This part of the monograph discussed the interrelated nature of textual structures and discussed that linguistic aspects of text construction contribute to discourse-level features, which, by providing a thick description of language use, have implications for the understanding of genre construction. The relevance of EAP teaching, as a bridge for students to achieve membership in the discourse community, was also discussed. Here, I pointed to the special difficulty NNS writers have in constructing texts for publication in academic journals. Then the progress of written discourse analysis was discussed starting from the 1970s, when Halliday and Hasan’s (1976) still most widely used cohesion analysis method was conceived. I pointed out the need to focus on the interactive role of the reader and writer in the composition process, emphasizing the complexity of the task of composing linear texts on the basis of non-linear thought processes. A model that captures non-linear processes is Lautamatti’s (1987) TSA model; therefore, the basic idea of text representation used in the analytical approach will be applied in the development of the analytical tool for reference analysis to reveal new features of reference so far undiscovered.