這個研究的目的是要探討台灣EFL(English as foreign language)學習者對於英語的字詞辨識歷程。全詞辨識或是序列的辨識一直是研究字詞辨識的主要議題，在這個研究中，在行為上使用詞長效果作為驗證字詞辨識歷程究竟是平行式或是序列式的指標，為了進一步瞭解學習對不同詞性處理的異同，生理上則使用事件誘發電位來直接測量EFL學習者辨識不同詞性刺激的腦部變化。 實驗一的目的是發展出簡易檢測英語能力，也就是流暢性的指標，使用詞彙辨識作業的紙筆測驗版本及傳統的詞彙翻譯作業，計算受試者參加校內英語測驗成績及兩項作業的相關，結果兩種作業皆與校內英語成績相關達顯著，顯示兩項作業皆可預測受試者的英語流暢度。 實驗二、三及四為檢驗EFL學習者的詞長效果，以傳統的詞彙辨識作業進行檢測。結果發現不論是在辨識真詞與假詞、真詞與非詞、或是在進行咬音抑制的情況之下辨識真詞及假詞，都出現顯著的詞長效果，顯示EFL學習者是序列式的進行英語字詞辨識。 為了檢視EFL學習者的字詞辨識發展階段，實驗五及實驗六分別以行為實驗及事件相關電位探討不同英語能力的EFL學習者之詞長效果及腦部變化，結果不論是在詞長效果或是腦部變化的反應形式都顯示高成就組與英語為母語使用者較為類似。 目前的研究對台灣EFL學習者的字詞辨識歷程提供了初步的結果，包括EFL學習者是序列式的進行字詞辨識，且其中高成就組與英語為母語使用者較為接近的反應形式。This research explored the word recognition process in English of EFL learners. Both behavioral and electrophisological arpproaches were adapted. In experiment 1, LDT (lexical dcision task) through paper-and-pencil and vocabulary translation task were used to index the fluency abilities in English . The performances in both tasks were highly correlated with the scores of English classification examination done by university autority of each subject. To observe the word length effect, in the experiment 2, the subjects were asked to distinguish words from word-like pseudowords. To further clarify that the property of the pseudoword stimuli was not the reason why the serial decoding was found in the experiment 2, in the experiment 3, the subjects were instructed to distinguish words from random consonant strings. In order to eliminate the possible naming effect causing word length effect in LDT task, in the experiment 4, the subjects were instructed to distinguish words from word-like pseudowords while performing a concurrent task of articulatory suppression. A significant word length effect in LDT task was observed in all the three experiments. In experiment 5, to explore the developmental curve of English word decoding of EFL learners, subjects were recruit by their fluency. Word length effects were notable in the high and the low achievers when performing the same task in experiment 2. In experiment 6, ERPs were employed to examine the process of English word recognition. The results using N400 as an effect showed a pattern similar to native spaers only for high achievers but not for low achievers. The current research shows a robust and consistent word length effect and the pattern of word length effect and ERP component were more native-like in high achievers than in low achievers.
