Abstract
The development of logical connectives (operations) on fuzzy sets constitute a central issue in the theoretical and applied pursuits of the area. Taking into account the semantics aspects of the logical connectives, it is not surprising that they require a careful context-dependent analysis. We propose a new neural network model of these operations which address two of their new and important conceptual features such as (i) a compensatory character and (ii) spatial distributivity. It is also shown that many of the real-world problems exhibit a clear logical underlying skeleton whose structure lends directly into the discussed neural network model.