Integrated Learning of Chinese

From LearnLab
Revision as of 10:45, 28 September 2008 by Liuying (Talk | contribs)

Jump to: navigation, search

Summary table

  • Node Title: Integrated Learning of Chinese: reading, perception and production
  • Researchers: Ying Liu, Charles Perfetti, Min Wang, Suemei Wu
  • PIs: Ying Liu, Charles Perfetti, Min Wang
  • Others who have contributed 160 hours or more:
  • Post-Docs: Connie Guan
  • Graduate Students: Derek Chan
  • Study Start Date Jan 1, 2008
  • Study End Date July 31, 2009
  • LearnLab Site and Courses , CMU Chinese (Classroom and Online)
  • Number of Students: 100
  • Planned Participant Hours for the study: 300
  • Data in the Data Shop: experiments have not started yet


  • Learning second language is a challenge to learners. It is more so for English speakers to learn Chinese. The unique Chinese character writing system and tonal features are fundamentally different from English and thus presents a unique obstacle to learning by English speakers. In our model of reading Chinese, orthography, phonology and meaning are universal constituents and critical knowledge components that should be learned and integrated (Perfetti, Liu, and Tan, 2005). Working together with the CMU Chinese online course, the present studies will explore how to facilitate the integration by training both perception and production skills. The specific methods to be tested will be using multiple learning systems including learning orthography, pronunciation and meaning together through complementary visual, auditory and motor modalities. Integration factors affect the learning curve are examined in three studies.


Integration; Constituents; Orthography; Phonology; Meaning

Research question

  • How does integration of language constituents lead to robust learning?
  • Does writing Chinese lead to better integration and more robust Chinese reading?
  • Does the combination of writing and typing lead to more robust learning via better integration?


  • Our previous work on Chinese learning has focused separately on character reading (Liu, Wang, and Perfetti, 2007; Liu, Perfetti, and Wang, 2006), tone perception (Wang et al, under review), syllable production with “talking head” (Massaro, Liu, Chen, & Perfetti, 2006), and cotraining of characters (Liu, Perfetti, and Mitchell, in preparation). Most of above studies were implemented through PSLC Chinese online course, and we will continue to do so for all studies in the present project plan.
  • There have been various findings from above studies. The character reading study found that explicit learning of radicals facilitates the learning of character meaning. Tone perception study found that visual contour plus pinyin provided the best learning curve over one semester. Syllable production study suggested that the synthetic talking head “Bao” provided larger improvement on vowel production than audio only. The cotraining study showed significant advantage for “paired” learning, in which both visual font and auditory sound of a character were presented sequentially in one trial.

Dependent variables

  • Normal post-test: measures of partial character recognition (i.e., the accuracy and response time on recognizing the partial characters, the whole of which the students have learned in training), lexical decision (i.e., the accuracy and response time on deciding whether the character is a real character or not) and dictation task (i.e. the quality of character-writing given the English meaning cues or the sound cues of the characters)
  • Transfer: measure of lexical decision (see the definition above) on the novel characters

Independent variables

  • Integration of reading and writing vs. reading only


  • Production with automated feedback, production only, vs. no production
  • Early integration vs. late integration


  • Integration is crucial for robust learning;
  • Integration is most robust when the involved knowledge components are already refined.


The first experiment is under programming. Testing of students will start in late March of 2008.


These studies test hypotheses about the effectiveness of targeted integrated instruction which are based theories from both Refinement and fluency cluster and Coordinative cluster. Each study has a specific rationale. (That is, integration is not a general virtue, but inherits effectiveness according to specific assumptions about the relation of component knowledge to specific tasks of reading, perception, and production.) The studies share a general approach in the use of the Integrated Chinese Tutor (ITC). In terms of the assistance dimension, our assumption is the initial learning of a novel orthography along with a new phonological system places high demands on novice learners. The decomposition strategy represents a high level of assistance at this stage of learning. We are not testing the implication that, at advanced stages of learning, students might benefit from less assistance in the form of non-decomposed language units.



Further information

  • Perfetti, C.A., Liu, Y., & Tan, L.H (2005). The Lexical Constituency Model: Some Implications of Research on Chinese for General Theories of Reading. Psychological Review, 112, 43-59.
  • Liu, Y., Wang, M., Perfetti, C.A. (2007) Threshold-Style Processing of Chinese Characters for Adult Second Language Learners. Memory and Cognition, .
  • Liu, Perfetti, C.A., & Wang, M. (2006) Visual Analysis and Lexical Access of Chinese characters by Chinese as Second Language Readers. Linguistic and Language, 7(3), 637-657.
  • Massaro, D. W., Liu, Y., Chen, T. H., & Perfetti, C. A. (2006). A Multilingual Embodied Conversational Agent for Tutoring Speech and Language Learning. Proceedings of the Ninth International Conference on Spoken Language Processing (Interspeech 2006 - ICSLP, September, Pittsburgh, PA), 825-828.Universität Bonn, Bonn, Germany.