalized sets of exchanges, like marriages, Bourdieu's analysis is thoroughly appropriate. The act of giving and receiving is calculated and involves voluntary participation by all parties. It can be negotiated and is measured and finessed throughout the process. Each side jockeys for some position which allows them the `upper hand' in symbolic capital terms, if not in materially economic ones. Bourdieu's use of the term `violence', nevertheless, is contentious. In Pakistan guests are sacred. The manner in which one treats one's guests is a reflection of one's own value as a family and as an individual. Moreover, this belief seems to be widespread throughout the Islamic world; guests are in a very real sense the guests of Allah. Stinginess towards a guest of Allah would, of course, be an insult tantamount to blasphemy. An individual may well attempt to garner some accumulation of symbolic capital, or impose a debt on another individual as a means of effecting more control over some aspect of life, but the term `violence' may be misleading. I would be more comfortable with a term such as symbolic aggression but of course this may encompass far too many acts. So while I am not entirely sure the term is appropriate, I find the concept apt, so will accept the former in order to retain the latter.
The religious phenomenon of langar (public kitchens attached to Sufi shrines), mirrors the dég ritual. Langar are organised around shrines to pirs (saints). In the langar the `giver' and `receiver' become symbolically, if not materially, confused. The pir is the redistibutor of goods and food provided by his murid (devotees). He is not perceived to be a `receiver' because he does not `need' the gifts provided. Indeed Werbner states that saints will not accept ` sadqa (gifts to the poor), but only nazrana, a religious tributory gift of money, valuable objects or food given as a mark of respect or in gratitude for a blessing bestowed' [1990: 271]. Werbner reports that the recently deceased Zindapir of Ghumkol Sharif in Pakistan was a vegetarian and a perpetual or infinite giver [Werbner 1990: 272; 1998: 107]. Like God, a true saint needs no sustenance from his fellow humans. The saint's sustenance comes directly from God himself. In other
words, a saint, like God, cannot be a receiver. His status and position preclude the possibility of him assuming a subordinate role with other individuals, therefore even if he materially receives, the exchange must be redefined to negate the receivership of the saint. There is clearly a sense of status and prestige associated with giving and receiving. To what extent one may categorise this as `violence' is ambiguous, as indeed, is the particular status or prestige assigned to the givers and receivers. The act of receiving is an admission of `needing' which is itself a result of `imperfection'. The case of pirs and their `good faith economy' suggests the status associated with `giver' and `receiver' may be open to manipulation in order to shift the direction of status and prestige. Indeed, the intellectual gymnastics required to transform giving and receiving into some other category provide entertaining examples which lend additional support to Mauss' seminal thesis (see Laidlaw 2000 for the ways Jains in northern India may receive without being `receivers').
In the following case study, however, I believe that the term `violence' is not only appropriate from the outside analyst's point of view, but that it also resonates with an indigenous understanding of the process. The violence is not directed at any of the participants of the `gift', but rather at a co-equal of the giver who may or may not partake in the ritual. The ritual is an act of aggression and an expression of rivalry. The receivers, in this case the individuals and groups who actually eat the rice, may be equated to the scoreboard in a sports event. They, along with the number of cauldrons prepared, are the objective measure by which the aggression (or violence) may be deemed to have `succeeded' or `failed'.
During my field research, I attended many dégs, which were held for many different reasons. Some of them were modest and some were very lavish. All were imbued with religious symbolism and ritual, regardless of the stated `reason' for the dég. Most dégs involved fewer than 10 large cauldrons, some as few as two. All of them were divided equally between sweet and savoury rice. Every dég was open to all guests. Anyone who showed up on the day was given rice, so long as quantities lasted. The largest dég I witnessed in the village, reputedly the largest ever to take place in the village, lasted from just after sunrise to just before sundown. Somewhere between 25 and 30 large cauldrons of rice were prepared for everyone in the village. Villagers who could not walk to the dég area, had rice brought to them by children. Every household has some representative helping to eat the rice. The reason for this dég was kept particularly secret. The host was my best friend and host in the village, Malik Asif Nawaz.
Malik Asif and I kept few secrets from each other (at least I kept few secrets from him and he was very indiscreet with me), but on this occasion he did not choose to share his thoughts. Speculation on why Malik Asif was being twice as generous as anyone ever had been ran wild: he was taking a second wife; he was going to have another child; I was going to take him to Europe with me; he had arranged a Pakistani wife for me; I was converting to Islam; or he had managed to get compensation for his land destroyed by the nearby motorway construction. None of those explanations were true. Later Malik Asif gave me some clues as to his reasons. He felt that Allah had wanted him to feed his village so he did. The dég was in honour of his God. I accept his stated reasons (and those of others) as partial explanations. The reasons people give for such social phenomena, however, should not be taken as complete or thorough explanations (or even descriptions) of the social impact.
Bourdieu's symbolic violence is already an indirect expression of aggression so it may seem odd that I choose to use the term indirect symbolic violence. It might be more accurate to adopt a terminology similar to English kin terms for cousins: symbolic violence once removed (or twice removed, etc.). Symbolic violence is an indirect way of expressing `violence' against a giver, which is not the focus of this analysis, but rather the `violence' directed against someone who may not form a part of the giving directly. So the `target' of the violence is removed from the event, hence the cousin terminology. I find `symbolic violence once removed' an awkward term, however, so I have opted instead to risk redundancy for the sake of linguistic ease.
The interesting part of the dég is that from the giver's point of view, the receiver is somewhat incidental in his assertion of dominance. That is to say, the giver requires a great many receivers but domination over the receivers is not his primary goal. The giver's target group, so to speak, is the very few individuals in the village who are in a position to sponsor such events of this magnitude.
There are several reasons why I believe this to be true, both emicly as well as eticly. First, the efforts involved in swaying someone's primary loyalties far exceed a single gift-- even a highly valued one such as a good meal of rice. Typically landlords spend months cultivating sharecroppers and servants of other landlords in order to get them to work for them. The actual `endebtment' or domination over a villager usually coincides with some problem of the villager which the landlord is in some position to alleviate. Within the context of the dég, all wealthy men are somewhat obliged to host them on occasion. As I said before, the concept of izzat is extremely important for Punjabis, indeed for all Pakistanis as far as I know. The last thing any landlord wants said of him, is that he is miserly or stingy. Feeding villagers is therefore not an attempt to control villagers or endebt them, but rather an act which satisfies an obligation which exists because the landlord is dominant. In this sense, then the dégs are not attempts to exert domination over villagers.
Nor can they be seen as merely expressions of domination over villagers since in a large dég the entire village may freely participate. Other landlords and their children will also enjoy other people's dégs. Moreover, the act of eating rice at someone's dég does not incur a debt to the host of the dég. I say this with some confidence because hundreds of people may eat the rice and on no occasion did I hear about any landlord reminding someone that he had eaten at his dég to influence the person in the future. Other favours and gifts, however, do get brought up in the future in order to pressure individual's and groups into cooperation or compliance. So indigenously, there is not a sense that eating a man's rice at this ritual incurs any future obligations. Nor does it incur an obligation of reciprocity to feed a landlord. The obligation to feed one's guests is ever present in the Punjab and does not become heightened because one has received food. The presence of villagers heavily in the debt of other landlords and of complete strangers who have no connection to the village, such as truck drivers or passing pastoralists, goes further to negating the idea that the dég may be seen as some kind of expression of dominance over the receivers.
Furthermore, there is no real obligation to attend the dég of the landlord to whom one is closest. If one is in the village then there is certainly positive social pressure to attend; the events are festive and provide a welcome break from the routine of village life so villagers will certainly go to the dég if they have any
connection to the host. If for any reason they must be out of the village, the host does not interpret absence as a withdrawal of loyalty or any other expression of animosity. The exception to this would be if villagers boycotted a dég en masse in which case that would most certainly be construed as an expression of disloyalty and lack of respect.
There is, nevertheless, an expression of aggression involved in dégs. In order to understand this I will look more closely at the relationship of Malik Asif to his biraderi (patrilineage), which given the incidence of patrilateral cousin marriage effectively means his cognatic relatives as well as the majority of his affines. One of the activities of Punjabi landlords is either claiming a relative's lands or defending against rival claims to their own lands. There is frequent land `poaching' between cousins. While the concept of sharik, or cousin, does not seem to have the same animosity in the Punjab as does the equivalent tarbur among Pukhtuns [see Barth 1959; Lindholm 1982], there is tension associated with sharika. One's sharik is both one's close family and most frequent rival. Malik Asif had been engaged in a very serious land dispute with his fraternal cousins during the previous year [see Lyon 2002; 2003; Forthcoming] for a more complete discussion of land disputes and conflict resolution strategies]. He had also recently settled a multiple generation dispute with one of his father's fraternal cousins (though the dispute is not settled in such a way that it cannot arise again when the elder disputant dies). Between close family like this there is very little room for any physical violence. They may yell at each other and certainly engage in some serious rumour mongering and backbiting, but serious physical violence is culturally unacceptable. They must therefore find other ways to exert dominance. Hence my application of Bourdieu's symbolic violence. The personal authority landlords may assert over each other must be in `compliance with the values recognized by the group' [Bourdieu 1990: 129]. In this case, the group is restricted to the landlord families, which do not allow their members to try and kill or beat each other except in very limited circumstances. Indeed on the few occasions when intra-family tension has mounted to the level of physical violence, the entire biraderi become involved to resolve the dispute. The social pressure placed on family members to refrain from physical violence is such that it is an extremely rare occurrence.
When Malik Asif inherited his land he was a student in Punjab Government College in Rawalpindi. Within the first year of assuming control of his father's land he estimates he lost several hundred kanal of land (1 kanal = 1/8 of an acre). In the 16 years that passed between his father's death and the time of my field research he managed to recover some, but not all of that land. In addition, he has made a considerable effort to increase the profitability of his land by increasing his yield through the use of chemical fertilizers and irrigation. Five years prior to my arrival in the village he began a very ambitious project to build a citrus orchard. Citrus orchards are one of the symbols par excellence of landlord wealth and prestige. There is at present, only one citrus orchard in the village, and that belongs to one of the men that Malik Asif believes was responsible for taking much of his family land. This orchard belongs to his maternal uncle (who is nevertheless part of the same patrilineage). Relations between these men are very close and they frequently work together but there remains a good deal of tension and `jockeying' for control. Malik Asif, being 20 years younger than his mamou (maternal uncle) has limited room for open rebellion. He must abide by the `values recognized by the group' [Bourdieu 1990: 129]. One of those values stresses obedience and respect to elders. This does not stop Malik Asif from being an ambitious landlord however and wanting to assert his own position of `controller' or dominant landlord.
The `violence' expressed in Malik Asif's dég is therefore directed at two generations of his own family. First he is making a direct challenge to the elders of his family, whom he perceives as being responsible for his own loss of land and wealth. Second, he is making a statement to those of his own generation that he intends to be the most influential landlord in the village when the current elder generation retires or dies.
This is achieved by raising the stakes of the `giving' within the village. By hosting such a large dég Malik Asif has raised the level of expectations. To be sure, no one expects to have such a large dég on a regular basis, but it has become the objective measure by which other dégs are evaluated. Among Malik Asif's generation, none of them can truly hope to host such a dég. While they could probably pay for such an event, it requires a great deal of organizational cooperation which Malik Asif spent a long time cultivating. They have, by and large, not cultivated such an organizational network and so are at a disadvantage in hosting such an event. The only landlords who have both the wealth and the organizational ability are the elder generation-- and probably only two men among that generation. Of those two men, one is the mamou who Malik Asif feels is responsible for his own loss. The dég inflation affects them far more than Malik Asif's fraternal cousins. They could host such an event so if they do not then the explanation is that they are stingy or miserly. I have not heard of an equally large dég being hosted since my departure but I expect that it may happen, if only to demonstrate that the elder generation is as generous as Malik Asif.
Where Bourdieu's symbolic violence seems most appropriate is precisely between these men. The constraints of family are such that they cannot easily assert dominance, as landlords, over one another in any physical way. They must then make use of debts to control each other. If they can manage to out-give one another then they may successfully get one another in their debt, or more importantly, under their control. Giving between landlords is extremely costly however. Malik Asif could not hope to win control over his uncles by direct gifts, and it would almost certainly entail a serious sacrifice to his own family. By elevating the scale of dég giving Malik Asif is able to enter into a kind of `giving' competition with men who are far wealthier and more influential than himself. The challenge, or symbolic violence, is made indirect and deniable. Malik Asif's uncles cannot openly accuse him of challenging them but they remain obliged to respond by hosting their own dég or suffer the consequences of allowing Malik Asif to be seen in the village as the most generous landlord.
The attraction of such a challenge, or expression of dominance, is that if it succeeds it does not disrupt family harmony and if it fails there is no serious loss of izzat . The family members who decline to engage in the dég inflation can continue to host their own, more modest dégs and will be seen to assume their obligations to both the villagers and to Islam. Those that choose to host equal, or larger dégs, may be able to assume the role of `most generous' landlord but they will not detract from Malik Asif's generosity. He will be seen as one of the generous Maliks, if not the most generous.
The dég express rivalries in ways which are socially sanctioned. They are not effective means of garnering power but they are necessary expressions of power. I have included an analysis of this excessive dég because it demonstrates some of the tensions inherent in relationships of equality in Pakistan. Where individuals try to assert equality there will be tension. This event, and others like it, provide clues as to why Pakistanis so readily assume roles of asymmetry even when there is no immediately apparent need to do so, at least from
an outside observer's viewpoint. This example also provides further evidence for Lindholm's assertion that the differences between Pukhtun and Punjabi culture are ones of scale rather than kind. The dég provides a condoned mechanism for expressing collateral aggression, which, as in this case, may often be agnatic rivalry. The underlying principle of tarburwali can not, therefore, be understood as the exclusive domain of Pukhtun culture. It is an important and vibrant motivator of behaviour in Punjab, but like Pukhtun society, the repercussions of physical violence are normally considered excessively detrimental to the interests of all landowning clans; consequently Punjabis employ symbolic mechanisms for expressing their equivalent of tarburwali .
From the preceding argument I make two further generalizations which may be applied to wider anthropological debates. One, Bourdieu's concept of symbolic violence is less useful when examining patron/client relationships, but would seem to have more currency when dealing with groups or individuals that may be seen as co-equals. Two, the intended `victim' or target of symbolic violence may not be outside the giver-taker dyad, indeed a receiver or taker may not even be necessary.
The first generalization is drawn directly from the example of dég rituals. The relationship between patron and client is such that excessive giving from the patron can hardly be considered symbolically violent or even an attempt to assert control. It is the patron's obligation to give. It is the client's obligation to serve. If a servant works an extra three hours in a day should one consider that as an attempt to increase his patron's endebtment to him or herself? That is the nature of the relationship and any symbolic violence should be identified in other areas. Where the notion of symbolic violence does seem useful however is between equals who may be vying for some endebtment on the part of another person. In this case, it is imperative that individuals consider the obligations associated with every gift. As in the situation that Eglar  describes, where people may refuse part or all of a gift if they feel its value exceeds their ability to reciprocate, a gift may indeed be a declaration about hierarchical relationships. That situation pre-exists the patron/client relation however and may be seen as an attempt to establish such a relationship. Once the relationship is in place the notion of violence expressed through gift giving loses its utility.
The second generalization may be the one which is far easier to relate cross culturally and hearkens back to Maussian notions of reciprocity and the gift; an individual's, or a group's general level of expenditure constitutes a challenge to others of the same economic position. Specifically, the expense of a dég and the act of escalating that amount is, in fact, a challenge to other potential dég hosts. Not only does it eliminate poor people from the competition, it makes it more expensive for fellow landlords to compete in the stakes for the position of `most generous' landlord. Beyond the level of the village and this specific ritual, the manner in which a person spends his or her money makes a statement to others of the same economic group. Thus the former Prime Minister Nawaz Sharif's ostentatious personal spending offered a direct challenge to those of similar economic situations. Not only did he have a new road built through Lahore from his home to the motorway, he had parks built, he bought property in Britain, he and his family drove expensive cars and wore expensive jewellery and clothing. All of this excessive spending was directed at themselves, so there was not even a `receiver' to keep score (though in other expenditures there were receivers of Nawaz Sharif's generosity). I argue that this kind of spending is no less an act of symbolic violence, but of an indirect kind. Nawaz Sharif escalated the cost of what it means to be an important and influential man in Pakistan, thereby limiting the number of individuals who may reasonably consider participating. Out
spending in that category of persons is extremely dangerous. Spending on oneself is far safer and any challenge which may be inferred can be denied.
Ultimately these issues lay at the heart of the feasibility of human society. How do groups of people manage to live together at all? What prevents human beings from trying to kill each other whenever there are conflicts of self-interest? These symbolic challenges that individuals engage in provide interesting examples of the tensions that exist between individuals who may be striving for equality or dominance yet for allow peaceful coexistence. Not all challenges offer non-physically violent escapes, of course, and they require other explanations. In this paper I have attempted to demonstrate a cultural mechanism for avoiding physical violence while simultaneously challenging someone else's right to one kind of capital.
Lindholm, C. 1982. The Swat Pukhtun Family as Political Training Ground. In Anthropology in Pakistan: Recent Socio-Cultural and Archaeological Perspectives (eds) S. Pastner & L. Flam. Ithaca: Cornell University South Asia Program, South Asia Occasional Papers and Theses No. 8.
--. 1998. Langar: Pilgrimage, Sacred, Exchange and Perpetual Sacrifice in a Sufi Saint's Lodge. In Embodying Charisma: Modernity, Locality and the Performance of Emotion in Sufi Cults (eds) P. Werbner & H. Basu. London: Routledge.
1. I would like to thank Peter Parkes, Susan Bayly and members of the Public Culture in Theory and Practice Research Group at the Department of Anthropology, University of Durham for thought provoking comments on drafts and ideas contained in this article. All faults or weaknesses are, of course, my own.
2. Izzat is one of those fundamental concepts which confound simple translation. It may crudely be translated as `honour', however, the English term, with all of its ambiguity, fails to capture the importance of control and hospitality embedded within it [see Stewart 1994 for a review of the ambiguity of the term `honour' in Europe]. It is not dissimilar to the Pukhtun concept of pukhtunwali [see Ahmed 1980: 3; Lindholm 1982: 25-26] but has not been fetishised in the same way as an explicit defining Punjabi code.