This paper provides new data to tease apart existing analyses of the wide scope behavior of nominative objects in the Japanese potential construction. An object can get nominative or accusative Case in the Japanese potential construction (Kuno 1973). Significantly, only nominative objects can take scope over the potential suffix (Koizumi 1994, Nomura 2005, Sano 1983, Tada 1992, a.o.). There are at least three possible analyses to capture the wide scope behavior of nominative objects. First, nominative objects move to the Spec of TP, in which case nominative objects c-command the potential suffix (Case-movement analysis: Koizumi 1994, Nomura 2005, a.o. cf. Tada 1992). Second, nominative objects are base-generated in a position above the potential suffix (base-generation approach: Saito and Hoshi 1998, Takano 2003, a.o.). Third, the focus particle in nominative objects undergoes covert A’-movement (covert A’ movement approach: Bobaljik and Wurmbrand 2007, Takahashi 2010. cf. Sano 1985). There is one set of data concerning adjuncts, which favors the latter two approaches (Bobaljik and Wurmbrand 2007, Saito and Hoshi 1998, Takahashi 2010). We provide a new set of data to distinguish the two approaches. In particular, we show that the wide scope behavior of nominative objects is subject to LF intervention effects, which have been observed for Wh-constructions (Hoji 1985). As LF intervention effects are often analyzed in terms of covert movement of Wh-phrases (Beck 1997, Hoji 1985, Tanaka 1997), we conclude that movement is implicated in the wide scope behavior of nominative objects. As the base-generation approach involves no movement, we are lead to choose the covert A’-movement hypothesis over the base-generation approach.