Abstract This article takes a usage-based perspective on the partitive genitive construction in Dutch (iets moois, ‘something beautiful’), which has previously drawn scholarly attention from a theoretical perspective, due to the challenges it presents to Dutch nominal morphosyntax. We will argue that a good understanding of the construction at issue cannot circumvent the enormous variation in the expression of the genitive marker. Within the wide variation space, regular patterns can be discerned, which we uncovered by using mixed-effects logistic regression. This approach allows us to assess the precise contribution of internal factors (e.g. length of the adjective, or the type of quantifier) and external factors (e.g. regional variety, or register), as well as their interactions. This article has three objectives then: first, it wants to contribute to the description of Dutch syntax, second it aspires to advance methodological standards in grammatical investigation, and third, it makes a theoretical plea for a usage-based perspective, with full recognition of variation.

中文翻译：

---

荷兰语中的宾语所有格的多元分析。将定量数据纳入理论讨论

摘要本文采用基于用法的观点来研究荷兰语（iets moois，“美丽的事物”）中的分形格结构，由于它对荷兰名词性形态句法提出了挑战，因此先前已在理论上引起了学术界的关注。我们将争辩说，对所讨论结构的充分理解不能绕过遗传标记表达的巨大变化。在宽泛的变化空间内，可以辨别出规则的模式，我们通过使用混合效应逻辑回归发现了这些模式。这种方法使我们能够评估内部因素（例如，形容词的长度或量词的类型）和外部因素（例如，区域变化或注册）的精确贡献以及它们之间的相互作用。那么，本文有三个目标：第一，