Subparts of contrastive topics and their relevance for the syntax–information structure interface

This paper presents two experimentally confirmed observations about contrastive topics (CTs) in German: subparts of CTs can appear in the left periphery, and if a wide contrastive topic contains both a direct object and a directional PP, then only fronting of the DP is compatible with a wide VP contrast interpretation. These observations are reminiscent of the data presented by Fanselow & Lenertov´a (2011) about subparts of focus, who use them to argue against cartographic approaches that predict a one to one relation between fronted phrases and IS categories. My claim is that the CT data basically support this argument against cartographic approaches, but that it can not be subsumed under Fanselow & Lenertov´a’s analysis for sub-part of focus fronting, because their model predicts that movement of contrastively accented elements should be less restricted than it is in the data. A potential independent reason for this restriction is explored, suggesting that the unrestricted movement of subparts of CTs predicted by Fanselow & Lenertov´a can after all be observed under the right pragmatic circumstances.