Why a "word order difference" is not always a "word order" difference: a reply to Weyerts, Penke, Münte, Heinze, and Clahsen.

Abstract

We present evidence that the supposed processing advantage for an SVfinO word order over an SOVfin word order in German argued for by Weyerts, Penke, Münte, Heinze, and Clahsen (2002) is supported by neither experimental nor theoretical evidence. Specifically, we show (a) that the frontocentral negativity for an SOVfin in comparison to an SVfinO word order in Weyerts et al.'s Experiments 2 and 3 is reducible to more general differences in the electrophysiological responses elicited by nouns versus verbs in a sentence context, and (b) that the P600 difference between the two word orders in Experiment 2, as well as the reading time differences in Experiment 1, result from the fact that the two supposedly ungrammatical conditions actually differ in their degree of ill-formedness. We conclude that there is no evidence for a processing disadvantage for SOVfin, thus reconciling Weyerts et al.'s results on German sentence processing with the grammatical regularities of German.