The Statistical Measurement of Agreement

theorems of O'M*.41 This means that von Wright's system is a subsystem of O'M*; and that some account is also taken of Bohnert's suggestions. It would be pointless to try to characterize O'M* further in such a brief compass; formal systems always require intensive study if their is to be fully appreciated. If the explanatory power of such constructs seems highly abstract or remote, the situation should be compared with contemporary physical theories, where mathematical theories much more recondite than those presented here have a direct and immediate bearing on empirical research. And just as many mathematical developments were initiated by empirical problems, so the development of O'M* was motivated by the hope of throwing light on some empirical problems in small-group research. More specifically, in experiments currently being conducted in the Yale Interaction Laboratory, opportunities are available for studying, under well-controlled conditions, small groups in the process of developing normative structures. A clear and accurate account of these interactional processess calls for a precise and rigorous conceptual framework within which to characterize the behavior of the group. Small-groups research is of course only one, and not necessarily the most important, area of application of such formal systems. It is hoped that the foregoing discussion will have suggestive value for workers in other sociological fields, and will stimulate interest in current research in mathematical logic, especially modal logic. This research is not remote from the daily affairs of sociologists. For instance, whenever instructions are given for filling out a questionnaire, commands expressing obligations are involved. More generally, any adequate sociological theory must encompass, in our opinion, the concepts norm, obligation, etc. It is therefore a matter of importance to develop sound techniques for analyzing norms and systems of norms.