Companionate marriage in India: the changing marriage system in a middle-class Brahman subcaste


Department of Anthropology, London School of Economics and Political Science, Houghton Street, London WC2A 2AE, UK.;


The Eighteen-Village Vattimas are a Tamil Brahman subcaste. They were formerly rural landlords, but today they mostly belong to the urban middle class. In recent decades, the Vattimas' marriage system has changed markedly. Child marriage has ended and the age of marriage has since risen further. Close-kin marriage is no longer preferred, although subcaste endogamy remains the norm. Nevertheless, the education and employment of individuals, and their personal compatibility, have now become crucial criteria, and young men and women are involved in arranging their own marriages. Among the Vattimas (like other Indians), a form of arranged, endogamous companionate marriage has now developed, which plays a fundamental role in reproducing both caste and the middle class in contemporary India.


Les Vattimas des Dix-huit Villages forment une sous-caste de brahmanes tamouls. Ces anciens propriétaires terriens ruraux sont aujourd’hui pour la plupart intégrés à la middle class urbaine. Depuis quelques dizaines d’années, leur système matrimonial a considérablement évolué. On ne marie plus les enfants et l’âge du mariage ne cesse de reculer. Le mariage entre proches parents n’est plus privilégié, bien que l’endogamie au sein de la sous-caste reste la norme. En tout état de cause, l’éducation, l’occupation professionnelle et la compatibilité personnelle des futurs époux sont devenus des critères essentiels, et les jeunes gens des deux sexes s’impliquent dans l’arrangement de leur propre mariage. Parmi les Vattimas, comme chez d’autres Indiens, on voit apparaître une forme d’endogamie arrangée, qui joue un rôle fondamental dans la reproduction de la caste aussi bien que de la classe moyenne dans l’Inde contemporaine.

Since the nineteenth century, Brahmans in Tamilnadu, south India, have migrated in large numbers from villages to towns and cities, as they have entered new fields of modern, professional employment, from administration and law to engineering and information technology (IT). Many Brahmans have also moved from Tamilnadu to other parts of India and, in recent decades, to foreign countries. Thus by the end of the twentieth century, a traditional elite of Brahman landlords had become a predominantly urban, middle-class group with a significant presence across India and overseas (Fuller & Narasimhan 2008a).

There are approximately 1.5 million Tamil Brahmans, mostly belonging to two sections: Smartas or Aiyars and Sri Vaishnavas or Aiyangars. Aiyars are divided into four subcastes: Vadama, Brahacharanam, Ashtasahasranam, and Vattima. The Vattimas, numbering around 30,000, are the smallest subcaste and are subdivided into the ‘Eighteen-Village Vattimas’ and several smaller units. The Eighteen-Village Vattima subcaste (as we call it) is the principal ethnographic subject of this article.

Endogamous subcastes presumptively defined by a conventional number of villages (such as eighteen or twenty-four) are fairly common in Tamilnadu, although actual numbers often differ; thus in the Eighteen-Village Vattima case, sixteen villages are named by almost everyone, but the list is completed by picking from several more.1 Half-jokingly, too, Chennai (Madras), or even New Jersey, is sometimes called the nineteenth village. Tippirajapuram (TRP), a Vattima village near the temple town of Kumbakonam, was the main fieldwork site, although we have also visited other villages and interviewed Vattimas in Chennai and other Indian towns and cities, as well as in America.

Eighteen-Village Vattimas generally recognize that the close ties seen as distinctive of their small community are linked to endogamy, but some people claim that marriage outside the subcaste is now common as old traditions are abandoned. ‘It's now a computer world. Now we have love marriages’, said one Vattima woman in Chennai, even though that remark applies neither to her own married children nor to Vattimas in general, for the majority of marriages are still arranged within the subcaste. Yet significant changes have occurred during the last two or three decades. Child marriage has ended and the age of marriage has since risen further, for both males and females. Children are normally brought up in the parental home, not in their mother's brother's home, as they used to be. Most people no longer prefer close-kin marriage, and marriages to non-Vattimas, as well as love marriages, are less rare than they were. Arranged, endogamous marriage still remains the norm, both ideally and statistically. None the less, the mode of arrangement has changed, because the educational qualifications and employment of individual men and women, and their potential happiness as congenial partners, have become the principal criteria for selection. Young people are also now actively involved in arranging their own marriages. Indeed, among the Vattimas – in common with other Tamil Brahmans and other Indians – a form of arranged, endogamous companionate marriage has developed, which reflects and expresses their status as a middle-class subcaste, so that both (sub)caste and class are conjointly reproduced through the marriage system. This development, often overlooked because the contrast between arranged and love marriage is overstated (cf. Donner 2002; Mody 2002: 224-7; Pache Huber 2004: 252-9; Raj 2003: 108-9), is a significant feature of urban middle-class society in contemporary India.

Child marriage

According to Brahmanical tradition, marriage is a kanyadana, a ‘gift of a virgin’ who is ritually pure, and parents who fail to marry their daughters before the onset of polluting menstruation commit a grave sin. Among Tamil Brahmans, pre-pubertal marriage used to be the norm and most girls actually were married as young children. The marriage was consummated and a girl went to live in her husband's house after reaching puberty. For many girls, this did not happen until they were 15 or more, but some joined their husbands when only 11 or 12; their situation was often miserable, especially if they were confronted by an exploitative mother-in-law and a much older, unsympathetic husband, as Leela Gulati's grandmother was (Gulati 2005: 108-11). Particularly tragic was the position of young girls who became widows. An elderly Vattima man recalled his elder sister who was widowed when 14 and died aged 91, having spent almost all her life observing the austerities of inauspicious widowhood. ‘What did my sister do wrong? Why was she made to suffer like that?’, he asked, before castigating such ‘orthodoxy’ as ‘stupid, unwanted and unacceptable’.

Appalling suffering of this kind fuelled the campaign against child marriage in the Madras Presidency and the rest of India that eventually led to the Child Marriage Restraint Act (known as the Sarda Act), which became law in 1929 and fixed the minimum age for marriage at 14 for girls and 18 for boys. (Since 1978, the respective ages have been 18 and 21.) After the Sarda Act, child marriage declined among Tamil Brahmans, and by the late 1940s it was fairly rare. Yet child marriage continued among the Eighteen-Village Vattimas; the last occasion we know about was in 1973, when a boy of 18 married a girl of 12, but in the 1950s and 1960s it remained prevalent. Tables 1 and 2 show the age at marriage of married Vattimas living in TRP in 2006. All nine women aged 70 or over were married aged 13 or under in the 1930s or 1940s despite the Sarda Act, but ten out of nineteen of those aged 50 to 69 were equally young when married in the 1950s or 1960s. The minority of women aged 50 or over who married after puberty often did so because their parents were too poor to afford a respectable marriage. Husbands were (and are) older than their wives, but about half the senior men in TRP were also married before they reached 20.

Table 1.  Age at marriage of married Vattima women in TRP.
Age at marriageWomen by age
≤13  37919
14-19 113 5
20-41321 7
≥25  11 2
Table 2.  Age at marriage of married Vattima men in TRP.
Age at marriageMen by age
≤13    11
14-19  13711
20-24   347
≥25 245 11
Total 25111230

Everyone in TRP knew that child marriage contravened the Sarda Act, but the law was never a serious obstacle. A post-puberty marriage is not seen as a true kanyadana and hardly any Vattimas in the older generation – apart from a few people who have known child widows personally – actively condemn child marriage even now. Their attitude can be illustrated with two partly contrasting cases.

Lalitha, born in 1934, was married at 12 to Srinivasan, born in 1927, a landlord and accountant in TRP. Lalitha is her husband's MFBSD.2 She completed six years of education and started to cohabit with her husband when she was 20. Child marriage, said Lalitha, created affection and intimacy, because in the years between the wedding and consummation, partners saw each other only intermittently, so that their desire to be together grew. Women were taught about household work by their mothers-in-law and Lalitha appears to have had a good relationship with hers. Women, claimed Lalitha, also learned when young to adjust to their husbands' needs and never got angry with them, whereas among married couples today there is no adjustment and too much anger. Lalitha's own daughter, who lives in America, was also married at 12 (around 1968) and joined her husband at 20 after graduating from college. Lalitha accepts, however, that times have changed and believes that girls should now marry between 18 and 21, although her eldest granddaughter recently married a vegetarian north Indian, also resident in the US, when she was 24.

Lalitha was not the only elderly woman to reminisce happily about her own pre-pubertal marriage. None the less, more women talked about its drawbacks, and then the mother-in-law often played a crucial role. Sarada, born in 1934, was married at 7 to a boy aged 15, who grew up to be a landlord in TRP. Sarada had eight years of schooling. She began to cohabit with her husband when she was 15, but still had very little contact with him. Sarada remembers that a young wife such as herself was confined to the kitchen and the rear of the house, while her husband spent his time on the veranda at the front, but because his father normally stayed in the front room and a wife could not appear before her father-in-law, she could not walk through the house to see her husband, even when not working for her mother-in-law. Owing to such restrictions, young couples could rarely talk to each other privately at home, and several old people, both men and women, still regret this aspect of their early married life. Sarada, like other Vattima wives, was even prevented from serving food to her husband, because her mother-in-law kept this role – with all its connotations of nurturing intimacy – for herself. Furthermore, it was extremely difficult for Vattima couples to have sex, because the husband's parents, especially his mother, strictly controlled their sleeping arrangements. Hence the development of intimacy between often very youthful partners was almost impossible. Sarada recalled that Vattima couples only got together in secret for a minute or two, which is when their children were conceived; other elderly women half-jokingly reminisced about secret sexual activity in the cowshed at the back of the house. These sexual restrictions are commonly said to explain the small size of many Vattima families, although they were actually no more severe than those reported by Gough (1956: 838) for other Tamil Brahmans in nearby Kumbapettai village in the 1950s.

Under such circumstances, unless they were very fortunate like Lalitha and Srinivasan, young couples rarely developed close mutual relationships. Consequently, said Sarada (probably talking about herself, as well as other women of her generation), child brides often grew up to dislike or even hate their husbands, and vice versa, and they then had to suffer each other for a lifetime. We should make it clear that some women had good relationships with their mothers-in-law and spoke of them warmly. In the eyes of Sarada and many other older Vattima women, however, the root of their problems as young wives was a domineering mother-in-law, not the age at marriage.

Sarada, indeed, was not against child marriage, and her own eldest daughter, born in 1954, was married like her at 7 (to her MBS). Sarada's second daughter, born in 1961, was married at 13, within three weeks of menarche, and Sarada remembers her tears when this daughter's first period began before her marriage arrangements could be finalized. Her third daughter, born in 1968, married only after completing her college education; by then, it seems, Sarada accepted that times had changed and her youngest girl should not marry so early.

Ending pre-pubertal marriage was a crucial change among the Vattimas, which brought them into line with the majority of Tamil Brahmans, who had done so earlier. Young women's marriages were still fixed by their parents and senior kin, but giving up child marriage was one step towards the fuller consultation that would emerge in the 1980s and 1990s. The change also enhanced the decision-making powers of young men, who could now make a more informed choice of bride among adults, rather than immature girls. Interestingly, however, neither Sarada nor anyone else gave any specific reasons for abandoning child marriage, although several people mentioned the growing importance of girls' education. In fact, some Vattima child brides had continued their education after marriage; moreover, married female students are not uncommon in some Indian communities today. None the less, the majority of people in modern Tamilnadu have regarded child marriage as incompatible with good girls' education. Of course, this education may be only preparatory to marriage; thus Srinivas noted that colleges were ‘respectable waiting places’ for single girls, and ‘among the urban middle classes belonging to the higher castes, the possession of a basic degree is now regarded as essential for obtaining a “good” groom’ (1977: 234). In the mid-1970s, Srinivas's comments still did not fully apply to Vattimas, but by then even conservatives saw child marriage as an anachronism inconsistent with the modern, middle-class way of life – itself indexed by improved female education – that was developing among them. The informant who attributed the demise of child marriage to the ‘force of time’ (kalattin kattayam), a feeling that the custom's time was over, was probably right.

Post-marital residence and the mother's brother

In the past, even after starting to cohabit with her husband, an Eighteen-Village Vattima wife still lived in her parents' home, normally also her brother's home, for as much as six months a year or even more after bearing a child. This unusual residence pattern meant that a woman's ties with her parents and brothers were hardly weakened after marriage, unlike other Tamil Brahman wives, who were always more fully absorbed into their husbands' families. Moreover, in sharp contrast with other Tamil Brahmans, Vattima children normally used to be mainly brought up in their mothers' brothers' homes by their mothers and maternal uncles, not their fathers. Thus adult men – as if they belonged to a matrilineal kinship system – had a closer relationship with their sisters' children than their own. Crucial decisions on a child's behalf – for example, about marriage or a boy's education – were primarily an uncle's responsibility. Decisions about where to raise a boy were often influenced in practice by the location of a good school. Otherwise, unless they had no maternal uncle, the majority of Vattimas in TRP aged 60 or over spent their childhood in their mothers' brothers' homes. A generation or two ago, however, the system changed, so that most adults alive today have brought their children up in their own conjugal homes. Yet many older people, who remember their childhood fondly, regretfully say that when children are raised only in their parents' home, it promotes selfishness and disrespect. Several men specifically told us, too, how much of their success in life they owed to their maternal uncles, rather than their fathers.

The change in post-marital residence – like the end of child marriage – was not explained very precisely by our informants, although urban migration has certainly been significant. In Chennai and other urban centres, Vattima couples have always normally lived with their children in nuclear families, sometimes supplemented by one or more old parents. These nuclear family households then progressively became the standard pattern, even in villages, although spending school summer holidays with a maternal uncle in a village remained a common custom until very recently.

A Vattima woman staying in her parental home was obviously apart from her husband and his mother a lot, so that she could escape the latter's control. But when a woman was in her marital home, her mother-in-law often ruthlessly exploited her in the kitchen, and many older Vattima women still complain about their mistreatment as young wives. A daughter-in-law, commented one elderly woman, was supposed to work from early morning into the night. Another said that she was ‘tortured’ by her in-laws, and yet another remembered that many young wives cried when they went home, asking their parents why they had been ‘pushed into a well’ or ‘drowned in a river’. Almost always, though, they were told that mistreatment was normal and they should put up with it in silence. Very few Vattima men admitted to us that their wives suffered, but one who did remembered that his mother would keep spare food for his sister, not his wife, and he acknowledged the Tamil proverbial truth that a daughter is precious like butter and a daughter-in-law is despised like lime. Gough's data (1956: 837-44) suggest that Vattima women's relationships with their mothers-in-law were not exceptional among Tamil Brahmans, however, so that the impact of the unusual residence pattern is unclear. Today, though, partly because Vattimas usually live in nuclear families and partly because new wives are older and more assertive, women are much less likely to be exploited by mothers-in-law.

Preferential close-kin marriage

The Eighteen-Village Vattimas – in common with most south Indians – formerly had a preference for marriages between cross-cousins or between uncle and niece (elder sister's daughter marriage). Our quantitative evidence derives from genealogies collected in TRP (mostly five or six generations deep), which include numerous urban residents as well as overseas emigrants related to people in the village. The genealogies were not collected from every Vattima in TRP, and some are fuller than others. None the less, the data summarized in Table 3 are revealing and show that out of 389 marriages, 94.6 per cent were endogamous unions between Eighteen-Village Vattimas (though a handful involve Vattimas of other subcastes), whereas only 5.4 per cent included a non-Vattima partner. Marriages between genealogically traceable consanguines represented 10.0 per cent of the total. Several more united partners identified as affines, some of whom probably were consanguines, although the links have been forgotten.

Table 3.  Frequency of endogamous, close-kin, and exogamous marriages among TRP Vattimas and their relatives.
Marital partnerNumberPercentage
Vattima (endogamous)36894.6
FZD (first cousin)133.3
FZD (classificatory)51.3
MBD (first cousin)102.6
MBD (classificatory)61.5
‘Wrong’ (FMBD)10.3
All close kin3910.0
Non-Vattima (exogamous)215.4

For marriages uniting immediate plus more distant (classificatory) kin, the respective rates are 4.6 per cent for men marrying a FZD, 4.1 per cent a MBD, and 1.1 per cent an eZD. (In one ‘wrong’ marriage, a man married his FMBD, terminologically his MZ.) These percentages are lower than those reported for other groups, Brahman or non-Brahman, in Tamilnadu (Beck 1972: 253; Good 1991: 88-9; Gough 1956: 844; Kapadia 1995: 62; Trautmann 1981: 218). One reason must be that preferential close-kin marriage has declined throughout south India in recent decades, so that our data show a lower rate simply because they include so many recent marriages. The last close-kin marriage in our genealogies occurred in 2002; we know of only two others since 1990.

Even today, some people, especially among the elderly, believe that close-kin marriage is desirable, but the majority now positively prefer marriages between partners who are unrelated or only distantly related, for much the same reasons as other south Indians. Concern about ‘inbreeding’– specifically, fear that the offspring of closely related parents may suffer congenital diseases or disabilities – is often mentioned, most forcefully by those who married close kin and have children with serious health problems. But there are also other factors. Some older women say that a wife was treated better if her mother-in-law was a close relative, but other younger women argue that an unrelated mother-in-law is preferable, because a wife is then under less pressure from her own family to tolerate any abuse. Another reason is that close-kin marriage often reinforced alliances between landowning families and prevented property from devolving outside them, but this no longer matters to the many urban Vattimas who have sold their land. Most commonly, though, as we discuss below, people insist that education and employment are now the key factors in marriage arrangement, and that partners should be suited to each other. In fact, a cousin may be qualified on all these counts, as was so for the marriage in 2002 when Shankara, an IT consultant, married Gauri, his FMZDD (FZD) from TRP, a mathematics graduate who was keen to go with him to America. Hence the growing salience of education, employment, and congeniality does not logically preclude preferential close-kin marriage. None the less, most south Indians, including Vattimas, dismiss cases like Shankara and Gauri's as exceptional, and typically insist that the old criteria must be displaced by the new ones in marriage arrangement (cf. Fuller 2003: 73-4; Kapadia 1995: 57-67; Mines 1994: 121).

Subcaste endogamy

Table 3 shows that only twenty-one marriages out of 389 (5.4 per cent) include a non-Vattima partner. Fifteen partners are other Aiyars and the rest include one Aiyangar, one Karnataka Brahman, one Tamil non-Brahman, one Punjabi, and two unidentified ‘non-Vattimas’. We also know of some Eighteen-Village Vattimas married into other Vattima subcastes, traditionally seen as lower status. The first exogamous marriage probably occurred in the early 1980s. The others have taken place sporadically over the last twenty-five years, and nearly all the couples live in Chennai, cities elsewhere in India, or overseas. Our genealogies may omit some earlier marriages to non-Vattimas, but they certainly show that exogamous unions were and still are unusual.

In Mumbai (Bombay), where a sizeable Eighteen-Village Vattima community has existed since the 1940s, informants suggested that the proportion of marriages to other Brahmans is higher than in Tamilnadu, and the same may apply in Bangalore. On the other hand, most Vattimas in Mumbai and Bangalore on whom we have reliable data have married within the subcaste, often to people from the villages, or Chennai or other Tamil towns. Moreover, the most striking feature of all our conversations with Vattimas – in villages, Indian cities, or America – is their pervasive assumption that they can rely on their now global network to find suitable partners within the subcaste.

Many anthropologists have reported the widening of endogamous boundaries in modern India as subcastes merge or are ‘telescoped to form a single entity’ (Srinivas 1977: 233). Srinivas gave as an example educated, urban, middle-class south Indian Brahmans, who are satisfied if their children find partners from any south Indian Brahman group. Most Eighteen-Village Vattimas are now prepared to accept other Tamil Brahmans, but normally only if a suitable partner from their own subcaste cannot be found. Thus the Vattimas provide a clear, if unusual, example of how subcaste endogamy can remain the dominant norm, in both preference and practice, even for a group with so many members now living in cities or overseas.

The only significant exception concerns Vattimas who have been raised since childhood in the US, as opposed to those who went as adult migrants. As yet, very few of the former are old enough to be married, but those who are have spouses from various backgrounds. Furthermore, all the Vattimas we met in America pragmatically accepted that their children would eventually make their own decisions unconstrained by traditional rules, even though parents firmly hope that they will at least marry Indians who share the same cultural background (cf. Bhatia 2007: 125; Khandelwal 2002: 151-6).3 Almost certainly, though, subcaste endogamy will break down in the US because Vattima second-generation Indian Americans will have a diverse range of partners.

Ancestral status and family background

In the past, ancestral status, which is connected with village social structure, was a critical factor in marriage arrangement among the Eighteen-Village Vattimas. Several villages are regarded as the most important owing to the size of their Vattima population and the landed wealth of their leading families. Konerirajapuram – with the most Vattima households, many wealthy landlords, and a famous, ancient Shiva temple – is the premier village, but the richest has been Mudikondan, whose grandest landlords built mansions on ‘London Street’, so named for its former opulence. In addition to these two, Anandatandavapuram, Sengalipuram, TRP, and sometimes Vishnupuram also count as important villages.

In TRP, there are six ‘big families’ (periya kudumbam), each with its ancestral house; one was the first Vattima family to settle in the village, but the others have also been there for a long time. Each of the six families used to have large landholdings, although a couple of them sustained considerable losses a generation or two ago owing to their heads' extravagant lifestyle. Even today, though, all these old landed families retain much of their ancestral status and respect for them has not faded away. Villagers know, for example, that the rich landlord shown in Figure 1 as the apical ancestor of ‘First’ family built TRP's Vishnu temple, whose hereditary trustee comes from ‘Third’ family. Subject to the rules of patriclan (gotra) exogamy, members of the big families mostly married each other or partners from comparable landed families in other important Vattima villages until two or three decades ago; each of these alliances, of course, reconfirmed the leading families' ancestral status. Outside TRP, our information about big families is inadequate, but a similar gap between them and poorer, less prestigious Vattima families existed in all the important villages, and possibly in some other villages as well. Overall, as far as we can tell, wealth and ancestral status were generally matched, so that the pattern of marriage alliance reflected socio-economic stratification within the subcaste.

Figure 1.

Tippirajapuram ‘big families’ marriages.

The partial genealogy in Figure 1 depicts some of the marital links between four of TRP's big families, as well as with a leading landlord family from Sengalipuram in ‘First’ family and a family descended from a famous, wealthy, nineteenth-century administrator in ‘Second’ family. When the marriages of today's senior generation were being fixed in the 1960s and 1970s, some men were solely landlords, but others had urban professional jobs as well, so that new criteria pertaining to education and employment were becoming relevant even for alliances between leading families. Since then, education and employment have become more significant, although ancestral status can still carry weight. Thus, for example, two high-flying young men from ‘Second’ and ‘Third’ families – an IT consultant in America and a bank executive in India – who are respectively the sons of an industrialist (married to the administrator's great-great-granddaughter) and a bank executive, married the daughters of a wealthy Vattima entrepreneur. He does not belong to an elite landed family, and the superior ancestral status of his sons-in-law's families was plainly important to him (and was seen to be so by other Vattimas).

In other recent marriages, however, ancestral status has sometimes mattered less than education and employment, so that the range of partners has widened. For example, two IT consultants in America, respectively the sons of a wealthy Mudikondan landlord and a smaller landlord in a lesser Vattima village, have married the daughters of a retired schoolteacher in TRP, who does not belong to a big family. (The younger daughter Gauri's marriage to Shankara was mentioned above.) The Mudikondan landlord's eldest son, a businessman in Chennai, has a wife from a well-off family in a small Vattima village, and his youngest son, a doctor in Chennai, married Amrita, the schoolteacher daughter of a very successful businessman in Mumbai, who talked about a ‘secure’ marriage for her. In that respect, the Mudikondan family's ancestral status probably counted for a lot, although Amrita said that her marriage was initially planned by her father's father, a big landlord in Sengalipuram, who wanted an alliance with his Mudikondan counterpart. After the horoscopes had been matched, Amrita went with her younger sister to meet her prospective husband, then a junior doctor in Chennai. He was planning to practise in Kumbakonam town, however, which worried Amrita, and her sister told her to consent to the marriage only if she really liked him, not because their grandfather wanted it. But after they talked together, Amrita was sure she wanted to marry the doctor and later he decided to practise in Chennai anyway, which Amrita prefers over Kumbakonam.

The case of Bhavani shows another combination of features. Bhavani, a young woman from another of TRP's big families, is an only child, reputed for her beauty and impeccable manners. She has an engineering degree and works for a leading finance company in Chennai. But some of Bhavani's family members are chronically ill. An initial proposal came from a man whom Bhavani rejected because of worries about his job, and his family then started to gossip about her and her ‘mad’ family. The gossip spread in TRP and then the young bank executive mentioned above intervened; he called for an end to the slurs and announced that a good husband would be found for Bhavani. In 2007, she became engaged to a Vattima businessman's son working in IT in the US, with whom she communicated by email and telephone before agreeing to marry. This case particularly shows that unfortunate family circumstances are potentially damaging even for someone with many personal assets like Bhavani; it also shows that when one big family's head publicly supported her, this not only helped to find her a suitable boy, but also reasserted these two families' pre-eminence in TRP.

These (and other) cases demonstrate that ancestral status often, but not always, still counts in marriage arrangements – as it also does in so many communities throughout India. Irrespective of ancestry, though, among all Vattimas – like most Indians – there is always a strong preference for alliances with families that are broadly similar in general family background and reputation.

Education and employment

Increasingly vital for marriage today are the qualifications and qualities of individuals. Vattimas particularly emphasize education and employment – much like the middle-class Maheshwaris in Rajasthan described in detail by Pache Huber (2004: 172-93), as well as millions of other Indians – so that every Vattima informant, when discussing grown-up children or grandchildren and their spouses or marriage prospects, invariably itemizes each individual's qualifications and career. People just take it for granted that any sensible conversation about marriage must also be about education and employment.

For men, who normally marry in their mid-to-late twenties or early thirties nowadays, education is assessed in relation to current or prospective employment, and these are by far the most important criteria. The prestige rankings for Vattimas are standard ones: degrees from elite colleges, such as the Indian Institutes of Technology or Management (IIT, IIM), are highly esteemed, and IT professionals and business consultants are probably thought to have the best jobs, though not all equally, for men working in top companies or with jobs overseas, especially in the US, tend to be preferred. Similar considerations apply to engineers, doctors, lawyers, bankers, accountants, businessmen, company managers, or other professionals. Even prosperous farmers and landlords are now generally seen as unattractive husbands, and many young women flatly refuse to consider them.

For women, who often married in their mid-teens a generation ago but now normally do so in their early twenties, the prestige rankings are almost the same. Educated, professional husbands prefer wives who are also educated – though normally not to a higher standard than themselves – and women with little schooling are very unlikely to find a well-educated husband. Educated wives are expected to be more congenial partners for educated men, and vice versa, but they may also work remuneratively, at least until they have children, and they can then undertake the vital task of assisting with their children's education. On the other hand, many men prefer ‘home-loving’ wives and do not want them to work outside the home, even if they are graduates.

Yet, growing numbers of women who have graduated from engineering and other professional colleges, or even IITs, have career ambitions, so that they refuse to marry men who will prevent them from studying for advanced qualifications or going to work. Sarada's first granddaughter, for example, has an engineering degree and agreed to marry only if she could finish her master's degree in the US, which she did in the south while her less than enthusiastic husband was on the west coast, where she now lives with him and has a job. Some girls, said Sarada, still listen to their elders' advice, but many others insist on delaying marriage until they have jobs. Others delay for different reasons. One Vattima woman was amazed that her recently married niece had previously rejected numerous well-qualified suitors from America owing to their ‘lack of maturity’. It was all very different in her young days, she expostulated, and neither she nor her sister knew what modern girls meant by ‘maturity’. Thus demanding, single women may clash with unsympathetic parents or other senior kin, but today they can rarely be coerced into unwanted marriages – a crucial development that needs further discussion.

Congenial couples and happy marriages

When child marriage was the norm, the couple-to-be were obviously not consulted at all. At first, however, even after child marriage had stopped, parents or other senior kin were firmly in charge. Sons had a limited power of veto over prospective partners, but little active role in decision-making, and daughters had no say at all; neither party normally met the other before the wedding. Among the Eighteen-Village Vattimas, this was the normal situation until the 1980s, and some people – especially women – recall the old days with horror. For instance, one woman in her forties lamented the fate of a childhood friend from one of TRP's big families, who was in love with a non-Vattima Brahman boy in the village. She would not have been allowed to marry him, but an acceptable husband could have been found. Instead, her father's father insisted on marrying her to a cousin, who was allegedly a promiscuous drunkard, and she then threatened to commit suicide. The girl's friends consoled her and convinced her not to kill herself, but after an always unhappy marriage she died in childbirth; only one son survived from her five pregnancies and to her friend she compared herself to a street dog endlessly laying pups. The girl's mother sadly admitted, ‘My daughter's life was spoilt by us’, because they (the girl's parents) were unable to stop a domineering grandfather from forcing through a marriage seen as consistent with the family's ancestral status.

Such a marriage would not occur today. Usually, when a marriage is in the offing, possible partners are suggested by relatives, friends, or other contacts, and then their photographs, horoscopes, or both are exchanged by the respective families. A Vattima banker, using banking jargon, said that ‘due diligence’ is hardly needed within their small community, for subcaste endogamy greatly reduces the risk of selecting an unknown, unsuitable partner. If a suitable match looks likely, the two families collect more information and start to discuss a possible marriage. The young couple are encouraged to contact each other by telephone, letter, email, or in person, so that the traditional Hindu ideal that ‘the girl and boy are strangers to each other’ (Mody 2002: 225) has been reversed. If the couple, and their parents and other close senior kin, are all in favour, an agreement to marry – an ‘engagement’ (using the English word) – is announced; the engagement ceremony, a recent development resembling a wedding reception, is distinct from the traditional ‘betrothal’ (niccayatarttam) ritual held on the same day. Between the engagement and wedding, the couple try to get to know each other better and meet if they can. Nowadays, unlike earlier, a young woman and her family can stop negotiations before the engagement without any serious repercussions for their reputation; girls are also generally thought to be choosier than boys, so that people believe that it is harder to find a bride than a groom. Although some older people disapprove of the new modes of consultation, forcing a marriage through against a young person's objections (as in the unhappy case mentioned above) is no longer possible, and parents and grandparents know it.

It is, of course, necessary to agree the marriage gifts and payments. The cycle of prestations made at marriage and other rites of passage is too complex to be discussed here. Briefly, though, a Vattima bride's parents give her a dowry (chidanam; stridhana) of gold, silver, and jewellery (plus household goods and other items) that is relatively more generous than among most Tamil Brahmans. Normally, they also give the groom some jewellery and cash. The bride's family pay most of the wedding expenses, whereas in the past the groom's family made a greater contribution.

Partly because some children, especially daughters, are so choosy, their parents may agree to a non-Vattima with all the right personal attributes; almost always, though, this partner must be a Tamil Brahman. Too much choosiness brings its own problems, however, and if a daughter remains single into her late twenties, the search for a groom may become urgent or even desperate, so that by default a non-Vattima husband becomes progressively more acceptable. Compared with a daughter, delay in a son's marriage is regarded less seriously, but a man still single in his mid-thirties may also have trouble finding a wife and is therefore more likely to marry a non-Vattima. Concerned about this matter, some Vattima parents strive to marry sons in their mid-twenties, and the average age of grooms actually appears to have fallen in recent years.

In 2005, Uma was preoccupied with the marriages of her two eldest granddaughters, both in their mid-twenties. The first had recently married a Vadama Aiyar, a doubly prestigious graduate from both an IIT and an IIM; because she is fair and pretty, said her grandmother, she found a good husband easily. The young couple now live in New Jersey. The second granddaughter graduated from a leading engineering college and works for a major IT company in Chennai. She had announced that she wanted at least an IIT graduate, not just an ordinary engineer. No suitable groom had been found, however, which was blamed on the girl's dark skin by Uma, who said that she would willingly consider another non-Vattima Brahman because ‘only education matters’. Uma was exaggerating, but her assertion highlights the importance of education as a marital criterion that sanctioned her fair granddaughter's marriage to a non-Vattima and would probably do so again when her demanding but dark granddaughter eventually found a husband. Our informants disagreed, however, about whether finding suitable partners within the subcaste – especially for exacting young women – has become harder. Some said that it has, so that marriages to non-Vattimas are becoming more frequent, whereas others claimed that because educational standards have uniformly improved, the ratio of well-matched boys to girls was more even now than a decade ago.

Not only do parents today recognize that they cannot force their children to marry, they mostly accept that listening to them carefully is wise, because divorce, albeit very rare, has become a potential risk. (The first Vattima divorce occurred on grounds of personal incompatibility in the 1980s; both divorcees remarried.) More positively, though, Vattimas, like other Tamil Brahmans and many other Indians, now regard the future couple's happiness as vitally important. In the past, the personalities of prospective partners scarcely mattered, because they were thought to have no real bearing on a marriage's success – though in retrospect, of course, people often realized that they did. Actually, congeniality is salient not only for the partners-to-be, but also for their parents, especially the groom's mother, whose assessment of potential brides carries special weight. Young women, conversely, sometimes express more concern about future mothers-in-law than husbands, who can cause less trouble. Today, most Vattima couples live separately as nuclear families, but very few women ever regard the relationship with a mother-in-law as unimportant, however far away she is.

The concern with future happiness is the primary motive for consulting young men and women, and encouraging them to communicate with each other before a marriage is finalized. Of course, they must still judge each other quite hastily, and men tend to place a lot of weight on physical attractiveness, so that fair-skinned, beautiful girls are most likely to impress potential husbands. None the less, even if the mutual evaluation is quick or done from a distance – as in Amrita's and Bhavani's matches – it is taken seriously and is designed to ensure that individual young men and women are compatible and will have happy, harmonious marriages. Nobody expects the couple to fall in love before the wedding, which would anyway imply premature intimacy, but people hope they will do in time, for ‘love between husband and wife is expected to grow as the relationship develops’ (Mody 2002: 225). Yet none of this means that the role of senior kin has been significantly reduced. Most young people accept that their elders have more relevant knowledge and experience, and many of them assume, with good reason, that their parents can probably find a better partner than they could do by themselves. Sometimes, indeed, probably more by sons than daughters, selecting a partner is mostly left to parents, particularly when the son is overseas.

Shankara and Gauri were mentioned above. When Shankara's marriage was being planned, several proposals emerged; as his father in Kumbakonam said, ‘Once your son is in the US, people start pestering you’. Shankara, he continued, was quite young (about 25), but ‘I decided to put an end to this pestering’. He chose Gauri, then 20, because he knew her through her mother in TRP, a cousin whom he had known since childhood and liked a lot. We do not know how much Shankara and Gauri were in touch with each other before marrying, but understand that he readily accepted her from America and that she was keen to marry him and emigrate.

Shankara's cousin, Arjun, is an IT and business consultant living in America with his wife Leela, a financial analyst. Arjun's parents live in Chennai; his father is a retired engineer. After graduating from IIT-Bombay, Arjun moved to America in 1996 and married Leela in 2002, when he was about 25 and she was a little younger. Leela has a degree in commerce and qualified as a chartered accountant in the US; her father is a professor in Trivandrum (Kerala's capital city) and her mother is a friend and relative of Arjun's mother. Four proposals came for Arjun, but his father – disliking ‘auctions’– considered each one carefully in turn. Arjun's parents met Leela several times and his father told us they decided that she was the right girl – a Vattima raised in a city who was ‘dashing’ and self-reliant, so that she would cope with starting her married life abroad. (Arjun's sister married a non-Vattima Brahman and his father is not particularly attached to Vattima tradition.) Arjun and Leela may have communicated with each other more than his father realized, but they did not see each other before the wedding. The point to emphasize here, however, is that Arjun wanted his parents to conduct all the important discussions and, like Shankara, he relied on them to select a wife with whom he would be happy.

‘Love marriages’, in which young people choose their own partners, are rare among the Vattimas, and we have no details about any cases. In our genealogies, a handful of marriages were described as love marriages, such as one between a Vattima woman and a Vadama Aiyar, both IT professionals in Chennai; this marriage, like one or two others to Brahmans, was seen as acceptable. The remainder were love marriages to non-Brahmans that are widely disapproved of, such as the Vattima woman in Chennai married to a Muslim (a particularly disreputable choice for many Brahmans) or the various Vattimas married to different kinds of Indians (including one Tamil non-Brahman) or white people in the US. In relation to Vattimas overseas, a ‘don't-ask-don't-tell’ mentality or even outright denial is not uncommon, so that parents in India may carry on searching for a bride for a son who already has an American partner, or they may claim that a daughter's white husband is a fair-skinned north Indian. The prevailing attitude towards love marriage means that some people try to keep quiet about it; on the other hand, as it is a popular topic for gossip, it is unlikely that we overlooked many cases.

Marriage among the middle class in India

Before discussing companionate marriage specifically, we must briefly review the literature on middle-class marriage in India. Sociologists of Western society distinguish between the ‘old’ middle class, mainly small businessmen and property owners, and the very diverse ‘new’, ‘white-collar’ middle class, in which the most significant division is between educated, qualified, professionals and the rest (Giddens 1973: 177-9, 186-7; Mills 2002 [1951]: 3-6, 63-76). Misra defined the small, pre-Independence middle class in India as broadly equivalent to the combined old and new middle classes (1961: 12-13), but most social scientists define the predominantly urban middle class as more or less equivalent to the white-collar, new middle class in the West (Béteille 2003: 76-80). This middle class expanded steadily after Independence and has grown rapidly since economic liberalization began in the early 1990s. The post-liberalization middle class is often, but not always, described as ‘new’ (e.g. Das 2002: 280-90; Fernandes 2006), and one of its important features is the sizeable ‘new-rich’ stratum at the top, which includes entrepreneurs, business executives, and skilled technical and professional staff in booming economic sectors such as IT and financial services (Fuller & Narasimhan 2007: 122, 125-6).

Tamil Brahmans were disproportionately well represented in the small middle class of colonial Madras and, though no statistics exist to confirm it, they almost certainly still are in the much larger urban middle class, including its new-rich segment, in contemporary Tamilnadu. More salient for our discussion of marriage, however, is that Tamil Brahmans today, including Eighteen-Village Vattimas, tend to perceive themselves mainly as an urban middle-class caste. Rural landlords are a diminishing minority, and although there are a lot of poor Brahmans, many of them are also white-collar workers. Moreover, even though the majority of middle-class people in Tamilnadu are non-Brahmans, they still tend to be attracted to Brahmanical cultural and religious values, which continue to shape middle-class values decisively, so that ‘Sanskritization’ often accompanies upward class mobility (cf. Fuller 1999: 36-7). All this means that the ‘class awareness’ (Giddens 1973: 111) of middle-class Tamil Brahmans or Tamil Brahman middle-class people – whichever is the better phrase – is persistently equated with the values, attitudes, and lifestyle that constitute their caste status or, more accurately, are deemed to do so. Throughout India, the urban middle class is disproportionately drawn from the upper castes (Deshpande 2003: 116-20, 146; Fernandes 2006: 104-6), but, for reasons that we cannot further discuss here, the equivalence between class and caste among Tamil Brahmans is unusually complete.

Béteille examines the family's role in reproducing social inequality in modern India (1991) and caste's changing meanings among the urban middle class (1996). On marriage, he comments, attitudes are ‘markedly conservative’ (1991: 15, 17), endogamy is still normal, and caste is rarely ignored, although endogamous boundaries have softened and sanctions against inter-caste unions have weakened (1996: 162-6). Given the importance of marriage in reproducing caste and class inequality, however, Béteille glides over the topic lightly (cf. 2003: 81-2).

Admittedly, the ethnographic evidence on middle-class marriage is fairly thin, but it is mostly clear about arrangement and endogamy. Vatuk's ethnography of predominantly high-caste, white-collar migrants in Meerut, Uttar Pradesh, in the 1960s – the first detailed study of urban middle-class kinship and marriage – confirms that almost all marriages were arranged and conformed to caste endogamy rules, although inter-subcaste unions had become more common (1972: chap. 4). The same was largely true among the mainly high-caste, middle-class people in Bhubaneswar, Orissa, studied between 1965 and 1987 by Seymour, though subcastes are not discussed (1999: 109-10, 195-203, 211-17); a large majority of middle-class women studied by Caplan in Chennai in the 1970s also had arranged marriages (1985: 42), which were by implication endogamous. Nishimura studied Nagarattars (Nattukottai Chettiars) in Tamilnadu in 1991, many of whom belong to the urban middle class; she confirms that Nagarattars are ‘extremely strict in maintaining their endogamous principles’ (1998: 74) and inter-caste marriages are exceptional (1998: 54). Among Beeri Chettiars in Chennai in the 1980s, ‘the vast majority of marriages’ were not only arranged within the caste, but also within each endogamous subcaste (gumbu) (Mines 1994: 121-2). Caste endogamy strictly prevails, too, among contemporary middle-class Tamil Gounder entrepreneurs (De Neve n.d.).

Pache Huber's detailed ethnography of marriage among middle-class Maheshwaris between 1993 and 2002 shows that the majority of their unions were arranged within the caste. The Maheshwaris' ‘endogamous zone’ has expanded as marriages across some subcaste boundaries and into other, similar merchant castes have become more acceptable, but that has not undermined their clear preference for marrying within their own caste (Pache Huber 2004: 86-9, 97-8, 265-6). Among IT professionals in Chennai, we found that nearly all marriages were arranged within the caste (Fuller & Narasimhan 2008b: 201-2). In Bangalore, Baas (2007) does report a trend towards ‘arranged love marriages’ in the IT industry that ignore caste. On the other hand, Upadhya and Vasavi's extensive research found ‘a pattern of social conservatism ... with regard to family and marriage’ and, despite a large minority of love marriages, there was no ‘weakening of caste endogamy’ even in this ‘cosmopolitan and mobile professional group’ (2006: 111-12).

The Izhavas in Kerala represent a counter-example, because aspiring ‘modernists’ in this traditionally low-ranking caste sometimes approve of love marriages as a route to upward mobility (Osella & Osella 2000: 81-2, 85-7). Yet in all the ethnographic literature, Donner's study (2002) of middle-class people in Calcutta in the 1990s is the sole well-documented case of relatively frequent love marriages –‘one's own marriage’ as opposed to an arranged marriage – although only some unions were inter-caste. Hence the overall picture is clear. For the majority of middle-class Indians, arranged endogamous marriage remains the norm, in both preference and practice. In most groups, inter-subcaste unions have become more acceptable and common, although in someof them – including the Eighteen-Village Vattimas – a strong preference for subcaste endogamy persists, notwithstanding the development of a companionate form of marriage.

The development of companionate marriage

Competent academic discussion of arranged and love marriage recognizes that the dichotomy is fluid or fuzzy. Particularly in the urban middle class, however, and not only among the Vattimas, the limited evidence suggests that a couple's prospective personal happiness has now become as important in arranged marriage as it is in love marriage. Although several anthropologists refer to this change (Donner 2002: 84-8; Mody 2002: 226-7; Nishimura 1998: 53; Seymour 1999: 211-14; Trautmann 2003: 1125; Uberoi 2006: 25, 32) and Pache Huber (2004: chap. 11) describes it in detail, they still tend to reproduce their informants' preoccupation with the opposition between ‘modern’ love and ‘traditional’ arranged marriage. In reality, though, whether or not parents take part in choosing their children's partners, middle-class marriage has progressively developed in recent years into a companionate form, ‘a bond between two intimate selves’, as Parry describes one such union (2001: 816).

In eighteenth-century England, according to Stone, especially among the middle and upper classes, there was a clear trend towards companionate marriage. Romantic love became an increasingly respectable motive for marriage (Stone 1990 [1977]: 190) and people ‘began to put the prospects of emotional satisfaction before the ambition for increased income or status’ (1990 [1977]: 217). Companionate marriage was linked to improved female education, both because educated women were likely to be better companions for educated, middle-class men, and because they demanded a greater say in selecting their husband and running their family (1990 [1977]: 228, 233). Copious research on the history of love, marriage, gender, and the family in England, Europe, and America has now shown that Stone's evolutionary schema is too monolithic. Yet it still provides a comparative perspective that helps us to identify a significant change among the contemporary Indian middle class.

In the modern West, companionate marriage is closely linked to individual freedom of choice, and Stone assumes that it must be so. He postulates ‘four basic options’ in matchmaking: first, parents and kin select partners without consulting their children; second, they select but allow children a veto power; third, a system ‘made necessary by the rise of individualism’, children choose but parents retain a veto; fourth, children choose and merely inform their parents (1990 [1977]: 181-2). In modern India, however, another option has emerged: parents and children together select the partners, motivated by an ideal of companionate ‘emotional satisfaction’ that is not premised on young people's unfettered personal choices. This Indian option partially reflects a rise in ‘affective individualism’ (1990 [1977]: 151) linked with relatively more concern than in the past with individual personality, self-expression, and free will, which is in turn connected with improved education (especially for girls), the rising age of marriage, and reduced gender inequality. Plainly, though, affective individualism in India is less radical than its modern Western counterpart, which rests on ‘possessive individualism’ and presupposes that ‘[w]hat makes a man [or woman] human is freedom from dependence on the wills of others’; this correspondingly means that individual freedom is compromised by any relationship with others (including parents) except one that ‘the individual enters voluntarily with a view to his [or her] own interest’ (Macpherson 1964: 263). In this respect, incidentally, a revealing contrast exists between the Indian case and the Turkish one discussed by Hart (2007), in which the distinction between love and arranged marriage is also often blurred in practice, although love marriage is the modern ideal, children are expected to find their own partners, and people insist that love-matches are universal. In the Indian middle class, however, a form of companionate marriage created through arrangement has become the modern ideal, whereas love marriage definitely has not, except in a small ‘progressive’ segment (Uberoi 2006: 29).

In the eyes of almost all Eighteen-Village Vattimas, endogamy contributes positively to marital companionship, because partners sharing the same subcaste culture are most likely to form a strong bond. Other Tamil Brahmans, whose culture is similar, may be suitable partners if no Vattima can be found. The same logic about caste culture and companionship prevails among the Maheshwaris (Pache Huber 2004: 98, 266), for example, and is undoubtedly very widespread in India. Yet the Indian logic is not peculiar: Westerners, for instance, also generally assume that congenial marriage partners will have a lot in common. The point here is simple but important: (sub)caste endogamy is functionally consistent with companionate marriage and is not a survival from an earlier epoch before an Indian middle class and its new forms of marriage arrangement emerged.

At the same time, endogamous companionate marriage, in which partners are primarily selected with reference to their own educational qualifications and employment, and their potential happiness as a compatible couple, is a system that reproduces both caste and class, specifically ‘middle classness’ as social practice and cultural discourse. The Eighteen-Village Vattimas are an exceptional social group, but to look at arranged endogamous marriage in relation to caste or subcaste alone is a general mistake, because companionate marriage as it now exists also plays a fundamental role in the emergence and reproduction of the middle class, which remains a predominantly upper-caste section of society, even as it rapidly expands in twenty-first-century India.


  • Research was carried out among Eighteen-Village Vattima Brahmans in Tippirajapuram from September 2005 to March 2006, in the US in September 2006, and in Chennai and other Indian cities in January-April 2007, August 2007, and January-March 2008; Haripriya Narasimhan did most of the research, although Chris Fuller accompanied her for part of the time. This article was written by Fuller, although we have discussed it together extensively and it represents our joint views. We thank the Economic and Social Research Council, which has supported all the research. We are grateful to Marie-Louise Reiniche, who first told us about the Vattimas and gave us material that she had collected. We also thank Geert De Neve, Henrike Donner, Filippo Osella, and participants in a workshop at the University of Sussex for useful comments on earlier versions of this article.

  • 1

    The sixteen villages are Anandatandavapuram, Arasavanangadu, Konerirajapuram, Kundalur, Mandai, Mappadugai, Maratturai, Molaiyur, Mudikondan, Palur, Semmangudi, Sengalipuram, Tattattimulai, Tediyur, Tippirajapuram, and Vishnupuram. In addition, Anaikuppam, Puliyur, Sittanvalur, and Tuttukudi are listed quite frequently, but some others only occasionally.

  • 2

    Standard kinship abbreviations are used throughout: B = brother, D = daughter, F = father, M = mother, S = son, Z = sister, and e = elder. All individuals' names are pseudonyms.

  • 3

    Like other middle-class Indians in the US, Vattimas reside in suburban areas, mostly the outer suburbs of large cities throughout the country. Those we met live in: Chicago; Dallas; Edison, New Jersey; Milwaukee; Philadelphia; Pittsburgh; and Syracuse, New York.

C.J. Fuller is Professor of Anthropology at the London School of Economics and Political Science. His most extensive research has been in the Minakshi temple in Madurai, Tamilnadu. His books include The renewal of the priesthood: modernity and traditionalism in a south Indian temple (Princeton University Press, 2003) and The camphor flame: popular Hinduism and society in India (Princeton University Press, 2004, revised edition).

Haripriya Narasimhan is Research Fellow in Anthropology at the London School of Economics and Political Science and her Ph.D. research at Syracuse University was on poor women's decision-making about healthcare in Tamilnadu. Before starting their current research project on the Vattimas, the two authors carried out research among the middle class in Chennai.