ScienceVersusStampCollecting

 

Identifying and describing critical factors is a key step toward problem-solving, but it's only the first step. Next, one has to link those factors together and gingerly start to hook some numbers to them — relative weights, speeds, probabilities, and so forth. That's how to turn a handwaving exercise into a testable theory.

Malcolm Gladwell's recent book The Tipping Point: Why Small Changes Can Make a Big Difference is a good example. It's an engagingly fast read, smoothly-written, full of striking metaphors and case studies. (Some pieces of the book were published earlier as essays in the New Yorker.) In brief, Gladwell argues that many social phenomena — crime waves, fashion fads, teen suicides, success and failure of businesses — are (like the spread of disease) governed by nonlinear relationships, and so they can be profoundly affected by slight shifts in a few parameters. (Hard to disagree with that!) The book describes several elements which may be key to understanding social epidemics:

  • transmissive personality types:

  • "Connectors", who know and interact with lots of people * "Mavens", who gather, organize, and share information * "Salesmen", who communicate persuasively
    • * sticky ideas — things which are addictive, physically or psychologically, and which thus persist

    • small groups — ~150 people or fewer, the maximal size for continuous tight bonding and communication

    Good concepts, all.

    But The Tipping Point never quite gets past storytelling to the essential next stage of knowledge. To build a theory, one must:

    • link — take critical concepts and put them together into a system, with cause and effect relationships, feedback loops, sources and sinks, delays, etc.
    • quantify — assign numbers (if only tentatively) to nodes, coupling factors, and timescales
    • predict — apply the model to new situations and see how well it works
    • refine — modify and improve the theory based on experience

    This process doesn't have to be arcane or even very mathematical. But it does need to be done if there's going to be progress in understanding a situation. That's the difference between natural philosophy and philately.

    Tuesday, June 20, 2000 at 20:36:35 (EDT) = 2000-06-20

    TopicThinking - TopicScience - TopicPhilosophy - TopicSociety

    (correlates: TomorrowSinger, NanoMarkets, FifthDisciplinarians, ...)