Abstract
This thesis brings German data to bear on the question of what types of parallelism in linear order and syntactic structure are needed between the antecedent and target conjuncts of Gapping sentences, looking at a range of possible word order variations in each of the conjuncts. As a language with pervasive use of Scrambling, Topicalization, and verb-second ordering – which, in turn, is a significant factor responsible for its relatively free word order – German allows the investigation of a range of alternative word orders and syntactic positions in the antecedent and remaindered material of Gapping, and so is an ideal language in which to investigate these issues. \r I will present a set of empirical syntactic generalizations at work in the data, including most notably a requirement that an argument that is scrambled in the antecedent conjunct must be the counterpart of an overt remnant in the target conjunct, as well as a number of limitations for the possible positions of a remnant whose antecedent counterpart is topicalized. However, it appears to be the case that such syntactic restrictions may be overridden by linear parallelism traits such as matching linear order between the arguments of the antecedent and target conjuncts or a canonical ordering of target conjunct remnants. The mechanism by which linear order seems to improve on otherwise ungrammatical utterances is most likely due to separate processing issues. Independently of the syntactic and linear order generalizations, this thesis presents a (to my knowledge) new and mostly unstudied, body of data, which can be used in future research regarding issues of prosody and information structure, processing, and naturally, continued work on Gapping and word order variation, particularly Scrambling.