9353184 deBossenet The proposed research will develop further the symbolic manipulation system (SMS), which is a unified approach to knowledge representation, automated reasoning, and simulation of human conversation. With a symbolic language (SL) to build databases, the project will simulate the mental associations of human beings when responding to one another in conversations. The associations usually are based on deductive or probabilistic relations. The proposed research will extend SMS theory to include non-standard association and inference with the flexibility to operate in a conversational mode. The symbolic manipulation system will also handle imprecise and incomplete information in its inference operations by using SL to represent ambiguity, vagueness or incompleteness. The significance of the proposed research would be the advancement of expert systems through the establishment of a nonstandard theory of association and inference to address problems connected with the elicitation of imprecise or incomplete knowledge. ***
