no cls, but many alternatives this week

1 view
Skip to first unread message

Jelle Zuidema

unread,
Oct 31, 2011, 9:49:01 AM10/31/11
to complin...@googlegroups.com

Dear all,

There is no CLS this week, but there are two other events involving computational linguistics that might be of interest:

- Thursday 3 November 2011, Master of Logic defense, Andreas van Cranenburgh
Title: Discontinuous Data-Oriented Parsing through Mild Context-Sensitivity
Date and time: Thursday 3 November 2011, 13:00
Location: Room D1.114, Science Park 904, Amsterdam
Supervisor: Remko Scha

- Friday 4 November : ACLC seminar, John Nerbonne (CLCG, University of Groningen)
Title: "One of structuralism's dreams"
Abstract:

Structuralists famously observed that language is "un systÍme o˚ tout setient" (Meillet, 1903:407), emphasizing that the system of relations of linguisticunits was more important than their concrete content. The modern study of speech has not vindicated the structuralists on this point, as the contemporary emphasis lies on instrumental studies of pronunciation and auditory perception.
Instrumental phonetics and laboratory phonology are leaders in this direction.

But modern computational linguistics and corpus linguistics emphasizes that there is a wealth of information in distributions.  This study attempts to derive content from relations, in particular phonetic (acoustic) content from the distribution of alternative pronunciations used in different linguistic varieties.
It proceeds from data documenting language variation, examining six dialect atlases each containing the phonetic transcriptions of the same sets of words at hundreds of sites. We collect the correspondences via an alignment procedure, and then apply an information-theoretic measure, pointwise mutual
information, assigning smaller segment distances to segments which frequently correspond. We iterate alignment and information-theoretic distance assignment until both stabilize and we evaluate the quality of the phonetic distances obtained
by comparing them to acoustic vowel distances. For all dialect datasets (Dutch, German, Gabon Bantu, U.S. English, Tuscan and Bulgarian) we find strong signicant correlations between the induced phonetic distances and the acoustic
distances, illustrating the usefulness of the method in deriving valid phonetic distances from dialectal pronunciations.

Location
Spuistraat 210-212
1012 VT  Amsterdam
Room: 420


Reply all
Reply to author
Forward
0 new messages