My friend has a theory that invisible pink unicorns are orbiting Earth. I claim that according to Karl Popper's falsifiability criterion, this theory is not scientific as it is not falsifiable. My friend retorts: "Sure it is! My theory rests on the theory of gravitation, as gravity is what makes objects orbit Earth. If the theory of gravitation is falsified, then my theory is falsified. Thus my theory is falsifiable, and therefore scientific."

Is my friend's argument valid?

  • 1
    Your friends argument also relies on the premise that not only do unicorns exist, but that they are pink. I'd also say it relies on the premise those unicorns can survive the vacuum, cold, and radiation of space... but they didn't claim they were alive. – Uueerdo Feb 24 '20 at 21:48
  • 4
    He is half right. Pink invisibility aside, since the unicorns are apparently postulated to have mass they can, in principle, be detected from gravitational influence on other bodies (just like dark matter). Falsifiability is one condition for a scientific theory to be worth consideration, but it sure isn't sufficient. Another one is that it should actually be useful for explaining something else. – Conifold Feb 24 '20 at 21:59
  • Hm... Can an invisible object be said to have a color? – Ted Wrigley Feb 24 '20 at 22:15
  • 1
    @TedWrigley: Sure, with an alpha of zero (which maps to full transparency). – Nat Feb 25 '20 at 7:57

I've heard it said that string theory isn't falsifiable, and so should not be regarded as science (in my amateur opinion, this argument is not good since falsifiability is not the only criterion of science or rationality, but never mind that). If what your friend is saying is correct, then string theory is falsifiable; all you need to do is falsify quantum mechanics or relativity, for example. But this isn't what people usually mean by calling a theory falsifiable.

When people say that a theory is falsifiable, I'm pretty sure they mean something like this: Let S be the conjunction of all currently accepted scientific theories, and let T be some additional theory not contained in S. To say T is falsifiable means T and S imply new observations that weren't implied by S alone, and those new observations are testable and could, for all we know, turn out to be false. (Of course, this makes falsifiability relative to moments in time, since S is always changing, but this is fine; some things are testable today that were not testable in the past. And presumably there are some things that we cannot test today that we may be able to test in the future.)

Returning to your friend's invisible pink unicorn theory, we should ask him what new falsifiable predictions his theory (T) makes that were not already implied by our theory of gravity (S).


Interesting question. ;)

For me, this is where the rubber meets the road:

My theory rests on the theory of gravitation...

Does it really? I would argue that gravity is a secondary consideration here. What leaps out at me is the postulated existence of invisible pink unicorns.

If we decide to embrace their existence as a credible hypothesis or theory, then we can ask additional questions, including where they live. If they occur in space but near Earth, then it's only logical that they should orbit Earth.

Falsifying the theory of gravity would indeed shoot down this particular theory. But what is the likelihood of the theory of gravity being falsified? And if it is NOT falsified, does that one element of the equation mean that all the other elements (e.g. pink, invisible, unicorns) are true, or even logically plausible?

It's kind of like making something out of a series of defective parts, then adding one non-defective part (e.g. gravity) and saying,

You could theoretically find something wrong with the only part of my formula that's endorsed by scientists, therefore all those defective parts floating around in space make a lot of sense.

In related news, the idea that not falsifiable = not scientific is nonsense. This formula falls flat on its face in the realm of historical science, for example. In fact, some of Karl Popper's ideas regarding history are so bizarre I'm astounded that anyone takes him seriously.

Falsifying an idea can be an elegantly simple way of identifying a flawed theory. So if someone can explain to me how a particular theory can be falsified, then I'll thank them for the tip. But when someone says "Your idea is nonsense because it can't be falsified," I just yawn and walk away.

That's mind control, not science or philosophy.

Does Science Need Falsifiability? What scientific idea is ready for retirement? (Answer: Falsifiability)

Falsifiability is 'just a simple motto that non-philosophically-trained scientists have latched onto.'

  • LOL - Thanks for upvoting my answer before I finished it. I hit the publish button by mistake. ;) – David Blomstrom Feb 25 '20 at 0:56
  • Thanks for asserting that the falsifiability criterion should not be taken as gospel. I often find it at odds with how science is actually conducted, and a scientific straitjacket if it were to be taken literally, not merely as a guideline. In particular, I find that it deals poorly with uncertainty, for example when the object being studied is a complex dynamic system where the predicted output is uncertain. To say that a meteorological hypothesis is unscientific because it can't say with 100% certainty whether it will be a thunderstorm tomorrow, is preposterous. – matthiash Feb 25 '20 at 10:11
  • Yes, falsifiability and Occam's razor are commonly used by propagandists to attack conspiracy theory (which is essentially history). Obviously, there's credible conspiracy theory and non-credible conspiracy theory, but falsifiability and Occam's razor tell us that conspiracy doesn't even exist. In other words, no politician has ever lied. ;) – David Blomstrom Feb 25 '20 at 10:27

tl;dr Yes, your friend's theory is marginally falsifiable and therefore marginally "scientific", where "scientific" means "prone to scientific scrutiny". Of course, this doesn't mean that it's correct, nor is its falsifiable surface area particularly interesting.

My friend retorts: "Sure it is! My theory rests on the theory of gravitation, as gravity is what makes objects orbit Earth. If the theory of gravitation is falsified, then my theory is falsified. Thus my theory is falsifiable, and therefore scientific."

Yeah, sure, this is accurate-ish (ignoring the necessary-but-not-sufficient issue).

To be clear, when we say that theories must be falsifiable to be scientific, what we mean is that theories must be falsifiable to be prone to scientific analysis. Your friend's theory is marginally falsifiable, and so it's marginally prone to scientific scrutiny. For example, as your friend correctly notes, their theory could be falsified if gravity is disproven.

Of course, a theory being scientific (i.e., prone to scientific scrutiny) doesn't make it correct.

Nor does a theory being scientific (in the sense of being prone to scientific scrutiny) make it scientific (in the sense of being verified by scientific analysis).

Also, we're currently looking for such invisible-but-gravitational unicorns.

Note: "Scientific" means something different from normal here.

Folks probably think of "scientific" theories as those supported by scientific research.

In the context of falsifiability, "scientific" means "able to be scrutinized by scientific research".

For example, say Alice asserts that Luna is, in fact, made of cheese – a fine white cheddar, to be precise. Then:

  1. Alice's theory is "scientific" in the sense of it being prone to scientific analysis.

  2. Alice's theory is not "scientific" in the sense of being supported by scientific analysis.

Note: There's a framing ambiguity complicating the situation.

Consider a student handing in their homework:

Student: Here's my research paper!
[The teacher glances over the paper.]
Teacher: This was supposed to be a referenced research paper, but I don't see a single citation!
Student: Check out the cover page! I cited my birth certificate, state ID, passport, and social-security card to reference my name, giving me 4 authoritative sources!

Is the student's paper "referenced"?

Yes, it is! The student did include references, so the overall paper is, in fact, referenced.

That said, the teacher's likely to be unimpressed as the significant bulk of the paper isn't referenced, despite the totality of it being referenced.

Likewise, your friend's theory about invisible pink unicorns has a significant bulk of it that may be unfalsifiable. So while your friend may correctly point out that the totality is falsifiable, this isn't a defense against the observation that a good bulk of it isn't.

Note: Falsifiability is a very low bar.

Unfalsifiable theories have also been called "not even wrong".

We say that they're "not even" wrong because being unfalsifiable is worse than merely being wrong. Wrong theories were at least meaningful enough to get to the point where they could be called wrong; by contrast, not-even-wrong theories fail to even get that far.


First, to clarify that's a hypothesis, not a scientific theory. The reason why I think that argument sounds off is because evidence for or against a scientific hypothesis is usually of the sufficient and necessary nature. For example if I think protein A allows an organism to eat chemical B then I support the hypothesis by removing the protein and demonstrating how the cell no longer eats chemical B (it is necessary), the sufficient case is obvious because it does grow when it has the protein and I can go the extra mile and show that it catalyzes the reaction etc. In this case gravity is necessary for the hypothesized creature to exist, but not sufficient (gravity existing doesn't imply the creature existing) so it's not convincing evidence and doesn't show falsifiability (not that I think science is defined by falsifiability).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.