Who said: The shortest trajectory between two realities is complex?
I'm not sure of the citation, but the idea is that even when working on "real" problems the best way is to pass through "complex" problems (e.g. using residue theorem to compute complicated real integrals).
It seems to be the French mathematician Jacques Hadamard:
It was translated into English by J. P. Kahane (who passed away recently):
For other details, you can look at this page.