pendent work follow, paced and ordered by other workshops. Some of these are 

 designed only for technical people in order to subject the work to criticism and to 

 expose it to a larger technical audience who often have significant advisory roles in 

 policy making., Later workshops focus on a larger community of managers, decision- 

 makers, and citizens. Throughout, the rules are to make everything as transparent as 

 possible, to provide an interactive environment, and to modify the analysis, models 

 and evaluation as new questions and suggestions emerge. 



The prime lessons: a small organization with the core tightly organized and the 

 participants more loosely integrated can address not only simple but highly complex 

 resource and environmental problems; a great multiplication effect occurs through 

 the network of participants that reduces the central budget, accelerates communi- 

 cation, and provides an early warning of problems. Sanity, innovation, and learning 

 are encouraged by the rhythm of intense short periods of interdisciplinary and policy 

 analysis, interspersed with independent consolidation; the scheduling and focus of 

 each workshop sets the deadlines and pace. And finally, every effort must be made to 

 provide opportunities for self-discovery by all actors. 



Connecting the Parts of a Model 



Submodels are the parts of the full model. They are chosen to include variables 

 which interact tightly, in a complex manner and at similar scales of space and time. 

 The goal is to divide the problem into submodels such that relatively little 

 information needs to be communicated between them. Those interconnections are 

 absolutely key, for from them come many of the unexpected policy effects as social, 

 economic, resource, and biophysical aspects combined. They generate those 

 surprises, crises and opportunities that challenge so much of resource and 

 environmental management. 



In every workshop some of the experts push to represent their submodel in 

 exquisite detail. They are understandably motivated by scientific rather than policy 

 interest. But that leads to a level of complexity and detail that typically prevents 

 linkage of submodels. Carl Walters resolved that with the innovation of the 

 "Looking-Outward Matrix." The notion is deceptively simple. Do not let the expert 

 tell you what information he can provide. He cannot be expected to know what other 

 experts or policy makers need. Rather ask him what he needs from other experts' 

 submodels. That leads to a matrix that identifies the variables and units that each 

 submodel needs from others. Hence, the interconnections between the parts are 

 identified from the start. Reading the table one way identifies the inputs that a 

 submodel will receive. Reading the other way identifies the outputs that others 

 require. In addition, each sub-group knows the actions that need to be incorporated 

 and the indicators that have to be generated. The definition of inputs and actions and 

 of outputs and indicators goes a remarkable distance in defining the contents and 

 scale of each submodel. And it gives an overview of the structure of the system that, in 

 some workshops, has been all that was required to better order and focus the research 

 and policy effort. 



Prime lesson: Many interdisciplinary and "contracting-out" modes of analysis 

 defeat the policy purpose because the component parts of the studies can never be 

 interconnected. The solution is not to ask the expert what he can do for you; ask him 

 what he needs from others. The results are used to structure the constraints imposed 

 on each component analysis so that they respond to the policy needs at a relevant 

 level of detail. 



Phase III. The Proof of the Pudding is in the Eating 



By 1974 we had developed effective ways to bridge gaps between disciplines, 

 methods and concepts, between analysis and policy design and between expert, 

 manager and policy maker. Equally important, we had learned how nui to bridge the 



85 



