- so in this system is cosine defined using exponentials and i? Is i defined as sqrt(-1)? Do we just not worry about how there are two square roots of -1? Do we not care that e^x - log(y) involves branch cuts, and is thus either multivalued and/or undefined somewhere? Maybe there's some consistent algebraic theory of these operations that treats them "formally", ignoring issues like branch cuts?
I'm asking you (and the world) because I couldn't quickly find the answers in the paper.
