Small projections such as small n (cf. Carstens (2000, 2001), Radford (2000), and Ticio (2005)) and small v (cf. Larson (1988), Hale and Keyser (1993), and Kratzer (1996)) have been argued to dominate lexical projections such as N and V. For Korean, however, these projections have not been extensively studied. This dissertation examines the properties of small projections in Korean. I argue that small projections indeed exist in Korean; the evidence is from Bound Nouns (BNs) and Dependent Verbs (DVs).

Specifically, the BN swu is merged in the lexical N position and there is small n on top of it which assigns theta-roles as in (1)a. The more grammaticalized BN kes, on the other hand, further involves movement from the lexical N to the functional small n as in (1)b. Empirical evidence is drawn from historical facts about when the BN kes and swu acquired their functional structures: their frequency skyrockets sometime around 19c in the Korean National Corpus (KNC).*

I also argue that there are two types of DVs. Some of them (e.g., -e nay, -e noh, -e twu, -e ga, -e tay, and -e mek) merge in a functional position within the verbal domain (e.g., small v). However, in this case, morphemes such as -e/a and -ko merge in InnerAsp between small v and V as in (2). Also, based on MacDonald (2008), Watanabe (2006), and van Koppen (2005), the presence of InnerAsp in Korean is scrutinized. Other DVs (-e o , -ko iss, -e iss, -e po , -e cwu, -ko siph, -ko mal, and -e peli) are more grammaticalized and merge in the functional domain outside of the verbal domain (e.g., OuterAsp/Mod) as in (3). This claim is empirically supported by an experiment showing that we find two clusters among DVs (based on a cluster analysis from the KNC).*

*Please refer to dissertation for diagrams.