Abstract Dependency distance (DD) and hierarchical distance (HD) measure syntactic complexity linearly and hierarchically, reflecting comprehension and production preferences, respectively. Using syntactic dependency treebanks of English e-commerce live-streaming by American (CAL) and Chinese (CCL) hosts, this study compares the two hosts’ syntactic complexity via mean dependency distance (MDD) and mean hierarchical distance (MHD). The results show significant MHD differences between CAL and CCL but no MDD differences. Both treebanks follow Zipf’s law in DD/HD distributions, with short sentences dominating. While nine dependency types overlap in the top 10 frequent relations across three sentence lengths, some dependency types, such as discourse and aux , are more frequent in CAL. Notably, CCL exhibits significantly higher MHDs, a finding that is mainly attributed to structures like complex sentences, coordinate clauses, that -clauses, and quantifier phrases. The findings further suggest a universal tendency to minimize syntactic complexity in English live-streaming, with native language transfer primarily influencing hierarchical complexity (MHD).
Lin Wang (Tue,) studied this question.