中文讀者對於中文字形表徵字音的構形知識：一個探討表音覺識的研究

Chinese Readers' Knowledge of How Chinese Orthography Represents Phonology

根據Shu等人（Shu, Anderson, & Wu, 2000）的研究，當一位中文讀者的識字量增多時，他能夠發展出一種被稱做表音覺識（phonetic awareness）的後設語言知識。亦即，該讀者能夠瞭解中文字構形表徵語音的原則，並根據此原則猜測生字的讀音。本研究將延續Shu等人的研究，並根據「語言學習的統計分佈理論」（statistical model of language learning;下文簡稱「統計學習理論」）進一步探討表音覺識的發展和內容。根據統計學習理論的觀點，一位語言使用者會根據其所接收到的語言訊息，建構訊息與訊息之關係的機率分配表徵。而這些機率分配不但會影響該使用者的語言行為，更代表了他對語言結構的內隱知識。本研究分析了包含於國小一到六年級的國語課本，以及一個成人字庫中的中文字。分析結果顯現，不論哪一種字庫中的左右排列雙部件字，右部件提供文字讀音訊息的機率均比左部件高。但分析也發現，左部件在約7%的左右排列雙部件字中提供字音訊息。另外，本研究發現，有一組文字不但可以成為其它中文字的部件，而且具有高表音性；亦即，當作為文字部件時，它們經常為文字讀音提供有效線索。由此觀之，中文是一種深層文字，其構形和讀音之間具有一種不完全對應的機率關係。如果統計學習理論是對的，則中文讀者會在文字學習的過程中，習得中文字之形音對應的機率關係。本研究以兩個實驗來檢驗上述的想法。實驗一以國小學童為受試者評估其識字量，並收集他們對兩類假字的標音反應。實驗結果顯示，當受試者的識字量增多時，他們會發展出下面的知識：（一）有些中文字做為部件時，具有高表音性，（二）左右排列雙部件字的右部件和左部件都可能提供字音的訊息，但右部件提供字音訊息的機率較高。在上述的知識下，本研究發現中文讀者會發展一種以右部件猜測文字讀音的便捷策略；實驗二以大學生為實驗對象，驗證了這種想法。

Phonetic awareness refers to a Chinese reader's knowledge of the principles governing the orthography-phonology relationships in Chinese characters (Shu, Anderson, & Wu, 2000). Shu et al. suggested that a phonetically aware reader can infer the pronunciation of an unknown character from its constituent components. In particular, they observed that a Chinese reader uses the right component of a left-and-right arranged two-component character to infer the character's pronunciation. However, analysis of seven character groups used by elementary school students and adults showed that in a left-and-right arranged two-component character, both components may provide cues to its pronunciation. The analysis also found that certain simple characters have higher validity in representing phonology than others. According to the statistical model of language learning proposed by Saffran, Aslin and Newport (1996), a learner can acquire an understanding of the statistical nature of linguistic input through repetitive use of the language. For Chinese characters, the mapping of orthography to phonology is imperfect and probabilistic. Thus, according to the statistical learning model, a Chinese reader needs to understand the statistical nature of the mapping between the pronunciation of a character and its components. We tested this assumption in two experiments. The participants in the first experiment were second-, fourth-, and sixth- graders. Their vocabulary sizes were assessed, and their guesses about the pronunciation of a group of pseudo-characters composed of two left-and-right arranged components were recorded. The experiment results indicate the following. (1) The so-called "position strategy" is an over-simplification of how a Chinese reader uses a character's components to infer the character's pronunciation. (2) A Chinese reader knows which simple characters are more likely to represent the phonology. (3) A Chinese reader knows the probabilities that the various components of a character represent the phonology. That is, the reader knows that the right component is more likely to provide cues to the character's pronunciation than the left component. (4) As a Chinese reader's vocabulary increases, the way he/ she uses a character's components to infer the character's pronunciation may best be described as the so called "position strategy". In the second experiment, the concept of the "position strategy" was tested further and verified, using college students as the participants.