obnoxio@BRAHMS.BERKELEY.EDU.UUCP (08/04/87)
In article <1731@tekig5.TEK.COM>, chrisa@tekig5 (Omega) writes: > A scientist, for example, may believe that a certain phenomena can >be explained by a certain theory because past experience tells him that >it 'sounds' right. However, a good scientist will not commit himself >to that theory until there has been sufficient experimental proof of >it. Be careful: you are walking across a very difficult philosophical mine- field. Does such a thing as "sufficient experimental proof" ever exist? Does it matter whether a scientist is right or wrong? That is, I'll agree that a *careful* or *cautious* scientist avoids committing himself, but I don't think that always makes for "good" science. > For example, the theory of relativity explained a lot of peculiar >phenomena (the orbit of Mercury for one), but it would not be accepted >until it was sufficiently proved in experiments. Ah, one of my favorite examples of how interesting science gets vis-a-vis "truth". General relativity (GR) was Einstein's greatest achievement. As a theory, it was discovered perhaps 30 years ahead of its time, solely because of Einstein's personal determination to get a relativistic theory of grav- ity. The experimental evidence in its favor was extremely meager at the time. The only thing that was firmly correct was the Mercury perihelion precession. The light bending was, in the popular press, the dramatic clincher, but in actuality the always difficult measurements of it made for wildly varying measurements over the decades. Not the sort of stuff that makes for conviction. What is fascinating is that about 10-15 years or so before GR was found, an obscure German physicist came up with his own unusual gravitomagnetic theory, that languished in obscurity until after GR, and was only then noticed to give the exact same formula for perihelion precession as GR. This theory became most popular among anti-Semitic physicists who were more than happy to reject anything of Einstein's. This theory, if made coherent with SR, then led to a modified (== incorrect) prediction. There were also other theories, geometric like Einstein's, that were known to agree with GR on the major tests. Moreover, there was one prediction that GR apparently got *wrong* at the time. (Something about Venus--it's in Eddington.) Over the years, it was finally realized that the experimental figures were incorrect, but at the beginning, this discrepancy was something of a nuisance. Given the weakness and confusion of the experimental evidence at the time, and the number of competing theories, GR's wide acceptance can seem rather astonishing at times. This acceptance relied primarily on Einstein's pre- eminence, the intrinsic simplicity of the theory (something that was not really appreciated until the sixties), and probably most importantly, the massive irrelevance of the theory. It couldn't *matter* to scientists, on a practical level, whether it was true or not! All this changed in the sixties. Both theory and experiment exploded, and nowadays GR is considered as well-established as Newton's theory ever had been. ucbvax!brahms!weemba Matthew P Wiener/Brahms Gang/Berkeley CA 94720
sarge@thirdi.UUCP (Sarge Gerbode) (08/05/87)
In article <8708040902.AA21998@brahms.Berkeley.EDU> obnoxio@brahms.berkeley.edu (Obnoxious Math Grad Student) writes: >In article <1731@tekig5.TEK.COM>, chrisa@tekig5 (Omega) writes: >> A scientist, for example, may believe that a certain phenomena can >>be explained by a certain theory because past experience tells him that >>it 'sounds' right. However, a good scientist will not commit himself >>to that theory until there has been sufficient experimental proof of >>it. > >Be careful: you are walking across a very difficult philosophical mine- >field. Does such a thing as "sufficient experimental proof" ever exist? >Does it matter whether a scientist is right or wrong? That is, I'll >agree that a *careful* or *cautious* scientist avoids committing himself, >but I don't think that always makes for "good" science. > When you talk about something being "sufficient", you must answer "sufficient for what?". If a scientist believes in a particular theory, then clearly any evidence (not necessarily *experimental*, by the way, as in astronomy) there is was *sufficient* to convince him. The same evidence may not be sufficient to convince another, or to convince most scientists in related fields. Sufficient truth does exist, when you specify who is to be convinced. A child (or a student just starting in a subject) requires little or no proof. An extreme skeptic requires a great deal of proof. A bigot could be said to require an infinite amount of truth. The criterion of sufficiency of truth that makes the most sense in order for something to be widely accepted in scientific circles is just this: "Sufficient to convince most scientists". What this specifically amounts to is what most scientists currently find convincing, whatever that happens to be at any particular point in the history of science. A "good" scientist is one who will ultimately be most convincing to fellow scientists, the one who gets the last laugh, so to speak. -- "Absolute knowledge means never having to change your mind." Sarge Gerbode Institute for Research in Metapsychology 950 Guinda St. Palo Alto, CA 94301 UUCP: pyramid!thirdi!sarge