It is argued that the chiral partners of the lowest-lying hadrons are hadronic molecules and not three-quark or quark-antiquark states, respectively. As an example the case of a_1 as the chiral partner of the rho is discussed. Deconfinement -- or as a precursor large in-medium widths for hadronic states -- is proposed as a natural way to accommodate for the fact that at chiral restoration the respective in-medium spectra of chiral partners must become degenerate. Ingredients for a systematic and self-consistent in-medium calculation are presented with special emphasis on vector-meson dominance which emerges from a recently proposed systematic counting scheme for the mesonic sector including pseudoscalar and vector mesons as active degrees of freedom.