[The engineer] must identify the significant and critical problems, but in his education, problems have been predetermined and assigned. He must deve… - Jay Wright Forrester

It is to be hoped that those who believe they already have some different model that is more valid will present it in the same explicit detail, so that its assumptions and consequences can be examined and compared. To reject this model because of its shortcomings without offering concrete and tangible alternatives would be equivalent to asking that time be stopped. But the world will continue to turn. We always use the most acceptable model at any point in time. But how should we proceed so that the most acceptable model is also the best one that is available? We should try for three things. First, the best existing model should be identified at each point in time. Second, the best currently existing model should be used in preference to traditional models that may be less clear and less correct. Third, aggressive effort should be devoted to a continual improvement in the available models of the world system.

It seems traditional for explicit models of social systems to be greeted by vague criticisms about their lack of perfection. Instead, we need equally explicit alternatives with a demonstration that the alternative leads to a different and more plausible set of conclusions. By proposal and counter proposition our understanding of social systems can advance.

In spite of the tentative nature of the world model described here, various conclusions are drawn from it. Man acts at all times on the models he has available. Mental images are models. We are now using those mental models as a basis for action. Anyone who proposes a policy, law, or course of action is doing so on the basis of the model in which he, at that time, has the greatest confidence. Having defined with care the model contained herein, and having examined its dynamic behavior and implications, I have greater confidence in this world system model than in others that I now have available. Therefore, this is the model I should use for recommending actions. Those others who find this model more persuasive than the one they are now using presumably will wish to employ it until a better model becomes available.