> > Knowledge is Power > > What do you think knowledge is? (besides power) :-) > Dan M. > > you got me there, dan. i could have said > reason is the path to knowledge, and power, > but even that is often not true... Given the framework of the discussions we have had on electric cars, trains, and other forms of transportation, as well as the nature of economics, it seems that empirical knowledge is the type of knowledge that is worth referencing. Other forms of knowledge, such as self knowledge, is also useful, but I think I'm biting off quite a bit in considering empirical knowledge.
The easiest form of empirical knowledge to both attain and to illustrate, is scientific knowledge. There are two parts to the development of scientific knowledge. The first is the acquisition of good data. The second is the successful modeling of those data. The start of physics is a good example of this. The very start of it was the collection of data regarding the orbit of the planets. These data required more and more complicated systems of epicycles to depict the data. Copernicus suggested a heliocentered system, instead of the geocentered system of Ptolemaic.and it was simpler. While the Ptolemaic system required n epicycles, Copernicus could get by with n-1. So, there wasn't much in the way of an advantage for either system. If anything, Ptolemaic had a bit of an advantage, since he didn't have to explain why we didn't feel the earth move. Galileo collected additional data of many kinds. In particular, he was the first one I know of to do directed experimentation, instead of simply observing. His inclined plane work.his attempt to find the limit value of sliding blocks on inclined planes was critical. As was looking for implications of the Copernican vs. Ptolemaic system, and trying to determine which was superior by looking at the observations. Kepler made a critical advance by trying a different model than either Copernicus or Ptolemaic: he used ellipses instead of circles. They didn't require epicycles..they describe the orbits near perfectly. Newton, of course, pulled things all together. He had a system that explained both the motion of blocks on inclined planes, and gravitation. When his theory required planets to behave as though their mass was at a point at their center of mass, he developed calculus to be able to determine that this was a valid assumption for gravity. And, when asked how gravity worked, this "spooky" action at a distance, he simply said "I have no hypothesis." He successfully modeled observations, he didn't have to find the truth. About 10 years ago, a Dutch physicist (Patrick Van Esch) made a statement that summed this up: "The most important step in the development of science was when it was decided it wasn't about the truth." Now that I've wandered down old physics stories which you probably know too, let me use them to get to my points. In a sense, I'm re-expressing the scientific method.but in a way that can be useful when things are not as clear as they are in physics. The first thing we see is the utility of precise quantitative data. If there were not precise observations of the motions of the planets.there would be no difficulty with the old model. If Galileo didn't have a telescope, that allowed more precise determination of objects in the sky, he wouldn't have seen Saturn's rings, or Jupiter's planets. The second is the directed experiments. The nature of the direction is absolutely critical here. A good scientist does just find ways to support his/her hypothesis. Rather, they look for implications of the various hypotheses and try to find places where the various models produce different results. Let us first look at the verisimilitude of Aristotle's view. He stated that heavier objects fall faster, that objects tend to slow down and stop if there isn't something that keeps them moving along, and objects fall at a constant velocity. Well, just look. A rock falls faster than a feather. Arrows slow down in flight. Drop anything that falls slowly enough so that you can time the fall rate, and it falls at a constant velocity. But, there must have been exceptions to this general rule that were observable in the time of Galileo. So, he dropped lighter and heavier cannon balls, rolled balls down inclined planes, etc. He did this while using timing mechanism, like water clocks, possibly a pendulum. Doing this, he was able to verify general rules like momentum, constant acceleration due to gravity, etc. The exceptions could/can be explained by the resistance of air, friction, etc. The types of experiments he did were valuable in a two fold manner. First is the obvious, he falsified some of Aristotle's rules. Second, he was able to test the source of the differences. For example, sliding blocks on inclined planes go slower than rolling balls. Balls and blocks on oiled planes go faster. Changing the surface of the plane would also change the velocity. Thus, there is considerable evidence for friction. Also, changes in oiling and the surface of the plane do not change the nature of the sliding object. Further falling blocks and balls don't show the same differences in rate as sliding blocks and balls. The direction of the experiments should be clear here. They helped Galileo separate the various effects. His experiments tended to separate out different effects (e.g. momentum, gravity, friction) so he could look at them in isolation. One very good way to look at this is that he did limit value experiments..he couldn't eliminate friction, but he could reduce it and look at the limit value for motion without friction from the data he obtained with various degrees of friction. The third is the desire for generalization. There wasn't a separate rule for inclined planes, falling balls, arrows, etc. He tried for one rule to fit them all. Doing this doesn't help just because it's easier to remember one set of rules. Rather, it allows one set of experiments to inform the interpretation of other sets of experimentation. For example, if gravity were universal, the planets should behave as though there was no friction on them. The fourth is development of models that fit the data. For centuries, circles were associated with perfection, and thus heavenly objects must move in circles. For nearly a century after Copernicus, the geocentric solar system was very complicated. One way to see this is to try to fit a real elliptical orbit with circular motion..epicycles become de rigor. Kepler tried a different model, an ellipse. The data fit the ellipse well.not perfectly, but well. We can look at this as checking for the hidden assumptions that stop one from accurately modeling data. I've found in engineering that, once we found the solution, we could look back and see the hidden assumptions we made. Next, the time it took from Copernicus to Principia Mathematica (almost 150 years) helps us understand the best way of dealing with uncertainty, as well as later difficulties in matching the orbit of the moon to predictions obtained from Newton's laws gives a good feel for how one deals with contradicting observations and predictions. Finally, during this time, there were many experiments with significant errors.as there always are. One has to deal with contradicting evidence, while not knowing what the source of the differences is. I've seen this all the time.95% of data points fit, and 5% are outliers. One repeats the experiment, and fits the 95%. One has guesses for the source of the outlier, but not being able to repeat the experiment, one drops them. Sometimes, the outliers are actually important. There is a variable that's important, but which is not well addressed by the experimental protocol. So, one rejects outliers like this while modeling results, but keep them in the rejection bin. If there is a repeated pattern, outliers are removed from the rejection bin looked at again, and either considered, or returned to the bin. This has been longer than I intended, but forms a basis for what I wish to argue for. Next, I'd like to address the use of this in engineering, and finally the use in answering empirical questions in areas where good science just cannot be done. Dan M. _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
