Phonology Seminar Schedule
UCLA
Department of Linguistics
Spring 2006
The Phonology Seminar
(Linguistics 261) will meet Spring Quarter 2006 in Campbell Hall 2122,
the
Tuesday,
April 04, 2006, 4-6 p.m.
Chacha
(Leonard) Mwita
“A
Metrical Analysis of a Kuria Praise Poem”
Tuesday,
April 11, 2006
We’ll
read and discuss two (short) papers by Gunnar Hansson. You can download them here.
·
Tone
and Voicing Agreement in Yabem
·
Long-Distance
Voicing Agreement: An Evolutionary Perspective
Tuesday,
April 18, 2006
Jeff
Heinz and Shabnam Shademan will each give a WCCFL practice talk.
Jeff’s
title: “Learning Phonotactic
Grammars from Surface Forms”
Tuesday,
April 25, 2006
Peter
Graff
“Generalization
according to natural class: Evidence from an artificial language
experiment”
You
can read Peter’s abstract at the end of this schedule.
Tuesday,
May 02, 2006
Kevin
Ryan
“On
Vocative Accent”
Tuesday,
May 9, 2006
Andrew
Martin
“Phonotactic
constraints on word formation”
Tuesday,
May 16, 2006
Practice
talks for Manchester Phonology Conference:
Sameer Khan, Katie Schack
Titles
TBA
Tuesday,
May 23, 2006
Lauren
Varner
Title
TBA
Tuesday,
May 30, 2006
Ingvar
Lofstedt
Title
TBA
Tuesday,
June 6, 2006
Paper
reading: Eric Bakovic,
“Antigemination, assimilation and the determination of identity”
You
can download the paper here.
Peter
Graff
“Generalization
according to natural class: Evidence from an artificial language
experiment”
In past research Artificial Language experiments have mostly been used to explore learning biases in UG. I will claim that an Artificial Language experiment can produce evidence for the linguistic reality of natural classes.
Based on this idea an Artificial Language experiment was conducted. The inspiration for the setup was the Greek Consonant Shift, where voiced and voiceless-aspirated stops spirantized but voiceless unaspirated (lenis) stops remained unchanged. Thus, the Greek Consonant Shift can be interpreted as an example of a single feature change applying to two natural classes but not to a third.
For the experiment an Artificial Language with an aspirated/voiced/lenis stop system was devised. Four subjects were instructed to learn this language. It was hypothesized that, if a subjects, after acquiring the language, were instructed to spirantize the voiced stops, this would also alter their pronunciation of the aspirates but leave the lenis unaffected. Subjects were provided with a textbook and an audio CD, and given two weeks to learn the language. After that period of time they were recorded performing a reading task.
On the same day they were presented with learning materials for a variant of the language, called the “Northern Dialect”. In the variant voiced stops were replaced with voiced fricatives. Lenis stops, however, were left unchanged. Crucially, no instructions were given on how to pronounce aspirated stops. When asked to pronounce words with aspirates in the variant, subjects independently generalized spirantization of voiced stops onto the natural class of voiceless aspirated stops, which surfaced in significantly reduced closure duration. Their pronunciation of lenis stops remained unaltered.
The results of the experiment thus support the hypothesis and we can conclude that i) natural classes are a linguistic reality and not only a concept phonologists use to describe languages; ii) aspirated stops have more in common with voiceless fricatives than lenis stops do; and iii) Artificial Language Experiments are a valuable tool for phonological research.
