However, I don't agree that art is unconcerned with reality, nor that maths is. We learn to draw by looking at things in the world, we get our rules about anatomy and so on from there; we learn maths based - at least initially - on physical examples; we tell stories based upon common themes and situations. The basic rules of these things are drawn from their correspondence to reality; with what people have experience with; and form the meaningful grammar of the system. Art is always a language, and a language is always representative. Even when - as arts - we may bend a few rules and simplify certain aspects to lend emphasis or make it better suited to approaching a particular problem.

To hold otherwise seems to me to deprive them of a foundation to communicate any common meaning. It would be no better to learn maths, under that definition, than to spend your life insanely scraping crayons across a page. (Indeed, given that the symbols would also be arbitrary in their meaning, it would be hard to tell the difference.)

Is maths an art, is it a science?

It's a language of enquiry. To claim that it proceeds solely by a 1-1 correspondence to reality is to picture mathematicians off somewhere counting things and deriving pi from taking increasingly precise measurements of actual circles. (And, really, even physics proceeds with the help of a healthy dollop of imagination that forms the foundation of hypothesis.) To claim that it has no correspondence to reality is to reduce the whole exercise to nonsense.

I do not see how either position taken as an absolute is tenable.

If you're believing this, you've probably never done physics.