On the Syntax-Discourse Interface in Hungarian

In this paper, we investigate the syntax-discourse interface in Hungarian and propose the first steps of its formalization in LFG. Concerning the syntax, we propose that the c-structure is flat in Hungarian, and the preverbal domain in it is governed by the information structure. After examining the distribution of elements in the topic field and the Prominent Preverbal Position (PPP), we point out that they can be filled differently in neutral and non-neutral sentences, i.e. depending on the discourse. The PPP in non-neutral sentences can be occupied by an element/constituent referred to as the hocus, which cannot easily be accommodated in the i-structure assumed in the mainstream LFG literature. This is why a new architecture is put forth, building on the common features and not the discourse functions themselves, that can account for the Hungarian data more adequately. The syntax-discourse interface is then exposed via the mappings between the c- and i-structures.