What do falsifiability and verifiability actually mean as a precondition of a scientific hypothesis?

I have been hearing from science guys that unlike philosophical hypothesis or theory, to be a scientific hypothesis an educated guess must have to be falsifiable and verifiable. But I cannot comprehend falsifiable and verifiable conditions for some scientific hypotheses especially in astronomy field. For example,

Couple of scientific hypotheses exists about the shape of the universe,

  1. Universe is positively curved
  2. Universe is negatively curved
  3. Universe is flat

These are the well known scientific hypotheses as far as my knowledge (correct me if I am wrong). It seems to me, though I may be wrong, none of the above hypotheses are anyway falsifiable and verifiable because it is never possible for us to measure the geometry as the universe is expanding beyond our scope of study. For example, we can may be able to test on larger amount space as time goes on and technology advances, but to measure the actual shape of the whole universe we need to know theoretically the full picture of the universe which is not possible for us according to current understandings of science.

Now, there might be many such hypotheses which we cannot take under testable and verifiable condition yet we call them scientific hypothesis. My question is,

Is a scientific hypothesis really required to be falsifiable and verifiable? If so, what are the falsifiable and verifiable conditions for above mentioned hypotheses?

  • 1
    $\begingroup$ This is a question about the philosophy of science rather than astronomy. It probably belongs on Philosophy SE. $\endgroup$
    – user38308
    Commented May 2, 2021 at 16:58
  • 1
    $\begingroup$ The reason I put it here because confusion is limited mostly for astronomy related hypotheses. Also in philosophy site most people try to answer on philosophical aspects where my intention is to get the answer strictly on definition from scientific community and consensus. @tfb $\endgroup$ Commented May 2, 2021 at 17:54
  • 1
    $\begingroup$ I like your question! To keep it verifiably on-topic here I think you should narrow the scope to hypotheses in Astronomy/Cosmology. I see that you've chosen curvature as the example which is great, but maybe emphasize those specific topics rather than ask about all scientific hypotheses? Astronomy is a little different than some other fields of science in that it's hard to do controlled experiments and to reproduce/repeat them. $\endgroup$
    – uhoh
    Commented May 2, 2021 at 21:40
  • $\begingroup$ @SazzadHissainKhan - Your question is fundamentally philosophical and cannot be resolved by empirical research. We can of course poll scientists for how they would answer the question, but that is not going to be a rigorous answer. While it may be informative and important to hear what scientists say about this, remember that their answer will be philosophy of science as done by practitioners of the science, rather than professional philosophers. $\endgroup$ Commented May 2, 2021 at 22:02
  • $\begingroup$ @uhoh you can improve it if you wish. $\endgroup$ Commented May 3, 2021 at 3:22

1 Answer 1


[I am hesitant to post this answer, as I am unclear that this is a question about science rather than philosophy as I said in a comment. But I've written it now, so.]

First of all, all scientific theories rest on a mass of assumptions not all of which are testable. For example if I'm doing astronomy I'm assuming that I'm not just living in a simulation, or that, in fact, anything outside my own mind exists at all. I can't test those ideas, but I'm going to assume them or I can't do any astronomy because there's no point.

Without assumptions like that all you can do is maths: Fermat's last theorem is true whether or not Fermat or Andrew Wiles, exist, say.

But there are assumptions which are made at a level above these kind of things which also matter in astronomy. For measuring the spatial curvature of the universe one is very important: the cosmological principle.

The cosmological principle. Viewed on a sufficiently large scale, the properties of the universe are the same for all observers.

(From William C. Keel (2007) The Road to Galaxy Formation (2nd ed.), Springer-Praxis, via Wikipedia.)

Here an 'observer' doesn't mean 'someone on Earth', it means 'someone anywhere you like'.

The cosmological principle implies two things: that on large enough scales the universe is homogeneous (it looks the same everywhere) and isotropic (it looks the same in all directions).

The cosmological principle is not fully testable I think. We can check that the universe looks isotropic here, and people do this. If it's not then the cosmological principle fails. But if it does look isotropic then the principle still may be false: perhaps we are in some special location where the universe looks isotropic. If we could get far away (very far away) and test that the universe looks isotropic from at least two other places as well, then we could conclude that the part of the universe we've observed is also homogeneous I think.

But it could still be fooling us somehow: the cosmological principle is really just something we assume is true. It's a very reasonable assumption, I think, because what it's really saying is that where we are is not special: the universe is not, in fact, built around us. The cosmological principle is the inverse of the Ptolemaic model, where everything goes around the Earth, really, and even more it is an inverse of the 'nothing but me exists' model which I said I was going to rule out at the start.

Well, if we assume that the cosmological principle is true, then the bit of the universe we can observe is just the same as all the other bits of it on large enough scales. So, if we assume that general relativity is a good description of the universe on large scales (that's a lot more testable) and probably some other things which are testable then if we can measure the spatial curvature (or equivalently, the matter density) of the part of the universe we can see, we can conclude that this holds for all of it.

That is the sense, I think, that claims about the testability of measurements of spatial curvature are true.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .