One key problem for Computing with Words but often ignored or
discarded in Fuzzy Set Theories and in Fuzzy
Logic applications is that the same word can mean (slightly) different
things to different people, and (completely) different things in di
erent context. In this paper I will reflect on the first part of this
key problem, that is, how to build a model able to capture the meaning
of a word for group of people in speci c contexts.