The Problem of Empirical Equivalence

Despite the logical flaws we can now see with these definitions of science, the idea that science is reliable because it is tested against data is still a very powerful one. If there are two competing theories, we test them to see which one makes the correct predictions. So when Galileo looked through his telescope at the moons of Jupiter, he was testing two theories: one that says everything revolves around the Earth and his own theory that denied this. The moons of Jupiter were seen to revolve around Jupiter, not Earth, so Galileo claimed his theory was a better one than the old. This way of determining between theories seems self-evident to us: see which one works.

Nevertheless, this idea of testing theories to verify them assumes that the theories make different predictions. What if there are two theories that are equally elegant and equally explain the data but make the same predictions? How can you tell which one is true? When the method rests on testing predictions against empirical data, there is no way of determining between them. You can only hope that one day, with further research, some situation will arise in which the two theories will make different predictions. However, this raises another problem for faith in science. If there is only one existing theory whose predictions fare well against tested data, how can you be sure that there is not another equally good (or better) one that has not yet been thought of? Science is limited by the ingenuity of scientists. If no one has thought of the true theory, then no one knows the truth. How can we ever be sure that the true theory is not yet discovered? How can we be sure that someone will think of it?³⁵

The more confirming evidence there is, the more confident we can be that a theory is a good one, but it still might be wrong. Scientific theories come and go. Many scientific theories were successful in their day (i.e., they explained the observable phenomena and made confirmed predictions) but have been discarded. Medieval astronomy held that there were crystalline spheres carrying the planets around the earth. Medicine used to be based on four “humors,” the right balance of which ensured good health. Chemistry used to include a substance called “phlogiston,” whose effects are now explained by oxygen. Heat gain or loss was explained by a gain or loss of “caloric,” now thought not to exist. Electromagnetic waves were thought to move through the “aether.” Because most people are not very well educated in the history of science (or only know of its heroic “successes”), it is easy to lose sight of just how often scientific theories have changed. There is little, if anything, of the great discoveries of the scientific revolution that still survive unchanged. Science is always revisable, and it is very frequently revised. It is because science can always be revised in the light of new data that we have any confidence in it at all.³⁶

Until we know everything in the universe, we will not know whether our science is right. What is more, given our current limitations as human beings, we would not know if we knew everything in the universe. We could not be sure there was not more to know. As things stand now, the very way in which science proceeds (i.e., by testing theories against observational data) limits the amount we can know from science. There may always be something else we have not yet thought of.

Science and Community

We have examined the most basic and common understanding of how to define “science”: science is an empirical method of studying the world. But there is another definition: science is the people who do it. This may sound strange. Surely we should say that “scientists” is the word for the people who do it? That is true, but it is also a common usage for “science” to mean the people, too. Take for instance the simple phrase “science proves.” This is an odd phrase, for what is the subject that is doing the proving? It has to be a person or a group of people. Science does not prove anything; people do. If we are going to understand what “science” is, we must take into account that people do it.

Doing this, however, complicates the picture somewhat. When we move from purely methodological descriptions of what science “is” or “ought to be” to the real world of what scientists actually do, the picture becomes somewhat more messy. The world of science is not a matter of obedient robots going about following logical rules. It is a world of real people who can be creative or dull, law-abiding or unruly, innovative or boring, honest or dishonest, career-minded or dedicated to the pure pursuit of knowledge, and often all of these at different times. Whatever may be the logical niceties of scientific methodology, they are embodied in an industry of funded research, peer review, and journal publication. Working scientists have many complaints about their system. Peer review can put the fate of your research into the hands of a rival. The science that succeeds has to be first of all the science that is funded, and that itself is often decided by people with vested interests. Young scientists find it hard to get jobs, old scientists hoard power, and everyone complains about the facilities.

It was by comparing traditional theories of science with the real world of how scientific discoveries have happened that several theorists took a new approach to the philosophy of science.³⁷ It is not appropriate, they said, to write about the “logic” of theories, as if they exist as complete, formulated entities. Theories develop in a process of trial and error, intuitive leaps, discussion, and thought within a scientific community. Instead of looking at scientific theories in the abstract and trying to work out what the ideal theory “should” be, philosophers of science should instead be looking at the way scientists interact and come to new ideas, how their training affects their conclusions, and so on.

Pushing this to the limits, there has been of recent years a strand of philosophy of science that has focused on the sociological dimension of science. Some writers have concluded that science does not really deserve to make claims about “truth,” given what a thoroughly human and political enterprise it is. It is not just the darker side of human nature that promotes such skepticism about the reliability of scientific knowledge. Further investigation into science in practice can be confusing when it comes to establishing the criteria for truth. The difficulty in establishing certainty in scientific practice has led some authors to the extremes of denying that scientific knowledge is “knowledge” in the generally accepted sense at all. Rather, it is merely a social construction. A “fact,” then, is simply a statement that enough important people agree upon.

The most radical skeptical theories about science are not generally accepted. Between the two extremes (naive acceptance that science reveals facts that can be proved true on one side and total relativism on the other) we need to find a reasonable account of science. Science works on a day-to-day basis in a very organized, orderly, and rational manner. Results are tested against new data, and theories are subject to independent verification. What is the point, then, of recognizing the social aspect of science? It is that in order to test results at all, criteria must be set up that decide what constitutes a reasonable test. When theories are tested against data, someone has to decide which data are relevant. Someone has to decide which kinds of results are the most important. In the end, just saying “it works” is not good enough. It works in what context? Ptolemaic astronomy, which assumed stars, planets, and the sun revolved around the earth, worked very well for drawing up navigational charts. It accurately predicted positions of the planets and events like eclipses. When the theory was overthrown, it was not because the new one worked better in these ways. Rather, new criteria (e.g., mathematical simplicity) had become important.

Science does work, given these sensible qualifications of what “work” means. It works not just in its technological applications, but in gathering a remarkably complex, comprehensive and interlocking way of understanding the universe. Science achieves what western civilization wants it to achieve. What the analysis of scientific epistemology does teach us is that to claim that science can give us certain knowledge about everything, as popular rhetoric has claimed since the seventeenth century, is rather overstating the case. We must get away from naive statements that science “proves” this or that. It is not just overambitious but practically foolhardy to expect absolute certain proof of theories. Scientific conclusions are always tentative and revisable. This is the very strength of science.

For in keeping its theories tentative and revisable, science is able to retain some hope of discovering something about reality. It is not the grand claim of certain knowledge. It is a much more limited ambition, which many working scientists share. Science is a practical method for finding things out that involves hunches, rule-of-thumb heuristics, and training in the craft of research. It has its flaws, but as a general way of proceeding, it works rather well.