The phrase "restoring politics to political history" threatens to become the slogan for a new scholarly campaign. Kousser's recent elaboration of the theme is the most challenging, offering the author's usual theoretical and methodological inventiveness in combination with a sweeping indictment of recent work. Before the slogan gathers too many adherents, however, problems about the work of restoration need airing.' The urge to restore politics to political history arises from uneasiness that twenty years of studying mass electoral behavior has put at risk the "autonomy" of political history, threatening, as Kousser has it, to turn political history into a "mere branch of social history." Several other scholars are also concerned that studies of the social and cultural linkages within the mass electorate in relation to partisanship have rendered something called politics epiphenomenal. The implication is that there is a perennially separable category of human agency called politics which is not reducible to the social and cultural attributes of those who