If an n × n matrix A is singular, then the columns of A must be linearly independent. Is this true?
Invertible functions must be bijective
Invertible functions must have square matrices
Invertible functions must span R^n
Also, am I missing some other must conditions of invertible or singular functions?
What's special about invertible functions anyway?
No, it's exactly the opposite.
Yes.
Invertible linear functions can be represented by square (nonsingular) matrices.
Assuming that they are linear, and their codomain is $\mathbb{R}^n$, then yes.
Determinant, eigenvalues,...
They have inverse.