just another ‘‘collective group’’. Anglo-Saxon theories—with the exception of some
Hegelians in Britain from Thomas H. Green to Bernard Bosanquet—nourished a
deep distrust of the notion of the state and rather preferred ‘‘government’’ as the
central notion for institutional analysis.
The development of institutional theory after 1945 proved to be oscillating
between waves of neglect and rediscovery of institutions. The attempt to make
political scienceWnally scientiWc stood against the accepting institutional analysis
as the centre of research. The ‘‘new science of politics’’ in the USA used the term
‘‘institution’’ in the vague sense of neighboring social sciences, such as sociology or
anthropology, as ‘‘a pattern composed of culture traits specialized to the shaping
and distribution of a particular value (or set of values)’’ (Lasswell and Kaplan 1950 ,
47 ). The ‘‘behavioral revolt’’ was directed against the old institutionalism, but did
not avoid institutions altogether. Heinz Eulau ( 1969 , 1 , 158 ), a pioneer of the
‘‘behavioral persuasion,’’ developed a synthesis of ‘‘behavioral-institutional
research,’’ mainly concentrated in legislative and judicial studies. Whereas Eulau
critically worked on a theory of micro–macro-relations—in spite of the basic
individualism of this approach—later behavioralists frequently uncritically gener-
alized theWndings on the micro level in the macropolitics of institutions. The
‘‘epitaph of a successful protest’’ which Robert Dahl proclaimed in 1969 was
premature in the eyes of later analysts. John C. Wahlke ( 1979 ) in his presidential
address for the American Political Science Association ten years later was more
skeptical. After a quantitative analysis of review articles and research notes in the
American Political Science Review, he came to the conclusion that old-fashioned
institutional studies prevailed even in this journal which was considered to be the
‘‘battle organ’’ of the victorious behavioral revolt.
Behavioralism was accused of lacking theory-building.Systems theoryhoped to
heal this shortcoming. Systems theories in America had the virtue to develop—for
theWrst time since Weber—a generalized theory of institution, overcoming the
shortcomings of ad hoc theories in Europe. For Talcott Parsons ( 1959 ), deeply
inXuenced by Weber, institutional patterns, perceived in a demystiWed way, were
the backbone of social systems. Only in later variations of the theory of systems did
‘‘structures’’ become more important than institutions. They had, however, no
predetermined role. Similar functions within the system were completed by very
diVerent structures. The early Luhmann ( 1965 , 13 ), originally Parsons’ devoted
disciple but soon a defector who created his own autopoietic version of a theory
of systems, still used institutions and structures as synonyma: ‘‘Institutions are
behavioural expectations generalized in temporal and social dimensions, and thus
create the structure of social systems.’’ Systems theory created a new methodo-
logical terminology, but on the descriptive level it classiWed the traditional powers,
such as the executive and parliament, adding bureaucracy and parties. They got,
however, more scientiWc names such as ‘‘rule-setting,’’ ‘‘rule applying,’’ ‘‘rule
adjudicating,’’ and ‘‘rule-enforcing’’ institutions.
748 klaus von beyme