Articulatory and acoustic patterns in phonemic and phonetic nasalization
A single acoustic signal may be generated by many potential configurations of the vocal tract (the “many-to-one” problem (Maeda, 1990)). Because of this, acoustic information is an ambiguous reflection of the true configuration of the vocal tract. The use of advanced imaging technologies allows us to overcome the problem: we can now observe the physiological inputs associated with particular acoustic outputs to better understand the mapping between them. This allows us to infer articulatory strategies used by speakers to maintain phonological categories.
The production of nasal sounds occurs when the velopharyngeal port opens due to lowering of the velum, coupling the oral and nasal cavities (Stevens, 2000). Recent work has suggested that comparing oral and nasal vowel pairs only by velopharyngeal port opening is too simple, and that oral/nasal vowel congeners differ in oropharyngeal configurations as well, including distinctions in positioning of the lips, tongue, and pharynx (Barlaz et al., 2015; Carignan et al., 2011; Carignan, 2014; da Matta Machado, 1993; Rong and Kuehn, 2010; Shosted et al., 2012, 2015; Zerling, 1984, and others). In this talk, I discuss the application of advanced imaging technology, real-time Magnetic Resonance Imaging (rt-MRI) to the study of nasalization, specifically in Brazilian Portuguese (BP). BP contains both phonemic nasal vowels (which are distinct from oral counterparts in the phonological inventory) and phonetically nasalized vowels (that is, oral vowels that undergo nasalization via coarticulation). Therefore, a study of the articulatory and acoustic features and differences between oral, nasal and nasalized vowels is important for understanding how categories are maintained within the BP phonemic inventory. I present results detailing differences throughout the vocal tract in these categories, as well as in the acoustic domain. I also discuss the relationship between the acoustic and articulatory findings, in order to further resolve the many-to-one problem as it pertains to the study of nasalization.