According to Biber et al. (1991, 234) “chains of reference” are “sequences of noun phrases all referring to the same thing”. This same entity referred to might be altered, its reference may be narrowed or broadened, and its characteristics may change in the course of the discussion in the analyzed text. Therefore, to cover all these phenomena, in this book, cohesive chains of reference are defined as second-mention referring items all referring to the same entity represented by a presupposed element in the text, without which the other items would not be fully interpretable. A cohesive chain is made up of a sequence of cohesive ties. These chains may be short (involving not more than two ties) or long, consisting of more than two ties. When these chains refer to the same entity in the textual context, they are linear; however, linearity is broken if only a part or one specific aspect of the entity is referred to. In representing this phenomenon, the analysis may either identify it as a new chain, or may divide the chains to represent all the constituents of the presupposed element (e.g., a group of students  the girls // the boys; where the girls constitute only part of the original group). Or on the contrary, many chains of reference may be referred to by one umbrella term, which will unite all of them. These deviations from the linear can be represented by the referential cohesion analysis tool introduced in this book. Cohesive chains may be of any length, from two ties to any, up to the total number of sentences in a given text.