In the last decade a number of scholars concerned with patterns of change in the international system have turned to international law for indications of systemic trends. They have assumed that key rules and principles of international law reflect broad characteristics of the system while controversies over and changes in their content and status reflect the rate, extent, and direction of change in the system. For example, Stanley Hoffmann in an early contribution illustrated the transition from the stable international system of the nineteenth century to the revolutionary system of today by reference to changes in the principle of sovereignty. Morton A. Kaplan and Nicholas deB. Katzenbach discussed the same transition—in their terms from a balance of power to a loose bipolar system—by reference to the principle of nonintervention. Somewhat later Wolfgang Friedmann treated the principles of sovereignty and nonintervention together in light of systemic trends over the last century, while Richard A. Falk associated the substantive law relating to civil war with the changing position of the principle of nonintervention. More systematically, an article by William D. Coplin ranged over several critical areas of international law in describing the evolution of the traditional international system to its present form. Finally, in recent but cursory essays, Andrew M. Scott and Oran R. Young have broadened the context in which the principle of nonintervention is related to systemic trends.

In many ways the study of international integration is in what Thomas Kuhn has called the pre-paradigm stage of the development of science. This stage is characterized by disagreement on the entities to be studied and the definitions, concepts, and indicators to be employed. Controversy is widespread even on the simplest of issues and, what is more discouraging, this controversy does not take place in a theoretical-methodological context which facilitates its resolution. There is a general lack of common standards by which to evaluate knowledge, decision-rules to interpret evidence, and criteria to measure progress. Thus, what constitutes advance for one school may mean retreat for another, and efforts to exchange views and harmonize aims frequently only serve to further define existing cleavages. Given this disagreement on broad philosophical issues it should not be surprising that research efforts and presentations of “evidence” have had little compelling impact in altering theoretical convictions. It is also no accident that the two most promising efforts to date are not presentations of data but are attempts to fashion a generally acceptable framework within which to evaluate evidence. Leon Lindberg has urged the adoption of a systems perspective which would harmonize a variety of research efforts and would, as well, make findings roughly comparable. Joseph Nye has suggested that we must “disaggregate” the concept of integration before we can meaningfully answer the important questions facing us. I have already indicated how complementary Lindberg's and Nye's suggestions are, and I feel that implementation of their suggestions would aid in making research findings touch base with one another and thus impart to evidence its proper multiplying effect.

Students of integration place their money on different beasts. Some working with the neo-functionalist model bet on the hare, while others using the cybernetic model elect the tortoise. The difference arises in part from their disparate perspectives of the racetrack: Neo-functionalists, with their assumption of a short-cycle process, view only a segment of the track and sight the rabbit sprinting; students of the cybernetic model, assuming a long-cycle process, insist that the whole race be watched and that the runners be clocked at different stages. As bets are still being taken, we might reconsider the basis on which odds are set.

One of the most heartening trends in research on international organizations and international affairs is the current emphasis upon an empirical, as opposed to a merely theoretical, approach to an understanding of problems. This approach involves a series of distinctive elements. A hypothesis must be offered and its terms must be defined in ways that permit observation of events and collection of data on the presence or absence of the defined items. Data must be collected systematically, rather than eclectically, to ensure the consideration of those events which do not fit the hypothesis as well as those which do. Wherever possible data is reduced to numerical form and various statistical tests are applied to determine, for example, positive or negative correlations, and to compute the statistical significance of those correlations. Conclusions are then drawn on the basis of the results of the statistical operations. While such efforts are often time-consuming and tedious, even with the aid of a computer, the results of such an effort have the enormous value of demonstrably reflecting reality. Presumably, these results can be replicated by other observers and, since the methodology has been explicitly articulated, methodological defects can be detected and their effect on the results can be fully evaluated.