DISCOURSE AND IDEOLOGY: A CORPUS-ASSISTED CRITICAL DISCOURSE ANALYSIS OF EIGHT FLAGSHIP PAKISTANI POLITICAL SPEECHES (1988–2025)

This research applies a corpus-assisted Critical Discourse Analysis to eight flagship speeches delivered by leading Pakistani politicians between(1988(and(2025. Sixteen speeches (two for each leader) were cleaned, tokenized, and examined in AntConc to generate keyword lists, collocation tables, four-word clusters, and pronoun concordances. SPSS supplied descriptive statistics, chi-square tests, and a Ward dendrogram to validate lexical patterns. Findings show that each party cultivates a unique lexical fingerprint: PTI foregrounds corruption and justice, PPP stresses rights and climate, PML-N centers on development, and Musharraf’s military rhetoric highlights security. Collocational frames such as corruption + mafia or development + motorway turn neutral nouns into moral sign-posts, supporting Fowler’s view that ideology “travels in word networks” (Fowler,(1991). Four-gram slogans such as vote ko izzat do or Naya Pakistan ki tashkeel, act as memory hooks that compress complex agendas into rhythmic phrases, a mechanism Hart links to cognitive salience (Hart,(2010). Pronoun patterns reveal group construction: populist voices deploy we for virtuous citizens and they for corrupt elites, matching Van(Dijk’s model of ideological polarization (Van(Dijk,(2006). Quantitative tests confirm that keyword distribution is party-specific (p(<(.05), while the dendrogram clusters reformist and establishment leaders along lexical lines. Together these layers illustrate Fairclough’s claim that textual choices both reflect and reproduce wider power relations (Fairclough,(2013). The study contributes (i) the first open, balanced speech corpus for Pakistan, (ii) a mixed workflow that pairs AntConc counts with CDA reading, and (iii) practical guidelines for journalists, educators, and regulators who monitor political language. Limitations include the absence of Urdu originals and multimodal cues; future work should add aligned translations and gesture analysis. Even so, the project demonstrates that small, transparent corpora can expose the strategic engineering of ideology in bilingual political discourse.