英文為外語學習者理解英文隱喻的神經認知機制

Abstract

本研究旨在探討以英文為第二語言之學習者(second language learners)對於英文隱喻(English metaphors)的理解，以大腦事件相關電位(Event-Related Potentials)為研究方法，探討非母語使用者(non-native speakers)相對於母語使用者(native speakers)在認知處理歷程所產生的困難點。以往文獻多集中在母語使用者對於隱喻的理解。文獻中發現，相對於字面意義用語，隱喻字面意義用語，隱喻會引發較負的N400以及較大的LPC，這些結果顯示在提取正確的隱喻語義(metaphorical meanings)時，字面上的意思(literal meanings)也會同時、或是很短暫的被激發，這樣的語意衝突進而促使整體語意的重新理解與分析。相關研究指出，非母語使用者因為詞彙認識有限、個別差異（如學習的年紀、語言程度高低等）、文化異同等的影響而在處理譬喻性字詞(figurative meanings)時有一定程度的理解困難。在相關的第二語言學習者的行為實驗(behavioral experiments)更顯示，非母語使用者在理解譬喻性字詞時，字面上的意思在反應速度(reaction times)上較譬喻性字詞時間來的短。本篇主題以探討非母語使用者在神經認知機制上的差異與其可在教學上的應用。研究採用30位以中文為母語、英文為第二外語、托福至少達到100分的高英語能力受試者，刺激材料為英文的名詞性隱喻句子，句子包含三類：(1)表面語義句子(literal meanings)，如：Cholera is a plague that affects many people.；(2)隱喻性語義(metaphorical meanings)，如：Unemployment is a plague that affects many people.；(3)歧義(anomalous meanings)，如：Metal is a plague that affects many people.。結果顯示，歧義句相對表面語義句子的結果與母語人士的表現相近，句中的關鍵字(critical words)有N400和LPC的效果，表示其表面語義被激發，並嘗試根據已出現的句子做理解，至字尾有明顯的N400效果，顯現整句語義無法獲得合理解讀。隱喻性句子與母語人士的表現不同，在句中N400和LPC的效果不明顯，在句尾我們發現一個相對於表面語義句子的正向腦波。進一步我們將隱喻性的英文句子根據中文是否有跟英文相對應的說法分成兩類，結果顯示英文能力自我評量分數和托福成績在有中文相對用法的句子有相關，在有相關用法的那組句子所產生的語意連結較為豐富，產生較大的N400和LPC效果。本研究顯示，非母語使用者在處理隱喻性語義時，足夠的語義相關概念連結可以幫助理解隱喻性語義。

Understanding figurative speech in a second language (L2) is a bottleneck for L2 learning. However, little is known about the nature of these difficulties. Studies done with native speakers have shown that metaphor understanding involves retrieving metaphorical meanings shortly after or concurrently with literal meanings from semantic memory and additional controlled processes to resolve the conflict between literal meaning, figurative meaning and the context (indexed by a more negative N400 and a late positive component, or LPC, to metaphors relative to literal expressions respectively)(De Grauwe et al., 2010). Understanding which aspect(s) the challenge for L2 learners’ metaphor comprehension lies is therefore of pedagogical value. We assessed event-related brain potentials (ERPs) from proficient Taiwanese learners of English while comprehending English nominal metaphors familiar to native speakers as well as literal and anomalous sentences (e.g., Metaphorical: Unemployment is a plague that affects many people. Literal: Cholera is a plague that affects many people. Anomalous: Metal is a plague that affects many people.). Sentences were taken from Experiment 2 of De Grauwe et al. (2010), and were presented word by word on the screen (duration: 400ms; ISI: 100ms; same presentation speed was used for native speakers in De Grauwe et al.). Participants had to judge whether the sentence made sense at the end of each sentence. Thirty proficient Taiwanese learners of English were tested (mean age: 23.4, range 21-29, 16 males); all scored 100 or above on a TOEFL test taken within 2 years. Participants were 78% accurate on average in the plausibility judgment test (Literal: 82%, Metaphorical: 72%, Anomalous: 79%). ERP results from correct trials showed anomaly effect similar to what was previously observed in native speakers—compared to metaphor and literal conditions, anomalous condition elicited, on the critical words underscored in the examples, reliably more negative N400s (300-600ms) and more positive LPC responses (800-1000ms), and on sentence final words, more negative N400s. By contrast, ERP responses to metaphors were reliably more positive than literal sentences only on the sentence final words. Further analysis on critical words showed that, compared to the literal condition, metaphors that have a corresponding expression and therefore an existing conceptual mapping in Chinese did elicit more negative N400s and larger LPCs, as was seen before in native speakers. These effects were reliably correlated with proficiency indices such as self-rated proficiency and TOEFL scores, with higher proficiency correlated with larger N400 and LPC effects. These results suggest that L2 speakers may be less sensitive to the semantic relations between the source and target concepts and the literal and metaphorical interpretations so as not to show differential responses on the critical words but delay the conflict resolution process till the end of a sentence. However, existing conceptual links in the native language aid the comprehension process and enable L2 speakers to engage qualitatively similar mechanisms as do native speakers. Our results thus suggest that emphasizing the conceptual mappings between the literal and metaphorical interpretations may facilitate metaphor understanding in L2, which in turn can help build higher L2 proficiency.