In the past two decades or so, a number of researchers from various fields within linguistics have turned their attention to interpersonal phenomena, such as the linguistic expression of speaker opinion or evaluation (also called stance or appraisal), or the encoding of subjectivity in language and its diachronic development (subjectification/subjectivization). Many linguists have offered categorizations of evaluative meaning, based on authentic discourse data, but no connection has been made with cognitive approaches to appraisal processes. This paper offers a first meta-theoretical exploration of such issues. It compares dimensions of evaluation that have been identified in linguistic and cognitive studies, and also examines how psychological research into basic emotions can be related to linguistic research on affect. On the basis of these comparisons a proposal for a new classification of evaluative meanings is made. The focus of this paper is on only one aspect of the highly complex phenomenon of evaluation, namely potential evaluative dimensions, although other relevant issues will also be touched upon, including the pragmatics of evaluation (evaluation and context).