Redução de vogais postônicas mediais na variedade de São Carlos (SP)

Abstract

In this doctoral research, we focus on non-final post-stressed vowels, which occur in antepenultimate stress words. Our main goal is to investigate, in the light of Articulatory Phonology, the evident presence and the apparent absence of these vowels. The evident presence of non-final post-stressed vowels is when there are evident visual cues of these vowels in the oscillogram and in the spectrogram. The apparent absence of non-final post-stressed vowels is when there are not evident visual cues of these vowels in the oscillogram and in the spectrogram. We understand both phenomena of evident presence and apparent absence as two faces of a reduction continuum of non-final post-stressed vowels. Our data were collected from two experiments applied to native speakers from the Paulista town of São Carlos. The experiments required oral word production. They allowed us to analyze the hypothesis that the apparent absence of non-final post-stressed vowels tends to be favored by the non-visualization of target words spelling (in this case, the target words are antepenultimate stress words). As regards results, we verified that: (i) in the corpus of this research, the evident presence of non-final post-stressed vowels predominates over the apparent absence of these vowels; (ii) the hypothesis mentioned earlier was not confirmed; (iii) the preceding consonant context and the constriction degree of tongue body gestures involved in the articulation of non-final post-stressed vowels were suggested as the main factor groups that tend to influence the apparent absence of these vowels; (iv) the following consonant context seems to have relevance on this phenomenon; (v) the increase in the duration of preceding consonants tends to influence the duration of the vowels more than the increase in the duration of following consonants; (vi) there is relation between the duration of the vowels and their weakening (we understand weakening as a reduction phenomenon that can be manifested by periodicity loss and an increase in the irregularity of voicing in terms of duration and amplitude of glottal pulses; and (vii) weakening associated to durational decrease tends to be greater for non-final post-stressed vowels than for stressed vowels. Furthermore, we identified, through acoustic analyses, some evidences that can favor the hypothesis that cases of apparent absence of non-final post-stressed vowels can be interpreted as realizations of these vowels totally covered by gestures of adjacent consonants. We compared certain results we obtained for the apparent absence of non-final post-stressed vowels with results obtained by Lima (2008), Ramos (2009) and Silva (2006). These authors investigated what they understand as deletion of non-final post-stressed vowels. From the comparisons, we observed both convergences and divergences between our research and the respective researches by Lima (2008), Ramos (2009) and Silva (2006). Finally, we discussed the concepts of apparent absence, deletion and resyllabification. In doing so, we argued that: (i) apparent absence, contrary to what can be understood as deletion (whether according to traditional theories such as Structuralist Phonology, generative phonologies and Variationist Sociolinguistics or according to Articulatory Phonology), seems not to trigger resyllabification; and (ii) the apparent absence of non-final post-stressed vowels seems not to be influenced by the possibility of violating phonotactic constraints of Portuguese as a function of the hypothesis of what is understood by deletion of non-final post-stressed vowels according to Articulatory Phonology (when no articulatory gestures referring to a non-final post-stressed vowel are activated during speech production).