A Corpus Analysis of Hate Speech in Pakistani Political Discourse on Twitter

Abstract

This paper is a corpus-based quantitative study on language change that has arisen during national political events, such as the no-confidence movement against Imran Khan, the Prime Minister of Pakistan. This event, including many others, has given rise to the hate speech in Pakistani community online. In this paper, language change is discussed in terms of hate speech, collocation patterns, frequency of words (Nouns, Adjectives), N-Grams and Urdu hate speech words. A certain political time period is selected to make a comparison of the use of language and explore whether any variations have transpired during the political upheavals. These dynamics are explored because the frequency of hate speech words may suggest that how much aggressive the online political discourse has become. Moreover, it may also suggest that what kind of image is presented to the rest of the world through such usage of words. To find out the case, a data of 50,000 tweets was gathered using Tweetarchivist.com. This data was analyzed through sketch engine. The major findings from the research, reported in this paper, show that many of the tweets, through their collocations and frequency, contain hate speech elements which convey a negative image of the country across the globe.