On an Archimedean analogue of Tate's conjecture