As mentioned in the literatures, the definition of the Berry phase is meaningful only for non-orthogonal states. However, in the topological insulators it is defined for quantum states of a matter which are orthogonal. How to justify this inconsistency?