>I'm graduating soon and have for some time been haunted by the feeling that mathematics is just a symbol-game.
I've been through that phase and, in my opinion, you just have to accept that math REALLY is a symbol-game. Now the question is: is it meaningful? And if yes, in what way and to what extent?
I believe math is "meaningful" in a broad and empirical sense. Yes, I'm using the adjective "empirical" in relation to math. Most math professors and authors (in my experience) try to peddle an ahistorical, acritical version of contemporary math as the be-all and end-all of deductive science, something that comes out of our minds perfectly formed, like Athena in full armor from Zeus' head. But that's not how it works, if you ask me.
In the first place, how do we pick our axioms? Experience, more often than not. Same goes for theorems: a mathematician is usually already persuaded that a theorem is true even before he manages to get a formal proof. "Conjectures" are about as important as "proofs" in math, perhaps even more important. First we conjecture that a certain theorem is true (and we arrive through this conjecture through non-deductive ways: experience, analogy, intuition, divine inspiration or whatever) and then we try to prove it. Deduction is, in a certain sense, almost an afterthought. A deductive proof is a confirmation, it puts the conjecture on a more solid basis and it wins for the conjecture the honorable name of "Theorem", but the proof itself, however important and worthwhile, is merely the final, rational step of a process that began on grounds not necessarily logical.
Archimedes' book "On Method" is a good example of how this works: in order to find the volume of the sphere and the cylinder, and the ratio between the cylinder and the sphere inscribed in it, Archimedes used physical models and compared weights (under the reasonable assumption that if two bodies are made of the same material of the same density, their weights are proportional to their volumes). Another example is the development of Projective Geometry that started with the observations of Renaissance artists (like Piero della Francesca).
Ultimately Math is no different than Physics: it provides models. The difference is that Physics limits itself to models of the physical world, while Math can provide models for a variety of other things as well (like economic models). In some sense Physics is simply applied Math, applied to the physical world.
When you consider Math in itself, it has essentially no meaning, and it becomes meaningful only in so far you have a model in the real world for your mathematical system. At least that's what I think. Platonists would probably disagree.