It is both curious and instructive that the flowering of democracy in the 1980s in Latin America and elsewhere coincided with the collapse into incoherence of the behaviouralist effort to place the understanding of politics in general, and democracy in particular, on a new scientific footing. Verba laments that ‘We have been studying political development for a long time but we are no longer confident of what it is or our ability to understand it’ (1985:28-9), while for Dominguez twenty-five years of research on links between regime type and level of economic development ‘have ended with a great deal less certainty on these matters than they began’ (1987:85). Wiarda, seeing no coordinating or integrating theory, describes an atmosphere of ‘unhappiness and disenchantment with the field’ (1985:7); and for Mayer, a long-time devotee of the empiricist revolution, comparative politics is ‘in a state of conceptual disarray, with little consensus on the nature or purpose of the field’ (1989:273). As Easton concludes, ‘students are no longer so certain about what politics is all about’ (1991:284). The heady days of optimism in the future of ‘political science as science’ are far behind us: the best Lane and Ersson can do is to review a number of correlations (a technique Luskin (1991:1037) likens to ‘a wet finger aloft in the wilderness’), admit that they ‘do not suffice for the derivation of either necessary or sufficient conditions’, and feebly suggest that in the future ‘single-factor models may have to be replaced by combined ones’ (1991:73).
If empiricist theory were our only guide, we would conclude