Steve Humphrey swan dives into hypotheses

By Steve Humphrey
Illustration by Andrea Hutchinson

Science is the search for patterns and regularities in the external, objective, physical world, which are described by general propositions. Scientists construct hypotheses, or theories, as candidates for true generalizations. But how can we know that these hypotheses are true? That is, what justifies our belief that our theories are correct? We can’t examine every instance of some general claim to verify its truth. So, how do we know it’s true?

Now, much ink has been spilled in philosophical literature over the issue of just what constitutes knowledge. There are even skeptics who argue that we can never truly know anything. But I think it is safe here to rely upon our ordinary intuitions about knowledge, which is that knowledge consists of justified, true belief. The key, and most controversial, element of this analysis is the notion of justification. What sorts of things can provide justification sufficient for knowledge?

This issue was addressed explicitly during the 17th and 18th centuries, a time when philosophers were inventing the scientific method. (By this time, they had abandoned togas.) There were two major schools of thought: continental rationalism, whose primary figures were Descartes, Leibniz and Spinoza; and British empiricism, represented by Berkeley, Locke and Hume. The rationalists were inspired by the amazing success of Euclid’s geometry. Starting with a few simple, obvious postulates – and using deductive logic – Euclid was able to prove theorems about the geometrical properties of the physical world. Since deductive logic is “truth-preserving,” our confidence in the theorems depended only upon the truth of the postulates, or “axioms.” This is a “top-down” strategy. Start with obvious truths and deduce claims about the world.

The empiricists, on the other hand, had more of a “bottom-up” strategy. Their view was that sensory experience provided the foundation for the discovery of general laws. We see a large number of swans and notice that they are all white. We then construct the hypothesis that all swans are white and test that hypothesis by observing more swans. This kind of logic is called “inductive,” and our confidence in our conclusions may be stronger or weaker depending upon the evidence (i.e., how many swans we look at), but it lacks the certainty that accompanies deductive logic. (Eventually, black swans were discovered in Australia, which made this example less appropriate.)

This dispute was finally resolved by Isaac Newton, who, using data gathered by astronomers over the centuries, was able to construct a powerful theory that was very successful in predicting and explaining various gravitational phenomena. During the 19th century, mathematicians were able to show that the axiomatic methods of Euclid could yield different, non-Euclidean, geometries when one of the not-so-obvious postulates was altered. It was recognized then that the true geometry of the world would have to be established by empirical means and not simply through deduction. (Non-Euclidean geometry led directly to Einstein’s general theory of relativity, his theory of gravitation.)

So, science basically consists in constructing hypotheses, or theories, and then subjecting them to empirical testing by making predictions based upon those hypotheses and conducting experiments or observations to discover whether the predictions are true. Next time, I will discuss some of the challenges to theory confirmation. V

Steve Humphrey has a Ph.D in the history and philosophy of science, with a specialty in philosophy of physics. He teaches courses in these subjects at the University of California, Santa Barbara, and has taught them at the University of Louisville.