From the introduction: "This paper concerns Yagua, a lowland language of Peru. It also concerns the choices speakers make as to how they will refer to or "code" participants in discourse. The body of this paper will be organized into two broad parts corresponding to these two most general concerns. In the first part, consisting of Sect. 2, I will describe the major formal devices used to code participants in Yagua. In the second part, Sect. 3 and 4, I will look at the use of those devices in a body of folkloric narrative texts. I will take as a starting point for the textual study the framework and methodology developed by Givón (1983a, b, c, d) for measuring "referential distance", i.e., the distance in number of clauses between one mention of a participant and its previous mention in the text. The general framework is outlined in some depth and several substantive modifications to the methodology are proposed. Deviations from the general predictions of Givón's framework will then be examined in detail in Sect. 4. It is found that there are two circumstances under which relatively "new" participants can be coded with attenuated coding devices (i.e., verb coding and enclitics), and seven circumstances under which relatively recently mentioned participants can be coded with full noun phrases. Each of these circumstances represents an area of further research into the functional factors underlying the choice of participant coding devices in discourse."
Payne, Thomas E.
"Referential distance and discourse structure in Yagua,"
Work Papers of the Summer Institute of Linguistics, University of North Dakota Session: Vol. 29, Article 7.
Available at: https://commons.und.edu/sil-work-papers/vol29/iss1/7