Abstract
Contrary to common textbook statements, the macroscopic electric polarization of a crystal cannot be unambiguously defined as the dipole of a unit cell. The only measurable quantities are either derivatives of the polarization (dielectric susceptibility, Born effective charges, piezoelectricity, pyroelectricity) or finite differences (ferroelectricity). The differential concept is a basic one in theoretical work too. I sketch here the basic theory and the state of the art in its implementation to real materials, both for induced and spontaneous polarization; for the latter case, I also propose a novel approach.