Reverse Agree Is Not The Only Way To Agree: Evidence From Korean Negative Concord

The purpose of this study is to examine whether Zeijlstra's (2004, 2008, 2012) reverse or upward Agree for Negative Concord (NC) can be still a viable option for Korean NC. Reverse/Upward Agree allows a Goal with an interpretable feature to check its c-commanding Probe with an uninterpretable feature (cf. Wurmbrand, 2011). Reverse/Upward Agree seems to obtain an advantage over standard downward Agree (Chomsky, 2000, 2001) in that it explains the negativity of a fragment negative concord item (NCI hereafter) under the Identity Condition on ellipsis, where a covert operator with an interpretable feature [iNeg] checks off NCIs with an uninterpretable feature [uNeg]. I, however, argue that the standard Chomskyan downward Agree (Chomsky, 2000, 2001), where multiple NCIs with the [uNeg] look down their matching Goal with the [iNeg], is still able to capture NC in Korean, and that downward Agree is better to explain distinctiveness of Korean NCIs that outscope negation and asymmetrically c-command it. This paper also hints that the interpretation of NCIs in a fragmentary environment is heavily dependent on pragmatics, rather than syntax, opting for Chung (2012).