語言障礙學生口語說明能力之微觀與鉅觀結構分析

本研究之目的在於以微觀和鉅觀的層面分析國小高年級語言障礙學生和正常發展學生口語說明能力的表現，並探討有效區分預測語言障礙的口語說明指標。以立意取樣和配對的方式選取基隆市、臺北市、新北市、台中市之高年級語言障礙學生和正常發展學生各20名進行研究。研究者以改編之遊戲或運動語言作業與提示表，收集每位學生的說明性語言樣本。從語意（詞彙量、詞彙多樣性）、語法（語句量、語句長度、語句複雜度、特殊句式）、錯誤分析（語詞、語句之錯誤）及迷走語（修正、重複、填補之內容）等四個面向進行微觀結構之分析；以修訂之說明能力評分架構評量學生口語說明之鉅觀結構能力。本研究主要發現如下：一、語意方面，語言障礙學生口語說明時使用之「總詞彙數」、「相異詞彙數」與「校正後相異詞的出現率」均顯著低於正常發展學生，顯示語言障礙學生口語說明時有詞彙量不足與詞彙運用困難的問題。二、在語法方面，語言障礙學生在說明時語法能力的表現不僅在量的方面較正常發展學生少，在質的方面亦有明顯的組間差異，包括:。在語句數量方面，語言障礙學生所表達出的單句數過少，較無法表達出詳細與豐富的內容；在語句的長度方面，語言障礙學生的平均主謂句長度顯著較正常發展學生短，顯示「平均主謂句長度」可以做為偵測華語學生語言障礙的語法指標；在語句複雜度方面，語言障礙學生在說明時較少使用嚴密性高的「主謂句」，表達內容的邏輯性與精緻度較同齡學生不足。另外，本研究也發現「主謂短語的數量」能偵測出華語語言障礙學生在說明時，語句階層性與複雜性不足的問題。在特殊句式方面，語言障礙學生說明性言談中「被字句量」不足的問題，是區辨語言障礙學生的一項語法特徵。三、錯誤分析方面，語言障礙學生在說明時單句中錯誤比率顯著較高。在個別錯誤的類型分析中，則是以「單句錯誤使用語詞的比率」和「單句語意關係不明句比率」等語意層面的錯誤明顯高於同齡學生。四、在迷走語方面，語言障礙學生和正常發展學生在「迷走語詞佔總詞彙比率」、「單句迷走語數」、「單位迷走語詞彙數」和「單句放棄語句比率」等計量分析結果沒有顯著差異，但在「總迷走語詞彙數」和「迷走語單句數」方面則是正常發展學生明顯較多。五、鉅觀結構方面，兩組學生的口語說明鉅觀結構能力有顯著差異。其中「準備工作」、「進行的過程」、「得分」、「結束」、「策略」、「專門性術語」等6項元素與「綜合比率分數」均能有效偵測語言障礙學生鉅觀結構能力的缺損，可以做為口語說明介入的參考指標。六、語言障礙學生和正常發展學生的區分預測指標之分析結果，在微觀結構分析層面以「相異詞彙數」、「平均主謂句長度」、「被字句數」之組合性指標有最佳之區分正確性。在鉅觀結構分析層面，則是以「策略」、「得分」、「進行過程」等3項之組合性指標具最佳的區分效果。最後，根據上述的發現，本研究提出未來研究及實務應用方面的相關建議。

關鍵字

口語說明能力；語言樣本分析；語言障礙；鉅觀結構；微觀結構

並列摘要

The purpose of this study was to examine the micro- and macrostructural performance of expository discourse in students with language impairment (LI) and typically developing age-matched peers (TD). Expository discourse is a challenging discourse for school-age children, and has recently gained more attention because of its academic importance. An effort has also been made to explore the linguisitc indicators for differentiating Mandarin-speaking students with LI and TD. Expository data were collected form 40 Mandarin-speaking students enrolled in fifth and sixth grade, using a favorite game or sport protocol. Language samples were analyzed for microstructural measures of semantic, syntax, error, and maze. The macrostructural components of expository discourses form students were also assessed using a revised Expository Scoring Scheme (ESS). The results showed that LI was significantly poorer on most of aspects of microstructural measures, including semantic measures (ie., total words, number different words and corrected type token ratio), syntax measures (ie., number of Chinese simple sentence, subject-predicate sentence, subject- predicate construction, bei-sentence, mean length of subject-predicate sentence) and error measures (ie., percentage of total errors per Chinese simple sentence, word error, unclear context). Unexpectedly, only two maze measures（total maze words, Chinese simple sentence with maze）were significantly higher for TD. LI students exhibited inferior syntactic and semantic abilities in expository discourse than their age-matched peers. The group differences were also observed on ESS ratio score and six ESS components. The performance on ESS and the scores of six ESS components distinguished the language delay in use of expository macrostructure for LI students from the TD group, and could be applied to clinical intervention for expository macrostructure skills. Combining three microstructural measures, i.e., number of different words, mean length of subject-predicate sentence and bei-sentence demonstrated the highest accuracy of distinctive prediction for classifying LI and TD. Furthermore, the three ESS components, i.e., “strategies”, “course of play” and “scoring”, were the expository discourse macrostructure measures demonstrated the highest accuracy of distinctive prediction, with correct identification of 95% for LI and 85% for TD. In conclusion, LI have shown both deficits in microstructure and macrostructure of expository discourse. Based on these findings, the suggestions for future studies and implications for using the expository discourse in the clinical assessment and intervention were discussed.