`Democracy' in Northern Ireland: experiments in self-rule from the Protestant Ascendancy to the Good Friday Agreement* JOHN McGARRY Department of Political Studies, Queen's University, 99 University Avenue, Kingston, Ontario, Canada K7L 3N6 ABSTRACT. Pierre van den Berghe has argued that democracy in divided societies can take five different forms: Herrenvolk democracy, ethnic democracy, liberal democracy, multicultural democracy and consociational democracy. My article argues that each of van den Berghe's five versions of democracy, or relatives of them, has been experimented with in pre-partition Ireland and Northern Ireland. While all have clear limits, the one that is most suited to Northern Ireland's conditions is consociational democracy. The article discusses some limits of the consociational approach in Northern Ireland but also defends it against common criticisms. Pierre van den Berghe (2002) argues that democracy in ethnically divided states can take several different forms. The state can be dominated by one group, as in a Herrenvolk or ethnic democracy; protective of individual equality, as in a liberal democracy; accommodative of minority group rights, as in a multicultural democracy; or co-governed by its different groups, as in a consociational democracy. Northern Ireland (and pre-partition Ireland) provides a useful test case for weighing the relative merits and feasibility of these options, as it has experienced all of them, or relatives of them. Ireland's eighteenth-century Anglican control system, the `Protestant Ascendancy', has several features in common with van den Berghe's concept of Herrenvolk democracy, although in Ireland only a small proportion of the Herrenvolk had the franchise. Northern Ireland approximated an ethnic democracy under the Stormont regime that existed between 1921 and 1972. * As usual I am indebted to Margaret Moore and Brendan O'Leary for their many helpful insights. Thanks also to Sammy Smooha, who asked me to write this paper and who offered several useful criticisms, and to the anonymous referees at Nations and Nationalism. Finally, I would like to acknowledge the help of the Social Sciences and Humanities Research Council of Canada (SSHRCC), the United States Institute of Peace and the Carnegie Corporation of New York, all of which have funded my research. Nations and Nationalism 8 (4), 2002, 451±474. # ASEN 2002