This paper aims to investigate the genuine features of so-called Case particles in Korean and to solve the locality violation of subcategorization in Particle stacking. Ga and lul are believed to be Structural Case markers since generative grammar was introduced in Korean linguistics. However, it is not difficult to find their non-Case use. Even if I adopt the assumption that they can be categorized as D like delimiter nun, I think their genuine features should be defined not in terms of Case preference but in terms of Case rejection. On the other hand, the locality violation of selection and subcategorization in Particle stacking is inevitable in the traditional syntactic structures built by the asymmetric projection. I solve this problem by replacing asymmetric projection with symmetric projection proposed in my previous study to Particle stacking structures.