I have lost count of the times I have been told that science has proved something or other, including the absurd claim that ‘science has proved creationism’. Apart from the fact that creationism has been shown to be incompatible with biology, geology, genetics, as well as physics, science cannot prove anything. While this may come as a surprise to many people, perhaps even a few scientists, that is the way it is.
This where we bump into the difference between science and non-science, which is one of the major questions in the philosophy of science, and is termed the demarcation problem. While most people would have an understanding that physics, chemistry, biology are science, where do we draw the line? While most people realise that creationism and astrology are not science, it is unclear whether economics fits the bill? Does psychoanalysis fit the bill? While most philosophers of science can often, but not always, point out what is science and what is not, coming up with an overarching principle has been all but impossible. This has played on the minds of philosophers of science for a long time and it was Karl Popper who came up with one of the most popular, if not universally accepted solutions. That solution is the concept of falsifiability. This states that no number of experiments or observations can ever prove a theory, but one experiment or observation can refute a theory.
A scientific hypothesis is the product of human imagination to explain a series of experimental results or observations. It is retained as a working hypothesis if it is shown to be consistent with available experimental results or observations. If so, then it enables scientists to make predictions about further experimental results or observations. This is a constant process of testing that all hypotheses undergo. If these further experimental results and observations are consistent with the hypothesis then the hypothesis is retained. If these are not consistent with the hypothesis, then it is likely that the hypothesis will be chucked out or at least modified (if that is possible).
An example of this was Einstein’s 1916* theory of General Relativity replacing that of Newton’s 1687 theory of Universal Gravitation. Although Newton’s theory was a very close approximation for the understanding of the movement and forces acting on masses in the laboratory, it also gave a reasonable approximation of the movement of most of the planets given the ability to determine these in Newton’s time. As measurement became more precise, it showed that there were some small discrepancies between Newton’s equations and the movement of the planets, especially in the case of the innermost planet, Mercury2.
Mercury has a very eccentric elliptical orbit around the sun, such that the difference between its most distant (aphelion) and closest (perihelion) position relative to the Sun is fairly large. Its orbit precesses, and this means that its closest approach to the Sun (its perihelion) shifts forward with each orbit. Newton’s theory predicted a precession of only half of what had been observed, and this had been known since the 1700s. The predictions from Einstein’s General Relativity matched the measured precession of Mercury’s orbit exactly3. This was a major success for Einstein’s theory.
Einstein’s theory also predicted that the angular deflection of light passing near the sun would be twice that predicted by Newton’s theory. To measure the angular deflection of light around the Sun, a total eclipse is needed so the moon blocks out the brilliance of the sun and will allow the perceived position of stars behind the sun to be seen. That eclipse happened in 1919 and British astronomer Arthur Eddington travelled to the island of Príncipe off the coast of West Africa, while astronomer Andrew Crommelin travelled to Sobral in northern Brazil, both to measure this deflection.
During this eclipse, the Sun sat in front of the Hyades, a cluster of bright stars in the Milky Way, in the constellation of Taurus. During the total eclipse, many stars would be visible near the eclipsed disc. This was key because the deflection of light predicted is greatest for stars observed close to the Sun. The stars’ positions relative to the Sun was recorded and measured on photographic plates, and then compared with reference plates showing the stars when the Sun was nowhere near the field of view. Any apparent shifts, caused by the Sun’s gravitational field, could then be calculated. The more stars measured, the better the chance the observers would have of correcting for any systematic or random errors. Despite a number of difficulties experienced by both astronomers, including poor weather and equipment malfunctions, they determined that the deflection of light was much closer to that predicted by Einstein rather than that predicted by Newton. The results were delivered at a joint meeting of the Royal Society and Royal Astronomical Society in London and it made front-page news around the world4. Einstein became supremely famous and in the public’s mind, the epitome of genius. The deflection of light around the sun has been confirmed subsequently by numerous observations centred around total eclipses. It keeps getting tested.
To get back to my point of science not being able to prove anything; I think Einstein put it well when he said: “No amount of experimentation can ever prove me right; a single experiment can prove me wrong.” If these eclipse observations by Eddington and Crommelin had not turned out to be as Einstein had predicted, then General Relativity would have perhaps faded into obscurity as another failed hypothesis. However, it didn’t.
*Sometimes you will see the date of publication of General Relativity given as 1915, and sometimes as 1916. It was delivered in a presentation to the Prussian Academy of Science in November, 1915, but did not appear in print until June, 1916.
- Popper, K., 1959. The logic of scientific discovery. Hutchinson, London.