Subpart of Focus Fronting Is Discourse Subordination

In this paper, we make three specific claims about German information structure and the interface between information structure and syntax. (1) First, the phenomenon known as ‘Subpart of Focus Fronting’ in German is in fact constrained by discourse context, contra the current literature (Fanselow & Lenertová 2011 on German and Czech). (2) Second, information structure should be present in the syntax, minimally via a discourse subordination feature on C (cf. Asher & Lascarides 2003). (3) Third, methodologically, the Question-Answer paradigm alone is not sufficient to diagnose the complete information structure of the answer; topicalization may occur out of a putative focussed phrase. These claims speak to the ongoing debate about the place of information structure in the syntactic derivation. A growing body of literature asserts that no discourse contextual features at all should be present in the derivation. This literature has roots in Chomsky (1995)’s Inclusiveness Condition, which argues against sentence or discourse-level features being present in the numeration. Since then, the strongest formulation is Horváth (2010)’s ‘Strong Modularity Hypothesis’, which hypothesizes a complete disconnect between information structure and syntax. For this hypothesis, all apparent syntactic effects of information structure (such as topicalization, focus fronting, etc.) are in fact due to prosodic constraints, optional prosodic movement (cf. Zubizarreta 1998’s p-movement), or null information structural operators. A crucial datum for the Stong Modularity Hypothesis is ‘Subpart of Focus Fronting’ (henceforth, ‘SFF’), a phenomenon reported in German and Czech. Fanselow & Lenertová (2011) present data such as (1a–b). They assert that the two utterances (b) and (b') are completely optional in context, because both are felicitous responses to the question What happened? This assumes that a question completely controls for the information structure of the answer. The movement in (1b') is both syntactic Āmovement but also constrained by prosody: only the linearly leftmost stressed phrase may undergo movement. Taking all this into consideration, Fanselow & Lenertová argue that prosodic operations may apply in the syntactic derivation. Furthermore, they suggest that all apparent information structural effects in the syntax are in fact due to this optional prosody-driven movement.