Sophia Kao: dissertation defense (July 28)

Sophia Kao defended her dissertation with the title "Phonological Learning Bias in Tone Patterns" on Friday, July 28. Congratulations!
The relationship between typological asymmetries and acquisition of phonological patterns has been a controversial topic in the field of phonology. This dissertation approaches the issue by focusing on the source of typological asymmetries involving tone patterns, and the role that typological naturalness plays in the learning of patterns that are common vs. patterns that are logically possible but are rare or unattested. The target tone patterns of this study were tone elision and contour tone formation patterns, which are common strategies for dealing with tones that are stranded as a result of elision of a vowel. I provide data from a typological survey of thirty tone languages of Africa that elide a vowel when two vowels come together by concatenation of words, leaving a stranded tone that may associate with a neighboring vowel to form a contour tone, or may displace the tone of a neighboring vowel. The preferences for particular patterns within both tone elision and contour tone formation revealed by the typological survey provide the basis for the experiments discussed in the dissertation.
I present evidence from four artificial language learning experiments that examined participants’ ability to learn the cross-linguistically common vs. uncommon or unattested patterns within both tone elision and contour tone formation. Participants, who were speakers of either English or Mandarin Chinese, were exposed to a tone pattern that is either typologically common or uncommon, and were later tested to see whether they learned the tone pattern or not. They were also tested on novel stimuli they had never heard before to see whether they applied the learned knowledge to new forms.
The results showed better learning performance for the common patterns than for the uncommon or unattested patterns by the English speakers. The Mandarin speakers were able to learn the typologically uncommon patterns, but still favored the common patterns in generalizing to new forms. The fact that participants who were exposed to the common tone patterns exhibited better performance both in learning the pattern and in generalizing the pattern to novel forms conformed to the hypothesis that typological frequency reflects a learning bias toward phonological patterns that are cross-linguistically more common.