Issue of Monday, April 28th, 2008
Friday, May 2, 3:30 PM
University of California, Santa Cruz
The role of dispersion, focalization, and articulation in vowel system simulations
Since the seminal work of Liljencrants and Lindblom (1972), a key testing ground for functional, evolutionary, or emergentist approaches to sound systems has been the typology of vowel inventories (e.g., Lindblom 1986, Schwartz et al. 1997, and de Boer 2000). An important innovation of Schwartz et al.’s Dispersion and Focalization Theory (DFT) was calculating the optimality (“energy”) of a vowel system as a weighted combination of:
(i) dispersion: minimization of the auditory distance between vowels (as in Liljencrants and Lindblom 1972) and (ii) focalization: maximization of the importance of “focal” vowels such as [i] and [y] (cf. Stevens’s (1972) quantal vowels).
In this paper, we report results of new vowel system simulations, following the original DFT calculations of Schwartz et al. for the optimality of a given vowel system. However, our algorithm for selecting candidate systems for comparison explores the search space more effectively, allowing for more thorough and accurate computation of DFT’s predictions.
Our results for DFT differ significantly from those published. Specifically, we find a greater number of optimal systems throughout the entire range of possible parameter settings in DFT. Some of these are attested, meaning the DFT does better at modeling the facts than it was originally thought to do. Other optimal systems are unattested, and these help us better determine the parameter space within which the model performs well. We discuss implications and further work.
Monday 4/28, 11.30AM
“Obligatory redundancy in discourse: presupposition, antipresupposition and non-asserted content”
For more information about the remaining schedule of the Syntax-Semantics Reading Group, please visit our website: http://web.mit.edu/linguistics/groups/synsem/index.html
Come and join us for this week’s Ling-lunch talk:
Speaker: Elena Benedicto
WHEN: May 1, 12:30-1:45
Monday, April 28, 5-6pm
Presenter: Chiyuki Ito
Title: “Analogical Changes in the Accent of Sino-Korean Words in Yanbian Korean”
This paper presents the results of the analysis of accent changes in a corpus of c. 8,000 lexical items. The major results include different analogical paths depending on the word classes (native HL -> LH, Sino-Korean LH -> HL), an Island of Reliability effect (Albright 2002) based on coda consonants, a model of accent changes based on weighted constraints employing Jager (to appear)’s Stochastic Gradient Ascent learning algorithm, and notable deviations from the general LH -> HL trend which take into account sonorant vs. obstruent onsets and token frequencies.