The Lack of Sex Differences in Declining Endogamy in Hyderabad, India

Discussions of Indian social structure that emphasize the patriarchal, patrilineal family system and the maintenance of caste and family purity through arranged marriages have encouraged sweeping assumptions about the control of women by men. Such assumptions underlie predictions that within a given community more men than women will "marry out." Two hypotheses predicting greater control of women's than men's marriages are tested and largely invalidated, utilizing historical data documenting seventy-five years of declining endogamy among members of the Kayasth caste of Hyderabad City, Andhra Pradesh. The empirical data, however, provide two significant findings about exogamous marriages: siblings rather than sisters are the appropriate units of analysis, and there is a progression within families from intersubcaste to intercaste marriages, with both types of exogamous marriages increasing.


JOURNAL OF ANTHROPOLOGICAL RESEARCH marriages. It suggests two hypotheses about the control of women which can be tested by examining Hyderabad Kayasth marriage patterns by sex.
The first hypothesis is that there will be a sex difference in exogamous patterns over time; that a higher proportion of men than women will marry outside subcaste and caste boundaries. This differential is sometimes referred to as "men-based" and "women-based" exogamy; a tighter control of women's marriages within endogamous groups in the caste supposedly results in lower rates of "women-based" exogamy (Rakshit and Dasgupta 1971:89; Wiebe and Ramu 1971:117).
The second hypothesis is that subcaste endogamy is being replaced largely by caste endogamy, so that a hypergamous system of seven ranked subcastes underlies the new Kayasth marriage patterns. This hypothesis predicts that as intermarriages proceed between members of these subcastes, daughters will be given to higher ranking subcastes and taken from lower ranking subcastes. The observation that North Indian castes with this same kinship system do tend to marry hypergamously, lends initial support to this hypothesis (Vatuk 1969:110). For such a system to work there must be a local group of "pseudo," or illegitimate Kayasths to supply the bottom group with wives; there is such a category in Hyderabad. The Srivastava subcaste has an unbounded category into which individuals of doubtful ancestry have moved, establishing claims to be Kayasths (Leonard n.d.). These "Srivastavas" are indeed a source of wives for higher subcastes. Those of the highest status must then be able to marry their daughters elsewhere, perhaps outside the city, to persons of even higher rank; that possibility also exists (Leonard 1978:271).
If these hypotheses should prove to be correct, they would support a view of the caste system which sees women being controlled and used by men in a system of arranged marriages, a system which ranks men. This view of India's caste system appeals to structuralists and others using Levi-Straussian models of kinship, in which women are passive objects exchanged by men (Levi-Strauss 1963:47, 1969 Of course there have been changes in the condition of women in the twentieth century, particularly among urban castes. Until the last generation or two, for example, Kayasth women in Hyderabad were secluded, educated in Hindi if at all, and married young; they did not work outside the home. The increase in education for girls, in their age at marriage, and in their employment outside the home, both before and after marriage, have brought striking changes; but these changes would not necessarily conflict with continuing control of women's marriages so as to construct a hypergamous system among the seven subcastes. Table 2 gives a breakdown by sex and subcaste of all Kayasth marriages in each of three time periods from 1900 to 1975. This table is necessary for a careful interpretation of some of the data on women. Instances of marked imbalance by sex are underlined for some of the subcastes in the earlier two time periods, and for the Nigams in all three time periods; subcastes where the sex ratio is imbalanced are in fact the smaller subcastes, the ones which have frequently arranged marriages with members of their own subcaste outside of Hyderabad City. I have interpreted the underrepresentation of women evident in Table 2 as substantiating the reported practice of extralocal endogamy: daughters married and settled outside the city were often omitted from contemporary oral accounts used to construct family genealogies and marriage networks. These undercounts of women who made extralocal, endogamous marriages should be kept in mind when interpreting the following tables.1 Table 3 breaks the Kayasth population down into the seven subcastes; it is immediately apparent that the subcastes vary greatly in their marriage patterns. Subcaste endogamy was still 82 percent for the Mathur subcaste, from 1951 to 1975, while it was only 11 percent and 16 percent, respectively, for the Nigams and Asthanas. For most of the subcastes, subcaste endogamy has been replaced by caste-level endogamy (Table 3- Table 4 should help us address the initial hypotheses predicting differential, hypergamous marriage patterns by sex as subcaste endogamy declines in this urban Indian caste. It appears to me that, in the aggregate, and taking into consideration the undercounts of women evident in Table 2    and women married extralocal Kayasths of other subcastes. The proportions of men and women who married local Kayasths of other subcastes in the most recent time period are 31 percent and 35 percent, respectively; notice that this differential is not in the predicted direction. It is only in the category of non-Kayasth marriages that a difference by sex occurs which might indicate stronger control of women's marriages; this discrepancy will be addressed after a discussion of the possibility of a hypergamous marriage system by subcaste within the Kayasth caste. The similar proportions of men and women who marry Kayasths of other subcastes might result from the very large number of Mathur Kayasths, a number large enough to smooth out sharp differences by sex in the smaller subcastes. These aggregate proportions may also disguise patterns of hypergamy within the caste category. When I broke the data in Table 3 down further, by sex, as a partial check on these possibilities, however, the results did not suggest hypergamous patterns of any strength or clarity. But that complicated table (not included) was an inadequate test for such patterns; Table 5 focuses specifically on the local marriages among members of different Kayasth subcastes, and should enable us to discern such patterns, if they are there.

INVESTIGATION AND FINDINGS
There are in fact some imbalances evident here.2 In particular, the Asthanas received many more brides from local Kayasths than they gave in the second period, and the Mathurs gave brides but did not receive them in the same period. These figures substantiate local oral tradition: the Asthanas, newly arrived in the late nineteenth century, preferred not to marry in Hyderabad until recently, and the Mathurs initiated intersubcaste marriages only because they had "too many girls." But Table 5 as a whole gives evidence of the general practice of exchange marriage by Hyderabad Kayasths; i.e., they give and take daughters from the same families. They exchange with families of equal status, rather than giving daughters to families of higher rank and taking them from families of lower rank. Both the kin groups of the nineteenth century and the broader marriage networks of this century feature exchange marriages rather than the construction of hypergamous relationships.3 On one measure, that of individuals marrying outside the Kayasth caste, there was and is a clear difference by sex (see Table 4). No Hyderabad Kayasth woman (compared to eleven men) married a non-Kayasth in the first twenty-five-year period; the numbers and proportions of Kayasth women who married outside the caste in the other two time periods are still significantly lower than those for Kayasth men. Table 6 presents specific data for the thirty-four women and sixty-four men who have married non-Kayasths. Kayasth women began to make such marriages later and in lower proportions than did Kayasth men, and the variety of spouse categories for women is not as great-there are no non-Indian husbands, in contrast to the six foreign wives. Reflecting a lingering allegiance to the political culture of the former state of Hyderabad, most of the non-Kayasth partners are from old-city, Hindustani-speaking families. Despite Hyderabad City's current status as the capital of a Telugu-speaking state, no husband is a Telugu speaker and relatively few partners speak South Indian languages. There are more Muslim partners than one might expect. Love marriages constitute an increasing proportion of these marriages outside the caste.4 Love marriages were attributed by Hyderabad informants largely to school, work, or neighborhood contacts. Not surprisingly, other  I would argue, also, that this sex difference in intercaste marriage is explained by different educational patterns by sex. Intercaste marriages result increasingly from study in coeducational institutions of higher education, both in India and abroad, and opportunities for such study were extended earlier and more frequently to sons. The effect on marriage, then, is indirect, rather than being due to direct exercise of control over women's marriages. The most interesting feature of these marriages outside the caste is that they are not random, individual deviations from a new norm of caste-level endogamy.
By tabulating information about the thirty-four women from thirty-one families who have made intercaste marriages between 1926 and 1975, two patterns become evident ( Table 7). The Mathur Kayasths exemplify the "first generation": marriages outside the subcaste really only began after 1950, so that these women are almost all members of the first generation to marry outside the subcaste. Of the eleven sets of Mathur parents, only one set (9 percent) represented a marriage out of subcaste (the wife was a Kayasth of another subcaste). The rest of the Kayasths exemplify 360 JOURNAL OF ANTHROPOLOGICAL RESEARCH a second generation expansion of the range of partners for children of previous exogamous marriages; 65 percent of these sets of "other Kayasth" parents themselves married out of subcaste. Even more significantly, not only these particular women but their siblings, as well, are marrying outside their subcaste and caste. As the proportion of intersubcaste marriages rises, then, we can expect the proportion of intercaste marriages to rise as well; and when one child of an exogamous marriage marries out of subcaste, the chances are that a majority of the siblings will do so also. Of the seventy-eight married children from the non-Mathur sets of parents, the great majority (sixty-five, or 83 percent) have married out of subcaste; thirty (39 percent) have married members of other Kayasth subcastes; and thirty-five (45 percent) have married out of caste (twenty-three women and twelve of their brothers). For the Mathur group, of the forty-six children married, thirty-four (74 percent) have married out of subcaste; twenty (44 percent) have made intersubcaste marriages, and fourteen (30 percent) have made intercaste marriages. Clearly, intersubcaste and intercaste marriages are both increasing-and they are related.
The local-extralocal dimension of these marriages outside the caste also differ by sex. Marriages of Kayasth women to non-Kayasths outside of Hyderabad contrast strongly with marriages of Kayasth men to non-Kayasths outside the city. All eight daughters married outside Hyderabad to non-Kayasths from 1926 to 1975 had their marriages arranged; all were married to Indians in northern and western India. Six of these marriages have occurred since 1951 (and of those, three were of sisters married to Arya Samajists in the Punjab). Of the eleven men who made extralocal marriages to non-Kayasths, the four who married before 1951 brought back Indian brides from other parts of India; one of these was an Arya Samaj marriage, the only arranged marriage of the eleven. Of the seven who have married non-Kayasths outside Hyderabad between 1951 and 1975, five married foreign women they met while studying and working abroad, while a sixth met and married a Coorgi, in