Chomsky`s (1986) theory of barriers is problematic since it attempts to explain both the CED and wh-island effects in an identical way. Ochi (1998, 1999) and Agbayani (1998, 1999) argue convincingly and independently, based on Chomsky`s (1995, fall class lectures) two-step movement hypothesis, that the CED effects should be accounted for by Move, whereas the wh-island effect should be handled by Attract. In this vein, this paper argues that the CED effects follow from Move, whereas the wh-island effect follows from Agree. In addition, it is argued that there is an Agree-Move parameter by which some languages such as Korean and Japanese tolerate Movement without Agree, against the claims of Chomsky (2000, 2001a, 2001b). As a consequence, it is shown why scrambling, unlike overt wh-movement, does not show relativized minimality effects.

• 김정석 | Kim Jeong Seog