Aggregation of linguistic labels when semantics is based on antonyms

In this work, we introduce aggregation operators for linguistic labels (this is, ordinal scales) when different experts (or information sources) use different domains to express their knowledge. The aggregated value is computed (i) building first a unified framework, (ii) transforming all the initial values into this new framework, (iii) aggregating the transformed values, and (iv) finally applying a reversal transformation. Transformations and all the constructions are based on assuming an existing semantics for all the domains. In this work, we consider the semantics based on the existence of an antonym (or a set of them) for each element in the domain. This is equivalent to a semantics based on negation functions. (C) 2001 John Wiley & Sons, Inc.