Science is distinguished from pure math in that it yields predictive information. The only publicly endorsed methods of evaluating such information are statistical. However, many consumers of science, including politicians and business leaders, face decisions for which it would be impractical or impossible to gather justification solely through such methods. For example, it would be unethical to subject humans to properly controlled social experiments, and statistical models won't converge for such chaotic systems anyway. Thus, people are forced to base some decisions, at least in part, on beliefs which haven't been scientifically proven.

    We can reduce the need for such arbitrary decision-making by establishing standards for alternate methods of identifying intelligent predictions. In 1999, I posted the book you see here proposing a set of such standards for a step-by-step method called "predictive deduction". I later added supporting essays for specialized audiences ("Adapting Formal Logic to Selection of Beliefs About the External World", "A Use of Predictive Deduction in Management Consulting" and "Some Hidden Assumptions of Statistical Projection"). In 2002, I developed a more rigorous formalism for predictive deduction and, with the help of Anand Chhatpar, started the Organization for Collaborative Leadership organization, an open community that improves decision-making and strategy world-wide by developing, explicating, and/or evaluating practical predictive logics, software, formal methods, and expectations about the future. My most recent material (along with that of others in the community) can be found there. 

    Many scientists already use predictive deduction subconsciously (much as computer operators use software without bothering to read its code). They classify it among procedures which, as a group, are called "intuition", an old and useful pillar of science. Formal predictive deduction differs from intuition in that its users can explain how they reached their conclusions. This allows peers to point out specific oversights when the method is misapplied, oversights that can be corrected while preserving the remaining valid elements of the prediction. Thus, the formal use of predictive deduction converges on accurate predictions, even without the help of auxiliary confirmation (i.e. as provided by experimentation).

[TABLE of CONTENTS] [SEARCH] [PRINT] [REVIEWS]
[CONTACTING THE AUTHOR] [YOU CAN HELP]

Copyright 1998-1999, Chris Lang
Last updated October 16, 2000