Cognitive mechanisms of creative meaning decoding for non-literal language—An eye-tracking study on Chinese logogriphs

This eye-tracking study investigates the cognitive mechanisms underlying creative meaning decoding in Chinese logogriphs-character-based riddles whose solutions require a metalinguistic reinterpretation of the riddle surface. Unlike figurative language, where non-literal meanings operate via semantic-pragmatic extension, logogriphs demand that solvers inhibit the salient literal interpretation and instead access low-salience metalinguistic properties (e.g., the interpretation of orthographic decomposition, spatial manipulation of character components). Experiment 1 revealed a three-stage processing sequence: initial dominance of literal meanings (0-1000 ms), competition between literal and metalinguistic interpretations (1000-2000 ms), and eventual resolution toward the correct metalinguistic solution (2000-4500 ms). Experiment 2 manipulated contextual primes (literal vs. non-literal contexts), demonstrating that non-literal contexts enhanced accuracy and accelerated resolution by cueing the relevant metalinguistic operation, whereas literal contexts reinforced literal interference and delayed metalinguistic access. These findings support the Graded Salience Hypothesis, showing that salient literal meanings are activated automatically regardless of context, whereas non-salient metalinguistic meanings require controlled processing and contextual facilitation. By establishing metalinguistic meaning as a distinct category of non-literal language alongside figurative meaning, this study extends current models of non-literal comprehension and provides new insights into the cognitive architecture of creative meaning decoding.