[lingtalks] Monday: Kie Zuraw (Linguistics Colloquium)

[Klinton Bicknell]

On Monday 28 April at 2pm, Kie Zuraw (UCLA; http://www.linguistics.ucla.edu/people/zuraw/main.html 
  ) will give a colloquium in the UCSD Linguistics Department, in AP&M  
4301.

:: Abstract ::

Mapping patterns and surface patterns

Zhang, Lai, and Sailor (to appear in Proceedings of CLS), in testing  
the productivity of Taiwanese tone sandhi, find an interesting  
difference between participants in Taiwan and those living in the  
U.S., and speculate that reduced language use attenuates frequency  
effects (and, in their case, allows a phonetic effect to emerge).  
Inspired by their findings, this talk revisits unpublished  
experimental data to suggest that weakened mapping frequencies can  
give way to surface frequencies.

The experiment—in collaboration with Phillip Monk—dealt with the nasal  
substitution rule of Tagalog, and was conducted over the web, with  
participants around the world. It was a reverse wug test (Berko 1958),  
asking participants to “undo” the rule, which is neutralizing. For  
example, hypothetical pa-mugnat could come from the root bugnat or the  
root pugnat. Similarly, n can derive from t, s, d, and [eng] from k,  
g, [glottal stop]. The rule has abundant exceptions, but the  
distribution of these exceptions is patterned (Zuraw 2000): voiceless  
consonants undergo the rule more than voiced, and “fronter” consonants  
more than “backer”. The experiment tested which type of frequency was  
a better model for participants’ choices. For example, bugnat is a- 
priori much more probable than pugnat, because more roots begin with b  
than with p. However, pugnat is aposteriori slightly more probable  
than bugnat, because the greater productivity of the rule in p-initial  
stems outweighs the a-priori advantage of b. Although the results  
seemed messy at first, when they were sorted according to the  
criterion that is suggested by the Zhang et al. results, a pattern  
emerged: the ratings of participants residing in the Philippines are  
better modeled by the a-posteriori probability, and those of  
participants residing abroad are better modeled by the apriori, though  
the correlation is still not as good as for the Philippines  
participants.

Some suggestive results from another experiment will also be  
discussed; in this experiment, all the usable subjects were residing  
in the Philippines. This was a normal wug test: given a novel stem  
(such as pugnat), subjects chose whether to apply nasal substitution  
or not. Among the stimuli were sonorant-initial stems, which cannot  
undergo the rule; although for these stimuli both choices presented  
were legal in isolation, the nasal-substituted form was illegal as a  
mapping from the stem that was given. Surprisingly, a small majority  
of participants chose the illegal mapping for at least one sonorant.  
This strategy is fairly well modeled by the relative a-priori  
probabilities of the choices presented. I speculate that the  
difference in strategies doesn’t only reflect a difference in  
attention or interpretation of the task, because although 64% of  
participants who reported speaking no Philippine languages other than  
Tagalog chose at least one illegal mapping, 81% of participants who  
did speak another Philippine language—which probably tends to impinge  
on their rate of Tagalog use—chose an illegal mapping.
-- 

For further information about the Linguistics department colloquia  
series, including the schedule of future events, please visit http://ling.ucsd.edu/events/colloquia.html 
  .