Clefts and preposition omission under sluicing

This work argues against analyzing the possibility of preposition omission under sluicing as categorically dependent on the availability of preposition stranding under wh-movement. In particular, I demonstrate the descriptive inadequacy of the cleft analysis proposed for Polish sluicing in order to account for the environments tolerating preposition omission. Data from four acceptability judgment studies are presented in support of this argument. These data indicate that the pattern of preposition omission in Polish is graded, and reflects the linguistic complexity of correlates and sluicing remnants. One way of capturing these findings is, I propose, by connecting the availability of preposition omission in sluicing to general cognitive mechanisms involved in anaphoric processing.

► Clefts are unacceptable sources for Polish sluices without prepositions.
► Preposition omission is affected by linguistic complexity of correlates and remnants.
► Cognitive mechanisms involved in memory retrieval predict these data.
► The Direct Interpretation approach to sluicing can capture the data.