The aim of this section is to test the validity and reliability of Halliday and Hasan’s (1976) cohesion analysis method in a pilot study, which is Stage 1 in the process of developing an analytical tool for Referential Cohesion Analysis (see Chapter 4, Table 10). With this aim in mind, a brief introduction of Halliday and Hasan’s (1976) taxonomy is necessary, as so far it has been most used to analyze cohesion in written discourse analysis research. This taxonomy was ground-breaking work at the time. Besides providing a comprehensive typology of markers of cohesion in texts, it suggested a “method for the analysis of cohesion in a text” (Halliday & Hasan, 1976, 329); however, there is no data about the reliability of its application to texts of the length of a research article. The description of this method of analysis is thus followed by a study of 10 RA abstracts to test the validity and reliability of the part of this method of cohesion analysis that is relevant for our present purposes, the analysis of referential cohesion. Similar descriptions of reference (e.g., Biber et al., 1991) will help restructure the taxonomy and clarify dubious cases. Based on the results of this study, the last Section (5.7) of this Chapter discusses questions that emerged during the validation study and endeavors to find ways to overcome theoretical and methodological problems to devise a framework specifically for Referential Cohesion Analysis.