Indeterminateness in qualitative and quantitative reasoning

A central issue of expert systems is the man-machine interface. One of its most ergonomic realizations is a natural language interface. Uncertainty and indeterminateness are main properties of a natural language. While there are many formalisms dealing with uncertainty, there still is a strong need of those modeling indeterminateness. Therefore a formalism is presented which (a) formalizes the semantic interpretation of a natural language and (b) provides an instrument for reasoning with information given in a natural language. After a short outline of the problem the theoretical foundations are discussed, showing that it can be solved by a statistical formalism based on a trivalent logic. Finally the formalism is exemplified and compared to related theories.