stern@bnl.UUCP (Eric Stern) (04/04/85)
With the reappearence of the old debate on whether all those who question scientific theory should be heard out or not, I thought I would clarify the situation of how theories get to become accepted, and why the current system acts to prevent the acceptance of false opinion and dogma. Along the way, I will discuss the related issues of experi- mentation and theories, and peer review of scientific work. In general, theories do not develop in a vacuum. There is usually a body of knowledge that requires explana- tion. This body of knowledge is acquired through experimen- tation. What experimenters do is to measure parameters of phenomena. Sometimes, they are verifying existing theories, or searching for deviations from theories, but sometimes they also perform searches for new phenomena independent of any theories. This provides the grist for the mill of new theories. When they perform experiments and make measurements, the answer they come up with is subject to independent verification. An experiment that nobody can reproduce is not considered to be valid. This is happening all the time. For instance, an experiment has been done at Stanford that seems to observe fractional charge, by a method based on the Millikan oil drop experiment, only this one uses charged niobium spheres. The result would seem to indicate the existence of free quarks, which have not been seen else- where. In fact, the Stanford experiment has been repeated at other institutions, with a null result. Because the Stanford result is not reproducible, free quarks are con- sidered to be unobservable. The result of an experiment is expected to include an experimental error. The experimenter estimates the magni- tude of deviation from the quoted value that he or she con- siders possible, given the nature of the measuring apparatus and procedure. The experimenter cannot cheat and quote a error that is too small, lest a future experiment make a measurement which does not agree within the error limits. If the error quoted is too large, to insure that future experiments will agree, the experiment is considered to be worthless in the eyes of the scientific community. After all, if there is such a large error on the measurement, they think, it is no better than no measurement at all. Once several experiments agree on a phenomenon, those measure- ments enter the general body of knowledge and become candi- dates for theoretical explanation. So, Joe Experimentalist discovers a new phenomenon and gets a Nobel Prize. There is still this matter of an unex- plained observation to deal with. How does one go about cooking up a theory? One cannot make up an arbitrary theory, there are guidelines that should be observed. A new theory can't directly invalidate an old theory. The old theory did perfectly well at explaining its set of phenomena. The new theory must include the old theory as a special case; it must reduce to the old theory in the limit of a regime where the old theory held. This point must be stressed. Old theories don't die, they just become incor- porated into new ones. The two great new theories of this century, special relativity and quantum mechanics, conform to this principle. Classical Newtonian Mechanics is the limit of special relativity considered at velocities small compared to the speed of light. Quantum mechanics reduces to classical mechanics in the limit of large quantum numbers. The two examples above illustrate something in physics, that, while is not a fixed law, is a general principle of operation. That is, new phenomena occur when new regimes are looked at for the first time. Quantum effects become noticable on the microscopic scale, and special relativity becomes apparent at high velocities. Another good thing to have in a new theory is general- ity or universality. Physicists hate special cases, and find the idea of a theory that just applies to one thing, to be morally repugnant. Good theories should explain many different kinds of effects without making special assump- tions about each one. Proliferation of theories and of assumptions is something to avoid. The best clue a physi- cist has, that a theory is not fundamental, or just not working, is when new assumptions, new principles, and new laws have to be continually added to a theory to make it fit the data. Let us look at this in practice. The geocentric views of planetary motion were very complicated indeed. The planets moved around the sky, but periodically, they would reverse their motion. Furthermore, each planet had a dif- ferent manner of behavior that varied with the time of year. Then Kepler came along. Kepler showed that from a heliocen- tric point of view, the sky is very simple. The apparent oddities of planetary motion arise from the motion of our own viewpoint. All objects orbiting the sun share a common pattern, a relationship between the orbit period, and the orbit distance that holds for all planets. Another example of universality relates to chemistry. In the 1700's and 1800's, many new chemical elements were discovered. They all seemed to have different chemical pro- perties. Mendeleev arranged the elements in a table, ordered by atomic weight, and found a repeating pattern. Elements separated by a certain amount in the table had similar chemical properties. Later on, it was determined that atoms consist of smaller particles, and it was the con- figuration of these constituent particles that determined the chemical behavior of the atom. But, the clue that something funny was going on was the proliferation of the number of elements, and the discovery of a pattern made it obvious that there was an underlying principle. The story is not over when a successful theory is found. The new theory continues to undergo tests to deter- mine its applicability in new situations. Even well esta- blished theories are tested anew, when improvements in experimental technology allow more sensitive measurements of a theory's validity. The Coloumb inverse square law for electric and magnetic fields, which was discovered about 1770 are tested as better techniques become available. A. S. Goldhaber wrote a Scientific American article in 1979 describing experiments which verify the inverse square law by observing the Earth's magnetic field. An observed devia- tion from the inverse law would imply a non-zero photon mass, which would have many consequences. An established theory is used as the basis to build more theories. Physics advances by building new structures on top of old ones. Conversely, new theories depend on the validity of previous theories. Every new theory verified acts to confirm the theories upon which it is built. With that preface out of the way, we can ask the real question that prompted this article: what are the odds that someone working independently of the "scientific establish- ment" discovers something wrong in a basic theory, such as classical electrodynamics? Well, it has to be pretty small odds. The "scientific establishment" has been working for over 100 years testing its own basic premises. If something was wrong with a basic theory, we would be seeing incon- sistencies in later work based on those same basic theories. In fact, something wrong *was* discovered in about 1890. Classical mechanics seemed to be inconsistent with electro- dynamics. That was just one problem. Radiation calcula- tions based on electrodynamics indicated that all the elec- trons orbiting nuclei would radiate all their energy and crash into the nucleus. This was the great "Ultraviolet Catastrophe", so named because the radiation given off would be high frequency, low wavelength, hence ultraviolet. The resolution of these two problems eventually led to the development of special relativity, quantum mechanics, and finally quantum electrodynamics. As it happens, the most precisely verified quantity known is the gyromagnetic ratio of the electron. This property of the electron has been measured to over eight significant figures and found to agree with the value predicted by quantum electrodynamics. The correctness of quantum electrodynamics implies the vali- dity of almost every previous theory including classical mechanics, classical electrodynamics, quantum mechanics and special relativity. So it seems, if past experience is any guide, physics is quite capable of recognizing its own prob- lems and dealing with them. Current physical theory is the result of several hun- dreds of years of work by people who didn't necessarily like each other, and includes results of thousands of experiments verifying the theories. Because of all this work, a physi- cist feels confident in rejecting the assertions of someone who gets the idea that physics is wrong. The views of those who feel that rejecting such notions is equivalent to the treatment given Gallileo and Capernicus may be answered by noting that prior to these visionaries, there was no esta- blished science. At that time, ideas of the physical world were dictated by preconceived notions having no basis in observation. Physics, in the final analysis is totally based upon experimental observations. There is no accepted physical theory which can't be demonstrated to be correct in the laboratory. Given the large body of knowledge that has been collected about the laws of the universe, and the demonstra- tion by several independent experimenters of the validity of the procedures used to accumulate this knowledge, the ines- capable conclusion is that the theories which have been developed to explain the observations must reflect the true behavior of the universe. Eric G. Stern Dept. of Physics SUNY StonyBrook StonyBrook, NY 11794 stern@bnl.arpa stern@bnl.bitnet ...!philabs!sbcs!bnl!stern