Parasitic Gaps and ATB in Polish.

The paper examines ATB and parasitic gap structures in Polish in order to determine whether they can be conflated into a single phenomenon. Three available approaches to these two constructions are outlined and evaluated as to their applicability to Polish data. It is argued that the approach postulating the treatment of parasitic gaps as a special case of ATB put forward by Huybregts and van Riemsdijk (1985) and Williams (1990) is problematic as it does not specify how parasitic gap constructions are assigned the coordinate status. The second approach arguing in favor of subsuming ATB gaps under parasitic gaps, advocated by Munn (1992) and Franks (1993, 1995), is more advantageous. Franks's analysis is scrutinized in detail, as it directly deals with Polish. It is argued that there exist ATB and parasitic gaps which violate both Franks's thematic prominence condition and his case identity requirement. It is suggested that mere morphological case identity is not sufficient and should be supplemented with identity in abstract Case, perhaps along the lines of Dyla (1984). Next, we examine the third approach, proposed by Postal (1993), suggesting that parasitic gaps and ATB gaps do not represent a single phenomenon and therefore should be kept separate. It is pointed out that the differences between Polish parasitic gaps and ATB gaps are not uniquely characteristic for these two types of structure and that is why they cannot serve as sufficient evidence for claiming that the two examined constructions are instances of distinct phenomena. A conclusion along these lines is reached independently by Hornstein and Nunes (2002) on the basis of English and Portuguese data. Their analysis generally turns out to be applicable also to Polish ATB and parasitic gaps, and only the sentences where case mismatch occurs or where thematic prominence is violated require a separate explanation. [ABSTRACT FROM AUTHOR]