In this paper, I argue that the phenomena of multiple scrambling in Japanese are best captured by assuming different derivations, depending on whether they involve long‐distance or clause‐internal scrambling. I argue that long‐distance multiple scrambling involves remnant VP scrambling, on the assumption that long‐distance scrambling necessarily produces a focus chain, which thus prohibits separate applications of scrambling in multiple scrambling configurations. The evidence comes from (i) long‐distance multiple scrambling of wh‐phrases that behaves like wh‐movement, and (ii) long‐distance multiple scrambling of quantificational phrases (QPs) that behaves as if these QPs constitute a single QP. As for clause‐internal multiple scrambling, I argue that it may involve separate applications of scrambling as well as remnant VP scrambling. The evidence comes from scope facts that involve two scrambled QPs: when the two QPs undergo remnant VP scrambling, they both take scope in their original positions; when they are separately scrambled, they show a different pattern of reconstruction with respect to scope.

中文翻译：

---

如何在日语中应用多重扰码

在本文中，我认为日语中的多重扰乱现象最好通过假设不同的推导来捕捉，具体取决于它们是否涉及长途或子句内部扰乱。我认为长距离多重加扰涉及剩余VP加扰，假设长距离加扰必然产生焦点链，从而禁止在多个加扰配置中单独应用加扰。证据来自（i）长距离多重扰乱何‐表现类似的短语何‐运动，以及 (ii) 量化短语 (QP) 的长距离多重扰乱，其行为就好像这些 QP 构成单个 QP。至于子句内部多重加扰，我认为它可能涉及加扰和剩余VP加扰的单独应用。证据来自涉及两个打乱QP的范围事实：当两个QP进行剩余VP打乱时，它们都在原来的位置上占据范围；当它们被单独打乱时，它们在范围方面表现出不同的重建模式。