refinement on the basis of improved knowledge (see Morrow and Hull, 1996; Asquith, 1999). Of course, the application of design principles is also filtered through the political processes through which institutional design decisions are made, so that these choices involve more than a straightforward application of generic knowledge in a specific situation. The applicability of design principles also changes over time and across contexts and thus proves to be contingent (Weinstein, 2000). For example, based on the discussion of the empirical cases presented at the 1985 Annapolis meeting discussed in Chapter 1, Ostrom (1986:611) proposed that institutions that had developed simple rules were more likely to survive. Specifically, the factor discussed was “The development of a clear-cut and unambiguous set of rules that all participants can know and agree upon.” The logic on which this was based is that the “fewer rules used to organize activities, relative to the complexity of the activities, the more likely that individuals can understand them, remember them, and follow them, and the more likely that infractions will be interpreted by all as infractions” (Ostrom, 1986:611). This is obviously a highly contingent principle that has to be tailored to the complexity of the resource system itself, the cultural heterogeneity of the users, and their communication patterns. Slavish adoption of any stylized version of a design principle is unlikely to be a successful strategy (see Steins et al., 2000). Evidence exists that the “simple-rules” principle applies most strongly to institutions that engage large, diverse groups with weak community ties; small, tightly linked groups sometimes can function quite well with complex rules, provided that the users understand them well (Berkes, 1992).

How can generic knowledge be of practical value?⁶ We do not expect that it will be prescriptive in the sense of providing a standard set of procedures that tells practitioners exactly what to do in particular situations. However, generic knowledge is useful to practitioners when they combine it with detailed knowledge about the situation at hand. Generic knowledge also has diagnostic value for practitioners. It describes the characteristics that determine the actions that will be effective. After a practitioner has accurately diagnosed a situation, knowledge about what works in which situations comes into play more strongly.

Even with a perfect diagnosis of a situation, however, there are several reasons why generic knowledge cannot be expected to provide detailed prescriptions for action. First, generic social science knowledge will never be as solidly established as, for example, a law of physics. For one thing, human actors can defy the laws said to govern their own behavior; for another, conditions continually change in ways that may invalidate conclusions from past experience. The principle of uniform laws across time and space, so central to the intellectual program of the physical sciences, is not realistic in developing theories about human behavior. Second, the many tradeoffs involved in any decision make general knowledge an imperfect guide to action. All the desired aspects of success cannot be achieved all at once, and choices must be based on tradeoff or compromise. Often, resource sustainability is not the only outcome relevant to practitioners. They must then