The Importance of Being Not-Obviate

In this paper, I discuss the notion of obviation in the language isolate, Ktunaxa,2 and in the Central Algonquian language, Meskwaki. While there are important differences between these two languages with respect to obviation, my concern in this paper is with a feature of obviation found in both languages—namely, that proximate shifts are often not accompanied by proximate switches. That is, while a proximate shift introduces a new proximate, the previous proximate need not, and frequently, does not, become obviative when a new proximate is introduced. Although this fact is well known, I believe it has a consequence that has been under-appreciated. I will argue that an implication of proximate shifts not being proximate switches, is that proximate shifts fail to track discourse prominence in a narrative. By 'track discourse prominence', here, I mean, systematically identify the most discourse-prominent referent in the narrative. It is widely assumed that when a proximate shift occurs, the new proximate becomes most discourse prominent at that point in the narrative, and remains so until the next proximate shift. Nevertheless, the central aim of this paper is to argue that this assumption should be rejected. Specifically, I will argue that when a proximate shift does not include the previous proximate becoming obviative, the previous proximate retains its proximate status and that, therefore, the status of the new proximate is not more discourse prominent than the preceding proximate; instead their relationship is one of discourse parity. Because such proximate shifts without switches are frequent, it will often be the case that a proximate shift does not produce a new most-discourse-prominent referent. Hence, simply tracking the new proximate in successive proximate shifts will not systematically identify the most discourse-prominent referent in a narrative