Description of the data structure used in the semantic file of the MIND system (Management of Information through Natural Discourse), and of the procedures for manipulating information stored in the file. The MIND system consists of nested and chained modules of high-level programming language statements; it is relatively easy to modify, either for improvement or for adaptation to specialized applications. The major features of the data structure are: (1) It is a net whose nodes represent conceptual entities and whose edges represent relations that hold between entities. (2) Some nodes of the net are variables, and are used in constructing general statements and deduction rules. (3) Each conceptual entity is represented by exactly one node in the net from which all information concerning that entity is retrievable. (4) Nodes can be identified and retrieved either by name or by a sufficient description of their connections with other nodes. The use of the system to experiment with various semantic theories is demonstrated by examining several questions of current linguistic theory. 175 pp. Ref.