How to handle uncertainty in the analysis of complex situations? How to juggle a constellation of alternative hypotheses? How to derive actionable advice to offer a harried decisionmaker? Technologists and software developers, bless them, don't have the area expertise to build the tools that the specialist-analyst needs — so the analyst had better learn to talk to the technologist.

Epistemology — the study of knowledge itself — is the true foundation of powerful information-handling tools. Substantive experts don't need to know the mathematics of uncertainty, but they do need to at least speak some of the language. Then they can guide the toolsmith-methodologist toward useful action. A basic vocabulary of these ideas is also a splendid way to lift a discussion up from the "yes it is" / "no it isn't" level to a higher-dimensional space — where all sides can see how the uncertainties and alternatives balance out. Moreover, knowing something about the underpinnings of knowledge gives the thinker a new stock of metaphors ... and thereby enhances the ability to formulate and solve tough problems.

There are a host of disciplines that help wrestle down uncertainty, including:

• Probability & Statistics — discrete & continuous probability distributions, means, standard deviations, dependent & independent variables, conditional probabilities, and error propagation
• Combinatorics — permutations, combinations, multivariate experiment design, clustering & similarity metrics, and heuristics for scenario generation
• Logic — deduction, induction, syllogisms, fuzzy logic, logic programming, and idea-mapping techniques to assist in structured argumentation
• Inverse Methods — reverse-engineering, matrix inversion, back-propagation, etc.; see Bypasses by Z. A. Melzak (enlightening reading on these and related topics at the graduate level ... but anyone should feel free to skip the equations and explore Melzak's ideas of metaphor and transformation in literature and language)
• Curve-fitting — data modeling, error estimation, and key variable identification; see Mathematical Methods That [Usually] Work by Forman J. Acton (ideas on numerical methods, readable at the advanced undergraduate level, but with many nonmathematical parables of universal relevance)
• Noise & Random Perturbations — power spectra, correlation functions, and pattern discovery in unclean data
• Game Theory — from rock-paper-scissors to Mutual Assured Destruction (MAD), two-person zero-sum & beyond, prisoners dilemmas, minimax, etc.; see The Compleat Strategyst by John Williams (highly entertaining, with many stories and funny illustrations; needs only high-school math or less)
• Information Theory — bits of data, entropy, & evidence; see The Recursive Universe by William Poundstone (fascinating popular-level exposition, with chapters on cellular automata, self-reproducing systems, and deep concepts of information)
• Systems Analysis — sources, sinks, valves, delays, positive & negative feedback loops, attractors & instabilities, critical paths & chokepoints; see The Fifth Discipline by Peter Senge (a self-improvement and applied math short-course disguised as a business book ... powerful and important concepts presented in engaging fashion)

This quick tour of the epistemological engine-room will help a captain pilot the ship of thought with more precision. A little understanding of how the machinery works will also assist in diagnosis and recovery when things go awry. The bottom line: clearer thinking about complex issues.

(The above snapshot of a work-in-progress is derived from a talk I gave today to a small class. More to come!)

Thursday, August 10, 2000 at 21:17:01 (EDT) = 2000-08-10

TopicThinking - TopicScience - TopicPhilosophy - TopicOrganizations

(correlates: Twitter Poetry, MinimaxStrategy, BooksToConsider, ...)