Populist polarization in postcolonial Philippines: Sociolinguistic rifts in online drug war discourse

Abstract

Social psychological scholarship on populism explains polarization processes in terms of individual differences and group‐level divisions. However, predominant approaches often elide wider historical contexts, implicitly assuming the structural features of Western settings. Invoking insights from postcolonial psychology, we posit that populist polarization dynamics in the Global South are structurally shaped by colonial histories. Turning to the Philippines under populist President Rodrigo Duterte, we examine polarized online constructions of his controversial war on drugs. Employing a discourse‐historical approach, we uncover a structurally bifurcated view of populist polarization. We specifically unpack sociolinguistic rifts across vernacular discourses of community security uttered in Filipino versus cosmopolitan discourses of democratic integrity uttered in English. Our findings expand prevailing scholarship by illustrating postcolonial discursive possibilities of populism from above and resistance from below. We conclude with insights for studying populist polarization at the nexus of local and global inequalities.

Bots and online hate during the COVID-19 pandemic: case studies in the United States and the Philippines

“…First, we discuss the problem of hate speech on social media. We highlight the importance of shifting from the prevailing concern with classification to applied settings of characterization [80,84]. Second, we link our view of hate speech to the literature on bots and information maneuvers.…”

“…Scholars note that many kinds of social bots display distinctive behavioral patterns to achieve concerted political or economic ends, including spamming, increasing the perceived following of a politician, or sowing discord [83,87]. The literature also points out that some botssometimes called cyborgs-are only partially automated, such that certain messages they send can attain a higher level of sophistication due to intermittent interventions by humans during real-time interactions [27,80]. For the purposes of this work, and in line with the operationalizations embedded in tools described in Sect.…”

“…For user-level analysis, we obtained the average hate speech score assigned to all tweets by the account within the time interval under consideration. Our model utilized handcrafted linguistic features using the Netmapper software [11,79,80]. These features included lexical counts of pronouns, abusive terms, exclusive terms, absolutist terms, and identity terms, among others based on prior scholarship linking language use to psychological states [63,76].…”

“…Adopting a tiered structure, BotHunter progressively leverages account information, network information, and dynamic features to generate performant bot predictions. It has been successfully used to understand various domestic and international influence campaigns [14,16,78,80].…”

“…This work bears several implications for understanding online hate speech in the context of the pandemic and beyond. We pivot from extant technical approaches of classifying hate speech [7,35,53], to theory-informed frameworks for characterizing it in the context of large-scale social interactions and potential information maneuvers [44,80,84]. We also reflect on the value of taking a global approach to computational social science, especially in the context of international issues like COVID-19, with its universal yet also distinct and unequal impacts in societies worldwide [31,51,54].…”