This paper presents a descriptive model of the melodic characteristics of expressions of politeness attenuators in Catalan, based on the acoustic analysis of a spontaneous speech corpus and its subsequent perceptive validation. The model allows the pragmatic analysis of the phenomenon to be covered, as well as offering future guidelines for its teaching. In addition, the results of the analysis reflect the extensive character of the phenomenon, which surpasses the strict limits of idiomatic code, and allows inter-linguistic analysis. The paper is the product of the collaboration between the Melodic Analysis of Speech and Fonocortesía research projects