This paper is concerned with how the so-called double NP constructions in modern Chinese can be categorized. The double NP constructions are generally taken to be double-object sentences. However, their syntactic and semantic behaviors suggest that these constructions include three distinct structures as follows. (1) Doubl-oject sentence(SVO1O2) of ``NP1+V+NP2+NP3``←[NP1+gei(給)+NP2+V+NP3] / [NP1+V+NP3+gei(給)+NP2] / [NP1+V+gei(給)+NP2+NP3](2) Jianyu(兼語) structure sentence(SV“O/S”VO) of ``NP1+V+NP2+NP3``←[NP1+V1+“NP2(jianyu)”+V2+NP3](3) Appellative predicate sentence(SVO) of ``NP1+V+NP2+NP3``←[NP1+V+“NP2+NP3”] Among others, I focus on the nature of appellative predicate sentences (SOV) and propose that the object of these sentences, i.e. “NP2+NP3”, is a small clause formed by combining two NPs. I also show that NP3 of these sentences is [+vocative] in nature, which accordingly imparts distinctive characteristics on the appellative predicate sentences and makes them have relatively restrictive syntactic and semantic inner relations.