Worth pointing a few bits out.
In mathematics, and hence pretty much most of modern physics, theory has a couple of very specific meanings that are different to the rest of science. One “theory” is used to describe an area of study. As **Asymtotically fat **notes, there are whole areas of study like String Theory. Or in other areas of mathematics say Catastrophe Theory, Graph Theory, Number Theory. Also, some people get confused between Theory and Theorem. A Theorem is something that has been proved to be true within the scope of the axioms it is defined within. For instance, the Unique Prime Factorisation Theorem is regarded as a truth. For numbers as we defined them, no number has more than one set of prime factors. This isn’t something that is determined by experiment. Mathematicians are 100% sure that no number with more than one set of prime factors exists, and any attempt to find one experimentally will never succeed, for all values of “never”.
Mainstream science has mostly adopted Karl Popper’s view on the nature of scientific theory. A theory is a scientific theory if it can be falsified. That is, you can create the theory, and also come up with a set of discoveries or experimental results that can be actually be measured or discovered, that should they be found to be true, will invalidate the theory. A theory that you can never work out how to disprove is useless.
So, early successes of Popper’s ideas can include Einstein’s prediction of the offset of a star’s location during a solar eclipse. Relativity is falsifiable. It makes predictions, and these predictions can be tested in such a way that should they fail, the theory is accepted to have been invalidated. Counter to this is pseudo-science and non-science. Homoeopathy is pseudo-science. The tennets of homoeopathy don’t accept the idea of statistically valid double blind trials, and any scientific experiment is dismissed with various forms of special pleading. Popper was particularly scathing of Freudian psychology and Marxist theory. Both of which claimed to be scientific, but which moved the goal-posts every time an invalidating result was found.
Darwinian evolution fits as real science, as it makes specific predictions about what won’t be found in the fossil record. In particular it makes predictions that differ from things like Lamarkian evolution, or creationism. Creation Science isn’t a scientific theory, because it invokes miraculous intervention to explain away any finding that contradicts it. So it falls into the pseudo-science camp.
Thus modern scientific theories are always accepted as in some way tentative. They are cast in a form that allows for their disproval. Those theories that survive concerted attempts at disproval, and which are useful become mainstream science. The poster boy is probably general relativity - it makes some seriously counter-intuitive and strange predictions, and scientists continue to probe the bounds of these predictions, and it has survived every one of them without even the slightest hint of doubt. And yet the EPR paradox tells us that we don’t know the entire story.