主题：【原创】浅谈汉字与拼音文字的比较 （上） -- 人间树

A REVIEW OF RELEVANT LITERATURE

English orthography looks better than the Chinese orthography in terms of relating the script to speech. Written Chinese is different from that of English in that the former is a logographic system. In the Chinese logograph, each character is a pronunciation unit. Some researchers (e.g., Au, 1992; Rozin et al., 1971) proposed that reading Chinese characters does not involve the processing of phonological recoding. That is to say, phonological information is not involved, or at least not as important as reading an alphabet such as English.

This statement has been criticized in the past few decades (Leong, 1973). Critics argue that about 85% of all Chinese characters used today belong to a category called "radical compound" (Wang, 1973, 1981). A radical compound is composed of two parts: a "radical" and a "phonetic." The radical presents the semantic or categorical information (e.g., an object made of metal; a person ... etc.); and the phonetic provides the phonological information.

Previous studies on reading efficiency have focused on the measure of eye fixations (pauses made by a reader's eyes along the lines of print). Just and Carpenter (1980) has used an aggregate measure (gaze duration) that sums all fixations beyond some minimum. It is more typical in eye movement research not to summate over fixations. According to Kliegl, Olson, and Davidson (1982), the procedure of Just and Carpenter (1980) produces misreadings of fixation time that cause problems for theoretical modeling.

To examine fluency in reading English and Chinese texts, Just, Carpenter, and Wu (1983) conducted experiments in their laboratory. College students were asked to read short scientific texts and the eye-fixation was monitored by computer. The English version of the texts was particularly revealing about word decoding and processes that translate written symbols into mental concepts. The Chinese version of the experiment also focused on word encoding processes because these processes were most likely to be influenced by the properties of a particular orthography.

The experiments examined the eye-fixation of 13 native Chinese speakers as they read Chinese translations of the 15 scientific texts. The Chinese readers had received their college undergraduate education in Hong Kong or Taiwan. Their reading ability was compared with their American counterparts, although the two groups of readers differed in several ways. For example, the Chinese students were bilingual and generally had more undergraduate or graduate education than the American students.

The texts used in this study were translations of the scientific passages excerpted from Newsweek and Time magazines, presenting generally unfamiliar information on scientific or technological topics. The translations into Chinese were made by a native Chinese speaker who was also fluent in English. Every attempt was made to produce Chinese texts comparable in style and difficulty to the English versions, by translating the texts sentence by sentence and by using comparable vocabulary.

The researchers found that, in English, native readers spent more time on long words than on short words, and in Chinese, the gaze duration on Chinese words increased with the number of characters in a word. The interpretation given to this finding in the English readers is that encoding processes operate on letters or letter groups and that the letters or letter groups are processed once at a time. As a result, longer words take longer time to encode. The word-length effect in Chinese cannot be identical because of the difference between the two orthographies.

In another important study, Gray (1956) compared the eye fixations and reading speed of a group of 78 adult readers of 14 different languages, including Arabic, English, Hebrew, and Chinese. Most of the readers were university graduate students and native speakers of their original language. They were currently living in the United States but had received all or most of their elementary and secondary education in their native country.

The readers were presented two texts that had been translated into their native languages. The translated versions varied from language to language in the number of words and lines of print. For example, in English, each of the two experimental passages contained approximately 150 words and covered 13 lines of print. In Chinese, each took approximately 140 words, 5 lines, and covered a much smaller space. These differences reflected variation in space (e.g. number of lines) that was considered to express the same concept in different languages. Nevertheless, the differences were also indicative of the large variation among the orthographies.

Controversially, Gray concluded that orthography had little effect on the nature of the eye fixation of the readers, although there was some slight variation in the average number of eye fixations and their average duration. For example, the average number of fixations was 1.6 words/fixation for English, 2.5 words/fixation for Chinese, and 1.3 words/fixation for Hebrew. Gray explained these differences with accidental differences among the various samples of readers rather than with the orthography. In conclusion, the general finding of Gray's study was that at a very general level, fluent reading was similar across different languages and orthographies. In other words, orthography had little effect on the nature of the eye fixation of the readers, although he found some slight variation in the average number of eye fixation and their average duration. Gray attributed the differences to (1) the fact that the linguistic unit of words varies from language to language, and (2) accidental differences among the various samples of readers rather than to the orthography itself.

As the study by Just, Carpenter, and Wu (1983), Gray's study was flawed in at least the following five ways. First, his sample consisted of only international graduate students studying in the USA who spoke their native languages as well as English as their second language. The subjects' proficiency in English varied, and their proficiency in their native languages may also vary since they left their native country at different ages. Second, the sample size of only 78 subjects was small. The 78 subjects represented 14 different languages, with an average of 5-6 readers attending each language. Third, there was no control of educational levels. It was claimed that the readers had received all or most of their elementary and secondary education in their native language, but their native language reading proficiency should vary contingent upon the amount of education they received before they left for the USA. Fourth, reading speed was not objectively and straightforward measured. There is no doubt that eye fixation is a measure of reading speed, but a more straightforward and efficient measure of reading speed should involve subjects' reading long passages for which time is monitored and comprehension is checked. Fifth, in Gray's study, subjects' decision of how carefully and quickly to read was not controlled for. The variation in eye fixation could be a consequence of the differences in the readers' attention rather than those in orthography.

Previous researchers have tried to investigate linguistic factors that affect reading speed, but without much success. Mann (1986) and Willows (1974) used developmental data in their studies of the effect of phonemic awareness on children's later reading skills. They found that phonemic awareness was a good prediction of children's later reading skills. In English, there is little doubt about the importance of phonological information in the process of word perception. In other words, English language allows its readers to process the information from phonological script to comprehension in an easy way, as long as the reader is able to identify the phonemes.

On the other hand, phonological recoding was thought to be unnecessary in reading Chinese characters. Some researchers have intuitively thought that reading Chinese may not require any phonological information (Tzeng, 1994). By definition, it is "possible" to get the meaning onto a logograph without the mediating of phonological information anyway.

Studying the differences in various orthographies and factors that influence reading efficiency, previous researchers have made significant contributions to our understanding of the issues. Some empirical studies (e.g. Just, Carpenter, & Wu, 1983; Mann 1986, 1988, 1989; and Willows, 1974) have supported the hypothesis that reading efficiency differs from language to language in accordance with the type of orthography, while some others (e.g. Gray, 1956) have not been able to make the connection. However, previous research on the effect of type of orthographies on reading efficiency has at least three weaknesses: (1) reading efficiency (or speed) was not straightforwardly measured, (2) extraneous variables (such as the subjects' educational level and their proficiency in the examined language) were not adequately controlled for, and (3) the sample sizes were not big enough for reasonable generalizations.

To study the reading efficiency without the above-mentioned problems, an attempt was made by Everson (1988) to investigate the variation in reading speed and comprehension from logographic and alphabetic orthographies. Everson selected 60 first year Chinese learners from the United States Air Force Academy to read a same passage printed in either in pinyin (romanization) or characters. Results of the study indicated that the learners of Chinese as a second language read the romanized form of Chinese faster and with better comprehension than they read the passage in the form of characters.

Explanations by the researcher for the finding are in the difference between romanization and the character. Different from the characters, the romanized form of Chinese denotes the four tones, which are distinguishable phonemes, capitalizes the first letter of a proper noun, and writes syllables of one word together. Results of the research might have told us that romanization is more efficient for reading than characters. The research might have provided supporting evidence for the Chinese language reform. However, the finding might also be a function of the subjects' first language, English, which is an alphabetic orthography. People tend to be more used to the script they were more familiar with. Second, the finding might also be a function of instruction in which pinyin was taught more than the characters. Further, subjects of the research were all first year students of the language, and more studies with higher level Chinese learners might be needed to investigate the variation in reading efficiency between romanized and characterized scripts.

Some similar research was conducted for the Japanese language. Researchers compared subjects' response speeds to Kanji (logographs, Chinese characters) and Kana (phonetic graphs) among Japanese readers of various levels, and found that the subjects' response to Kanji was three times faster than that to Kana. Those researchers argued that three steps have to be involved in reading an alphabetic script like Kana: graph, sound, and then meaning, but in reading a logographic script, only two of the three steps are needed: from the graph to the meaning (Xu, 1996). This partially explains why reading Chinese is usually faster than reading English.

Eventually, only by comparing native readers of different orthographies can we find the reliable and valid variances in reading efficiency as a function of the orthography. The current research project is designed to exclude all the previously documented extraneous variables so as to conduct a more systematic study of the effects of orthography on reading efficiency.

This project is intended to use a more scientific design with large samples to investigate the reading efficiency in English and Chinese. Based on what we know about the orthography theories and previous studies, it is hypothesized that:

1. Native Chinese reading is generally faster than native English reading;

2. on a speed reading test, native Chinese readers score higher than native English readers in comprehension rates for the same reading contents;

3. when reading efficiency is calculated by comprehension score (number of questions answered correctly) divided by speed (number of minutes used to finish the reading), native Chinese reading is more efficient than native English reading;

4. with comprehension rates being controlled for, native Chinese reading is generally faster than native English reading; and

5. with speed being controlled for, the comprehension rate in native Chinese reading is higher than that in native English reading.