This paper aims to investigate the correlation between “sentence topic (ST)” and “discourse topic (DT)”, both based on an “aboutness” relation. It also compares the degrees of correlation between French and Chinese, considered respectively as a subject-prominent and a topic-prominent language. Kendall's correlation tests, Poisson models and z-tests were conducted using manually annotated textual data. The results confirm a positive correlation between STs and DTs in both languages. Although the correlation coefficients are higher in Chinese, z-tests' results do not show a significant difference between the two languages. We explain that the two languages differ not only in the more or less frequent topic-comment constructions, but also in the choice of more or less accessible referential markers which represent the STs, or in the more or less tacit way that STs of different sentences are related to each other.

中文翻译：

---

法语和汉语句子主题与话语主题的相关性：注释问题与对比分析

本文旨在研究“句子主题（ST）”和“话语主题（DT）”之间的相关性，两者都基于“关于”关系。它还比较了法语和汉语之间的相关程度，分别被视为主语突出语言和主题突出语言。Kendall 的相关性检验、泊松模型和z检验是使用手动注释的文本数据进行的。结果证实两种语言的 ST 和 DT 之间存在正相关关系。尽管中文的相关系数较高，但z检验的结果并未显示两种语言之间存在显着差异。我们解释说，这两种语言不仅在或多或少频繁的主题评论结构上有所不同，而且在代表 ST 的或多或少可访问的指称标记的选择上有所不同，或者不同句子的 ST 或多或少默认的方式有所不同。彼此相关。