In group decision making, the usage of linguistic information entails the requirement for computing with words, a methodology in which words from a natural language are the computational elements. Recently, a new approach based on a framework of granular computing has been applied to cope with linguistic information. Its advantage is that the distribution and the semantics of the linguistic values that represent the words, in place of being defined a priori, are established as defined by the

more »

... imization of a given optimization criterion. Once they have been obtained, the distribution and the semantics of the linguistic values are the same for all de decision makers involved in the decision problem. However, as a word means a different thing to different people, we present in this contribution a new approach that is able to obtain personalized linguistic values.