which is a linear system with unknowns a, b, c and d, having a unique solution
given by (2.5).
If the eigenvalues of A are not distinct then the formula for eA is very simple.
In fact matrix A is always diagonalizable (since ιA is Hermitian) and if ±ιλ are
the eigenvalues of A then it is possible apply the same formula (2.6) to find eA. In
this case eA is a first degree polynomial in A and proceeding as in Theorem 2.4
we find