The ability to anticipate a pattern is not the ability to understand its root causes. Both are useful but they require different skills. Science starts by observing, from careful observation across time or many similar systems, patterns emerge. Pattern recognition is the beginning of understanding because it can be used to predict unobserved things.
But pattern recognition has its limits. Pattern recognition is much better at interpolation than extrapolation. It also can result in absolutely weird, but useful ways of dealing with life. Anyone who understands computers knows this when their loved ones explain how they use the danged things. Beyond the amusing and ultimately non functional (A distraught Ms. Rabett on the line for you Professor) pattern recognition can also give you major investments into cold fusion and hydrino science, and even worse a refusal to deal with the viral origin of HIV. Engineers and amateurs are unfortunately prone to such enthusiasms. Give them credit the engineers have invented some really cool tools that empower pattern recognition.
Still, over time the organizing principles of the system become clear, either through observation or insights gained in other areas. At that point someone (aka Einstein, Newton, Dalton) states what the basic principles are and shows how you can both understand what is happening and predict what will happen in the future. To do this may require new tools, including mathematical ones such as calculus. You can go beyond theory and become entrapped in the beauty of those simple ideas that explain nothing, thus string theory. The best physicists (ask them) have become entangled in that ball of universes.
The shift from observation to pattern recognition and that to basic principles are the tipping points in science. The transition from science to beauty is the tripping point to metaphysics.
Different scientific fields have encountered these transitions at different times. The earliest were physical sciences: mechanics, electricity and magnetism. Kepler, Gallileo and others were the pattern recognition forerunners for Newton's synthesis. Faraday, Ampere and Co. did the dirty work for Maxwell. Physics is subject to drastic simplification and was best amenable to this type of change. Chemistry and biology were both pattern recognition sciences until very recently, chemistry until ~1980 and biology just now. Their continuing shift is driven by the explosion of computational capacity.
Climate is also at the boundary point between pattern recognition and theory. Here (and in chemistry and physics) a new wrinkle is being added by the availability of computers: the coupling of pattern recognition to improve theory. Meteorology has been a leader in this. Weather forcasting starts with mathematical theoretical models, but assimilates observational patterns into the mix to provide longer range forecasts.
Much of the venom in the climate wars occurs on the boundary between basic theory and pattern recognition.
Sunday, October 29, 2006