Tautology and theories

What is a tautology? Simply put it is a statement that is always true. More specifically, it is a statement that is true because of its structure. Usually such a statement is not informative. The easiest way to explain this is by example, so lets look on the following statements:

1. All apples are round.
2. All apples are either round or not round.

I have never seen an apple that wasn't round, so the first statement is a correct one. However it is not a tautology. It is perfectly possible for a square shaped apple to exist, you just need to make it grow inside a box. Therefore this statement is not always true.
The second statement is always true, and therefore a tautology, but it doesn't say anything. That is, if all we know about apples is that an apple is either round or not round we don't know anything about apples.
It doesn't mean that tautologies are useless, they have both use and importance in certain cases.

Lets look on tautologies in a more formalistic way. In the example above I assumed we have an object "apple" and a property "round". However, this is not necessary. All we need to write this example is two symbols: P, Q. If we rewrite the example using this symbols we get:
1. P->Q
2. P->(Q(or)(notQ))

I don't want to explain the notation, if you don't know it you are welcome to use wikipedia.
This notation is far better that the previous one. With this notation we are free to chose the meaning of the symbols. For example we can suppose that the meaning of P is x=7, and the meaning of Q is x+4=0. In this case 1 is usually false (but not always) but 2 is true.
From this we see that the second statement is completely independent from reality. It doesn't matter if we use it to tell something about apples or about mathematics, it will always remain true. This is the true meaning of being a tautology.

Sometime ago I read a post that claimed that all theories are tautologies. From the above it should be obvious that this is wrong. But there is a certain moment that causes this confusion. A theory is basically a collection of statements (theorems) that can be logically concluded from a certain set of prepositions. Those prepositions are divided in two groups - tautologies and axioms. (We don't really need to include tautologies, but because they are always true they get included automatically). Axioms are not tautologies. They are just "rules" that we choose by ourselves. In physics, those rules are based on reality and experiment, but this doesn't have to be the case. For example - the parallel postulate of Euclid. It seems natural to us to think that it is correct. But it is possible to built a geometry without it. Naturally in such a geometry all the statement that were derived from the parallel postulate are no longer correct (they might be correct, but not necessarily).

To sum up this discussion, a theory is correct as long as the axioms from which it was built are correct. However, if we try to apply it to a "world" in which the axioms are not true, the theory will not work. By the way, this is a major problem for physics and economics. They can build wonderful theory only to find out that the world we live in is different from their initial assumptions. In mathematics this is not a problem because there is no desire to describe our world, but to built a mathematical structure.
This also makes easy to explain the statement that a theorem once proved is true forever - a theorem is dependent only on its own conditions, so as long as they are satisfied the theorem will always be true.