What makes a chengyu? Native speaker intuitions for categorising Chinese idiomatic expressions

Abstract

'Chengyu' are distinctive formulaic expressions central to Chinese language and culture. While many chengyu originate in classic literary texts, these expressions are widely used in contemporary communication, featuring in everything from Chinese literature and political discourse to advertisements and everyday conversation. Thousands of chengyu are listed in authoritative dictionaries and explicitly taught in the Chinese school curriculum, serving as markers of Chinese proficiency and cultural literacy. Yet despite their cultural prominence, extensive documentation, and formal transmission, we know little about ordinary Chinese speakers' understanding of chengyu. This thesis investigates how native Chinese speakers conceptualise 'chengyu'. It adopts a mixed-methods approach with three sequential research components: linguistic analysis, a chengyu rating study, and an experimental acceptability task. Analysis of acceptability judgements, native speaker ratings, and questionnaire responses from over 250 participants reveals that Chinese speakers do not share a common understanding of what constitutes a chengyu. Chengyu are better characterised as a culturally salient prototype category with intuitively recognised yet variable boundaries. Semantic compositionality, structural patterns, and individual backgrounds all influence which conventionalised four-character expressions are considered chengyu, challenging both broad dictionary classifications and narrow scholarly perspectives. Notably, speakers with tertiary qualifications and strong Chinese knowledge develop an increasingly conservative concept of chengyu, favouring expressions with idiosyncratic structures, opaque meanings, and cultural content. This research advances theoretical understanding of multiword expressions (MWEs) and linguistic categorisation, and offers practical insights for psycholinguistic research and language pedagogy. Chengyu exhibit a counterintuitive pattern whereby expression characteristics that facilitate learning -- semantic compositionality and structural symmetry (Liu & Cheung, 2014) -- undermine perceived authenticity. This inverse relationship between learnability and authenticity challenges MWE frameworks developed for incidentally acquired formulaic language, demonstrating how formal acquisition and cultural salience create distinct processing dynamics. For psycholinguistic research, these results underscore the importance of validating stimuli with study participants rather than relying on authoritative classifications. For language pedagogy, the findings motivate teaching approaches that distinguish culturally valued formulaic expressions from merely comprehensible ones. By documenting the shared expectations and individual variation that allow chengyu to remain salient across generations, this research provides an empirical foundation for studying how culture-specific linguistic categories are mentally represented, culturally transmitted, and effectively taught. Liu, L., & Cheung, H. T. (2014). Acquisition of Chinese quadra-syllabic idiomatic expressions: Effects of semantic opacity and structural symmetry. First Language, 34(4), 336-353. https://doi.org/10.1177/0142723714544409 Show more