Las ecuaciones del locus y el punto de articulación en español

Abstract

This paper explores the possibilities offered by locus equations for classifying stop consonants regarding place of articulation. Locus equations, originally conceived by Lindblom (1963), are linear functions obtained from a correlation between two F2 values of the postconsonantal vowel: F2 values at the first glottal pulse and the F2 values at the midvowel nucleus. An experiment with ten subjects (five male and five female), which produced a series of [p]-[b],' [t]-[d], [k]-[g] tokens for the 5 Spanish vowels, showed that, beyond individual and coarticulacion variation, each place is sharply identified by a locus equation. A later discriminant analysis confirmed this point: using either both slopes and y intercepts or slopes alone, a 100% correct classification followed. It seems thus that beyond the variation inherent to speech acts there are acoustic invariants explaining the communicative fact.