Erik Moberg ã:
A Theory of Democratic Politics
1.1 - THE NATURE OF THE THEORY
The theory presented here, although verbal rather than formalized, aims at being a logical, deductive system. In essence this means that a number of propositions are logically derived from a limited set of basic propositions, which thus are used as axioms. Ultimately and ideally all propositions should, of course, be empirically tested. If this leads to verification that is good so far, if not that is a problem for the theory.
Here, in this context, the theory should however rather be tested for logical consistency. It is of utmost importance that propositions which are presented as logical consequences of other more basic propositions really are so. To the extent that this is not case the theory, considered as theory, is a failure. If the theory does not form a logical system in the sense described it just is not a theory.
There are, I think, at least two dangerous pitfalls involved in an undertaking like this one. The first is the risk of presenting, without noticing it, circular arguments. The other is the risk of mistaking common knowledge of real matters for logical conclusions. In what follows I have, at my best, tried to be aware of these pitfalls, but I am not sure that I have succeeded in avoiding them everywhere.
At last it should be emphasized that a theory is more than an instrument for producing hypotheses. Even if it were possible, somehow, to produce hypotheses for empirical testing by other means than from a theory, the theory will still always be a necessary part of a scientific enterprise. The reason is that the logical order, and by that the predictive and explanatory power, which a theory brings to a set of empirical findings cannot be achieved in any other way. In this sense theory supports empirical results as much as empiricism supports theory. Perhaps it was something like this which the great British astronomer and physicist Arthur Stanley Eddington (1882-1944) had in mind when saying that "One should never believe any experiment until it has been confirmed by theory" (Quoted by Steven Weinberg, 1993, p 101).