Abstract: The deconstruction analysis shows that the definition of personal information inside and outside the domain can be divided into "generalization" and "generalization + enumeration" in terms of formal structure. In terms of elements structure, it is divided into "identification utility" and "correlation relation". In terms of risk structure, it is divided into general and special types. Currently, the concept of personal information faces the dilemma of uncertainty, such as difficulty in identifying standards and demarcating risk levels. The problem lies in the continuous expansion of the scope of personal information driven by technology and cognitive iteration. To address this dilemma, it is necessary to examine existing approaches that focus on identifying information, so this paper proposes a dual connotation framework including referential component and descriptive component that is more consistent with the nature of personal information, and analyzes its operating logic. Clarifying the dual structure of the connotation of personal information has practical value in clarifying the objective basis of identifying personal information, clarifying the internal logic of the conceptual family schema of personal information, clarifying the risk considerations of internal differentiation of personal information, and distinguishing between personal information and personal data.