>>   TOC


For the last thirty years, syntacticians have struggled with scrambling structures without reaching a consensus about the mechanisms involved. In this paper I will argue that this situation is due to two reasons: first, most of the discussion revolves around data that provide only an obscured view on scrambling, and secondly, many analysts have pondered the question of what kind of movement result in scrambling structures without having shown that movement is involved at all. Furthermore, it appears at times as if the concept of scrambling was interpreted rather loosely, such that any unusual word order could be subsumed under the general heading of "scrambling", which makes it even more difficult to clearly identify general principles.

I will thus focus the discussion here on the overt ordering of VP/IP-internal arguments. So far, I have no opinion about how adjuncts are to be treated formally, and I will therefore ignore them completely. Furthermore, I will try to use "purified" data wherever possible, i.e. structures without quantifiers, negation, or distinctions with respect to definiteness etc. I believe that all of them introduce independent constraints on any syntactic structure, which may be difficult to separate from "purely syntactic" mechanisms. 

In the end, I will commit to a base generation account of scrambling and combine, among others, analyses proposed by Haider (1993) and Bayer and Kornfilt (1994) to develop a multidimensional model of sentence structure, which strongly resembles the "tier" concept from autosegmental phonology. While this analysis will remain sketchy, I hope at least to raise questions that may warrant further research.

>>   TOC

Philipp Strazny 1997