Trial and error
Experiments are essential to test theories. But how reliable is the empirical basis of science?
Precise observations are both the starting point and goal of many research projects. In the 19th century, for example, it was noticed that the planet Uranus does not orbit the sun perfectly according to Newton's laws. As a result, astronomers and mathematicians John Couch Adams and Urbain Le Verrier independently came up with an explanation: the orbit of Uranus is being disrupted by a new planet that is even farther from the Sun. They even calculated the trajectory of the unknown object, making their theory easily testable. If they're right, it should be possible to find the planet.
That's how it was. After a short search, the astronomer Johann Galle reported to Le Verrier in 1846: "The planet whose position you have calculated actually exists." Newton's celestial mechanics was not wrong, as one might have assumed from Uranus' orbital anomalies, but rather was impressively confirmed with the discovery of Neptune (however, it was later unable to explain an anomaly in Mercury - only Albert Einstein was able to do this, see Part 2 of the Series). The philosopher Karl Popper later saw it as a prime example of how science should work: researchers boldly formulate precise theories and subject them to rigorous testing.
But such a test, which decides the weal and woe of a theory, an "experimentum crucis", is rare…
