3. Combination of approaches for crosscutting questions. Most important and interesting administrative questions, of course, do not reside neatly within one of these cells. Rather, administrative problems increasingly spill across two or even three of the issues. That leads to a clear conclusion: obtaining a full picture of administrative issues usually requires several different analytical approaches. One approach has the virtue of asking an important question that other approaches would neglect.
4. Combining approaches rarely produces stable answers. Our experience, both pragmatically and theoretically, is that such cross-fertilization is difficult. Each major approach has taken great pains to separate itself from the others. Just as the Hamiltonian and Madisonian traditions have long struggled for the soul of American politics, it is unlikely that competing approaches to administration based on these disparate traditions are likely to find an easy accommodation. Cross-fertilization among the approaches is essential; common ground among them is unlikely. From the inevitable conflict, however, can still come useful insights, just as the struggles between Hamiltonian and Madisonian traditions have yielded useful -- if often untidy -- patterns of governance.
5. Profound normative and educational differences. The inevitable struggles among the disparate approaches produce starkly different normative prescriptions, from strengthening bureaucracy to weakening its hand, from encouraging market competition to improving the power of government executives. These normative differences likewise produce widely varying approaches to education. Some approaches, especially the organizational ones, seek to give clear guidance to impart to future government managers. Others, such as those focused on the political environment, seek theoretical understanding of why bureaucracies behave as they do, but are largely divorced from public service education. Even approaches pursuing the same goal -- notably the organizational approaches of public administration and public management -- have sharply different views of what is important, how what is important can be learned, and how to transfer that learning to students. Cross-fertilization undoubtedly would bear much fruit, but the search for common normative or educational ground is likely to be contentious. Subdisciplines, as well as whole professional schools, have been founded on the principle of exceptionalism, not commonality. That fundamental self-definition would be hard to shake.
These comparisons and contrasts mirror the ancient dichotomy between politics and administration on which the field cut its teeth. The study of administration, of course, has moved far beyond the early search for a clear line between politics and administration -- that is, for a Hamiltonian vision based on bureaucracy as an independent actor. In discarding that well-meaning but poorly focused search, however, other approaches have risked committing the same error. Within the rich politics of the American political system, it is no more possible to solve the problem of bureaucracy by embracing the Madisonian tradition than by following the Hamiltonian vision. Americans' strange love-hate affair with bureaucracy is the direct result of America's conflicting traditions about what role bureaucracy ought to play. We want bureaucracy to be powerful enough to do -- effectively and efficiently -- what needs to be done. We also want its power carefully constrained so that it poses no threat to personal liberty ( Goodsell 1985).
The academic study of administration began by asserting that the phenomenon was worthy of study in its own right. In making the case, early public administrationists overdrew that argument by ignoring important complexities, especially Madisonian influences and the effects of other political forces on bureaucratic power. The reform efforts, however, sometimes replicated the mistakes of early public administration by ignoring the Hamiltonian traditions and the independent power of bureaucracy. Having solved the first problem of Wilson, Goodnow, and the other early founders of public administration -- that administration had become too important to ignore -- public administration could not solve the second -- the troubled relationship between administration and politics. Administration is both a dependent and an independent variable, a product of Hamiltonian and Madisonian traditions. It is virtually impossible to study administration without viewing administration as the result of one of the competing political traditions. It is certainly impossible to produce an adequate theory based on one set of choices. Thus, only a catholic approach, fully informed by the cross- fertilization of competing traditions, is likely to advance our understanding of administration. That is the lesson that the struggles over the politics-administration dichotomy ultimately teaches.
I am deeply indebted to the following colleagues who generously read the manuscript and made many, many, suggestions for