I don't know that I agree that just because we find complex numbers convenient to solve problems from QM to third order polynomials that makes them somehow 'intrinsic' or discovered.
They are a construct which can be used to solve problems - much in the way a negative number can be used to represent a debt.
We define the behaviour of a negative number - for example that 'multiplying' two together gives a positive
We define the behaviour of complex mumbers - for example that dividing is multiplying a complex number by its conjugate
These definitions give us answers that work for us
if they did not work we would seek other mathematics that did work.
Maths is as Marcus Du Sautoy likes to say all about patterns
It is these patterns that are fundamental to nature - maths is a formalisation of techniques and representations that allow us to identify these patterns, predict them and represent them.
So when complex numbers were invented it was as a tool that solved a problem (finding the roots of third order polynomials) what then happened was this tool gave us ways to represent other objects (like elctronic circuits or quantum mechanics) - this gave us a deeper pattern and showed us a connection between these things where we had not previously expected one.
Sometimes these connections are so unexpected that we are shocked by them into an almost mystical sense of wonder
Like Eulers' identity that e to the power i pi is minus 1
http://en.wikipedia.org/wiki/Euler%27s_identity
these are the connections between the patterns that we discover