The discipline of political science deliberately abandoned history sixty years ago. In a quest for a true “science” of politics, it turned to physics and especially psychology for models of what a science ought to be like (Jensen, 1969). Rigor in defining and measuring variables and in the statement of theories and testing of hypotheses was sought, plus a method that could predict, and thus control, the future. History seemed unnecessary, for the terrible cataclysm of the Great War seemed to have erased or smothered the impact of the past. It would take political scientists a half century to rediscover history, and the goal of this paper is to examine the route of one of their most important explorers of the past, Walter Dean Burnham.

It has now been over twenty years since The American Political Science Review published “The Changing Shape of the American Political Universe” by Walter Dean Burnham (1965). This remarkably rich work is at once a study of historical variations in citizen electoral behavior, of the partisan realignment of 1896 and the realignment process in general, and of the connections between voter behavior and the patterns of American politics and public policy. It also raises fundamental epistemological issues about the relationship of micro-level and macro-level phenomena in politics—especially the inherent limitations of single-shot public opinion surveys or of a focus on a single political period in understanding the full range of possibilities for citizen involvement in a democratic political order. Unlike many scholarly works which rise meteor-like to prominence then rapidly vanish, “Changing Shape” remains at least as influential today as it was twenty years ago.

Anyone who has tried to teach undergraduates about the election of 1896 should instinctively appreciate Walter Dean Burnham’s enormous contribution in making sense of that election and its aftermath. Waged between two rather uninteresting men over issues that defy easy understanding, the presidential contest of 1896 hardly stacks up with those of 1860 and 1932 as “critical” in the casual sense of the term. Perhaps if William Jennings Bryan had defeated William McKinley, or, better still, if the glamorous Theodore Roosevelt had been the victorious Republican candidate, or, best of all, if Roosevelt had won and immediately started a major war, the election of 1896 would more readily appear to have been the transforming event that modern scholars contend it was. But, alas, McKinley won, waited over a year before reluctantly waging even a minor war, and proved unwilling to make any significant departures in domestic policy. Compared to the election of Abraham Lincoln or that of Franklin D. Roosevelt, the contest of 1896 appears trivial. Who can blame undergraduates for yawning over the “Battle of the Standards”?

1. Was there any such thing as the “System of 1896”—or indeed any other “system” in American political history, for that matter? Professor McCormick’s paper suggests that he would clearly answer this question in the affirmative. For essentially he argues not that no such system existed, but that much too heavy an interpretative burden in my analyses has been placed on 1896 itself. I am sure that, so far as this is concerned, Professor McCormick is right. Yet there do seem undercurrents in this friendly critique of the broader doubt as to whether there is any substantiality at all to the model which I first attempted to sketch out quite a few years ago. On any one of a variety of specific details, I would have to plead, if not guilty, then at least nolo contendere. Yes, I undoubtedly placed too great a burden on 1896 itself, which inevitably should and does create scholarly doubts. Yes, the original sketch which attempted to link elite motivation with underlying power structure and those with electoral politics in and after the 1890s was painted in very vivid colors indeed. This may tell a dramatic story, but it may assume a greater conspiratorial intentionality and future-prediction on the part of elites than was actually there. But once we deal with issues such as these, is there anything left? Professor McCormick, a friendly critic, seems to think there is, but undercurrents of doubt remain.

The recent popularity of community studies among investigators of nineteenth-century social history in the United States owes much to convergence of interests since the early 1970s among four broad groupings of historians: labor and radical historians concerned with class-formation; historians of women and the family; immigration historians; and urban historians concerned with the transformation of spatial and social structure. Stressing the importance of the interrelationships between their subjects, historians with these interests have tended to see the community study as the best means of describing the interrelationships fully and concretely. Howard Chudacoff expressed this perception when he characterized books on the artisans of Newark and on the iron and textile workers of Troy and Cohoes as “community studies of the best type, for they combine working class history with perspectives on family, ethnicity, mobility, stratification, ideology, technology, politics, and … show the importance of interactions between place and behavior” (Chudacoff, 1979: 535).