In spite of the success of electron theories to account for the ferroelectric phenomena, ideas in classical crystal chemistry are still useful in unraveling complex behavior of multicomponent solids. In particular they provide a natural explanation of high susceptibilities of mixed-ion systems based upon the concept of competing interactions. However, they are largely phenomenological and empirical. It is suggested that the concept of atomic level stresses can bridge the gap between the crystal chemistry approach and first-principle calculations in a most meaningful way. Also some novel experimental methods, which directly probe certain aspects of the electronic state relevant for this phenomenon, are discussed. ©2000 American Institute of Physics.