Humanities and Social Sciences

Linguistics Program

Staff Research Page

Dr Marija Tabain

Major Research Areas

• acoustic phonetics
• articulatory phonetics
• cross-linguistic phonetics and phonology
• coarticulation
• Australian languages
• articulatory prosody

Current Projects

The effects of prosodic structure on segmental articulation

(Collaborator: Pascal Perrier of the Institut de la Communication Parlée in Grenoble, France). In recent years, it has become clear that the prosodic structure of an utterance influences the articulation and acoustics of individual consonants and vowels. Broadly speaking, stronger prosodic boundaries induce greater hyper-articulation, and a more prototypical segment (e.g. a more peripheral /a/ vowel); while weaker boundaries induce a more "lax" articulation and more ambiguous segment (e.g. a more centralized /a/ vowel). I have used EMA (electro-magnetic articulography) and acoustic analysis techniques to examine the articulation and acoustics of the three point vowels in French. Results from my studies have provided evidence that speakers have an auditory/acoustic goal in mind when articulating a speech sound, rather than an articulatory goal, since different speakers adopt different articulatory strategies while achieving the same acoustic goal. Thus, a natural manipulation of speech via the grammar of prosodic structure has led to important insights into speech production strategies.

The effects of consonant phoneme inventory on speech production.

(Collaborators: Gavan Breen of the Institute for Aboriginal Development, Alice Springs; and Andy Butcher of Flinders University, Adelaide). There is evidence to suggest that the greater the number of vowel phonemes in a language, the greater the precision required in the production of any given vowel. By contrast, the smaller the number of vowel phonemes in a language, the greater the variability in the production of any given vowel. However, the vowel space is described in a way very different from descriptions of consonants spaces. Simply speaking, most vowel spaces can be described along two dimensions: F1 and F2 (or front/back and high/low). However, consonant spaces are not so easily described, and consonants may differ on many dimensions. My work on Aboriginal languages aims to provide a better picture of how consonant phoneme spaces work. The multiple places of articulation in these languages (especially in the coronal series) provides a rich inventory of consonant contrasts. To date the effects described provide only a small part of the overall picture. For example, the articulation of velar consonants is much more forward in Aboriginal languages than in Germanic languages, which have far fewer places of articulation. Also, while F2 (the second resonance of the oral cavity) is sufficient to describe place-of-articulation contrasts in most languages, this is not true for Aboriginal languages, which have a greater number of place contrasts. And as a final example, the CV sequence is much more tightly controlled than the VC sequence in most languages of the world, but both the CV and the VC sequence are equally well controlled in Aboriginal languages. All of these examples are due to the imperative to maintain place-of-articulation contrasts in languages which have multiple place contrasts in the consonant series.

Return to Dr Marija Tabain

Return to Linguistics Research Page