Anotace:
This paper is a contribution to the approximate reasoning theory. We present an alternative to the well known Mamdani rule, which we call the \emph{perception-based logical deduction}. The leading assumption is that fuzzy IF-THEN rules are interpreted as genuine linguistic sentences consisting of the, so called, \emph{evaluating linguistic
expressions}. Evaluating expressions determined on the basis of an observation in a concrete context are called \emph{perceptions}. The latter and sets of fuzzy IF-THEN
rules, called \emph{linguistic descriptions}, are used in derivation of a conclusion. The perception-based deduction is demonstrated to behave in accordance with human way of reasoning. A lot of simulations as well as real applications corroborate that it works very well and is easy for the user because of natural language comprehension.