University of Chicago Linguistics Colloquium

Keren Rice, University of Toronto

What determines morpheme order in the Athapaskan verb?

March 5, 3:30-5:00pm, Cobb 201

Abstract:

The surface order of morphemes in the verb word of Athapaskan languages has traditionally been considered to be idiosyncratic, stipulated by a template. In Rice 2000 I argued that what I called semantic scope plays an important role in the ordering of morphemes. Here I extend the account of morpheme ordering in the verb word, focusing on a series of problems that arise if scope alone is involved. I argue that if phonological factors are also taken into account, a systematicity to the complexities of morpheme ordering in the verb emerges, with morphemes being segregated by their phonological shapes, and, within these phonologically determined groups, scope plays a major role in the ordering of morphemes. I examine the principles that control the ordering in light of recent claims that functional principles such as parsability are key to morpheme ordering.

For future colloquia, please visit: http://linguistics.uchicago.edu/newsevents/colloquia.shtml

After our holiday hiatus, this year’s colloquium series resumes with the first of four (so far) scheduled talks for the Winter quarter. Tania Ionin of the University of Illinois at Urbana-Champaign will be presenting on “The scope of English indefinites: an experimental investigation” (abstract here).

As usual, this talk will take place on Thursday at 3:30 p.m. in Cobb 201, with department tea immediately following. We hope to see many people there as we kick off another quarter!

For our final colloquium of 2008, we are delighted to host Alicia Wassink, associate professor in the Department of Linguistics at the University of Washington, who will be giving a talk on “The Development of Sociolinguistic Competence in Children” this Thursday, December 4, from 3:30-5:00 p.m. in Cobb 201. (See abstract here.)

If you’re sorry to see the last of the Autumn colloquia, fear not! Here are the remaining scheduled talks for the 2008-2009 series, which will resume on January 15:

January 15: Tania Ionin, University of Illinois at Urbana-Champaign

March 5: Keren Rice, University of Toronto

April 2: Adam Albright, MIT

April 30: Teresa Satterfield, University of Michigan,

May 7: Sun-Ah Jun, UCLA

May 14: Nick Fleisher, Wayne State University

May 28: Shigeto Kawahara, Rutgers University

The Department of Linguistics was pleased to host Georgetown University’s Paul Portner last Thursday at our inaugural colloquium of the 2008-2009 series. This week, we will be having Diane Brentari of Purdue University giving a talk on tracking phonological emergence in sign languages.

We eagerly anticipate the scheduled speakers for the Autumn colloquia and invite you to join us for the remaining sessions, which generally commence at 3:30 p.m.

October 23: Paul Portner, Georgetown University

Two Problems about Permission

October 30: Diane Brentari, Purdue University

When does a system become phonological? Grammatical regularities at the interfaces

November 4: Matthias Brenzinger, University of Cologne

Changing roles for African languages in the past, present, and future

*Note special date, time and place: 4-5:30pm, Harper 103

November 13: Duane Watson, University of Illinois at Urbana-Champaign

November 20: Luis López Carretero, University of Illinois at Chicago

December 4: Alicia Wassink, University of Washington,

The Development of Sociolinguistic Competence in Children

See you there!

Phonological Models in Automatic Speech Recognition

Karen Livescu
Toyota Technological Institute at Chicago
Location: Cobb 201
Time: 3:30pm

Abstract:

The performance of automatic speech recognizers varies widely across contexts. Very good performance can be achieved on single-speaker, large-vocabulary dictation in a clean acoustic environment, as well as on small-vocabulary tasks with fewer constraints on the speakers and acoustics. One domain that is still elusive is that of spontaneous conversational speech. This type of speech poses a number of challenges, among them extreme variation in pronunciation. I will describe efforts in the speech recognition community to characterize and model pronunciation variation.

The most thoroughly studied approach is augmentation of a phonetic pronunciation lexicon with phonological rules. Despite successes in a few domains, it has been surprisingly difficult to obtain significant recognition improvements by including those phonetic pronunciations that appear to exist in the data. I will advocate an alternative view: that the phone unit may not be the most appropriate for modeling the lexicon. I will describe approaches using both larger (e.g. syllable-sized) and “smaller” (e.g. articulatory) units. In the class of “smaller” unit models, ideas from articulatory and autosegmental phonology motivate multi-tier models in which tiers have semi-independent behavior. I will present a particular model in which articulatory features are represented as variables in a dynamic Bayesian network.

Non-phonetic pronunciation models can involve significantly different model structures than those typically used in speech recognition, and as a result they may also entail modifications to other components such as the observation model and training algorithms. At this point it is not clear what the “winning” approach will be. The success of a given approach may depend on the domain or on the amount and type of training data available. I will describe some of the current challenges and ongoing work, with a particular focus on the role of phonological theories in statistical models of pronunciation (and vice versa?).