Language complexity reveals the ability to use a wide and varied range of sophisticated structures and vocabulary. Although different languages compose complexity differently, complexity measures such as the T-unit have typically been based on clause subordination, which may underrepresent complexity and threaten the validity of studies. This study argues that an organic complexity measure should avoid the assumption of clause subordination and instead consider the typological features of the target language. Therefore, this study proposes the TC-unit in recognition of the topic chain as the underlying unit of Chinese complexity. It further validates TC-unit-based measures by investigating how accurately they predict proficiency group membership. Discriminant analyses of L1 and L2 Chinese speakers’ spoken (N = 115) and written (N = 116) output elicited from a designed timed online test, revealed that TC-unit-based measures classified proficiency group membership with high efficiency (61.2–75.7 per cent). Mean length of terminable TC-unit proved the most effective indicator of spoken Chinese syntactic complexity, while mean length of terminable TC-unit and single TC-units per terminable TC-unit in combination proved the most effective for written Chinese syntactic complexity.

中文翻译：

---

中文的有机句法复杂性量度：TC单元

语言的复杂性揭示了使用各种复杂结构和词汇的能力。尽管不同的语言构成复杂性的方式有所不同，但是诸如T单元之类的复杂性度量通常基于子句从属，这可能不足以表示复杂性并威胁研究的有效性。这项研究认为，有机复杂性度量应避免假设从句，而应考虑目标语言的类型学特征。因此，本研究提出了将主题链识别为中文复杂性的基础单元的TC单元。通过调查预测熟练组成员的准确性，它进一步验证了基于TC单位的措施。母语为L1和L2的语音的判别分析（N  =  115）和书面（N  =  116）输出是从设计的定时在线测试中得出的，结果表明，基于TC单元的措施对熟练组成员的分类效率很高（61.2-75.7％）。可终止TC单元的平均长度被证明是汉语口语句法复杂性的最有效指标，而可终止TC单元的平均长度和每个可终止TC单元组合的单个TC单元的平均长度被证明对书面汉语句法复杂性最有效。