A Computational-Augmented Critical Discourse Analysis of Tweets on the Saudi General Entertainment Authority Activities

Abstract

This study used both computational tools in the form of a machine learning predictive model (Support Vector Machine) and a critical discourse analysis model (Van Dijk’s ideological square model) (Van Dijk, 1993, 2008, 2009) to fulfill three objectives: (1) clustering the Saudis’ Twitter-based opinions and sentiments regarding the entertaining and recreational activities run by the Saudi General Entertainment Authority (GEA); (2) offering empirical evidence on how computational linguistic methods could be implemented for offering a reliable conceptual framing of such opinionated big data; and (3) outlining the central themes generating ideologically motivated polarity in Saudi public opinion and the macrostrategies through which this polarity is textually instantiated and actualized. Toward fulfilling these objectives, we designed a purpose-built corpus of 9378 tweets based on five trending hashtags, covering the period between 2020 and 2022. Findings affirmed the efficacy of synergizing the Support Vector Machine model and the ideological square model in clustering and interpreting the target tweets. Based on the output discourse features and thematization of the tweets, two main groups with different ideologically motivated perspectives were identified. This ideological polarity was achieved through the use of two macrostrategies: positive self-presentation and negative other-presentation. These findings may prompt policymakers to reconsider current (mis)practices in order to achieve long-term sustainable development goals.