Developing an annotation protocol for evaluative stance and metaphor in discourse: theoretical and methodological considerations

Abstract The process of identification and annotation of evaluation has received a lot of attention in recent years. However, given the complexity of the topic, the discussion of some of the central issues is still ongoing. The present article contributes to this debate by presenting an annotation scheme that is designed for the identification and annotation of evaluative stance in a corpus of four English genres, namely, newspaper discourse, political discourse, newspaper scientific popularization and fora. A 4,862-word corpus was sampled from a larger 400,000-word corpus compiled within the research project STANCEDISC on the study of stance in discourse varieties. The scheme posits a series of ad hoc categories designed to optimise the transparency, reliability and replicability of the identification, annotation and analysis of evaluative stance. The categories are as follows: parts of speech (Noun Phrase, NP; Adjectival Phrase, AP; Adverbial Phrase, ADVP; Verbal Phrase, VP), function (classifying, predicational and attitude), metaphoricity (metaphoric and non-metaphoric), and value (positive and negative). The aim of this paper is to explain the scheme together with the theoretical justification of the categories and the methodological procedure adopted, and to illustrate the implementation of the scheme by discussing examples taken from different genres.