隱喻熟悉度以及個人心像能力對於理解動作隱喻之影響：事件相關電位研究

Joint Influence of Metaphor Familiarity and Mental Imagery Ability on Action Metaphor Comprehension: An Event-Related Potential Study

本研究操弄隱喻熟悉度及個人心像能力，藉以探討動作感覺經驗對於動作隱喻理解的影響。我們記錄受試者閱讀表達具體動作的句子、熟悉度高的隱喻、熟悉度低的隱喻，以及抽象句子時的腦波反應，並以視覺意像生動程度量表(VVIQ-2)測量受試者的心像能力。實驗結果顯示，心像能力與腦波圖像效果（mental imagery effect，200～750毫秒）呈現正相關。高心像能力者在處理熟悉度低的隱喻時所引發的圖像效果（350～550毫秒）較處理熟悉度高的隱喻所引起的效果持久（350～450毫秒），顯示出他們傾向利用自身的動作感覺經驗來理解隱喻，只是熟悉度會影響心像模擬的時間長短。相反地，低心像能力者在理解隱喻時，並未引發圖像效果，而是在理解熟悉度低的隱喻時，引發較早的不匹配效果(N400 mismatch effect) （200～350毫秒），此效果反應出他們傾向採用普遍的語意處理機制，運用字面意來幫助隱喻理解。在此研究中，兩種理解方式皆受到隱喻熟悉度的影響，且即便採用不同的理解方式，亦皆可成功地獲取隱喻的抽象涵意，然而，這兩種理解方式是否會影響到不同層面如：對話或篇章結構內的隱喻理解，尚待未來的研究多加探討。

The present study investigates whether an individual's mental imagery ability, in addition to metaphor familiarity, affects the degree of sensory-motor involvement during action metaphor comprehension. We assessed participants' mental imagery ability using the Vividness of Mental Imagery Questionnaire-2 (VVIQ-2) and recorded the participants' event-related potentials (ERPs) while they read (1) Literal, (2) Familiar Metaphor, (3) Unfamiliar Metaphor, and (4) Abstract sentences. The ERP mental imagery effect (200~750 ms) in the Literal relative to the Abstract condition was reliably correlated with participants' VVIQ scores. A median split based on the VVIQ scores showed that high-VVIQ participants elicited ERP frontal imagery effects that were more prolonged in the Unfamiliar Metaphor condition (350~550 ms) than in the Familiar Metaphor condition (350~450 ms), suggesting that people with high imagery ability tend to routinely recruit sensory-motor experiences to facilitate metaphor comprehension, and to a greater extent for unfamiliar metaphors than for familiar ones. On the other hand, low-VVIQ participants with less effective mental imagery abilities showed no imagery effects in either metaphor conditions, but an early posterior N400 mismatch effect (200~350 ms) in the Unfamiliar Metaphor condition. The results suggest that low-VVIQ participants tend to rely on more general semantic access mechanisms during metaphor comprehension that detect the semantic mismatch between the unintended literal meanings of the metaphors and the context. Both processing styles are affected by metaphor familiarity and lead to successful metaphor comprehension in the current study. However, whether these processing styles lead to comprehension differences for metaphors appearing in discourses or conversations will require further research.