We will present and analyze Perception-based Logical Deduction (PbLD). It is an inference method aimed at the treatment of fuzzy IF-THEN rules as linguistically expressed genuine logical implications. We will propose global and local variants of PbLD, introduce PbLD with fuzzy inputs and study it from the point of view of the interpolativity property. In the last part of the presentation, a Linguistic Fuzzy Logic Package for the R software will be presented with focus on searching for association rules, size reduction of rule bases, and running the inference based on PbLD.