BELIEVE   ME   NOT!    - -     A   SKEPTIC's   GUIDE  

Empiricism

We are often led to believe that new theories are derived in order to explain fresh data. In actuality this is never the case. Theories are proposed to explain experimental results, which are always reported in an intermediate state of digestion somewhere between the raw data and the general explanatory theory. Data are merely meaningless bits of information and are often disregarded entirely unless and until their custodian (usually the Experimenter who collected them) translates them into some empirical shorthand that allows their essential features to be easily appreciated by other people. This is not always a simple task. Kepler, for instance, accumulated a large body of information in the form of observations of the positions of planets and stars as a function of time. In that form the data were incomprehensible to anyone, including Kepler. First he had to extract the interesting part, namely the positions of the planets relative to the Sun, from raw data complicated by the uninteresting effects of the Earth's rotation and its own annual trip around Sol, which required both a good model of what was basically going on and a lot of difficult calculations. Then, with these "reduced" data in hand, he had to draw pictures, plot different combinations of the variables against each other, and generally mull over the data (presumably scratching his head and thinking, "Now what the hell does this mean?" or his contemporary equivalent) until he began to notice some interesting empirical generalizations that could be made about his results. Of course I don't know exactly how Kepler went about this, but I do know the experience of turning new data over and over in my mind and on paper until some consistent empirical relationship between the variables "leaps out at me." And I am very impressed with the depth and delicacy of Kepler's observations.

Note that the Empiricist^10.4 has not explained the observed behaviour at this point, merely described it.^10.5 But a good description goes a long way! One should never underestimate the importance of this intermediate step in experimental science.