BRINGING KASHMIR CLOSER: AN AI-ASSISTED PROXIMIZATION ANALYSIS OF IMRAN KHAN’S 2019 ADDRESS TO THE UNITED NATIONS GENERAL ASSEMBLY

Abstract

This study examines how the former Prime Minister of Pakistan, Imran Khan, used language to turn the Kashmir dispute into an urgent concern for the international community in his address to the United Nations General Assembly (UNGA) on 27 September 2019. Earlier studies of the address have mostly listed its themes or described its cohesion; less attention has been paid to how it builds a sense of approaching danger. The study combines Fairclough’s three-dimensional model of critical discourse analysis with Cap’s spatial–temporal–axiological (STA) model of proximization, and adds two AI-assisted procedures: first-pass annotation of proximization markers by a large language model, checked by the researcher, and lexicon-based sentiment scoring with VADER. The data is a published transcript of the address (2,575 words, 212 sentences). Of all sentences, 57.5% carry at least one proximization marker, rising to 65.0% in the Kashmir section, which takes up just over half of the address. Islamophobia is framed mainly through clashes of values, while the Kashmir section fuses all three strategies and builds towards a temporal climax around the analogy of Munich 1939. The address contains no hedging modals (may, might) and keeps direct address to the audience (you) almost entirely for the Kashmir section. The sentiment tool broadly agreed with the manual reading but scored a third of the proximization sentences as neutral, including the Munich analogy. The study argues that Khan proximizes from the periphery, moving his audience’s historical memory towards Kashmir, and that AI tools serve critical discourse analysis only when their output is read critically.

Keywords: Critical discourse analysis; proximization; Imran Khan; United Nations General Assembly; Kashmir; AI-assisted discourse analysis; large language models