Philosopher of science Karl Popper.

What Gets to Count as Science?

Sam Enright
Oct 29, 2019 · 6 min read

Much of modern science sound bizarre and completely goes against common sense. At the same time, a great deal of quackery and gobbledygook purports to be “scientific”. Given this situation, it would be useful to know if there was something concrete separating science from non-science. This is something philosophers call ‘the demarcation problem’.

The philosopher of science Karl Popper famously proposed the criterion of falsifiability: an idea is scientific if and only it can be unambiguously proved wrong. To Popper, the reason why theology and psychoanalysis are not scientific is that you can always weave a story to connect the dots between observed data. The same observation (like “avoiding one’s mother”) could be evidence both for and against a hypothesis (like “my patient has an Oedipal complex”) being true.

This scene from ‘The Life of Brian’ illustrates the trouble with unfalsifiable claims: once you are convinced of their validity, there is little anyone can say to change your mind.

‘Falsificationism’ runs into some problems when we encounter ideas in modern physics, astronomy, and cosmology — ostensibly sciences! — which stretch what counts as falsifiable. For instance, the “strings” in string theory are one-dimensional loops which exist on the scale of a Planck length (Around 10⁻³⁵m). Observing one such string would require a particle accelerator 300,000 light-years in circumference. Even if the EU could work that into its budget, it’s certainly not going to fit in Switzerland! If these strings exist, we’re almost guaranteed to never observe them. In fact, if we did observe them, that would be in violation of the theory.

Or consider the many-worlds interpretation of quantum mechanics — a favourite to many science-fiction authors. It predicts that all possible states of the quantum wave function (which describes the probability of finding a particle in a particular state) are realised in the multiverse. But if many-worlds is right, there’s no possible way we could access these parallel universes, so we wouldn’t know if it were true. Is that falsifiable?

If you were a cynic, you might claim that these highly speculative theories are a waste of time, or even pseudoscientific. Would you be right? Will the multiverse theory one day join the ranks of Islam and Hinduism as one of the world’s major faiths?

Even asking this question contradicts a tenant of scientific thinking: either a statement is true, or it isn’t. The universe consists of strings, or it doesn’t. Rejecting an idea out of hand because of some a priori philosophical principle is as unscientific as it gets. What’s also often omitted is that Popper thought that his criterion would only apply if an idea were testable “in principle”, and so the wild ideas of string theory are fair game.

Problems with falsificationism

Popper held Einstein’s relativity up as an exemplar of a theory which made definite, falsifiable predictions. One prediction of general relativity was that the universe should be expanding or contracting, which Einstein modified because he thought the universe was static. So even in this case, the falsifiability criterion is not so clear-cut. How do we take into account the fact that scientists can revise their theories? Physicists don’t throw out ideas and start from scratch when an experiment doesn’t go to plan — nor should they! Even if falsificationism were true, it is an unworkable way to actually do science in practice.

Another challenge comes from the fact that our very existence might mean that our data self-select. For example, one problem in modern physics is explaining why the pressure exerted by the vacuum of space — which is driving the expansion of the universe — is so tiny. Someone might point out that, if this value were much larger, we wouldn’t be here to observe it. The universe would have expanded too quickly, and our solar system would have ripped apart before the Earth could form. This is called anthropic (Greek for “human”) reasoning. We ought to be very cautious with anthropic reasoning: if we take it too far, we could use it to explain away anything. “Of course the experiment gave X result: if it gave Y result we wouldn’t exist!” doesn’t constitute a real explanation. This attitude was satirised as far back as 1759 in Voltaire’s Candide, in which the protagonist Professor Pangloss is convinced that he lives in the “best of all possible worlds” and that the purpose of noses is to hold up spectacles. If we assume that the very facts which we are trying to observe are requisites for our observation to occur, then science is dead.

We must also understand that no scientific idea can make (falsifiable) predictions in a vacuum. Even something as foundational as Newton’s theory of gravitation rests on a menagerie of background hypotheses, like “there exist particles”, “they have mass”, etc. Therefore, no individual hypothesis is falsifiable in and of itself, but only with reference to a set of background or auxiliary hypotheses. When a theory’s claim is falsified, logic can’t tell you which is at fault: the theory itself or its set of background assumptions. And so, as most philosophers of science agree, this kills the idea that there is a line in the sand which separates science from non-science. Sniffing out bullshit will always be an exercise in careful, nuanced, thinking.


The dominant school of thought which Popper’s falsificationism displaced is called inductivism. This assumed that “the future resembles that past” and that the more times a hypothesis is observed as holding true, the greater certainty with which we can say it is true. If all of the 1,000,000 times we have observed the night sky, the gravitational attraction between bodies has been proportional to the product of their masses and inversely proportional to the square of the distance, the chances that the pattern will fail on the 1,000,001st observation are negligibly small. So, we call it “true”. Induction rests on an assumption of uniformity: that the laws of nature are the same everywhere, and stay consistent across time. This raises its own problems, perhaps best summarised by Bertrand Russel in The Problems of Philosophy:

And this kind of association [of uniformity] is not confined to men; in animals also it is very strong. A horse which has been often driven along a certain road resists the attempt to drive him in a different direction. Domestic animals expect food when they see the person who usually feeds them. We know that all these rather crude expectations of uniformity are liable to be misleading. The man who has fed the chicken every day throughout its life at last wrings its neck instead, showing that more refined views as to the uniformity of nature would have been useful to the chicken.

What use is usefulness?

Considering the “usefulness” of a scientific theory in explaining the data may turn out to be the best predictor of whether or not it is true. While classical philosophical methods like that of Plato rested on deduction (inference of a particular case from a general rule), and pre-Popperian philosophy of science on induction (general rule from particular cases), a more cautious delineation of the scientific enterprise rests on abduction. This is a third category of reasoning, based on finding the most plausible explanation for a set of observations.

String theory, in spite of its difficulties, is the only physical theory to successfully unify electromagnetism and the nuclear forces with gravity. This alone makes the theory highly fruitful and worth pursuing — all assertions of its truth value aside. So long as an idea is grounded in reason and contributes to the progress of science, I see no reason why we shouldn’t pursue it. If string theory and the multiverse help scientists to understand the world, they will continue to be explored. If they are too vague or some better idea comes along, they’ll be discarded. Either way, experimentation will be the ultimate guide. As the American physicist Richard Feynman put it:

“It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.”

Medium's largest active publication, followed by +586K people. Follow to join our community.

Sam Enright

Written by

The Startup

More From Medium

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade