The goal of this paper is to suggest a feature-based analysis for the licensing of null subjects found in languages. After reviewing the uniformity principle suggested in 1980s, we investigate the possibility of more insightful description on the distribution of null subjects based on Holmberg and Roberts (2013)'s research. Then we show that the typological classification of the null subject constructions among languages helps to provide a universal rule that applies to explain the full, partial and radical pro-drop phenomena among languages. Finally we suggest a feature copying or feature sharing analysis for null topic and null subject constructions in Korean.

This paper is a revision of my earlier suggestion (2006) that the form of NPs is determined after all the features are specified for the computation under the assumptions of Multiple Spheres Hypothesis. Adopting the DM-based notions of morphological shaping (Heinat 2006, Safir 2011), I develop the proposal, arguing that the choice between the ‘-self’ form and the pronominal form is determined after the computation of root NPs is accomplished by Free Merge operation in multiple Membranes (Im 2013).

The cross-linguistic divergence of copula constructions does not necessarily imply that there is no universal structure for the constructions. Since the introduction of the notion, "small clause", the degree of categorial differentiation for copulas found in many languages has no longer been a bothersome phenomenon for linguists. Some recent works on the copula structure based on Merge theory claim that one of the pair {XP, YP} must move to satisfy the canonical order of the construction in the spirit of LCA. This paper attempts to explain how we can choose the one from the pair that must escape from the first Merged structure. Under the assumptions of Multiple Membrane Hypothesis (Im 2013), and employing the cartographical order of functional categories (Rizzi 1997, Cinque 1999, Starke 2001, 2006, among others), I argue that the syntactic object left in small clause has its own intrinsic [Focus] feature that induces the VP-internal location of the object at the interfaces. The morpho-phonemic realization of √BE is followed to meet the legibility condition at the interfaces.

Pseudogapping is treated as a combination of movement and ellipsis by many researchers. After reviewing some arguments based on movement analysis, I argue for a merge-based approach with the assumptions of Multiple Sphere Hypothesis (Im 2004-2012). The remnants in VP are not moved and elided but merged in Ω-sphere evading the PF deletion rule posited with the notion of E-GIVENness (Merchant 2001).