10 SOCIAL HIERARCHIES For unto everyone that hath shall be given, and he shall have abundance; but from him that hath not shall be taken away even that which he hath. — St Matthew 25: 29 Like gender and age, caste in the Indian subcontinent refers to ostensibly inborn, ascribed characteristics. In theory, changing one's caste membership is as difficult as changing one's gender. The concept of social classes is different, and not only because classes exist in many different kinds of societies, while caste is usually associated with Hinduism and India. Although many social scientists have demonstrated that most people in class societies take over their parents' class membership (if one is a working-class child, it is highly likely that one remains in the working class), there is a great deal of mobility between the social classes, both in theory and in practice. In many societies, class membership is emically considered an achieved and not an ascribed status. The relationship between caste and class is complex, and both concepts are difficult to define accurately. Let us begin with the concept of caste. THE CASTE SYSTEM The caste system encompasses aspects of both 'culture' and 'society'; that is, it is both a symbolic system associated with Hinduism, and a set of rules and practices regulating social organisation, interaction and power in Indian society. The caste system can be defined as a system dividing all of Hindu society into endogamous groups with hereditary membership, which are simultaneously separated and connected with each other through three characteristics: separation regarding marriage and contact; division of labour in that each group, at least in theory, represents a particular profession; and finally hierarchy, which ranks the groups on a scale dividing them into high and low castes. The caste system thus entails a ranking of people according to ascribed statuses, it provides rules regulating the interrelationships between members of different castes, and it creates mutual dependence of the castes through 142 Social Hierarchies -, ., the division of labour, which implies that certain tasks can only be carried out by members of specified castes. Regarding ideology and religion, the caste system is based on notions of ritual purity and impurity, which serve to justify the segregation and division of labour between the castes. The variations in ritual purity imply, among other things, that a member of a high caste will be polluted if he or she eats food prepared by a member of a low caste, and that only members of the Brahmin caste are entitled to lead religious rituals. Each caste has its own rules for good conduct; for example, high castes tend to be teetotallers and vegetarians. VARNA AND JÄTI It is common to think of the caste system as a hierarchy dividing the entire Hindu population into four main groups, the varnas (a Sanskrit word meaning 'colour'). The Brahmins (priests) have the highest rank, followed by the Kshatriyas (warriors and kings), the Vaishyas (merchants) and the Shudras (artisans and workers). Outside the varna system proper, at the very bottom of the ladder, are the so-called Untouchables, labelled thus by the British because a high-caste person had to go through an elaborate purification ritual after having touched such a polluting person. Members of the three highest varnas are called 'twice-born' because they have gone through a ceremony entailing spiritual rebirth. Such a description of the caste system, although it is not incorrect, is simplistic and ultimately misleading. First, it should be noted that there are also non-Hindus who belong to castes; India's approximately 130 million Muslims have their own castes (usually low ones), and the many ethnic groups sometimes called 'tribals' tend to be classified and treated as Untouchables. Some of them, partly for this very reason, have converted to Christianity or Buddhism. Caste, moreover, also exists in non-Hindu societies in the Indian subcontinent, from Buddhist Sri Lanka to Muslim Pakistan. Furthermore, and more significantly, this fivefold partition of the Indian population is highly abstract and has a relatively modest significance in daily life. The Indian social anthropologist M.N. Srinivas wrote already in the early 1950s: 'The real unit of the caste system is not one of the four varnas but jäti, which is a very small endogamous group practising a traditional occupation and enjoying a certain amount of cultural, ritual and judicial autonomy' (1952, p. 24). There are thousands of jätis in India; they all represent an hereditary profession or craft (even if it is no longer practised) and have longstanding commitments to the other jätis in the area as well as a special place in the ritual hierarchy. Most jätis are relatively small, but some are large and internally differentiated groups with asymmetrical alliance patterns between the lineage segments. 144 Small Places, Large Issues At an intermediate level between the jätis and varnas, we find the all-Indian occupational castes, for example the Lohar caste of blacksmiths, which encompasses jätis all over India. The practical function of the varna system is mainly to make it possible for members of local jätis to locate themselves in relation to jätis in other parts of the country; additionally, it represents a fixed, abstract hierarchy and value system. 'Untouchables' (now called Dalits, meaning 'the oppressed ones') do not have a place in the varna system proper, but they do belong to jätis. Members of the lower jätis, comments Srinivas (1952, p. 30), have always tried to improve their rank 'by adopting vegetarianism and teetotalism, and by Sanskritizing [their] ritual and pantheon'; this process of caste-climbing, he adds, has led to the spread of a remarkably uniform value system in the subcontinent. There seems, in other words, to be widespread agreement concerning values and criteria for distinguishing a low caste from a high one. THE JAJMANI SYSTEM The traditional jäti-based division of labour in Indian villages is called the jajmani system. It consists of a set of traditional rules about the exchange of products and services between the members of different castes. In other words, each caste has specific commitments towards the others. Seen from a systemic perspective, one may say that the village is upheld as a social system thanks to the interdependence between the castes; seen from an actor perspective, one may say that it creates significant structural constraints on individual opportunities. The jajmani system is ideologically connected with and justified through religion, and thus contributes to maintaining notions about purity and impurity and about relative rank within the caste system. The fact that members of the sweeper caste actually sweep is interpreted as evidence that they are actually polluting; at the same time, they have to sweep because they are perceived as polluting. In this way, the social and symbolic aspects of caste are interconnected, and contribute to reinforcing each other by creating a correspondence between ideology and practices. Traditionally, little money circulated through the jajmani system, which largely consisted of direct exchange of goods and services. Frequently, no exact account was kept of these prestations, since each caste had specified duties toward the others. In modern India, it can be difficult to make the jajmani system function according to traditional practice. First, the monetary economy has made it possible - and this is now widespread - to buy all kinds of services and commodities from people with whom there is no j ajmani relationship. Second, changes in Indian society have created a large number of new occupations which are not legitimated through the jäti system. Third, the incorporation of Indian villages into the capitalist market and state bureaucracy has weakened the ties between the jätis. Urbanisation has made Social Hierarchies 145 the jajmani system impractical. It did integrate the social system in villages with a stable economic system and division of labour, but it does not work properly in a large city like Mumbai (Bombay), with a high economic rate of change, huge in-migration and a very complex division of labour. This does not imply that every connection between jäti and profession has vanished in Mumbai; it rather means this connection is more tenuous, ambiguous and open to manipulation than in the traditional village. CASTE AND SOCIAL MOBILITY The relationship between caste and other criteria for social rank may be complex in other, more 'traditional' settings as well. F.G. Bailey's classic studies of 'caste-climbing' and political conflict in the village of Bisipara (Orissa, eastern India; see Bailey 1968) reveal some of this complexity. His perspective is largely an actor-centred one, while most research on India has tended to focus on systemic properties. In Bisipara, the jajmani system was still more or less intact when Bailey carried out his fieldwork in the 1950s, but some disturbing elements had entered from outside. Notably, the village had been integrated into the modern Indian state in the sense that important political career opportunities were now available to members of the low castes. Bailey describes different kinds of political conflict and competition within the village, and we shall look more closely at two: caste-climbing and caste conflict. A person has three possible ways of improving his or her situation. He or she may try to change caste membership (which is exceedingly difficult in a small village), try to improve his or her own caste's relative rank, or dismiss the entire caste system and try to make a career outside it. The distiller caste in Bisipara had improved its economic condition steadily for decades, but this did not entail an improvement of its ritual purity. One may, in fact, be well-off and ritually polluting and vice versa, although there tends to be a correlation between wealth and ritual purity. The leaders of the caste thus tried to convert their economic capital into ritual purity. First they had to purify their own practices through what is generally known as San-skritisation. They ceased to perform typical low-caste rituals (such as animal sacrifice) and adopted Brahmanic rules in other respects as well, becoming vegetarians and so on. For the distillers to be recognised as a high caste, however, they also had to improve their position within the jajmani system, and this was the most difficult part. Each caste which entered into a jajmani relationship with the polluting distillers risked being surpassed by them in the local hierarchy; on the other hand, the distillers were able to pay well for services. Eventually they succeeded - after having passed through several rungs on the ladder -in having Brahmins perform ritual services for them and were thereby 146 Small Places, Large Issues considered a clean caste, but at the cost of a considerable amount of money. They converted a high economic position into a high ritual position. This kind of social mobility, we should note, does not challenge the caste hierarchy as such: rather, it openly endorses it. The type of political conflict spoken of as caste conflict is of a different kind. The Pan caste was a lowly, ritually polluting and poor caste whose leaders were inspired by Gandhi's notions of caste equality. On several occasions they had tried to enter a temple reserved for the high castes, but had been evicted. They then built their own temple and declared themselves a clean caste. This strategy was not accepted by the other castes in Bisipara, who saw it at a blatant breach of rules, and it did not help in improving the socially defined rank of the Pan caste. To the Pan caste, the strategy followed by the distillers was not feasible for economic reasons. Thus some of them began to follow a third course of action, trying to improve their rank through a rivalling value hierarchy, namely the public service. By exploiting the quotas for the ritually impure in local government, they succeeded in climbing socially within a hierarchical scale where criteria other than caste membership and ritual purity were relevant. From this sketch of social mobility in an Indian village, we see that there are three scales of rank, functioning partly independently of each other: the economic system, the caste system and public administration. Different resources are at stake within the respective systems, but all of them entail power differentiation, are legitimated through different principles and are partly incommensurable. The Brahmins of Bisipara did not recognise the Pan rise to prestige through public administration; however, through climbing in local government, members of the Pan caste became able to exert some power over Brahmins. It should also be noted that while it proved possible to convert economic power into ritual purity, it is not possible to convert political power (in the public sector) into similar prestige. This seems to be caused by the workings of the jajmani system, which combines economic and ritual factors but does not include the state sector. CASTE: A KIND OF SOCIAL STRATIFICATION? Research on caste has always been linked to the study of the Indian subcontinent. However, it has been argued that the concept of caste can fruitfully be transformed into a comparative concept with a wider scope. Let us consider two examples of this view. In connection with his study of politics in Swat valley, north Pakistan, Barth (1981) describes the system of stratification in the valley. Although the Pathans are Muslims, in this area they are so strongly influenced by their Hindu neighbours that they are divided into hereditary status categories Social Hierarchies 147 associated with varying degrees of purity, which are reminiscent of castes. The division of labour in Swat valley resembles a jajmani system: there are relations of interdependence and mutual obligations including 'saints', landowners, priests, craftsmen, herdsmen, peasants and 'despised groups' (which correspond to the Untouchables or Dalits in Hindu society). However, the 'caste system' of Swat is not related to, or justified through, a religious superstructure. On the contrary, there is a direct contradiction between the caste system and Islam, which teaches that all are equal before God. Since the Pathans are not Hindus, Barth therefore argues that their castes are chiefly an aspect of social stratification and of the division of labour. This view of caste as an aspect of social structure, which Max Weber also endorsed, has been argued against by Louis Dumont (1980 [1968], pp. 208-12). According to Dumont, to understand caste it is necessary to view it as an integrated part of a social and cultural totality; one cannot therefore talk of castes in isolation from the particular cultural context in which they have emerged. The presence of 'castelike systems' in non-Hindu societies is accounted for by the spread of some aspects of the caste institution. To Dumont, these 'imitations' of the caste system in Swat, among Christians in Kerala (south-west India) and among Buddhists in Sri Lanka and elsewhere, are encompassed by, and influenced by, Hindu culture without fully taking part in it. In sum, Dumont argues that caste is an aspect of Indian culture and has to be understood within a Hindu sociocultural totality. The very concept of 'stratification', Dumont has elsewhere argued (1986), is further an individualistic European concept which does not make sense in the hierarchical societies of the Indian subcontinent. Gerald Berreman (1979) has taken a more radical stance than Barth in trying to make the caste concept a comparative one. He argues, among other things, that there are castes in the United States, describing the American blacks as an 'impure caste'. Notwithstanding the official ideology of meritocracy, which holds that everyone controls their own destiny, he claims that blacks belong to a hereditary low-rank category, with low-ranking professions and polluting power if they touch members of the pure castes (whites). Before Berreman, Kroeber also argued the usefulness of such a wide concept of caste, which would clearly be unacceptable to Dumont, who sees it as intrinsically related to the totality of Indian society and culture. Dumont's perspective on caste is clearly a systemic one, while Barth and Berreman place greater emphasis on the actor's available options. In doing the latter, it becomes possible to find important similarities between blacks in the USA, Swat Pathans and Hindu villagers, although their respective cultures differ greatly. Berreman has also criticised Dumont's view of caste as a 'Brahmanic view' (Berreman 1979; see also Burghart 1990; Ouigley 1993) more or less uncritically reproducing the views of those in power. The former does not, in other words, see Indian society as a totality whose members necessarily are encompassed by a relatively uniform world-view. Others, including Pauline Kolenda (1985), have shown how members oflow 148 Small Places, Large Issues Louis Dumont (1911-99)taughtat the University of Oxford for several years, but worked in France from 1955. Although he published many books about India and about general sociological and anthropological questions from the 19 50s onwards, his influence in anthropology has been particularly strong since the early 1980s. A fundamental idea of Dumont, which is adapted from classical sociology (notably Durkheim and Tönnies), amounts to the notion that societies are integrated wholes and therefore qualitatively different from the sum of the parts that compose them. In his famous analysis of the caste system in India {Homo Hierarchicus, 1980[1969]), Dumont thus argues that this system has an intrinsic logic unique to Hindu culture. Whereas the individual is accorded a special value in Europe, the Indian individual is meaningful only in relation to an encompassing, holistic cultural system. In Essays on Individualism (1986), Dumont criticises modern social anthropology for neglecting such fundamental differences between societies, since anthropologists have tended to take it for granted that an egalitarian, individualistic ethos is universal. One of Dumont's most important contributions to anthropological theory is his presentation of value hierarchies. Here, he argues that every society is integrated according to specific values which are expressed at a variety of levels, and that some such values determine - they encompass - other values at lower levels. The highest value of European society, he claims, is the individual. Such encompassing values are seen as more fundamental than, and determine, other values in society, which may nevertheless well be opposed to them. castes consciously develop liberating ideologies in direct opposition to the caste system, many even converting to Buddhism, Islam or Christianity. CASTE IN MODERN INDIA The caste system has a religious, or spiritual, and a practical, social aspect. It has significance for the religious position of people and their ritual practices, for their marriages and alliances and for their possibilities in professional life. It is nevertheless impossible to maintain the caste system unaltered in contemporary India, and there are four main reasons for this. First, the introduction of new professions complicates the classification of people according to jati. Second, in many contexts wageworkers are hired on the basis of qualifications (achieved statuses) rather than caste. Third, Indian authorities actively try to level out the differences between castes through quotas for 'Scheduled castes' and 'Scheduled tribes' in the public sector. Fourth, urbanisation makes it difficult to classify the people one meets, and makes it possible for many Indians to escape from a stigmatised identity by moving to a city where nobody knows them. Social Hierarchies -, a q Caste outside India From 1840 to 1917, following the abolition of slavery in the European colonies, several million Indians were transported to remote colonies where they settled permanently. The cause for this mass migration was the need for fresh manpower in the plantations after the freeing of the slaves. Many have argued that these Indians, tempted by promising labour contracts, were virtually shanghaied and that their actual situation in the plantation colonies were scarcely better than that of the slaves had been. Hugh Tinker (19 74) has described the system of indentureship as A New Form of Slavery (see also Mintz 1974). Most of the Indians came from Bihar and Uttar Pradesh - some of the most impoverished parts of India - but a fair number were also Dravidian speakers from Tamil Nadu and Andhra Pradesh. The descendants of these Indians still live in the former plantation colonies, and they are particularly numerous in Mauritius (forming 65 per cent of the total population), Guyana (c. 55 per cent), Fiji (c. 45 per cent after the coups d'etat in 1987), Trinidad and Tobago (c. 40 per cent) and Surinam (c. 35 percent). There exists a large anthropological literature on this Indian diaspora, and many scholars have raised the question of cultural continuity and change. Regarding the caste system, for example, it was modified from the day of departure, since dietary restrictions were impossible to maintain on board ship, and since the division of labour in the colonies made the jajmani system obsolete. In some of the societies, notably in Mauritius, some castes are nevertheless still endogamous, but in others, such as Trinidad, the various subgroups have for most practical purposes merged into a single category of Tndo-Trinidadians'. Even in Trinidad, however, only Brahmins can become orthodox (Sanatanist) priests. In Mauritius, further, there has in recent j?ears been a revival of caste consciousness in political contexts, and the low castes have formed their own interest groups. The experience from the Indian diasporas, where neither jajmani systems nor panchayats (caste councils) have survived but where notions of caste continue to exist, seems to indicate that the caste system can be both flexible and adaptive, and that it is by no means certain that social and cultural change will eradicate it. On the other hand, it is certain that caste has a very varying significance as a criterion for rank and differentiation - both in India and among Indians overseas. To some persons, caste membership may define their place in society in great detail; to others, it may be relevant only at religious festivals and, perhaps, during election campaigns. 150 Small Places, Large Issues Within Hinduism itself, attempts have also been undertaken to eradicate the caste system. An important reformist movement, Aryanism, has worked towards this end since the nineteenth century, and Mahatma Gandhi tried to modify the caste hierarchy by, among other things, naming the Untouchables 'Harijans', which means 'children of God'. Today, Dalit organisations (rejecting what they see as Gandhi's patronising term for them) militate for the actual abolition of caste. Despite these and other attempts to abolish the caste system or at least mitigate its effects, it remains very much alive. The marriage ads in the Sunday Times of India, for example, are classified according to caste membership. Although the jajmani system may be on the wane in many parts of India, the caste system retains a number of functions in the ritual and social spheres. CLASSES AND STRATA Gender and age function everywhere as principles for the social differentiation and classification of people. Systems based on caste and/or class too are more complex in this respect and tend to have a more complex division of labour. The term 'social class' is usually applied to capitalist societies, although classes, strictly speaking, exist in other societies as well. The most influential theory of social classes was developed by Karl Marx in the mid-nineteenth century. In his very wide-ranging studies of historical societies, especially capitalist ones, the term 'class' had a privileged place, since the relationships between the classes, according to Marx, were decisive in historical change. Marx defined the classes in relation to property. The ruling class in any society is the one whose members control the means of production (land, tools, machinery, factories and the like) and who buy other people's labour power (that is, employ people). Below this class, one would usually find classes of farmers and independent craftsmen, as well as wageworkers who have to sell their labour power to survive. In modern industrial societies, we usually speak of three important social classes: the bourgeoisie, or capitalists, who own means of production; the petty bourgeoisie, whose members own means of production but do not employ others; and the working class, whose members sell their labour power. In addition, there are lumpenproletariats of unemployed, criminals, vagrants, etc., as well as an aristocracy whose members live off the interest from property. There are doubtless great systematic rank differences between people even in societies where equality is emphasised. In practice, class differences tend to be reproduced over the generations, so that children take on their parents' class membership, although there is always a certain social mobility. Whether or not such differences are necessarily connected with ownership of means of production, they are very important from a systemic as well as Social Hierarchies 151 an actor-centred perspective. It should nevertheless be noted that the majority of social scientists support a way of thinking about inequality which, contrary to Marxism, does not give priority to economic property as an explanatory variable. This is sometimes labelled the theory of social stratification, and is associated with Max Weber. Weber, writing half a century after Marx, argued that there were several, partly independent criteria which together gave a person a specific rank and that property was not necessarily the most important one. Political power and intellectual prestige could, for instance, be just as important in a given society. An important difference between perspectives on classes or strata concerns the significance placed on conflict. Class theory is nearly always a kind of conflict theory, seeing the conflicts between different classes as fundamental. Marx saw class struggle as the most central factor in social change, since successful class struggles eventually led to changes in the relations of production (property relations) and qualitative changes in the social order. Both Marxist thought about social class and other theories of social strata or classes have been criticised for being ethnocentric. Dumont's criticism of the wide-ranging analytical uses of the term 'caste' is representative of this kind of argument. Whereas some would hold that all societies are stratified and that concepts of classes or strata are therefore universally useful, others would stress that the concepts themselves are European and relate intrinsically to modern state societies. We return to some of the economic and political aspects of social strata and classes in later chapters. At this stage, we concentrate on class as a principle for social differentiation and classification. The following example, which shows the introduction of capitalism in a formerly feudal society, indicates that there may indeed be important interrelationships between economic change and cultural change. During the first half of the nineteenth century, an important shift took place in San Jose, Puerto Rico (Wolf 1969[1956]). Instead of largely growing food for their own consumption, the farmers increasingly began to grow one main product for the world market. The most important product was coffee, for which there was a growing demand. Several problems had to be solved, however, for the production to become profitable. The landowners had to increase their cultivated area, and thus they also had to find wageworkers. They also had to find a source of credit (a bank or similar institution) to fund the expansion. As long as land was abundant, it was difficult to find wageworkers, since people preferred to cultivate their own plots. Gradually, however, the coffee-growers established control over most of the available land and eventually it ceased to be free: all land now had to be purchased. This new situation led to the proletarianisation of a large number of formerly independent smallholders: they became dependent on selling their labour power. Wolf describes the confrontation between the two systems - the capitalist one, based on purchase, sale and interest, and the traditional one, based on 152 Small Places, Large Issues subsistence production - as a cultural conflict, a clash of opposing value systems. The locals described the Spaniards (most of the coffee-growers and money-lenders were Spanish) as mean and individualistic: 'They rob other people's money, but then they just sit on it. They don't spend it', and they 'don't give people subsistence plots on which to grow things to eat' (Wolf 1969, p. 178). The Spaniards, for their part, described the Puerto Ricans as lazy, drunken and unable to plan their life properly. When this new, specialised economy was fully developed, it became possible to distinguish between four social classes in San Jose: 1. Peasants. They own a little land, cultivated on a family basis, and sell the surplus on the market, but are unable to generate enough surplus to expand their production. 2. Middle-sized farmers. They own more land, buy labour to grow it and make a larger profit than needed to sustain their lives. 3. Rural proletarians. They own no land and earn their living through selling their labour. 4. Landowners. They are specialised coffee-growers, buy labour on a large scale and make large profits. Here the social classes are definitely defined in relation to the means of production. The hacienda owners rank highest; the propertyless lowest. In many traditional societies, we should note, it is impossible to rank people according to ownership of means of production, since land frequently cannot be sold or bought. Among the Dogon, we should recall, the village headman (the hogon) decides who is to cultivate which plot and land rights are tied up with kinship. In hunter-and-gatherer societies, there is no systematic difference in access to means of production; for example, all men have a bow and arrow. In Chapters 12 and 13 we look more closely into these major differences in the economic organisation of societies and their connections with political power and social organisation in general; at this point, it is sufficient to note how economic differences, and the social organisation of production (division of labour), have ramifications in the cultural sphere and engender important differences in the classification of individuals. 'CULTURAL CLASSES' In many contemporary societies, it may be difficult to argue that access to means of production is the main criterion for the class divisions, and in this respect, Weber seems to be right contra Marx. Notably, large population segments in industrial societies are public servants or 'white-collar workers' - they are neither capitalists nor workers. Many highly salaried directors of companies, for instance, own only a negligible number of shares in the firm they run. This may necessitate a less rigid concept of class than the one Social Hierarchies 153 developed by Marx, who wrote at a time when the main twofold division of Western societies into capitalists and working class was clearer than it is today. In Wolfs analysis of Puerto Rico, it transpires that actors rank each other according to their symbols of wealth: what makes a difference in the ways people classify each other could thus be whether someone owns a bedspread, whether the women of the household ride a horse or a mule, whether they eat their bananas plain or with milk. Such markers were, in Puerto Rico at this time, closely linked with differential access to property. However, this need not be the case. As the economist Thorstein Veblen has shown (1953 [1899]), Americans may strive to acquire status symbols, as a form of impression management, to give the impression that they are better off than they actually are. Veblen calls this kind of strategy conspicuous consumption. More recently, Pierre Bourdieu has developed a systematic theory of 'cultural classes', taking France as his chief empirical example (Bourdieu 1979; see also Jenkins 1993). A principal idea in Bourdieu's work is that power is connected with symbols, and that the ruling class in any society is by default the class which decides the ranking of symbols and the form of dominant discourse; in other words, the class that controls the criteria for good taste. Someone who knows the codes for decent behaviour, 'proper' speech, good taste in art and music and so on has a surplus of symbolic capital. Bourdieu admits that such differences are often connected with economic inequality, but he has analysed them as power systems in their own right. In many societies there are people, such as politicians and intellectuals, who possess a great deal of symbolic capital and wield considerable power without owning means of production. In his meticulously researched study of 'taste' in French society, Bourdieu stresses its social origin. Contrary to popular notions to the effect that taste is somehow inborn, his observations show: that the cultural needs are created by education: our study demonstrates that all cultural practices (museum visits, attendance at concerts, exhibitions, talks, etc.) and preferences within literature, painting or music are closely connected with the level of education (which is measured as academic title or number of years at educational institutions) and social origins. (1979, p. 1) Differences in taste thereby express 'objective class differences'. For example, Bourdieu shows that knowledge of classical music is strongly correlated with education and class background, and argues that the very definition of good taste is a manifestation of power which confirms and strengthens rank differences, as well as giving a certain prestige in itself. Just as an unclean caste in India may change its way of life in a bid to improve its rank in the caste system, 'upstarts' in modern class societies may try to appropriate as many symbols as possible that indicate good taste. Bourdieu calls this kind of strategy 'conversion of capital'; it may, in other words, be possible to 154 Small Places, Large Issues Pierre Bourdieu (b. 1930) was educated as a philosopher, but has mainly worked as a sociologist and anthropologist. His ethnographic regions are Algeria and his native France, and his most important work deals with the relationship between knowledge, culture and power. In La Reproduction (with Jean-Claude Passeron), he shows how the French educational system reproduces class differences through presenting bourgeois ideology as 'natural'. In Outline of a Theory of Practice (1977 [1972]) and Le Sens pratique (1980), Bourdieu develops a general theory and method for the study of implicit, often invisible power structures immanent in culture. In Homo Academicus (1988 [1984]), he identifies such power structures in French academic life, but his main empirical study remains Distinction (1984 [1979]), which is a wide-ranging treatise on taste and power in French society: how cultural concepts of good and bad taste express, and contribute to the maintenance of, particular symbolic power relations. (See also Jenkins 1993.) convert economic capital into symbolic capital (cultural prestige). In France, for instance, aristocratic titles may sometimes be purchased. The parallel with the rise of the distillers in Bisipara should be obvious. Although there is usually a clear connection between economic and symbolic capital, the two are not congruent: some have much of the former but little of the latter, and vice versa. This is why conversion may be an interesting strategy for actors who wish to increase their prestige. Whether the chief form of conversion follows one direction or the other depends, of course, on the dominant value system in society. COMPLEXITY IN SOCIAL DIFFERENTIATION A general principle in studies of stratification, class and social differentiation, is the rule of cumulation. This 'rule' holds that if someone is economically wealthy, he or she probably also has a good education, good health and secure employment. This line of thought has been well documented, particularly in sociological studies of modern industrial societies. As anthropologists, we nevertheless need to be aware of the great variation between societies concerning criteria for rank and perceptions of rank. Although wealth nearly always provides high rank, it is not necessarily more important than, for example, ritual purity. Advanced age may give high or low rank; female gender may be completely disqualifying or nearly irrelevant, and so on. Further - as indicated earlier in this chapter - there are often contradictions between different criteria for rank, which can be interpreted as conflicts between value systems or between principles for the legitimation of power. A classic example is the conflict, prevalent in many African societies. Social Hierarchies 155 between age and education. Old men in the village try to retain their power, which is legitimated through tradition, while young men returning from college may insist that their educational, achieved qualifications are superior and entitle them to greater power than their elders. Different criteria for rank, or principles for differentiation, thus do not necessarily overlap. The social hierarchy in Bisipara, for instance, may well be conceived of as a system based on no less than five pyramids which can be distinguished analytically and which influence each other and interact to varying degrees: caste, wealth, local government, gender and age. These principles for differentiation function simultaneously, but their relative significance can rarely be predicted. Sometimes their significance is situational. This means that in some kinds of situations, such as during a religious festival, caste membership is more important than any of the other criteria. In other situations, economic power may be the most important criterion, and so on. For example, does a rich woman from a relatively impure caste rank higher or lower than a poor Brahmin? It is impossible to give this question an unambiguous answer, but it hints at the complexity of social classification and differentiation. POWER AND THE POWERLESS The last two chapters have to a great extent dealt with power and influence. Social differentiation, whether it is based on gender, age, class or caste, creates and reproduces differences in power. Often such power differences may lead to revolt and protests among the powerless, and sometimes these revolts may lead to permanent changes in the power relations of society. The French Revolution is often cited as an example of such a change: after this important event in European history, the privileges of the nobility and royal family were eventually replaced by formal principles of equality and democracy. As this chapter has suggested, there may be quite varying notions within a society about justice, good and bad and, ultimately, what the world looks like. Societies are, in other words, internally differentiated, not only in economic and political terms but also in cultural terms. Yet certain fundamental values are usually widely agreed upon, whether they are tacit or explicit. Even people who seem profoundly oppressed frequently support the dominating ideology, even if it may be said to contribute to their oppression. Any ideology attempts to make a certain perspective on society appear 'natural': if it succeeds, people will perceive their own place, and the dominant hierarchy, as natural. This was the basic mechanism Marx had in mind when he wrote that the ruling ideas of society are the ideas of the ruling class. The distinction between actor perspectives and systemic perspectives is clearly relevant when we look at inequality and differentiation, and both caste and class systems can be studied profitably through a conscious 156 Small Places, Large Issues switching between the two perspectives. One is born into a caste and/or a class; the caste or class structure is a systemic property, but each actor relates to his or her position of relative power or powerlessness in an independent, unpredictable way. It is therefore necessary to grasp the duality of social process - it is simultaneously the product of agency and the objective condition for agency - in the study of power. This is shown in the next chapter. SUGGESTIONS FOR FURTHER READING Louis Dumont: Homo Hierarchicus, 2nd edn. Chicago: University of Chicago Press 1980 [1968]. Pauline Kolenda: Caste in Modern India: Beyond Organic Solidarity. Prospect Heights, IL: Waveland 1985. Karl Marx and Friedrich Engels: The Communist Manifesto. New York: The Modern Reader 1968 [1848]. 11 POLITICS AND POWER Politics is parasitical on other social relationships. — M.J. Swartz Politics is linked with power; both power that people exert over each other, and ways in which society wields power over people by imposing institutionalised constraints on their agency - constraints ranging from property taxes to torture and genocide. However, politics also has to do with the prevention of lawlessness and insecurity; that is, it concerns law and order, the implementation of the rights of persons, conflict resolution and social integration. Politics can be identified analytically in all societies, but by no means all societies that anthropologists have studied have political institutions distinct from other societal realms. In modern state societies, it may seem fairly easy to delineate what is politics and what is not. Political science, developed to study politics in such societies, deals with the formal political institutions; with a legislative assembly, local administration, voting patterns and other aspects of society recognised as political. In non-industrial societies, it may be far more difficult to single out politics as something distinct from the ongoing flow of social life. In industrial or post-industrial society, we think of politics as something they have; a specialised set of institutions. In societies with no centralised state, the political system may rather be seen as something intimately woven into other aspects of existence. Very often, in stateless societies, kinship and religion are in practice indistinguishable from politics. That institutional differentiation which is characteristic of modern societies is absent in many others (see for instance Godelier 1975). This implies that it would often be fruitless to look for identifiable political institutions which could be compared with, say, parliaments. Instead, political anthropologists have to look for the political decision-making mechanisms - they must find out where and how the important decisions are being made, who is affected by the decisions, which rules and norms govern political action, how hegemony is challenged, and which possible sanctions the rulers of society dispose of. A central problem in classic political anthropology, which was developed in Britain from the 1940s to the 1960s, was simply the question of how stateless societies were at all integrated: why they did not just fall apart due to lack of a central authority, how they managed to resolve conflicts and 157