A multifactorial aspectual analysis of variations in the completivemarkers <i>wán</i>  (‘finished’) and <i>hǎo</i>  (‘well’) in Mandarin

Abstract This study presents a multifactorial aspectual analysis of the Mandarin completive markers wán and hǎo . Using multiple correspondence analysis, conditional inference trees, and conditional random forests, we examine the distribution, prototypical features, and conditional importance of the constructions V wán and V hǎo across thirteen semantic and syntactic variables. The results indicate that eight variables, namely creativity, situation type, telicity, agentivity, concreteness, boundedness, animacy, and definiteness, play the most significant roles in distinguishing between these two constructions. In contrast, five variables, including semantic orientation, dynamicity, the aspectual marker le , plurality, and durativity, demonstrate comparatively weaker influence. We argue that specialization and generalization are the cognitive mechanisms underlying variation between the two constructions. The study proposes the Completive Aspect Hypothesis, refining Bybee’s classification of the completive aspect, and offers empirical support for Chinese completive-aspect typology and second language pedagogy.