This chapter further explores the bidimensional model of salience by probing forward-looking (imposed) salience. The forward-looking dimension involves two types of speakers’ guidance of hearer’s attention: active foregrounding of information, which is the case with collaborative repetitions discussed in chapter 5, and active backgrounding of information (forward-looking non-salience), which has drawn little attention in previous literature. It is important to distinguish imposed non-salience from imposed salience in the forward-looking dimension because, just as the speaker can intend certain entities to be more important than others, the speaker can also intend certain entities to be less important toward the development of the subsequent discourse. This active backgrounding must also be kept separate from natural decay in activation, which occurs when there is no continued reference to the information in the subsequent discourse. The discussion is based on the analysis of zero (post-nominal) marking of arguments and post-predicate placement of arguments in L1 Japanese conversation data.