In this thesis, I propose a research programme whereby prognostic critique, originally presented as part of the Discourse-Historical Approach to Critical Discourse Studies (CDS) (Reisigl & Wodak, 2016), is further developed and integrated into CDS praxis more generally, systematically and explicitly than it has been thus far. Prognostic critique is where the socially transformative agenda of CDS finds its culmination, as it expressly orients towards “the transformation of communication” (Reisigl & Wodak 2016, p. 25) via the creation of guidelines and recommendations to challenge, dismantle and replace socially problematic (e.g. racist, sexist) discursive practices. Explicit engagement with this level of critique within CDS, however, is relatively rare. In the few cases where such critique has been attempted (e.g. Kargl et al., 1997; Godin, 2021), recommendations for discoursal transformation rest on analysts’ interpretations and largely untested intuitions about what alternative discourses look like and how they may impact social action and attitudes in more ethically productive ways. In this thesis, I advocate for the integration of experimental methods to CDS work as a way to enable researchers to empirically examine the emancipatory potential (or lack thereof) of texts, so that any interpretations and prognostic recommendations may be informed by more systematic probing of the effects of alternative discourses on different audiences' attitudes and/or behaviours. As proof of concept of how this may be done, I report on the results of a small-scale prognostic intervention aimed at tackling Islamophobic prejudice and discursive practices in a UK context. Given the attested influence of news media discourse on social attitudes, this intervention explicitly engaged with the question of what alternative/counter-hegemonic representations of Muslim communities in UK news media could look like, and how their prejudice-mitigating potentials could be examined. A framework of discourse analysis was developed using descriptive parameters informed by Cognitive-Linguistic CDS (Hart, 2011b) so that the lexicosemantic makeup of these representations of Muslim communities could be described and accounted for. In particular, I focused on news-based narratives of contact between Muslim and non-Muslim communities, whereby episodes of intergroup contact and processes of attitudinal change are depicted as central themes. The impact of exposure to these texts on implicit attitudes (Greenwald & Banaji, 1995) was subsequently measured via an Evaluative Decision Task (Draine & Greenwald, 1998). Drawing on Contact Theory (Hewstone & Swart, 2011), it was hypothesised that exposure to such representations of intergroup contact would modulate Islamophobic prejudice. Though only partially successful in modulating intergroup prejudice as measured in this way, this work offers one example of how the more general model of prognostic critique advanced can be implemented. I also offer insights into learnings, limitations and areas of improvement to consider when applying this model of research.