The international system: bipolar or multibloc?

Important historical data that cannot be accommodated by the theoretical concepts intended to elucidate them pose one of the most troublesome problems for the political theorist. Especially in unsettled periods of historical transition, events have a discon-certing quality of eluding the ideal analytic types established for them. Approaches to the study of international politics which focus on the concept of system are no ex-ception. In what follows I shall be using the work-ing definition of an international system put forth by Stanley Hoffmann (1961, p. 207): An international system is a pattern of relations between the basic units of world politics, which is characterized by the scope of the objectives pursued by those units and of the tasks performed among them, as well as by the means used in order to achieve those goals and perform those tasks. This pattern is largely determined by the structure of the world, the nature of the forces which operate across or within the major units, and the capabilities, pattern of power, and political culture of those units.2 The systems approach is unusually flex-ible and can be adjusted to fit many and diverse analytical purposes, primarily be-cause a broad definition of system allows application in a wide variety of contexts which may differ greatly in terms of unit size, type of relationship, issues, and so forth. But the patterning of behavioral regularities, which is at the core of systemic treatments, usually is much less adaptable because it represents, in effect, the creation of an ideal analytic type. The formulation of typologies such as a balance-of-power system or a bipolar international system is always arbitrary in the sense that the theorist settles on system types, and the relationships they represent, either because they seem to approximate the real world or because they order the range of possible alternative systems in clear-cut and esthetically satisfying models. Of course, the needs of empirical investiga-tion cannot always be met by ideal analytic types because of the kaleidoscopic com-plexity of empirical phenomena. Even if a1 For their criticism and advice I am grateful