Comparing two experimental designs for the study of subject islands in Spanish

This paper investigates the phenomenon of subject islands in Spanish through the lens of two experimental designs, which we refer to as the subject/object design and the simple/complex subject design. Our study is the first one to directly compare these proposed designs in a single, controlled experiment. By comparing these methodologies, we aim to evaluate their effectiveness in capturing the nuances of subject island effects to establish the basis for future investigations in this area. Our findings reveal that both designs yield a significant degradation in acceptability for extractions from subjects, which is consistent with theoretical proposals that aimed to account for subject islands. We question, however, the assumption that a significant interaction, and by extension super-additivity, is the sole criterion for identifying island structures, given that the different factors tested (i.e., the position of the gap, the object/subject distinction, and the extraction/sub-extraction distinction) did not contribute to the degradation observed in subject islands.