Minimizing Spell-Out Material in the Phase

Based on the previous arguments for the SVO hypothesis and against the SOV hypothesis for Korean, I discover that languages attempt to reduce computational burden and thereby secure interface economy by minimizing Spell-Out material in phases throughout derivation. This emerges as the following broader generalization--minimize Spell-Out material as much as possible in each phase. I present an extreme case of this generalization obtained from Korean--by Spell-Out, don t leave anything in the phase until the last phase. It thus turns out that the generalization in Alexiadou and Anagnostopoulou (2001)--by Spell-Out, VP can contain no more than one argument with an unchecked Case feature--is one mild case on the Spell-Out scales, and that the EPP is a much milder one.