Actually, the best definition I've heard for math is that it's "A set of rules."
We can make up the rules to do whatever we want, and then test them to see if they work. There's different kinds of math, like algebra and calculus. Then there are more obscure ones that only apply to specific things like the movement of electrons, and nothing else. It's interesting how many kinds of math there are that you won't learn about in school!
The answer to "what is mathematics" (and what
isn't) has changed significantly over time. Pythagoras and his ilk would have listed logic,
including philosophy, and (planar) geometry - the total opposite of the times where set theory and Venn diagrams were thought to be the most basic form of math. Today, we're focused on analysis and calculus, the latter being extended from medieval math mainly by
infinitesimal calculus. Which modern philosophy has discovered as well, getting close and cuddly to math again after centuries of being something entirely different from math.
One thing that virtually all mathematicians will be able to agree on is that there is a concept of "pure" mathematics that plain doesn't care whether the models it churns out match anything in the universe, known or not. Case in point:
aperiodic tilings being a purely hypothetical thing from their inception in the 1960s until
Dan Shechtman first crystallographed a quasicrystal in 198x.
Bam!, application case.
Even more extreme, one somewhat restricted but popular definition of mathematics is "the science of infinity", as in,
being enumerable while
is not etc. - note that this puts computations like "2+2=4" outside mathematics proper. And,
newsflash: There isn't
anything in the (known) universe that's truly infinite.