Examining Readability of the Texts in the English Psychology, Management and Economics Textbooks from a Discourse Analysis Perspective

從篇章分析的角度探究大學心理、管理和經濟學原文教科書特定篇章之可讀性

This study examines readability from a discourse analysis perspective. The data for study are 15 most popular English textbooks on psychology, management and economics in Taiwan. The first study step is to key in all the units in the textbooks and measure each unit's readability via Coh-Metrix L2 Reading Index. Then the most and the least readable units are sorted out for analysis via the approach of Lautamatti (1987) and Sympson (2000), based on their indices. The goal is to analyze (1) both units' topical structure; (2) their information structure and information flow; (3) how each unit uses syntax to ground various types of information; (4) similarities or differences among the 3 least readable texts. It is found that readability is constrained by 3 variants: (1) complexity of the topical, information and syntactic structures; (2) if more global topical progression strategies are used to achieve continuity; (3) if the text's information flow fits the ''from accessible to new'' pattern and if identifying the topic requires inferencing skills and highly integrative understanding of the prior text? In terms of the impact on readability, the third variant is the strongest of all, and the first, the weakest. As a result, researchers should always prioritize the third variant, especially when it affects readability in a different direction from the other two. Last, this study agrees that the ''word overlap index'' in the Coh-Metrix L2 Reading Index is very effective in calculating a text's readability.

本研究企圖從篇章分析的角度切入探討影響可讀性指數的變數。研究的語料是大學心理學、管理學和經濟學三個學門各五本在台最常被採用的原文教科書。研究步驄首先將纊共十五本教科書所有的篇章輸人電腦，用Coh-Metrix L2 Reading Index的公式計算文本中每一篇章的可讀性指數，再根據計算出來的指數挑出其中最簡單和最困難的篇章，各選其中完整的一部分，用Lautamatti（1987）以及Sympson（2000）的理論架構，進行質注分析，比較（一）兩篇的主題發展。（二）兩篇的訊息結構和訊息流動情形。（三）兩篇如何運用句法結構來安置各種訊息。（四）分析三個學門裡最難的篇章彼此間的異同。研究結果發現篇章的可讀性受制於以下三種變素：（1）篇章的王題、句法和訊息結構的複雜程度；（2）用來達成主題連貫性的銜接策略的層次偏高；（3）訊息流動違反「從舊到新」的模式以及主題的辨識要求讀者進行高層次、整合性的推論。就箝制力強度而言，第一類變素的影響相對薄弱，第二類其次，第三類最為顯著、最能正確的反映可讀性指數。這三種變數合併一起考量可以提高預測的準確度，但前提是它們必須都是往同一方向作用，如果彼此的影響發生扞格時，則優先考慮第三個變素，以它為判斷的依歸。另外本研究也間接支持Coh-Metrix L2 Reading Index公式中的「字詞重複指數」確實是一個很有效的計算指標。