I studied applied linguistics at East China Normal University (华东师范大学) in a Chinese-language program from 2005 to 2008. The following is my master’s thesis abstract, in English and in Chinese. If there is interest, I may eventually put the whole thesis online. (There is no English translation of the thesis; it is written in Chinese.)
Abstract
A Study of Error in Foreigners’ Chinese Disyllabic Tonal Pairs: A Production-Focused Keyword-Based Experimental Model
This study proposes a new model of production-focused experiment model whereby an error analysis of foreigners’ tones is focused on pre-selected disyllabic keywords. The results suggest that certain tone combinations are more difficult for learners than others.
Section 1: Preface
Tones are an essential component of Chinese syllables. As such, they must be mastered by foreign students of Chinese, and yet tones remain one of the most difficult aspects of the spoken language for students of Chinese. This is a special challenge for the field of Chinese as a second language, and one which must be further explored.
Section 2: Previous Research on Foreigners’ Study of Tones
Foreigners’ struggle with tones is well documented, and has been explored in the past through several types of experiments, which include oral reading experiments, tone differentiation experiments, and production experiments. Of these three types, production experiments are most closely related to actual speech used in communication, and, as such, are also the model for this experiment. Chen Qinghai’s 2000 experiment is the model for tonal analysis of Chinese words in natural speech. With some revision, it is the basis for the current experiment.
Section 3: The Production-Focused Keyword-Based Experimental Model
Because foreigners’ pronunciation of tones in isolation does not accurately reflect their speaking ability, this experiment takes disyllabic tone pairs as the unit of investigation. Furthermore, this experiment does not seek to examine all the words used by the subject, but instead focuses on certain keywords. These keywords are all basic, disyllabic words which fall into one of three tone pairs: 1-1, 2-4, 3-2. The experiment involves two groups of subjects: elementary and intermediate. The experiment consists of showing each of the subjects a series of elicitation cards and asking questions about the content of the cards. The subjects’ replies will naturally involve the use of the keywords, and all replies are recorded for later analysis. The hypothesis of the experiment is that (1) tonal competency will be higher overall for the intermediate students, and (2) there will be a discrepancy in accuracy between the three pair types, with 1-1 highest in accuracy, and 3-2 lowest in accuracy.
Section 4: Experiment Results and Analysis
After all keyword data was tabulated, the results showed a relatively high accuracy rate for the 1-1 and 2-4 keyword tone pairs across both subject groups, and a relatively lower accuracy rate for 3-2, with elementary 3-2 the lowest. Further analysis showed that number keywords were skewing the accuracy rates, and when removed, the accuracy rates were exactly as predicted in part two of the hypothesis. The conclusion is that the relatively high level of familiarity the students had with basic numbers raised their tonal production accuracy rates. This implies that tonal accuracy is not only closely tied to familiarity with tone pair groups, but also with individual words.
Section 5: Shortcomings of the Experiment and Recommended Improvements
The keyword-based experiment model is worth further experimentation on a larger scale; the current exploratory experiment was rather limited in terms of number of subjects, number of tone pairs covered, and number of keywords examined. In addition, the current experiment’s use of computer tone contour analysis should be expanded upon and made more objective.
Section 6: Conclusion
In conclusion, this experiment’s main contribution was the keyword-based production-focused experiment model. More experimentation is needed to reach firm conclusions, and is certainly warranted in the arena of foreigners’ tonal competency.
Key words: tones, Chinese as a foreign language, foreigner, pronunciation, experiment, error analysis
Master’s Thesis by John Pasden, 2008
摘要
外国人的汉语双音节声调组合偏误研究–关键词模式的产出性实验
本文提出新的产出性实验模式，通过事先采选的双音节关键词来分析外国人在声调产出方面的偏误。结果显示，对外国人而言双音节声调组合的难度存在差异。
第一章，绪论
声调是汉语音节不可缺少的成分，因此学习汉语的外国人必须掌握汉语声调。但是对外国学生而言，声调仍然是汉语学习中最大的困难之一。这给对外汉语界一种特别的任务，而学者有必要进一步探索该问题。
第二章，以往对于外国人声调学习的研究
以往研究充分探讨了外国学生学习汉语声调的困难，包括几种实验类型，如朗读性实验、识别性实验、产出性实验。其中，产出性实验的对象最接近实际交际使用的话语，因此成为本实验的基本模式。在陈庆海2000年的实验的基础上，本文修改实验方法并分析了外国人产出的自然话语中的声调偏误。
第三章，关键词模式的产出性实验设计
由于单音节声调产出不能正确反映外国人的声调产出能力，本实验以双音节声调组合为基本实验单位。另外，本实验并非检验实验对象所有声调产出，而是检验若干关键词的声调产出。本实验关键词都是初级、双音节词，声调组合为1-1、2-4或3-2。本实验有两个实验对象组：初级水平学生和中级水平学生。本实验内容主要是给每个实验对象展示一系列引发卡片，并由实验助手提出引发问题，使实验对象产出包含关键词的自然话语，话语都被录音并分析。实验假说是:(1) 中级学生的声调正确性会普遍比初级学生的高，（2）不同双音词声调组合的正确率会有差别，1-1正确率最高，3-2正确率最低。
第四章，实验结果及其分析
实验数据显示：在初级和中级实验对象组中，1-1和2-4声调组合的正确率都较高，而3-2声调组合的正确率较低，初级组的3-2声调组合的正确率最低。通过分析我们发现：数字关键词的正确率数据存在着偏斜，而排除数字关键词的数据正好符合假说（2）。结论是：实验对象对基本数字较高的熟悉度提高了相关关键词声调产出的正确率。我们可以推论：声调正确率不但与声调组合的熟悉度密切相关，而且与个别关键词的熟悉度密切相关。
第五章，实验的不足与改进方法
“关键词为对象”实验模式值得在更多的实验中更大规模地使用，本文的探索性实验在实验对象、覆盖声调组合及关键词的数量等方面受到限制。现有的实验需更广泛地使用计算机进行调型分析，并更客观化。
第六章，结语
综上所述，本文的主要贡献是产出性的、“关键词为对象”模式。我们还需更多的实验以使结论更有说服力，并将成果应用于对外汉语教学实践。
关键词：外国人、双音节声调组合、偏误、关键词模式、产出性实验
John Pasden（潘吉）的硕士论文，2008年