How we Define Science

The word “science” derives from the Latin “sciens”, a substantive participle [verbal adjective] meaning “that which is known”. Hence one encounters terms like “political science” which use science in the sense of a field of study, but such terms should be viewed as hold-overs from previous usage in that the word “science” has come to have a more precise technical meaning. Specifically, science is the systematic explanation of observed, reproducible phenomena in terms of observed reproducible processes using the scientific method.

If a phenomenon has not been observed, one cannot attempt to scientifically describe it. For example, attempted descriptions of the nature of quantum gravity are regarded by physicists as currently speculative because so far no unequivocally quantum gravitational phenomena have been observed. As physicists, we strongly hope in future to either observe some form of quantum gravitational interaction or to eliminate the possibility of doing so, but this has not yet occurred. Similarly, laymen often talk about science “proving” either the existence or non-existence of G-d. Since G-d has no physical form however neither is possible because no specific physical phenomenon is associated with either the presence or absence of G-d. Moreover, the phenomena to which science can be applied must be reproducible so that one can test a proposed explanation. For example, some years ago, researchers Fleischmann and Pons claimed to have observed “cold fusion”, i.e., the process of nuclear fusion at a relatively low temperature, but ultimately the claim was rejected by the scientific community as a whole because the phenomenon was not unambiguously reproducible.

Similarly, once a physical phenomenon has been definitely observed, one attempts to explain that phenomenon scientifically by invocation of known processes. For example, if one finds an old coin in the carcass of a fish, a scientific hypothesis could not be suggestion that maybe the flesh of the fish magically transmuted into a coin because no one has ever observed such a process of transmutation to occur. If one speculates that a process occurs which has never been observed, one must first show that it happens before one can use it to explain other phenomena. An example in relatively recent scientific history is the case where Einstein first showed that mass and energy in nuclear reactions become interchangeable according to the formula E-mc^2 before he used that same formula to describe the role of energy in gravity.

Of course, the observation of a phenomenon and postulation of a process by which to explain it are not strictly separable. For the example above of establishing Einstein’s formula E=mc^2 for an interchange between mass and energy in the context of nuclear reactions, experiments first observed nuclear reactions in which mass and energy were not conserved separately. Then Einstein postulated a combination of mass and energy called mass-energy, to which mass m contributes an effective energy E=mc^2, in order to explain changes in the total mass in a reaction by means of mirrored changes in the total energy of that reaction. This explanation was then additionally tested after the fact, but the first two steps were virtually simultaneous. One must remember that the scientific method is a system of reasoning, not a cook-book procedure.

Nevertheless, the elements of use of the scientific method can be described as follows: 1. Given a particular observed phenomenon to be explained, one formulates a hypothesis, i.e., a possible explanation consistent with previous observations but as yet in principle untested. 2. One constructs and runs an experiment which unambiguously tests the hypothesis. Such an experiment must be reproducible by other experimenters using similar but different equipment and must eliminate ambiguities. Ideally, this latter point means that the experiment should involve the effect of variation of one and only one parameter on one and only one observed variable. In practice, one may not be able to experimentally effect only one variable but statistical methods when properly applied can allow one to construct a rigorous explanation of an observed variable or variables in terms of an experimentally modified variable or variables. One reproduces the same identical experiment repeatedly under identical circumstances. For each variable, the experiment should be run at least 100 times for qualitative results and 1000 for reliable quantitative results. Thus, for a one variable experiment, one must run the experiment one thousand times in order to reliably describe how the value of that variable depends on manipulation of the experimental parameter; for two variables, one must repeat the experiment a million times. This is why particle accelerators like those at CERN or Fermilab continuously run the same experiment over and over again thousands of times day, all day and all night, continuously sometimes even for years. 3. One compares the experimentally measured results with those predicted by the initial hypothesis. If the hypothesis is not consistent with the observed results, then that hypothesis must be discarded. If the hypothesis is consistent but only within strict limitations, then the hypothesis must be modified to apply within those limits. For example, Newton’s law of gravity only applies when the gravitational field is relatively weak; for example, it cannot explain the orbit of Mercury because that planet is so close to the sun that the gravitational effect of the sun’s energy cannot be neglected. Yet, Newton’s law works very well to calculate the orbits of bodies farther from the sun and so it is still used often scientifically- only under appropriate conditions.

“Science” is therefore not a set body of knowledge but a process of understanding the world. That process is never-ending so long as new things exist in the world to be observed.

1. For discussion of cold fusion: