[+/-Focus]/Subject Prominence, Feature Inheritance and the EPP

Miyagawa (2005, 2006, forthcoming) suggests that differences between "agreement" languages and "agreement-less" languages stem from different parametric valuation of a super feature (Fs) on C either as Φ -feature agreement or as focus agreement. A language like Korean, however, constitutes an interesting example in that it employs not only Φ-feature agreement but also focus and topic agreement, as essential agreement features to derive the mixed properties of subject prominence, focus prominence, and topic prominence. Grounded on these Korean facts, I will argue that languages cannot be strictly divided into "agreement" languages and "agreement-less" languages, and that the apparent differences in the aspects of agreement are due to the interplay of feature inheritance and EPP parametrization.