Pages

Saturday, February 20, 2021

Karl Popper and the Falsification Criterion

Charlotte Sleigh at Aeon has an interesting piece on philosopher of science Karl Popper (1902-1994), at least it's interesting until it becomes clear that Ms. Sleigh has no time for Popper's greatest contribution to his discipline.

Popper, in his book The Logic of Scientific Discovery (1935), offered a solution to what's called the demarcation problem. This is the problem of distinguishing between what constitutes valid science and what Popper calls pseudo-science. His solution is called the falsification criterion.

What he means by this is that to be considered a legitimate scientific theory there must be some conceivable set of observations which, if they were made, would show the theory to be false. In other words, a theory to be scientific must be subject to testing and disproof.

Of course, this doesn't mean that to be considered scientific a theory must be shown to be false. That would be an absurdity. Rather it means that in principle there has to be some way to test it, some test which it could conceivably fail. Scientific tests usually consist of predictions based on the theory. To the extent that a theory succeeds in satisfying the predictions, the theory is confirmed and strengthened although it never reaches the status of final truth or proof since it's always possible that some future observation will not conform to a prediction.

For example, suppose it's your hypothesis that all objects will fall to earth at the same rate regardless of their mass as long as air resistance is discounted. You test your hypothesis by simultaneously dropping two balls of different masses from a height, as Galileo is said to have done, and you discover that no matter how many times you do this the two balls hit the ground at the same time.

Your hypothesis has been confirmed, but had one of them, perhaps the more massive one, hit the ground first it would've indicated that the more massive ball fell faster and your hypothesis would've been falsified. Since we can imagine a result that would've falsified our hypothesis if that result had occurred our hypothesis is a legitimate scientific hypothesis.

Popper pointed out that it's the essence of science that it deals in theories (i.e. well-confirmed hypotheses) that are subject to being falsified. If a theory is impervious to falsification, if no realistic result could ever serve to show it to be wrong, if it's compatible with every possible observation, then it's not testable and therefore not a scientific theory. That doesn't mean that the theory is wrong, it just means that it's not a scientific theory. It could be a metaphysical hypothesis or pseudo-science, but it's not a theory that scientists should entertain in their capacity as scientists.

Popper's falsification criterion was widely believed to have been a great insight into the nature of scientific practice, but in recent decades, as Ms. Sleigh tells us, there's been grumbling. We might wonder, for instance, why a theory cannot be considered scientific if it can be verified but not falsified.

Suppose your theory is that there's life elsewhere in the universe. Your theory could be shown to be true if extra-terrestrial life were discovered, but it's hard to imagine any observation that could falsify the theory. After all, in order for the theory to be shown to be wrong every cubic centimeter of the universe would need to be inspected be sure that there's no life anywhere, yet despite the fact that the theory can't be falsified it seems as if it should be scientific because it can be verified if living things were to turn up somehow.

The difficulty can be resolved by simply recasting the theory. Rather than posit that there is life elsewhere in the universe we could posit that there's no life elsewhere in the universe. Now, the discovery of life on some distant planet, if such a discovery were made, would falsify the theory so the reformulated version passes Popper's test for a legitimate scientific theory.

Popper, in other words, argues that what scientists do, or should do, is seek to show their theories to be wrong, not to show them to be right.

One reason why Popper is falling into disfavor among some scientists, though, is that his falsification criterion would eliminate from the realm of science certain theories that some scientists have spent their entire careers investigating and writing about. String theory, the multiverse hypothesis, some aspects of Darwinian evolution and the origin of life, anthropogenic climate change and many hypotheses in psychology and social science such as critical race theory, have all been criticized as unfalsifiable and therefore unscientific.

Yet the people who work in these fields argue that the inability to test their theories shouldn't diminish their status as scientific projects. For them science is whatever it is that scientists are doing.

Popper has also been waved aside by those who wish to blur the boundaries between science and economics, politics and even philosophy and theology. Scientists who wish to use their prestige to speak authoritatively on matters outside their field don't like being reminded that they're outside their field.

One interesting consequence of the rejection of Popper's solution to the demarcation problem is that if falsification is not essential to the practice of science then hypotheses like intelligent design and special creationism cannot be dismissed as unscientific. Even the theological claim that God exists be placed outside the purview of science.

Indeed, if testability/falsifiability are no longer essential to the scientific enterprise it's hard to see how anything could be excluded from the scientific purview.

Here's a 9 minute video that explains Popper's falsification criterion: