Languages of Power: Metaphorical Grounding of Perceived Power in the Bilingual Mind

Abstract

Embodied cognition theories propose that abstract concepts are grounded in sensorimotor experiences (Barsalou, 2008), metaphorical mappings (Lakoff & Johnson, 1999), and social constructs (Pexman et al., 2023; Borghi et al., 2019). Perceived power is a culturally salient abstract concept, but its grounding across languages is largely underexplored. This thesis examined the neural and behavioural manifestations of perceived power embodiment in Chinese-English bilinguals across three experiments, investigating how power-related concepts are metaphorically associated with vertical space and physical size, and how such grounding varies between first (L1, Chinese) and second (L2, English) languages.
Study 1 employed an auditory source localisation task with event-related potentials (ERPs) in which participants reported whether power words (e.g., ‘king’, ‘servant’) and ‘location words’ (words associated with upper and lower locations, e.g., ‘sun’, ‘ground’) originated from above or below their sitting position. High-power words produced faster responses when presented from above (congruent) versus below (incongruent), whilst no such congruency effect was observed for low-power words. Congruent high-power words elicited enhanced P3 and reduced N400 amplitudes in L1 only but there was no embodiment effect for location words. English controls showed no embodiment effects in Study 1.
Study 2 investigated power–size association using a font-size judgement task. High-power words in bold-larger font prompted faster and more accurate responses than those in thin-smaller font, with stronger effects in L1 than L2. English controls showed a robust congruency effect, twice as large as that found in bilinguals operating in their L2 English. Low-power words showed no congruency effects in either group.
Study 3 examined the neurophysiological basis of power–size embodiment, comparing high-power words to matched control words unrelated to power (e.g., ‘daylight’). Larger fonts elicited enhanced P3 and late positive component (LPC) amplitudes for both power and control words, but font-size effects were significantly stronger in the case of power words. Unlike in Studies 1 and 2, congruency effects were comparable between L1 and L2, with L2 proficiency correlating positively with effect magnitude across all late ERP components. Differential ERP patterns across experiments reflect distinct processing mechanisms: N400 modulation in Study 1 indicated semantic integration difficulty when power–verticality metaphors were violated, while P3/LPC modulation in Study 3 reflected enhanced attention for metaphorically congruent power–size mappings. Perceived power embodiment varies systematically across metaphorical domains and languages, with physical size representing a more robust and universal grounding mechanism than vertical space. This may partly reflect modality differences (visual vs. auditory tasks) in the present design. Future research could test whether the robustness of size effects generalises across modalities. The asymmetric pattern between high- and low-power words supports a self-relevance hypothesis I propose, whereby concepts with greater personal relevance engage embodied representations more strongly than concepts with ambiguous self-relevance, extending multiple representation theories by incorporating individual relevance as a key modulator of abstract concept grounding.

Date of Award	25 Nov 2025
Original language	English
Awarding Institution	Bangor University
Supervisor	Guillaume Thierry (Supervisor) & Gary Oppenheim (Supervisor)