Traditional vs. Simplified

From the New York Times "Room for Debate" section

How a Computer Might Respond

Norman Matloff

Norman Matloff is a professor of computer science at the University of California, Davis, and is the author of KuaiXue, a software tool for learning Chinese.

The original rationale for simplification was to accelerate the learning process. But is this necessary today, given China’s much improved economic and social conditions? There may be no easy answer.

What’s certain is that converting from the simplified characters, or jiantizi, to the traditional characters, fantizi, would be a huge task, affecting everything from school textbooks to government documents to online systems. Automation of that process would present serious technical challenges.

The trouble stems from fundamental differences in the two character sets. The simplification process of the 1950s sometimes resulted in two different traditional characters becoming identical in simplified form. For instance, the traditional characters 發 (”develop”) and 髮 (”hair”) are both written as the simplified character, 发. When the software sees the latter, it must guess which of 發 and 髮 is intended. Typically the guess is made by analyzing context. Sometimes, the software can produce the occasional howler. A passage describing “loss of face” might be translated by the computer as loss of 麵 (”noodles”) rather than loss of 面 (”face”)!

So while most of the process could be automated, especially with more fine tuning in the software, much work would need to be done by hand as well.