In a number of Bantu languages, object marking is correlated with a definite or specific interpretation of the agreeing object DP, and similar claims about the semantic effects of object marking have also been made for Zulu (Nguni; S42). This paper examines these claims by applying a range of diagnostic tests for (in)definiteness and (non-)specificity to sentences with object-marked objects in Zulu. The paper’s first finding is that agreeing objects in Zulu can violate the uniqueness requirement that holds for definite expressions, and can therefore appear in contexts in which definite DPs are not tolerated. The second finding is that object-marked objects in Zulu can take narrow scope in relation to intensional verbs and negation, in which case they are interpreted as (scopally) non-specific. Object marking in Zulu therefore cannot be regarded as a morphosyntactic device to mark definiteness or specificity. Rather, it is suggested that the interpretative effects of object marking follow from information structure: agreeing DPs in Zulu are obligatorily dislocated and hence appear outside the focus domain (the vP). Consequently, an agreeing object in Zulu is incompatible with semantic focus, which implies that it can (but crucially, does not have to) be interpreted as denoting a discourse-familiar referent.

中文翻译：

---

祖鲁语中约定的对象可以是不确定的和非具体的

在许多班图语中，对象标记与一致对象 DP 的明确或具体解释相关，并且祖鲁语也提出了关于对象标记语义效果的类似声明（Nguni；S42）。本文通过对祖鲁语中带有宾语标记对象的句子应用一系列（不确定性）和（非）特异性的诊断测试来检验这些说法。该论文的第一个发现是，祖鲁语中的一致对象可能违反明确表达式的唯一性要求，因此可能出现在不容忍明确 DP 的上下文中。第二个发现是，祖鲁语中的宾语标记宾语与内涵动词和否定相关的范围很窄，在这种情况下，它们被解释为（范围上）非特定的。因此，祖鲁语中的物体标记不能被视为标记确定性或特异性的形态句法手段。相反，有人认为对象标记的解释效果遵循信息结构：祖鲁语中一致的 DP 必然错位，因此出现在焦点域 (vP) 之外。因此，祖鲁语中的一致对象与语义焦点不相容，这意味着它可以（但重要的是，不一定）被解释为表示话语熟悉的所指对象。