Some Further Notes on Party Dynamics
FROM THE GENERALIZATIONS about party behavior set out in the preceding chapter, two even more general statements can be drawn. First, the party system is dynamic, not static. Second, it is multidimensional, not unidimensional. These conclusions merit some further explication, because they challenge conceptions of the party system that have been widely accepted in both academic and popular discussion and that have been incorporated into the models of the party system presented in standard textbooks.
The generalizations also have a bearing on normative concepts of the party system--about how American political parties should behave and what constitutes a healthy party system.
Writing in 1959, V. O. Key, Jr., suggested that analysis of the political party system must take account of what he called "the time dimension." Characteristically, said Key, analysts of the party system have used static models, and these "may exclude from attention fundamental aspects of party behavior."1 A few years later, Donald E. Stokes called attention to a number of defects in the prevailing model of the party system used by both journalists and academicians; among the defects he listed were not only that the model was static, which Key had criticized, but also that it was unidimensional. In the model, voters and parties are thought of as arrayed along a single spectrum from left to right, from liberal to conservative, when in fact political competition is multidimensional. Stokes pointed out, for example, that the evidence from public opinion polls indicates no relation between a voter's attitude on social welfare____________________