Writing Systems


Having written a post here on typographic discrimination, when I read a recent article on word recognition, my first thought naturally was, What about Chinese?” But at the time I couldn’t find anything on the differences between how we read English and Chinese. Luckily, I finally came across Liu Ying’s web site. Liu Ying does research comparing brain activity while reading languages with different writing systems. This is a little different than the word recognition paper I linked to which tracks eye movement, but in a forthcoming paper, entitled Orthography to Phonology and Meaning: Comparisons across and within Writing Systems” (to be published in Reading and Writing) he, together with his co-author Charles A. Perfetti address some of the issues involved in making comparisons across writing systems.

Dr. Liu was kind enough to e-mail me his paper, from which I have quoted excerpts below. (UPDATE: Liu Ying provided me with a direct link to the paper. — Thanks!)

The paper starts with this:

Reading is fundamentally about converting graphic input to linguistic-conceptual objects (words and morphemes). Understood this way, a viable theory of reading must include attention to the nature of both the graphic input and the nature of the linguistic-conceptual objects to which they connect. On this formulation, learning to read is learning how one’s writing system encodes one’s language.

This is important, because different writing systems encode different information about language. For instance, English and German vary quite differently in their ability to encode phonological information. Despite Germany’s failure to implement spelling reform (or perhaps the reason for it?), German spelling seems to be much more reliable than English at encoding phonology:

English and German comparisons, for example, suggest that German children and adults can trust their orthography to implement a reliable grapheme-phoneme conversion, whereas English children and adults come to rely more on orthographic whole-word reading … Thus, German children read pseudowords nearly as well as real words, whereas English children do much more poorly on pseudowords.

Chinese, of course, is even more different from English:

Across writing systems, the potential for differences in reading is even more dramatic. Reading Chinese activates phonology, just as reading English and German does. However, unlike alphabetic systems, Chinese phonology is not activated by sub-syllabic connections, because there are none. It is activated at the syllable level, where whole characters and phonetic radicals–constituents of compound characters that are themselves characters—are associated with spoken syllables. Because the characters themselves have meaning, either as single syllable words or as single syllable morphemes that are constituents of multi-syllabic words, reading Chinese is a process that simultaneously (more or less) yields both pronunciation and meaning.

In short:

Chinese reading includes automatic phonology, but this phonology occurs with the recognition of a whole orthographic syllable unit.

The paper also looks at Korean Hangul which is different yet again, but a little too complex for me to go into here (especially since I don’t read Korean). I also feel quite unqualified to summarize the judgments about differences in brain activity when reading Chinese and English. I’ll just leave it here and hope that this post inspires some more blogging on the topic. I’d be especially interested to see something about the results of eye movement studies comparing Chinese and English.