PROBING THE REPRESENTATION OF NASAL VOWELS IN BRAZILIAN PORTUGUESE WITH LANGUAGE GAMES

Abstract

Abstract: In this paper, we demonstrate that Brazilian Portuguese phonemic nasal vowels are, as Mattoso Camara (1953, 1970) suggested, derived from an underlying sequence of oral vowel plus nasal consonant... or almost that. Two types of transformational language games reveal that Mattoso was 80% right: four of the nasal vowels are indeed derived from such a sequence, while nasal [a)] (strictly speaking, [å)]) is the result of an underlying, inherently nasal vowel. We also propose a further revision to Mattoso’s disegmental hypothesis for nasal vowels: rather than being a sequence of oral vowel plus nasal archiphoneme, we propose that they are specifically the result of a combination of oral vowel plus homorganic nasal glide (i.e. [j)] or [w)]). This nasal glide, however, is incompatible with \a)\, and moreover would disrupt the contrast between [å)], [å)j], and [å)w]. As a result, [å)] must be inherently underlyingly nasal, a result confirmed through a variety of external evidence. Keywords: phonemic nasal vowels, invented language games, Catete language game