Skip to Main Content
 

Global Search Box

 
 
 
 

Files

ETD Abstract Container

Abstract Header

The Representation, Organization and Access of Lexical Tone by Native and Non-Native Mandarin Speakers

Abstract Details

2015, Doctor of Philosophy, Ohio State University, East Asian Languages and Literatures.
This dissertation explores how lexical tone in Mandarin Chinese is learned and used during spoken word recognition by native and non-native speakers. This research begins with the hypothesis that lexical tone is not just one of an arbitrary set of unpredictable speech cues, but rather, like any other component of language, its frequency of occurrence and the predictability with which it co-occurs with other aspects of spoken language can be tracked and stored as statistical knowledge. This hypothesis challenges established theories of word recognition by predicting a dynamic contribution of tone; speakers learn to listen for and store tone information on the basis of the frequency and probability of syllables and tones co-occurring over time in speech. To explore the statistical learning of lexical tone and how speaker variability influences this learning, an artificial tonal language was tested over the course of four days. In this four-day training and testing paradigm, three groups of participants — 40 native Mandarin speakers, 40 native English speakers learning Mandarin as a second language, and 40 monolingual native English speakers — learned 130 CV+tone nonce words, each paired with a black and white nonce symbol. Unique CV syllables were each combined with four different tonal contours (directly comparable to those in Mandarin). CV syllable frequency (high/low) was crossed with syllable-specific tonal probability (a tone contour was most probable or least probable to occur with a specific CV) to produce four conditions. Participants were trained and tested with either a single voice or with four voices. Each day, participants’ word identification was assessed using a naming task and a word recognition task, which recorded mouse clicks and real-time eye movement responses to speech stimuli. All participants showed rapid, daily improvements across the four sessions. Results from the naming task indicate that all participants, regardless of their first language, began tracking syllable distributional properties, specifically frequency of occurrence. Less speaker variability seemed to facilitate the learning of statistical regularity in recognition. The results also suggest that learners first acquired separate syllable followed by syllable+tone representations. For each syllable type, participants with prior tonal experience were also able to track the probability of each tonal category; native Mandarin speakers and Mandarin L2 learners most accurately named the most probable tone given the syllable. The monolingual control group did not show sensitivity to tonal probabilities during the naming task, potentially due to their inexperience in producing different pitch contours. Results from the eye-tracking task showed all participants made use of syllable frequencies, tonal probabilities and their interaction. Specifically, on the last two days of testing, all three groups showed early anticipatory eye movements to infrequent syllables with most probable tones regardless of which tones were heard. These results corroborate recent findings that show native Mandarin speakers exhibit similar probabilistic processing for infrequent syllables, i.e., syllables with higher tonal informativeness. The speed at which listeners recognized tone’s dynamic informativeness depended upon the learner’s linguistic experience with tone: native Mandarin speakers showed the earliest sensitivity to the distributional properties of tone, followed by Mandarin L2 leaners and then monolingual English speakers. By the last day, listeners with three different experiences with lexical tone performed similarly in the eye-tracking task, navigating between careful signal-based acoustic processing and knowledge-based probabilistic processing. Taken together, these results indicate that statistical learning of tones takes place even with relatively short exposure to novel words, and less speaker variability in speech input may help listeners represent syllable frequencies and tonal probabilities more accurately. This dissertation concludes by discussing the theoretical implications for how the Mandarin lexicon is organized and how these findings affect models of spoken word recognition. Additionally, pedagogical implications of this finding are discussed, including strategies that allow L2 learners to capitalize on knowledge of the language’s distributional properties to better improve the acquisition and processing of Mandarin tones.
Marjorie Chan (Advisor)
Shari Speer (Advisor)
Mineharu Nakayama (Committee Member)
Kiwako Ito (Committee Member)
Chao-Yang Lee (Committee Member)
242 p.

Recommended Citations

Citations

  • Wiener, S. (2015). The Representation, Organization and Access of Lexical Tone by Native and Non-Native Mandarin Speakers [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1429827661

    APA Style (7th edition)

  • Wiener, Seth. The Representation, Organization and Access of Lexical Tone by Native and Non-Native Mandarin Speakers. 2015. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1429827661.

    MLA Style (8th edition)

  • Wiener, Seth. "The Representation, Organization and Access of Lexical Tone by Native and Non-Native Mandarin Speakers." Doctoral dissertation, Ohio State University, 2015. http://rave.ohiolink.edu/etdc/view?acc_num=osu1429827661

    Chicago Manual of Style (17th edition)