Abstract In this article, I argue, contra Cervoni (1991), Olivier (2007) and others, that French does have prepositions without an overt complement that are syntactically transitive. That is, these prepositions take a syntactically projected, phonologically unrealized pronominal complement, as first hypothesized by Zribi-Hertz (1984), who coined them “orphan prepositions”. I offer new evidence in favor of this view based on a number of tests that distinguish between various types of implicit arguments, among which (in)definiteness, the (un)availability of bound variable readings, scopal properties relative to operators such as negation and intensional verbs and the (un)availability of sloppy identity readings. I also show that the set of French orphan prepositions turns out to be much smaller than that envisaged by Zribi-Hertz in that some lexical prepositions can function as orphan prepositions in some uses but not others and some lexical prepositions have an intransitive use as well. The article also includes two appendices that lay out a number of reasons why orphan prepositions should not be assumed to be prepositions that have become adverbs through a process of grammaticalization nor should they be assumed to be relational definite descriptions that pragmatically link up to an antecedent via accommodation, as suggested by Olivier (2007).

中文翻译：

---

重新审视法国孤儿介词

摘要在本文中，我与Cervoni（1991），Olivier（2007）等人相反，认为法语的介词没有明显的补语，在句法上是及物的。也就是说，这些介词采用了句法上预测的，语音上未实现的代词补语，如Zribi-Hertz（1984）首次提出的，他将它们称为“孤儿介词”。我提供了支持这种观点的新证据，它基于多种测试，这些测试区分了各种类型的隐式参数，其中（在）确定性，绑定变量读数的（不）可用性，相对于运算符（例如，否定和意图动词和identity散的身份阅读的（不）可用性。我还表明，法国孤儿介词的集合比Zribi-Hertz设想的要小得多，因为某些词汇介词在某些用途中可以用作孤儿介词，但在某些用途中不能用作孤儿介词，并且某些词汇介词也具有不及物动词的用法。这篇文章还包括两个附录，列出了为什么不能将孤儿介词假定为通过语法化过程成为副词的介词的原因，也不应认为它们是关系确定的描述，它们通过语用学联系到先行词如Olivier（2007）所述。