This is an extract from the OPIP book. Previously, A(lice) and B(obby) discussed the importance of physicists being aware of the assumptions their theories are based on.
A: Apart from being right or wrong, are all assumptions created equal?
B: No, there are significant practical differences. One difference in assumptions is how easy it is to refute them. Assumptions don’t cause a lot of damage as long as they can get refuted quickly. The real issue is that some assumptions are very hard to refute, causing a lot of time spent in vain. Those we have to watch out for.
A: So if we have to make assumptions, at least we should go for the ones that can be more easily refuted?
B: Right. Refuting incorrect assumptions may even be more important than trying to come up with correct ones. An analogy: In business, there is the saying that to be a successful entrepreneur, you must have failed at least 10 times. If you look at it from this angle, it means that every failure is a step closer to success, as you need to get to your 10 failures as quickly as possible. Hence, it’s suggested that if you want to be successful faster, you need to increase your rate of failure. Translated to physics it would mean to celebrate refutations, as that gets us closer to theories that work.
A: You seem to like weird analogies.
B: Here’s an even weirder one: From literature I know that in the pick-up community (guys trying to pick up girls), an important rule is to quickly cut the chats that aren’t promising, as the night is short. The key insight here is that “Your enemy is not the rejection; your enemy is the 20-minute discussion that goes nowhere.”
A: From literature, I’m sure…
B: Could we please focus on the point I’m trying to make? We need to get good at spotting efforts that aren’t likely to succeed so that we spend the time on more promising approaches.
A: Refutations seem to be very close to your heart.
B: Yes, indeed. They are physics’ reality check. Without them, there cannot be progress in science, or even science at all. As Karl Popper said, falsifiability is the distinguishing mark that separates science from pseudoscience.[1]
A: Do physicists really need to be reminded of that?
B: Some do. There are several theories out there that are far from being proven or falsified. To be fair, that in itself doesn’t discredit such theories—maybe the next big theory is just so complex that it will take a while to make it falsifiable. As long as there is a strong drive to get to falsifiable predictions as quickly as possible, it’s acceptable. However, for my taste, some theories enjoy their presence far away from the falsifiability demarcation line a tad too much. In any case, not being falsifiable currently or in the near future is never a point on the pro side of any theory.
A: Is it always realistic to expect physicists to embrace refutations? For example, consider a situation where they’ve spent 10 years developing or understanding a specific theory. Then, several indications emerge showing that the theory is not correct. Do you think they’ll share your enthusiasm when you shout, “Woo-hoo, let’s celebrate another refutation!”?
B: You’re now referring to psychology and the individual aspect of progress in physics—let me come back to that a bit later.
A: Alright. Any other reasons why refutations are important?
B: Refutations are important because in science, there isn’t really anything else.
A: What do you mean?
B: The best we can ever do is refute an idea or theory. There’s no proof in science.
A: Really, no theory can be proven? What about the theory of gravity?
B: What exactly do you mean by that?
A: I mean that objects fall down to Earth.
B: It’s a good, practical rule that you can take for granted in your everyday life. Until, of course, the guy with the hot balloon shows up.
A: Well, that doesn’t disprove gravity. There may be countereffects stronger than the gravitational pull. There’s still gravity.
B: You’re right; I was referring to your rule that objects fall to the ground. There is always the possibility of exceptional cases that prove a theory wrong that worked fine in all cases before.
A: There must be some proof. For example, let’s take the discovery of Neptune, a planet that isn’t visible to the unaided eye. It was first hypothesized by Alexis Bouvard in 1821. He suggested this due to unexpected changes in the orbit of Uranus, indicating the gravitational influence of another planet. In the early to mid-1840s, John Couch Adams and Urbain Le Verrier mathematically predicted Neptune’s exact position, which Johann Galle confirmed with a telescope in 1846. If that’s not proving a theory, what is?
B: It’s not so much a proof rather than a refutation of our earlier state of knowledge that didn’t include the existence of a planet like Neptune. This may seem pedantic and hair-splitting, but it’s an important difference. Mathematics and logic may know proofs, but when it comes to science, we can never be entirely sure. Even if the evidence seems overwhelming, there could still be a flaw in our reasoning somewhere, or unknown phenomena we haven’t taken into account.
A: Hm. That makes it all so elusive.
B: Yes, it’s not a comforting thought, which is why it’s psychologically difficult to accept. We want to have clarity and consider things as true (or false). But it’s not accurate. That said, in everyday language it’s acceptable to talk about proofs, to keep it simple. The correct wording, “Corroborated to such a high degree that, for all intents and purposes, we can assume it to be correct until we find its refutation,” is a bit of a mouthful. Still, that’s what it is, strictly speaking.
A: What if someone claims “Zeus, the sky and thunder God, is real.” We cannot tell him “It’s you who has to prove it” because we’d be asking for proof, which we said is impossible.
B: You’re right, but in this context, we use the expression of proof in the practical sense of providing evidence. Or, looked at it the other way, if someone postulates something that goes against our observations, they need to refute our existing understanding (that there is no Zeus). To keep things real, we must base it on evidence. And evidence is about what is, not about what is not. It’s impossible to prove the negative. It reminds me of Bertrand Russell’s analogy about the claim that a teapot is orbiting the Sun somewhere in space between Earth and Mars.[2] It’s impossible to disprove, but that doesn’t make it right, nor respectable. I think Richard Dawkins’ statement, “I’m an a-theist for the same reason I’m an a-teapotist” sums it up well.
A: Alright. Anything else to be said in the context of refutations?
B: It’s worth noting the value of simplicity in scientific theories. While this isn’t a novel concept—with Occam’s Razor advocating for simpler solutions since the 14th century, and even Aristotle voicing similar sentiments[3]—most interpret these principles as implying a higher likelihood of such theories being true. However, there’s an often overlooked advantage to simplicity: such theories are easier to refute. This holds true in several ways. First, simple theories are more readily testable through experiments. Second, their simplicity reveals potential vulnerabilities or “points of attack” more conspicuously, facilitating their refutation. Last, as they are easy to understand, more individuals can (and are willing to) grasp and thereby challenge simple theories. This exposure to refutation is at least as important as the idea that simpler theories have a higher probability of being correct.
The book continues by analyzing the role of simplicity, unification, and beauty in physics. Buy it now.
—
[1] Popper, K. (1934 in German, 1959 in English): „The Logic of Scientific Discovery”.
[2] See Opip.lol/teapot. Similar classic analogies include the Flying Spaghetti Monster and the Invisible Pink Unicorn.
[3] Aristotle in Posterior Analytics: “We may assume the superiority ceteris paribus of the demonstration which derives from fewer postulates or hypotheses.”
