Gapping without gaps

This squib supports Kyle Johnson’s contention that ellipsis is not involved in gapping. However, I note several problems with his specific proposal, suggesting across-the-board movement is not the way to go either. These problems have to do with the symmetric nature of coordination in gapping constructions, and with the nature of gapping in Dutch, where the proposed analysis can be shown not to work. I end by proposing an even more radical ellipsis-free analysis of gapping.