以資訊理論探討注音與拼音的語音表徵特性及其對第二語言學習者的教學應用

Zhuyin vs. Pinyin: An Entropy-Based Study of Mandarin Phonological Representations for L2 Learners

中文的語音表徵系統對於第二語言學習者的漢語語音習得與語音教學設計具有關鍵影響。本研究運用資訊理論中的熵（entropy）指標，針對現今最常使用的兩種中文語音標註系統──注音符號與漢語拼音──進行系統性的量化比較，探討其語音透明度、結構複雜度與國際音標（IPA）之對應關係。研究透過香農熵（Shannon Entropy）、條件熵（Conditional Entropy）、互資訊（Mutual Information）與Kullback-Leibler散度（Kullback-Leibler Divergence）等指標，分析兩系統於音節、聲母與韻母三個層級的資訊分布。結果顯示，儘管三種表徵系統在總體資訊含量上相同，注音符號於韻母層面的熵值較低，且符號區辨性高，有助於初級學習者的音位辨識與穩定輸出；相對地，拼音系統與IPA的對應度較高，亦展現出在語音輸入與語音技術應用上的潛力。研究成果不僅提供語音表徵的新視角，也對華語教學中的標音系統選擇與教學策略提供實證依據與建議。

The phonological representation of Mandarin Chinese plays a crucial role in both second language acquisition and computational language processing. This study employs an information-theoretic framework (Shannon 1948) to quantitatively compare Zhuyin and Pinyin—the two most widely used systems for encoding Mandarin sounds—by examining their phonological transparency, structural complexity, and alignment with the International Phonetic Alphabet (IPA). Using entropy- based metrics such as Shannon Entropy, Conditional Entropy, Mutual Information, and Kullback-Leibler Divergence, the study analyzes each system at the syllable, initial, and final levels. Results show that while total phonological information is equivalent across systems, Zhuyin exhibits lower vowel entropy and greater symbol uniqueness, potentially benefiting early-stage L2 learners. In contrast, Pinyin demonstrates closer alignment with IPA and higher computational efficiency, making it advantageous for learners with IPA background and digital language applications. These findings offer a novel lens for understanding phonological systems and provide pedagogical insights into script selection for Chinese language instruction.