The analysis of newspaper discourse offers valuable insights into how society represents or misrepresents certain social participants and their actions. In view of the bias claimed to exist in journalistic prose (Bednarek, 2006; White, 2006), it is not uncommon to find evidence of the mistreatment directed towards particular minorities (Baker et al., 2008; Fowler, 1991). In this paper, the ideological stance associated with a specific minority group (i.e. homosexuals) is brought to the forefront in 2008, when Ireland’s vibrant economy took a dramatic turn for the worse. Incidentally, this coincided with homosexuality taking centre stage in Ireland’s political agenda, as 2008 marked the final stage of the long drawn-out debate on the Civil Partnership Bill. This paper is designed to offer insights into how evaluative language may reflect the mentality of Irish society in relation to the LGBT community. Martin & White’s (2005) appraisal theory is highly relevant and applicable for this purpose, as it covers the idea of social esteem, social sanction, personal attitude and appreciation, which can be powerful indicators of a society’s take on current affairs. The methodology employed here is that of corpus-assisted discourse analysis (Stubbs, 1996). The dataset comprises over 200,000 words taken from three different newspapers: Two tabloids and one broadsheet. Our dataset is annotated on the basis of the categories in Martin & White’s (2005) subsystem of attitude (affect, judgement and appreciation). The application of this taxonomy uncovers a remarkably negative stance towards the Irish LGBT community in the sample analysed. This is particularly evident in the predominance of evaluative and emotive language associated with the categories of negative judgement and affect. Previous research on the same sample, looking at metaphor, transitivity and modality (e.g. Bartley & Hidalgo-Tenorio, 2015), has cast light on how homosexuals are repeatedly discriminated against and vilified in the Irish public arena. This study confirms the results so far obtained through the analysis of evaluative language.