“…Similarly, the IRS is more user-friendly if the estimated relevance levels of the documents are supplied in a linguistic form (e.g., linguistic terms such as "relevant," "very relevant," may be used) rather than with scores. Following these ideas (Bordogna & Pasi, 1993), several fuzzy linguistic IRSs have been proposed using a fuzzy linguistic approach (Zadeh, 1975) to model the weights in the query and evaluation subsystems (Biswas, Bezdek, Subramanian, & Marques, 1987a, 1987bBolc, Kowalski, & Kozlowska, 1985;Bordogna & Pasi, 1993, 1995aDoszkocs, 1986;Kraft, Bordogna, & Pasi, 1994). In this context, the ordinal fuzzy linguistic approach (Delgado, Verdegay, & Vila, 1993;Herrera & Herrera-Viedma, 1997;Herrera, Herrera-Viedma, & Verdegay, 1996b) is a linguistic approach that allows us to overcome the limitations of the classical fuzzy linguistic approach (Zadeh, 1975), i.e., we do not have to explicitly establish semantic rules or syntactic rules (e.g., using a context-free grammar), thereby reducing, the complexity of the design for the IRS.…”