Arguing for remnant movement

This work will provide a number of arguments showing that certain comparative generalizations in Romance can best be captured if they are seen as reflecting the displacement of Remnant constituents. The empirical domains to be studied all involve interrogative syntax, in particular constructions like S(ubjec) Cl(itic) I(nversion) and C(omplex) I(nversion) that the ‘Government and Binding’ framework typically analyzed as head to head movement, I°/V° to C°. We shall show that they need to be reanalyzed as resulting from movement of Remnant constituents to various layers of the (highly split) left periphery, just as those French interrogative constructions in which wh-words and phrases seem to be standing at the right edge of the sentence. We shall not attempt to argue here that Remnant Movement should be allowed to exist for the simple reason that there is no need to do so; if displacement – ‘internal merge’– affects constituents, displacement of constituents which sub-constituents have vacated at previous stages in the derivation cannot be avoided. Extra machinery would be needed to ban that option, which might be required if languages never made use of it; since Germanic and Romance most emphatically do, that would be an extremely ill-advised step.