Читать книгу Geochemistry - William M. White - Страница 15

Science deals in only two quantities: observations and theories. The most basic of these is the observation. Measurements, data, analyses, and experiments are all observations in the present sense. An observation might be as simple as a measurement of the dip and strike of a rock formation or as complex as the electromagnetic spectrum of a star. Of course, it is possible to measure both the dip of rock strata and a stellar spectrum incorrectly. Before an observation becomes part of the body of scientific knowledge, we would like some reassurance that it is right. How can we tell whether observations are right or not? The most important way to verify an observation is to replicate it independently. In the strictest sense, independent means by a separate observer, team of observers, or laboratory, and preferably by a different technique or instrument. It is not practicable to replicate every observation in this manner, but critical observations, those which appear to be inconsistent with existing theories or which test the predictions of newly established ones should be, and generally are, replicated. But even replication does not guarantee that an observation is correct.

Observations form the basis of theories. Theories are also called models, hypotheses, or laws. Scientific understanding is achieved by constructing and modifying theories to explain observations. Theories are merely the products of the imagination of scientists, so we also need a method of sorting out “correct” theories from “incorrect” ones. Good theories not only explain existing observations, but also make predictions about the outcome of still unperformed experiments or observations. Theories are tested by performance of these experiments and comparison of the results with the predictions of the theory. If the predictions are correct, the theory is accepted and the phenomenon considered to be understood, at least until a new and different test is performed. If the predictions are incorrect, the theory is discarded or modified. When trying to explain a newly discovered phenomenon, scientists often reject many new theories before finding a satisfactory one. But long-standing theories that successfully explain a range of phenomena can often be modified without rejecting them entirely when they prove inconsistent with new observations. And as Carl Sagan once said, extraordinary claims require extraordinary proof.

Occasionally, new observations are so inconsistent with a well-established theory that it must be discarded entirely and a new one developed to replace it. Scientific “revolutions” occur when major theories are discarded in this manner. Rapid progress in understanding generally accompanies these revolutions. Such was the case in physics in the early twentieth century when the quantum and relativity theories supplanted Newtonian theories (Lindley, 2001). The development of plate tectonics in the 1960s and 1970s is an excellent example of a scientific revolution in which old theories were replaced by a new unifying one. A range of observations including the direction of motion along transform faults, the magnetic anomaly pattern on the sea floor, and the distribution of earthquakes and volcanoes were either not predicted by, or were inconsistent with, classical theories of the Earth. Plate tectonics explained all these and made a number of predictions, such as the age of the sea floor, that could be tested. Thus scientific understanding progresses through an endless cycle of observation, theory construction and modification, and prediction.

In this cycle, theories can achieve acceptance, but can never be proven correct, because we can never be sure that it will not fail some new, future test.

Quite often, it is possible to explain observations in more than one way. That being the case, we need a rule that tells us which theory to accept. When this occurs, the principle is that the theory that explains the greatest range of phenomena in the simplest manner is always preferred. For example, the motion of the Sun across the sky is quite simple and may be explained equally well by imagining that the Sun orbits the Earth as vice versa. Ancient astronomers could, for example, predict eclipses. However, the motions of the planets in the sky are quite complex and require a very complex theory if we assume that they orbit the Earth. If we theorize that the Earth and the other planets all orbit the Sun, the motions of the planets become simple elliptical orbits and can be explained by Newton's three laws of motion.† The geocentric theory was long ago replaced by the heliocentric one for precisely this reason. This principle of simplicity, or elegance, also applies to mathematics. Computer programmers call it the KISS (Keep It Simple, Stupid!) principle. In science, we call it parsimony, and can sum it up by saying: Don't make nature any more complex than it already is.