Gisbert Fanselow (University of Potsdam)

Gradient and non-gradient aspects of the superiority effect


This talk will report the results of a parallel acceptability rating experiment concerning superiority violations ("he wonders what who has proposed") carried out in seven languages (Czech, Dutch, English, German, Icelandic, Spanish, Swedish), and two further smaller experiments. The results show that

a. there is no or only a very small effect of placing a wh-object in front of a wh-subject in languages in which there is overt case marking (CZ,D,E,IS) irrespective of whether that property implies general free constituent order or not (as in Icelandic). In other words, a superiority effect shows up in those languages only in which grammatical functions are exclusively identified by position, which is in line with a large processing contribution to the superiority effect (in a crosslinguistic perspective). 

b. In Dutch and Swedish, the superiority effect is modulated by the animacy of the object, as one would expect if a major processing component is involved in the phenomenon. In Dutch and Swedish, violating examples can be found in corpora. In English, there is no effect of animacy on the strength of the superiority effect, and corresponding examples do not show up in the WWW. Therefore, the superiority effect seems to be grammatically triggered in English only. The most likely interpretation is a strong ban against in situ wh-subjects. 

c. We can show that there is a residual superiority effect caused by the crossing nature of a movement step in English but not in German. In German, there is no penalty for crossing movement even in the case of extraction from a finite complement clause crossing a matrix subject.