On the one hand, linguistic modifiers can be used to express imprecise values of variables which are described around a given fuzzy set. On the other hand, measures of comparison of fuzzy sets, such as measures of satisfiability or resemblance for instance, enable us to establish the proximity between two imprecise values of variables. A formal connection between modifiers and measures of comparison is presented, starting from studies about similarity-based systems, on the basis of simple types of modifiers and measures of comparison.