Although many of its practitioners refer to political science as a discipline, it may
be questioned whether it really is that. Unlike economics, for example, there is no
single set of assumptions about the behaviour of the actors in a system that is
accepted by all analysts; the common frameworks utilized by the vast majority of
economists have no counterparts in political science. On the contrary, the study
of politics has been strongly influenced by many other 'disciplines'--economics,
sociology, social psychology, philosophy, history, law, and, more recently, femi-
nist studies. In some cases, as with the behavioural revolution in the 1950s and,
more recently, with rational choice analysis, some of those advocating the adap-
tation of particular analytical frameworks from other fields have seen these
frameworks as capable of transforming the whole basis of the study of politics.
Utilizing them, their proponents argued, would make possible the creation of a
genuine political science--a discipline that could hope to rival the physical sci-
ences in the rigour of its causal explanations. Alas, such projects have come to
nought! Political science remains a 'market-place' in which different analytic
frameworks and different approaches compete with each other, without any of
them ever becoming dominant.

That judgement is reinforced by the evidence from the examination of parties
and party systems in this book. In the Introduction I said that a number of the
topics to be considered in subsequent chapters had been studied from one or
more alternative approaches--approaches which I called 'sociological', 'institu-
tional', and 'competitive'. It should have become clear by now that no one of these
approaches is demonstrably superior to the others. Rather what each has to offer
varies with the subject under consideration. Moreover, as was seen when consid-
ering the question of 'why party systems differ' (in Chapter 6), it would seem that
the utility of a particular approach may depend on the kinds of linkage evident
between voters and parties in the particular countries under consideration. (In
this case the institutional approach becomes more useful when social solidarity
is no longer the main factor binding voters to parties.)

That there have been alternative approaches employed by different researchers
has contributed enormously to our understanding of party systems. The early
disputes between 'institutionalists' and 'sociologists' have now generated

-377-