Advantages and Disadvantages of Processing Simplified and Traditional Chinese Scripts

September 1, 2007

The Chiang Ching-Kuo Foundation for International Scholarly Exchange (USA) has awarded BU Psychology Professor Catherine Caldwell-Harris a grant titled, "Advantages and Disadvantages of Processing Simplified and Traditional Chinese Scripts."

Most of the Chinese speaking world currently uses a simplified version of Chinese script. Simplified script is believed to be easier to learn because stroke-number and complexity were decreased during writing reform in mainland China in the 1960s, as part of a larger project to promote literacy in rural areas. However, the simplification process reduced semantic transparency and phonetic consistency. Many characters are now homonyms (the same character refers to two unrelated meanings, as in English words such as bow). In other characters, the components that afford clues to pronunciation and meaning have been lost.

Taiwan never adopted script reform and continues to use Traditional script. The existence of two scripts provides a useful window onto human pattern recognition abilities. While there is a growing body of work comparing reading processing in alphabetic and logographic (i.e., pictographic or character based) scripts, the two versions of Chinese scripts have never been directly compared for ease of learning, ambiguity resolution or ability to infer the meaning of novel vocabulary items from context.

Dr. Caldwell-Harris developed the project with the assistance of doctoral student Hui-wen Cheng, a native of Taiwan, now in her third year in BU's Applied Linguistics Program. When administering character recognition and reading tasks to participants who read with traditional script in Taiwan, Caldwell-Harris and Cheng will be collaborating with Dr. Su-Ling Yeh of National Taiwan University. When studying participants from Beijing, who have learned simplified script, they will work with Dr. Hua Shu of Beijing Normal University.

Script reformers in the 1960s focused on the difficult of memorizing a large number of characters, many of which may have 10 or more strokes. A simplified signal is advantageous when the goal is to reproduce the signal (as in writing), but a richer, more informative signal is an advantage when recognition can be used, as in for comprehension. One hypothesis to be investigated is that simplified script has advantages in the early stages of learning, but that traditional script offers more clues to meaning and thus is advantageous for advanced learners.

The researchers are also interested in how the advent of computers may have altered the information processing trade-off between recall (necessary for writing) and recognition (necessary for reading). With computers, a writer can type a phonetic component and 5-10 chinese characters pop up on the screen. At that point, humans' powerful pattern-recognition abilities will be able to select the desired character. This is a method of writing, unimaginable 40 years ago, in which stroke reduction may not be helpful and may even hurt recognition.

Caldwell-Harris and Cheng view their project as a blend of cognitive psychology, reading research and culture. Cognitive psychologists try to discover the basic properties and computational limits of human information processing. One question raised by their plan to compare reading across two scripts is how well readers can tolerate irregularities in a writing system. A second concerns the distinction between a writing system that evolves naturally through generations of use, and one that is designed (or modified) by a committee in a few years. Can a committee of linguists and writing experts design a system that fully takes advantages of human processing abilities? Cognitive scientists are skeptical of human engineering of languages and writing systems, because committees are frequently ignorant of how well cognitive systems can tolerate ambiguity and irregularity. A celebrated (or notorious) example is the failure of the gestural language called Signed English, designed by committee, to function as a successful communication medium for deaf individuals.

The researchers will construct a corpus analysis of Taiwanese (traditional script) elementary school texts, to complement an existing analysis of simplified script done by their Beijing collaborator Dr. Shu. The corpus analysis will reveal whether the two scripts differ in the proportion of high and low frequency characters which are phonologically regular and semantically transparent. A series of experiments is proposed to examine ease of reading and writing the two scripts by adult learners of Chinese as a foreign language, and studies of guessing of the meaning and pronunciation of novel vocabulary by Chinese high school students and adults.

This will be a new project for Dr. Caldwell-Harris, yet it draws on her decades of research studying visual word recognition. She has several papers on English word recognition, and two papers on Turkish word recognition. She has also has an on-going study of bilingualism which includes Mandarin-English speakers who grew up in the U.S. or overseas.

Conducting experiments on Chinese scripts will allow researches to make connections between cognitive principles and practical applications. An example application is that the people of Taiwan may want to consider script simplification, but the irregularities and ambiguities in the simplified script of mainland China suggest that Taiwan may want to develop its own simplification.

In the years to come more American citizens may want to learn Chinese. Americans who have grown up speaking Chinese in the home may want to make an informed choice about which script to learn first. Educators will be able to draw the researchers' comparison of the two scripts when teaching the Chinese writing system.

Mandarin speakers who have expertise in writing Chinese may assist or participate in the research studies by contacting Hui-wen Cheng at hwcheng@bu.edu or Dr. Caldwell-Harris at charris@bu.edu.