This study explores the ways in which academic writers employ expressions of attitude in the construction of evaluative stance in the introductory sections of research papers. The study draws on the theoretical base of Systemic Functional Linguistics (SFL) (Halliday, 1994; Halliday & Matthiessen, 2004), and in particular on Appraisal theory as a modelling of interpersonal meaning at the level of discourse semantics (Martin, 1992, 2000; Martin & Rose, 2003; Martin & White, in press). Attitude is explored from two perspectives: how it is expressed in the discourse, and what it is employed to evaluate. In addressing the second issue, the focus is on the general field (subject matter being constructed in the text) rather than on specific entities. The study is also concerned therefore with how different fields are identified in the texts, and how they relate one to another. The research contributes some significant dimensions to the modelling of attitudinal meanings in the register. Analyses reveal that the register of academic research writing is characteristically constructive of two fields, the knowledge domain being investigated and the research activity conducted in relation to that domain; that these fields are in a relationship of projection one to the other; and that each field is evaluated in quite different ways. The findings contribute at a theoretical level to an explanation of the apparent contradiction between the dual demands of persuasion and ‘objectivity’ in the register, and at a practical level provide a new dimension to frameworks for deconstructing and negotiating evaluative stance with novice academic research writers.