Browsing by Subject "Consonants"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Item Quantifying perceptual contrast: the dimension of place of articulation(2007-12) Park, Sang-Hoon; Harms, Robert T. (Robert Thomas), 1932-; Lindblom, Björn, 1934-This study investigates the role of perceptual distinctiveness in consonant inventories. While distinctiveness appears to play a role in the shaping of vowel systems, a literature review indicates that its status in consonant selections remains unclear. To address this issue I used speech materials recorded by a trained phonetician containing 35 CV syllables with seven places of articulation (bilabial, dental, alveolar, retroflex, palatal, velar and uvular) and five vowels: [i] [[epsilon]] [a] [[backwards c]] and [u]. Detailed acoustic measurements were performed: formant patterns at vowel onsets (loci) and vowel midpoints, transitions rates and burst spectra. To validate the speech material, comparisons were made with published data and with formant frequencies derived by means of an articulatory model. Perceptual data were collected on these 35 syllables. Multiple Regression analyses were performed with the coded dissimilarities as the dependent variable and with (combinations of) formant-based distances, time constant differences and burst differences as the independent variables. The results indicated that acoustic measurements could be successfully used to help explain listener responses. Optimal place sets were obtained from a rank ordering of the CV syllables with respect to 'individual salience' (defined as the sum of a syllable's perceptual distance to other places in the same vowel context) and from a replication of the Liljencrants & Lindblom systemic criterion of maximizing distances within all vowel pairs. Instead of the typologically prevalent pattern of [b d g], predictions were found to be vowel-dependent and to often favor CV:s located at the 'corners' of the acoustic F3-F2 space, viz., uvular, palatal and retroflex. This finding leads to a conclusion that distinctiveness alone is unlikely to account for how languages use place of articulation in voiced stops. For more successful attempts, future work should be directed towards defining and incorporating production constraints such as 'ease of articulation'.