This trope owes its existence in large part to equivocation — the popular definition of a "theory" is not the same as the scientific definition of a theory.
In colloquial speech, a "theory" is a broad term for any conjecture or explanation that hasn't been proven yet. In science, the word "theory" is more narrow: it is an explanation of a natural phenomenon. Crucially, a natural phenomenon is something whose existence we can observe. Gravity is such a phenomenon — after all, you've never seen anyone float off the Earth's surface. (As Dirk Gently would put it, they keep it on all the time, even on the weekends.) The theory of gravitation, therefore, would be a set of explanations for why nothing floats off the Earth's surface, consistent with all scientific observations and evidence. Importantly, a theory has been tested and reproduced, which leads to an increasing body of evidence (an explanation which hasn't been tested yet is called a hypothesis). And a theory has predictive power, meaning you can use it to predict what you should observe in a given situation. For instance, the theory of gravitation can be applied to all celestial bodies, and the math behind that should be able to predict that people and things aren't going to fly off the Earth's surface — and what do you know? It does!
But that's not good enough for some people. A theory is not definitively proven — if it were, it would be called a law. That gives them room to work with, and is why gravity is "only a theory!" Except "law" also has a specific scientific definition: it's a mathematical correlation that is repeatedly observed in many situations. The term usually applies to the physical sciences — e.g. Newton's laws, Ohm's law, the law of conservation of energy. In other words, a law is not a theory "made official" — it's a pattern of observations, of the kind that can be explained by a theory. You need both. As Rational Wiki puts it, a law of gravity will tell you that two objects will be attracted to each other and the exact magnitude of the force, but a theory of gravity will explain why they do that.
Since a theory describes an observable phenomenon, it can't be proven simply because it's limited by our ability to observe. Often, Science Marches On because it has discovered something new to observe, which isn't necessarily explained by the existing theory. This doesn't mean that the theory is wrong so much as it is incomplete. The new theory will explain not only the new observations, but also the old ones — with the same results with respect to the old ones as the old theory. Take gravity, for instance. It was first explained by Isaac Newton, whose model posited that gravity is a type of force which increases proportionally to the size of the object in question. And for centuries, it worked — for everything we could observe. But then as we started to look at very big objects like planets, and very small objects like subatomic particles, things started behaving unexpectedly. Newton's model couldn't predict their behaviour perfectly. So quantum theory and the theory of relativity posit that gravity is the result of objects themselves "bending" the fabric of space around them. Crucially, Newton's observations didn't change, and the new theories still reach the same conclusion that Newton did on the things that Newton could observe. One might say that the complex theories of gravitation simplify to Newton's theory at certain levels of observation.
This is how science works — when we get a new observation, we test our existing theories to see if they predict what we just saw. And if they didn't, that's when we change it. But people take this process to mean that Science Is Wrong — what's the point in believing in gravity if it's just going to be invalidated by a new "theory" in the future? But there are two problems with this line of thought. The first is that it ignores what we just explained — that new theories don't invalidate old ones so much as refine them. The second is that most people who believe Science Is Wrong don't believe it in good faith; they're looking for a hole in which to fit their own "theory", which they believe for self-serving reasons. And the alternative "theories" they posit have all sorts of problems:
- They tend to lack predictive power. Often, this happens because the "theory" is unfalsifiable — it seems to be perfect, but only because it's defined in such a way as to make anything possible. It's like saying that nobody flies off the Earth's surface because A Wizard Did It. The "wizard" can do anything, so it doesn't explain why the wizard did specifically this to us. The wizard could, in theory, turn gravity off at any moment, but we have no way to predict if and when he will. It's useless as a theory.
- They're very rarely held to the same standard as the theory they claim to supplant. Proponents tend to look for a single observation that the current theory can't predict and go, "Aha! Your theory is wrong!" Then they go on to say, essentially, "By default, my theory is right!", even though their theory has all sorts of observations that it can't predict.
- They hold tightly to ad hoc explanations. Ad hoc is worse than a hypothesis; a hypothesis is an explanation that's based on scientific research but just hasn't been tested yet, whereas an ad hoc explanation is just something someone came up with to explain an observation. If you try to poke a hole in someone's "theory", they'll come up with an explanation out of thin air that explains it away.
- They tend to be based on unobservable phenomena. If you're trying to establish that the Big Bang happened billions of years ago, The Fundamentalist might respond that "maybe God made the Universe so that it looks like it was made billions of years ago, but it really wasn't!", you end up with two problems. First, no theory can stand if it has to predict something that can't be observed; the whole point of a theory is to explain something we can observe! And second, such explanations are usually ad hoc — they come up after the fact to invalidate the existing theory by undermining the observations behind them. It's a "theory" that explains an observation by arguing that the observation itself, while valid, for whatever reason doesn't really exist.
All of this is borne from a poor understanding of science and the scientific method. When idiots are using the word "theory" for their ad hoc unscientific beliefs, it causes confusion. Richard Dawkins coined the term "theorum"note to distinguish the core knowledge of science (evolution, heliocentrism, the Earth being round) from both other scientific theories and the layman's conception of a "theory". However, the word hasn't caught on among scientists.
Compare Windmill Political: Depending on the setting and context, this trope can be for those who believe in gravity, those who don't, or those who believe in something else that the author is poking fun at by comparing it to gravity.
