Following Chomsky and Halle's (1968) work on the Germanic-Latinate contrast in the morphophonemics of English, Emonds (1986, 2000, 2005) argues that the English vocabulary is divided into two, i.e., (i) the primary vocabulary that consists mostly of native morphemes and (ii) the secondary vocabulary that consists mostly of foreign ones. Based on Emonds' notion of this division in English vocabulary, Kubo (1992) argues that the adjective class in Japanese is similarly divided into two subclasses, namely, A(djective)s and A(djectival) N(oun)s, and that As belong to the primary vocabulary, and ANs to the secondary vocabulary. In this talk, I will show that the synchronic distinction between As and ANs is rooted in the absence versus the presence of a null N at PF, respectively. More specifically, the category-ambiguous behavior of ANs, i.e., the NP-like external syntax and the AP-like internal syntax, can be accounted for by postulating nominalization at PF. Furthermore, I will show that this proposed analysis can be extended to the verb classes in Japanese.