HKIER

摘要

識字的歷程，基本上是在學習語音與文字之間的對應關係，或者說是學習如何將使用視覺的文字符號來表示語音，這個歷程又可被稱之為語音編碼。然而不同文字系統的表音透明度不一，也影響了讀者建立形音對應的層次與掌握的速度。本文透過跨語言研究的文獻回顧，探討中文閱讀習得歷程中，如何建立有效的形音對應。文獻指出中文形聲字的唸名表現，也有類似拼音文字研究中的一致性效果。亦即，音旁表音不一致的形聲字的唸名速度比音旁表音一致的字來得慢，正確率也比較低，並在左大腦下額葉，下頂葉，以及枕葉與顳葉交界處等大腦區域顯示較大的活化。此外，電生理的研究更發現有三個在時序以及認知功能上不同的成分，N170, P200與N400與中文字的形音轉換有關係，這些證據支持聲旁表音一致性的計算，從最早期的知覺特徵分析，到後期的語音與語意競爭，皆有涉入。更重要的是，聲旁表音一致性效果只有在該聲旁有很多的鄰項字時最為明顯。這些成果支持統計學習理論，也就是隨著字彙量的增加，讀者可以逐漸掌握中文形音義之間的統計對應關係或並發展出音旁覺識的後設語言知識。

關鍵詞：一致性；規則性；事件誘發電位；功能性磁振造影；閱讀習得

Abstract

To acquire reading skills, children need to learn a set of orthographic codes used in their writing system for representing the units of spoken language. Mastery of this mapping principle and decoding skill allows children to access thousands of words that are already present in their spoken lexicon. Based on the psycholinguistic grain size theory, reading acquisition in different languages, which vary in the consistency of mapping between orthography and phonology, can involve different psycholinguistic grain sizes (Ziegler & Goswami, 2005). This chapter reviews a series of studies to show how Chinese readers capture the mapping consistency between character and sound. Reading inconsistent/irregular phonograms involved longer naming time and greater activation in the brain than reading consistent/regular phonograms. The neural correlates responsible for the Chinese orthography-to-phonology transformation are very similar to what have been suggested for reading alphabetic writing system. Meanwhile, the ERP evidence suggested that the consistency effect in reading Chinese involved the early sublexical phonological computation and later lexical semantic competition. This is especially true for reading phonograms with large orthographic neighborhood (phonetic combinability). Taken together, those evidences support the interplay between orthographic density and the mapping consistency from orthography to phonology in the different stages of lexical processing and support the statistical learning mechanism of literacy.