Emotion recognition in italian political language for prefiguring crisis in the balance of the parties’ coalitions

Abstract

The purpose of this study is to describe the emotions that dominated the political vocabulary used on Twitter during the Covid-19 epidemic by the leaders of the major Italian political parties, including Forza Italia, Movimento 5 Stelle, Fratelli d’Italia, Italia Viva, Lega, and Partito Democratico. We developed a 4-step analysis model based on Natural Language Processing (NLP) that combines: (1) Exploratory Textual Data Analysis; (2) Emotion Recognition (ER) for Italian language; (3) cross-tabulation coding between leaders and emotions; (4) the multivariate approach of Correspondence Analysis in order to determine the associations between politicians and four categories of emotions: anger, fear, sadness, and joy. Specifically, we contrasted the language utilized by Covid-19 leaders throughout the first and second waves. The results reveal an intriguing shift in the emotions communicated through political discourse between the first and second waves. Giorgia Meloni’s position as head of Fratelli d’Italia shifted significantly from fury to fear and grief, aligning her more with centrists and foreshadowing a serious rift in the right-wing and the balance of coalitions in the political arena. Giorgia Meloni became Italy’s Prime Minister in September 2022. This political success shows that the sentiment analysis done in this paper could be a good way to predict the future.