Title: Prosodic focus within and across languages
Supervisor: Mark Liberman
Abstract: Here is the abstract in PDF form.
Dissertation: Here is the dissertation in PDF form.
Motivation: The fact that “purely” prosodic marking of focus may be weaker in some languages than in others, and that it varies in certain circumstances even within a single language, has not been commonly recognized.
Languages under investigation: American English, Mandarin Chinese, Seoul Korean, South Kyungsang Korean, Tokyo Japanese, Standard French, and Suzhou Wu.
Method: We used 10-digit number strings, read as connected individual digits grouped as (NNN)-(NNN)-(NNNN) in the style of American telephone numbers. We created sets of 100 10-digit sequences, designed so that each digit occurs ten times in each sequential position, and each pair of digits occurs once spanning each adjacent pair of positions.
This method has several advantages. First, syntactic and morphological revisions are completely ruled out, so that only prosodic modulation can be used to mark the focused item. Second, the materials are pragmatically uniform, so that all of the positions in the string are equally susceptible to focusing.
Production: Speakers read these digit strings in isolation as a background “broad focus” condition, and in a Q&A dialogue for contrastive focus, where someone asks for confirmation of a version in which one of the digits is incorrect, and the speaker answers with a string in which the wrong digit is given correctly (“No, the number is 215-417-5623”).
Perception: Listeners of each language heard only the phrase with the correction, and were asked to identify which digit was corrected.
Note: We are also now collecting data for standard Russian, Taiwan Mandarin, and Cantonese, and in the future we plan to collect for Standard Spanish. Since every modern language has a phone-number string, we can collect any languages for a wide range of data collection. Please email me at email@example.com if you are interested in working together with me.