The main purpose of this paper is to present a logical-mathematical modeling to the phenomenon of intonation perception in Brazilian Portuguese speech. This modeling is based on the standard automatic analysis of intonation developed by Ferreira Netto (2006, 2008, 2010); and it employs the principles developed by Hart, Collier and Cohen (1990), regarding the phenomenon of conjugation between perception and production of intonational curves. Furthermore, we apply thresholds of tonal differentiation, which are prescribed by Consoni (2011) to sing le words and to contextualized sentences. Threshold values ensure the perception of relativization regarding the intonation. Our modeling basically aims to create a conjugated system in which the values of F0 are processed according to a component called the mid-tone (FERREIRA NETTO, 2008). The thresholds of tonal differentiation operate as limits to the systematical laterality generated by the mid-tone; and they act as a restriction to variations. The modeling requires more studies for its better functionality, as well as a functional computer implementation in order to verify its applicability.
Phonetics; Intonation; Modeling; Phonology
