On an admissible sign of the static dielectric function of matter