Home Sociology Understanding Society and Natural Resources
A Pragmatic Modeling Philosophy
Practical problem solving requires the integration of three elements: (1) creation of a shared vision of both how the world works and how we would like the world to be;
(2) systematic analysis appropriate to and consistent with the vision; and (3) implementation appropriate to the vision. Scientists generally focus on only the second of these steps, but integrating all three is essential to both good science and effective management. “Subjective” values enter in the “vision” element, both in terms of the formation of broad social goals and in the creation of a “pre-analytic vision” which necessarily precedes any form of scientific analysis. Because of this need for vision, completely “objective” scientific analysis is impossible. In the words of Joseph Schumpeter (1954, p. 41):
“In practice we all start our own research from the work of our predecessors, that is, we hardly ever start from scratch. But suppose we did start from scratch, what are the steps we should have to take? Obviously, in order to be able to posit to ourselves any problems at all, we should first have to visualize a distinct set of coherent phenomena as a worthwhile object of our analytic effort. In other words, analytic effort is of necessity preceded by a preanalytic cognitive act that supplies the raw material for the analytic effort. In this book, this preanalytic cognitive act will be called Vision. It is interesting to note that vision of this kind not only must precede historically the emergence of analytic effort in any field, but also may reenter the history of every established science each time somebody teaches us to see things in a light of which the source is not to be found in the facts, methods, and results of the preexisting state of the science.”
Nevertheless, it is possible to separate the process into the more subjective (or normative) envisioning component, and the more systematic, less subjective analysis component (which is based on the vision). “Good science” can do no better than to be clear about its underlying pre-analytic vision, and to do analysis that is consistent with that vision.
The task would be simpler if the vision of science were static and unchanging. But as the quote from Schumpeter above makes clear, this vision is itself changing and evolving as we learn more. This does not invalidate science as some deconstructionists would have it. Quite the contrary, by being explicit about its underlying pre-analytic vision, science can enhance its honesty and thereby its credibility. This credibility is a result of honest exposure and discussion of the underlying process and its inherent subjective elements, and a constant pragmatic testing of the results against real world problems, rather than by appeal to a non-existent objectivity.
The pre-analytic vision of science is changing from the “logical positivist” view (which holds that science can discover ultimate “truth” by falsification of hypothesis) to a more pragmatic view that recognizes that we do not have access to any ultimate, universal truths, but only to useful abstract representations (models) of small parts of the world. Science, in both the logical positivist and in this new “pragmatic modeling” vision, works by building models and testing them. But the new vision recognizes that the tests are rarely, if ever, conclusive (especially in the life sciences and the social sciences), the models can only apply to a limited part of the real world, and the ultimate goal is therefore not “truth” but quality and utility. In the words of William Deming “All models are wrong, but some models are useful” (McCoy 1994). The goal of science is then the creation of useful models whose utility and quality can be tested against real world applications. The criteria by which one judges the utility and quality of models are themselves social constructs that evolve over time. There is, however, fairly broad and consistent consensus in the peer community of scientists about what these criteria are. They include: (1) testablity; (2) repeatability; (3) predictability; and (4) simplicity (i.e. Occam's razor – the model should be as simple as possible – but no simpler!). But, because of the nature of real world problems, there are many applications for which some of these criteria are difficult or impossible to apply. These applications may nevertheless still be judged as “good science”. For example, some purely theoretical models are not directly “testable” – but they may provide a fertile ground for thought and debate and lead to more explicit models which are testable. Likewise, field studies of watersheds are not, strictly speaking, repeatable because no two watersheds are identical. But there is much we can learn from field studies that can be applied to other watersheds and tested against the other criteria of predictability and simplicity. How simple a model can be depends on the questions being asked. If we ask a more complex or more detailed question, the model will probably have to be more complex and detailed. Complex problems require “complex hypotheses” in the form of models. These complex models are always “false” in the sense that they can never match reality exactly. As science progresses and the range of applications expands, the criteria by which utility and quality are judged must also change and adapt to the changing
|< Prev||CONTENTS||Next >|