Introduction to the philosophy of science
Karl Popper and falsification: disproof is more important than proof
Is it ever possible to prove a theory in science?
Say I have a theory that the Earth always spins in the same direction. My theory predicts that every day the Sun will appear to rise in the east and set in the west. So I can test my theory by going outside and watching the Sun.
How many times do I have to watch the Sun rising and setting before I say that my theory is proved? How do I know that tomorrow the Sun won’t rise in the west and set in the east or not rise at all?
In the 1920s the philosopher Karl Popper tried to resolve this problem by saying that you can never prove a theory to be true. All you can ever do is fail to prove a theory false, for example by never observing the Sun to rise in the west and set in the east. If I woke up one morning and did see the Sun rising in the west I’d have to abandon my theory that the Earth always spins in the same direction.
Don't try and prove a theory, try and disprove it!
Popper suggested that scientists should try really hard to knock a theory down. The harder you try and the more you fail the ‘fitter’ the theory is.
Here are two theories.
- The Earth always spins in the same direction.
- The Earth always spins in the same direction and takes on average 24 hours to spin once.
If the Sun suddenly moved from west to east then both theories would be ‘falsified’, in other words proved false.
But the second theory can also be falsified by showing the length of a day was, say, thirteen hours or twenty-seven hours. Popper would say that the second theory was better because it was more falsifiable.
But scientists don't work by falsification
Popper thought scientific progress was made when scientists invented bold and highly falsifiable (i.e. mathematical) theories.
Remember our theory that cathode rays are electrically charged particles. This theory predicts that cathode rays should be bent when they pass between two electrically charged plates. When this didn’t happen did scientists immediately abandon the theory? No they didn’t!
Thomson’s experiment pre-dated Popper by several decades but the key thing is that the history of science is not about falsification. Real scientists simply did not behave in the way that Popper suggested they should.
When scientists couldn’t bend cathode rays with charged plates they weren’t sure whether it was just something to do with their apparatus or whether it really was the case that their theory was wrong.
The demarcation problem: separating science from non-science
But the idea of falsification can still be used as a way of deciding how ‘scientific’ an idea is. For example one reason why astrology is not scientific is that its predictions are normally unfalsifiable.
You might read this sort of thing in your horoscope. ‘The Moon is in Libra, making this an ideal time to express your opinions and develop your creative ideas.’ How could you show, even in principle, that this statement was false? Or how about this: ‘The Moon is in Libra, making this a really bad time to express your opinions and develop your creative ideas.’
How could you choose between the two? You can’t. Both statements are unfalsifiable. For a theory to be scientific you must be able to imagine an experiment or observation of nature that could test it.
Science is an attitude
If you define science as ‘what scientists do’ then how could you tell if someone is a scientist? One way might be that in the long run they are honest enough to accept that their theories may be shown to be wrong.
As the Nobel physicist, great science communicator and bongo player, Richard Feynman, put it:
'It does not make any difference how smart you are, who made the guess, or what his name is – if it disagrees with experiment it is wrong.'
Thomas Kuhn and scientific revolutions
So if the history of science has shown that science doesn’t progress by Popper’s falsification how does it progress?
In the 1960s Thomas Kuhn, a Harvard physicist and historian of science, introduced the idea of scientific revolutions. Science didn’t progress brick by brick with each theory building on the one before. Instead every so often a whole theory was torn down and completely replaced by a new one.
The new ‘paradigm’ (meaning way of looking at the world) may not contain any of the ideas of the old one. For example Newton’s laws about how and why things moved completely replaced Aristotle’s.
Newton saw straight-line motion as ‘natural’. Circular motion of the planets could be explained using the ideas of mass, gravity and force. Aristotle saw circular motion of the planets as natural and didn’t need explaining. There was no concept of gravity or mass.
Kuhn was interested in how communities of scientists adopted a new paradigm. He saw scientific revolutions as having parallels with political ones. Scientific history was re-written by the revolutionaries to make it appear that the new paradigm was inevitable.
After a new paradigm had been adopted there was enough stability for scientists to do ‘normal science’. ‘Normal science’ means doing the sort of meticulous and detailed experimentation that inches the unknown into the known.
In this sense John Townsend's experiment giving a very rough value for the mass of an electron could be seen as revolutionary because it lent fairly conclusive proof to the theory that atoms were made up of smaller parts. Whereas Millikan's oil drop experiment was really part of normal science because all he was doing was trying to get a better value for a figure who's approximate size was already known.
For Kuhn the ultimate test of a scientific theory was whether it was accepted by the relevant community of scientists.
Paul Feyerabend and scientific anarchy
This idea was taken to its logical conclusion by the Austrian philosopher, Paul Feyerabend.
For Feyerabend there was no ‘scientific method’ that could guarantee scientific progress. Feyerabend argued that it was unfair to assume that science was superior to, say, voodoo without seriously examining voodoo’s aims and methods. So Feyerabend and Kuhn placed greater stock in the people who held theories than in the theories themselves.
Imre Lakatos and scientific research programmes
The Hungarian philosopher, Imre Lakatos, tried to rescue the notion that scientific theories could be true, even if no-one believed them. He was a great friend of Paul Feyerabend's, even though they had very different ideas.
He extended Popper’s ideas about falsification with what he called ‘scientific research programmes’.
At first glance a research program looks a bit like what Kuhn called a paradigm. A research programme consists of a ‘hard core’ island of certainty from which scientists can set off and do research.
The hard core is protected from falsification by the assumption that it is experimental details that cause problems not the core itself. Lakatos proposed that research programmes could be judged by how many new facts they predicted. Eventually a growing ‘progressive’ research programme would replace a shrinking ‘degenerating’ one.
We have used Lakatos’s ideas as our model for how science works in these lessons.
Science is about agreeing ways to be wrong
So we’ve just had a very quick tour of some of the most influential philosophers of science. Ironically many scientists may not have heard of Popper, Kuhn, Feyerabend or Lakatos. They are quite happy just to ‘do science’ without necessarily worrying about the validity of the ‘scientific method’.
So is science any better than voodoo? If no one can agree on what science even is why shouldn’t you just believe anything? Granted, there may be no such thing as absolute scientific ‘truth’.
But there is a key difference between people who ‘believe’ in science and those who believe in religion, spiritualism, witchcraft and magic. That difference is that in the long run scientists are prepared to change their theory if experiment shows that it can no longer be sustained.