Lexico-syntactic complexity in machine interpreting : a corpus-based comparison with human interpreting and translation

Abstract

The rapid development of artificial intelligence (AI) has significantly transformed cross-linguistic communication practices, with machine interpreting (MI) emerging both as a standalone solution and as an assistive tool for human interpreters. Despite technological advancements, few studies have systematically examined the linguistic characteristics of MI output compared to traditional human-mediated language production. This study investigates the lexico-syntactic complexity of MI in relation to human interpreting (HI) and human translation (HT) based on a comparable and intermodal corpus of Chinese-English language mediation. Using validated computational tools measuring 25 lexical and 14 syntactic indices, the findings reveal a general complexity hierarchy: MI exhibits significantly greater lexical and syntactic complexity than HI but remains less complex than HT. These systematic differences reflect both the cognitive constraints in HI and the architectural advantages and limitations of MI systems. By positioning MI between human interpreting and translation modalities, this study enhances our understanding of AI-mediated language production and its implications for interpreting pedagogy, professional practice, and technological development in applied linguistics.