Bivariate hypotheses may be too simple for understanding world politics. One response is to add more variables. If the goal is to explain a particular event then this is quite reasonable given that few events in the physical and human world have but one or two causes. The emphasis on context suggests that the mere “addition” of variables does not suffice, new kinds of relationships are necessary as well. Context as changing meaning and context as barrier are two different kinds of relationship that may be added to a conceptual tool box. In this book I have tried to construct a standard contextual tool and illustrate its use in a particular problem(s). But it is quite clear that customized tools work much better for specialized tasks.

Boyd and Iverson (1979) have already discussed a large number of statistical variations on the context as changing meaning tool, but there has been little corresponding theoretical development of the concept (hence the more theoretical nature of my discussion). The possibilities of the barrier mode remain to be exploited. Context as changing meaning models remain comfortably within the framework of general linear models (regression); barrier models seem to require different kinds of modeling techniques. In particular dynamic models appear appropriate; or perhaps with adaptation models could be stolen from hydraulics (after all the natural home of the metaphor).