This paper discusses the syntactic variation of modal predicates between structures with a nominative primary argument and those with an oblique primary argument. In the literature, this variation is related to a change from deontic to epistemic meanings, whereby epistemicity seems to be more commonly expressed by highly grammaticalized impersonal constructions. After having shown the weakness of this relationship, I suggest a new explanation for the variation of modal predicates on the basis of diverse ancient Indo‐European languages, such as Vedic, Ancient Greek and Latin, as well as of some of their modern descendants, especially Hindi, Modern Greek, and standard and colloquial Italian. I argue that modal predicates with an oblique primary argument are favoured for functions of necessity, while modal predicates with a nominative primary argument preferably express functions of possibility. This reflects the different meanings of the lexical sources of these predicates, that is, capacity or power for predicates of possibility, and lack or obligation for predicates of necessity, which also imply different degrees of agentivity and control.

中文翻译：

---

印欧语系中模态谓词的语法和语义

本文讨论了主格主论元结构和斜主论元结构之间模态谓词的句法变化。在文献中，这种变化与从道义意义到认识意义的变化有关，由此认识性似乎更常由高度语法化的非人称结构表达。在展示了这种关系的弱点之后，我建议对基于各种古印欧语言（如吠陀、古希腊和拉丁语）以及它们的一些现代后裔的情态谓词变异进行新的解释，尤其是印地语、现代希腊语以及标准和口语化的意大利语。我认为带有倾斜主参数的情态谓词更适合于必然函数，而带有主格主参数的情态谓词最好表达可能性函数。这反映了这些谓词的词源的不同含义，即可能性谓词的能力或权力，以及必然性谓词的缺乏或义务，这也意味着不同程度的施动性和控制性。