Following the spirit, but not the same mechanism, of Kayne's (1998, 2000) overt syntactic derivational analysis of focus-related elements such as only, also/too, and even in English, this paper explores the nature of association with focus in Japanese with special reference to the additive focus particle mo 'also' and the exclusive focus particle dake 'only.' In this paper, I will argue that, in spite of their attractiveness, at least the versions of LF focus particle movement analysis of association with focus entertained and discussed in Aoyagi (1998, 1999, 2006) are faced with theoretical/empirical problems, exploring another way of reframing Kuroda's (1965) attachment transformation analysis and spelling out an alternative analysis of association with focus in Japanese, which does not depend upon any LF-movement. More specifically, I will make the following points: (i) the additive focus particle mo 'also' can trigger both association with narrow focus and association with wide focus, and its association with focus should be best analyzed as involving overt XP movement to [Spec,FocP] via Agree between the Foc head mo and its focus associate; (ii) the exclusive focus particle dake 'only' can trigger only association with narrow focus, and its association with focus is to be analyzed as involving just Merge without any Agree operation. To the extent that this approach is on the right track, it shows that the empirical domain of association with focus in Japanese is not compatible with the GB/earlier Minimalist Program double-cycle computational system, but it can be compatible with the recent view of the single-cycle computational system in the Minimalist Program (Chomsky 2001b, 2004).