¹ These figures are taken from the Statistical Outline of India (Bombay: Tata Consultancy Services, 1986).Google Scholar

² Gupta, Jyotindra Das, Language Conflict and National Development (Berkeley: University of California Press, 1970), p. 33.Google Scholar

³ Powell, G. Bingham, Contemporary Democracies (Cambridge, Mass.: Harvard University Press, 1982), p. 45Google Scholar; Taylor, Charles and Hudson, Michael, World Handbook of Political and Social Indicators (New Haven, Conn.: Yale University Press, 1972).Google Scholar

⁴ Senas are groups of armed men.

⁵ Systematic sociological evidence for caste among the non-Hindus can be found in Singh, Harjinder, ed., Caste Among Non-Hindus in India (New Delhi: National Publishers, 1977).Google Scholar

⁶ For a fuller statement of the social cleavage theory of party systems, see LaPalombara, Joseph and Weiner, Myron, eds, Political Parlies and Political Development (Princeton, NJ: Princeton University Press, 1966)CrossRefGoogle Scholar; Lipset, Seymour Martin and Rokkan, Stein, ‘Cleavage Structures, Party Systems and Voter Alignments: An Introduction’, in Lipset, Seymour Martin and Rokkan, Stein, eds, Party Systems and Voter Alignments (New York: Free Press, 1967)Google Scholar; Rose, Richard and Urwin, D. W., ‘Social Cohesion, Political Parties and Strains in Regimes’, Comparative Political Studies, 2 (1969), 7–67.CrossRefGoogle Scholar

⁷ By ‘aligned’ party system we are referring to party systems in which there is a strong correlation between social and demographic variables and party preference. By ‘less aligned’ we mean to indicate weak correlations.

⁸ Lijphart, Arendt, Democracy in Plural Societies (New Haven, Conn.: Yale University Press, 1978)Google Scholar and Powell, , Contemporary Democracies.Google Scholar

⁹ Lipset, Seymour Martin, ‘Radicalism or Reformism: The Sources of Working Class Polities’, American Political Science Review, 77 (1983), 1–18.CrossRefGoogle Scholar

¹⁰ These data, slightly modified, are drawn from Powell, G. Bingham, ‘Voter Turnout in Thirty Democracies: Effects of Partisan, Legal and Socio-Economie Environments’, paper delivered at the Conference on Voter Turnout, San Diego, 05 1979.Google Scholar

¹¹ To be sure, substantial social diversity is not necessary for the development of socially-based party systems (there are several European countries with relatively little ethnic or racial diversity but very homogeneous parties along religious, language or class lines), but it is (usually) sufficient, and as the data in Figure indicate, the most socially diverse nations tend to have the most homogeneous parties.

¹² Kothari, Rajni, ‘The Congress “System” in India’, Asian Survey, 4 (1964), 1161–73.CrossRefGoogle Scholar

¹³ Kothari, , ‘The Congress “System” in India’, p. 1164.Google Scholar

¹⁴ See Lijphart, Arendt, Democracy in Plural SocietiesGoogle Scholar. Parenthetically, we would note that the social cleavage theory does not assume a ‘deterministic’ model of the Indian voter or the Indian party system. Beliefs and ideology have an autonomous role. People do express a vision of the public order in their vote, not just of group interests. However, we believe that social group conflicts are a major source of political beliefs and ideological orientations.

¹⁵ While State politics in India have been the subject of studies such as Weiner, Myron, ed., State Politics in India (Princeton, NJ: Princeton University Press, 1968)CrossRefGoogle Scholar; Narain, Iqbal, ed., State Politics in India (Meerut: Meenakshi Prakashan, 1976)Google Scholar; and Wood, John, ed., State Politics in Contemporary India: Crisis or Continuity? (Boulder, Colorado: Westview Press, 1984)Google Scholar, we believe that there has been no systematic articulation of the consequences of this State focus for the Congress as a type of party within a particular type of party system.

¹⁶ Seal, Anil, ‘Imperialism and Nationalism in India’, in Gallagher, John, Johnson, Gordon and Seal, Anil, eds, Locality, Province and Nation: Essays on Indian Politics 1870–1940 (Cambridge: Cambridge University Press, 1973), p. 25.Google Scholar

¹⁷ Bayly, C. A., The Local Roots of Indian Politics (Oxford: Oxford University Press, 1975), p. 4Google Scholar. Sarkar, Sumit, in his Modern India 1885–1947 (Delhi: Macmillan India Limited, 1983)Google Scholar, tacitly accepts the local bases of Congress support, but criticizes the Cambridge school for downplaying the role of ideology and patriotic motivation in the nationalist movement. While that debate has not been resolved we can fairly characterize the pre-independence Congress as a factional coalition tied together by an anti-colonial ideology that identified the Congress with the nation-state.

¹⁸ For this interpretation of the Congress see Kothari, ‘The Congress “System” in India’; Sheth, D. L., ‘Partisanship and Political Development’ in Sheth, D. L., ed., Citizens and Parties: Aspects of Competitive Politics in India (Delhi: Allied Publishers, 1975)Google Scholar; Diamond, Larry, Lipset, S. M. and Linz, Juan, ‘Developing and Sustaining Democratic Government in the Third World’, paper presented at the Annual Meeting of the American Political Science Association, Washington, DC, 1986Google Scholar; and Ahmed, Bashiruddin, ‘Dimensions of Party Support in Contemporary India’, paper presented at the Conference on the Comparative Dimensions of Indian Elections and Party Politics held at UCLA, 06 1987.Google Scholar

¹⁹ Sisson, Richard, The Congress Parly in Rajasthan (Berkeley: University of California Press, 1970), p. 212.Google Scholar

²⁰ For more on the concept of the ‘center’ see Daalder, Hans, ‘In Search of the Center of European Party Systems’, American Political Science Review, 78 (1984), 92–109.CrossRefGoogle Scholar

²¹ The data used in the analysis were culled from three post-election surveys covering the 1967, 1971 and the 1977 elections. The surveys were conducted under the auspices of the Centre for the Study of Developing Societies, Delhi. We are grateful to the Centre and the Department of Political Science at UCLA for permission to use the surveys for our analysis.

²² Frankel, Francine, ‘Compulsion and Social Change: Is Authoritarianism the Solution to India's Economic Development Problems?’ in Kohli, Atul, ed., The State and Development in the Third World (Princeton, NJ: Princeton University Press, 1986), p. 167.Google Scholar

²³ Moreover, the pattern in comparable earlier data from the 1967 election and later data from the 1977 campaign (see Appendix) is so similar to the 1971 pattern that it is unreasonable to believe that Indira Gandhi's election strategy significantly reformulated the way in which the parties con verted traditional social conflicts into electoral support.

²⁴ Class is a measure of the relative economic status of the respondents in the election surveys. This economic classification was based on the reported education, income and occupation of the respondents. A preliminary factor analysis revealed that these three variables could be used in conjunction with one another. The respondents were classified into three literacy categories dependent upon their years of schooling. With occupation, the algorithm was more complex. Since rural occupational hierarchies are based on land-holding those who lived in the rural areas were classified according to the amount of land they held. In the urban areas profession determined the respondent's position. Housewives were accorded the occupation of the head of the household. As a crosscheck, we found no difference in the reported family income of housewives (whose occupation was categorized by that of the head of the household) and that of respondents of similar occupation. We used family income as a measure of the financial status of the household. However, the division of the sample into income categories was based on family size as well. These three variables – education, occupation and income – were then standardized and added. The sample was then divided into thirds giving us the three categories of the composite variable we call class.

Caste is a measure of the social, not ritual, status of the respondent. Caste as Jati is the major form of social organization and identity at the local level. Caste, in our formulation, does not refer to the Varna categories, which are merely structural modes of organizing Jati categories and do not constitute self-conscious social groups. Jati is geographically bounded and rarely transcends the dominant linguistic culture of which it is a part. For purposes of comparison Jati had to be collapsed into an analytically useful form. The classification we have used was defined during an informal workshop held at UCLA during the summer of 1985. At that workshop, Ramasharay Roy of the Centre for the Study of Developing Societies in Delhi, Richard Sisson, William Vanderbok and Pradeep Chhibber went through archival material on most of the Jatis in each State and classified them into categories that reflected commonality of status in the State and simultaneously constituted meaningful social and analytical categories.

²⁵ There is evidence that the electoral coalition that formed the basis of Mrs Gandhi's victory was not only the minorities and the disadvantaged sectors, but also high caste, upper-status respondents in the rural sectors of the Hindi heartland who voted at higher levels for the Congress than in previous elections.

²⁶ Ironically, although the linguistic reorganization of States removed language as a ‘grievance’, the reorganization made language a permanent feature of the political landscape in India.

²⁷ Kothari, Rajni, ‘Introduction: Caste in Indian Politics’, in Kothari, Rajni, ed., Caste in Indian Politics (New Delhi: Orient Longman Limited, 1970), p. 24.Google Scholar

²⁸ There is both sociological and electoral evidence to substantiate our assertion that religious differences do not travel well across regional boundaries. Harold Gould in ‘Modern Politics in an Indian District: “Natural Selection” and “Selective Cooptation”’, a paper presented at the Conference on the Comparative Dimensions of Indian Elections and Party Politics held at UCLA, argued that in Faizabad district electoral mobilization operates along local communal lines – the level and point of intergroup conflict – and not along national ones. The 1971 post-election survey results provide further evidence of this phenomenon. In the ‘Hindi heartland’ the Muslim vote for Congress among those living in the urban areas was 16 percentage points higher than among their rural counterparts (79 per cent compared to 63 per cent). Moreover, only half of all Muslims in the South voted for Congress, whereas the corresponding percentages for the West and the East are 91 and 84 respectively. In no other region of India but the North do we really find evidence of this urban–rural divide among the Muslims, although this could in part be due to the not so large samples in other areas. Further corroborative evidence can be found in Table 5.

²⁹ Srinivas, M. N., Social Change in Modern India (Berkeley: University of California Press, 1972).Google Scholar

³⁰ We would like to thank John Zaller, whose comments made part of the following discussion possible.

³¹ The greater across-State average in the alignment of social and political differences (represented by the ability of religion, caste and class to explain 19 per cent of the variance in the vote among the States) is not a function of the exceptionally large explained variance in the four small-sample States. Of course these four States affect the average. Without them, the average across-State explanation of the variance is only 13 per cent. Our point, however, is that there is more structure within each State than there is at the national level. Consequently, the important feature of the data is not the absolute value of the correlations, but their significant differences.

³² The ‘adjusted’ score is an unweighted average of the degree of alignment within each State. A regression analysis of the relationship between the variance explained and the Alford alignment score (used in Figures 1 and 2) allowed us to obtain the adjusted value. Details of the procedure will be provided upon request.

³³ See Daalder, , ‘In Search of the Cenler of European Party Systems’.Google Scholar

³⁴ It could be argued that in 1977 there was a national position to oppose and the various opposi tion parties did form a coalition. Even then, there was a clear north-south divide, with the southern states, Andhra Pradesh, Karnataka, Kerala and Tamilnadu opting in favour of the Congress or its allies, thus revealing the regional basis of the Janata coalition. The fact that the Janata party leader, Ramakrishna Hegde, has been talking of a federation of opposition parties, instead of a united opposition, testifies to the territoriality of the party system in India.

³⁵ We are referring to the parties that are recognized as national parties by the Election Commission of India.

³⁶ We are referring only to those States which we dealt with in the body of the text.

³⁷ It is this organizational form that may explain the patron–client nature of the Congress today.

³⁸ ‘State’ leader implies not political leaders active in State politics but the fact that most national politicians have ‘State constituencies’.