The Curbing of Fake News: A three level cognitive approach

Liane Stroebel
Rwth Aachen

Abstract

The COVID-19 outbreak was accompanied by a massive "infodemic", making it difficult to identify reliable sources. While poorly edited fake news fundamentally differs from trustworthy information in the area of narrative structure (van der Linden, 2017), the situation is much different with fake news of higher editorial standards. Our goal is to develop targeted methods for identifying fake news using neurolinguistic approaches and big data analysis. The innovative character of this approach lies in the inclusion of linguistic elements beyond content analysis. In contrast to existing work, the focus is not on frequency analysis of keywords, metaphors or the interaction between the headlines and the content of the articles (Granik & Mesyura, 2017; Rashkin et al., 2017; Bourgonje et al., 2017), but on the identification of unconscious linguistic phenomena on three distinctive levels: the conceptual, the motivational and the iconic level; i.e. of speech material that is difficult to consciously manipulate or suppress. One advantage of this method, which should not be underestimated, is that untrustworthy sources fall off the grid, thus allowing for a targeted isolation of fake news on the internet in general and concerning the current pandemic in particular. Our corpus consists of online articles in three languages (English, German and Spanish).


Download Proceedings Book Volume1
Download Proceedings Book Volume2

Presentation