Propositions, projections and things

Introduction My interest in this topic arose from work that I was doing on attempting a description of the language used to report language (Thompson, 1994). It rapidly became clear that the commonly used categories were far from adequate, and I therefore turned to Halliday’s (1985) concept of projection as providing a much wider and more usable perspective than traditional ideas of ’reported speech’. However, I found that the data on which my description was based was throwing up phenomena that were not exactly covered by projection; and the close examination of projection involved in trying to apply it to a range of data led me to feel that certain aspects of the concept as presented by Halliday were to some extent unclear. This led to a ’tweaking’ of projection, a realignment of mainly already existing ideas, in order to make it work more effectively in the context of what I was trying to do. What I would like to do here is to briefly outline projection as seen by Halliday, indicate the reasons why I think it needs tweaking and how I propose tweaking it; and then look in a tentative, preliminary way at the implications for text analysis not only in terms of reporting but more broadly in terms of what I will be calling the propositional cline.