The Meaning of Populism: A Multi-country Text-as-data Analysis

• In Event: Measuring Populism and Its Impacts

Abstract

The rise of populism across Europe, the United States, and a number of other regions and countries has been recognized as one of the most important political developments of the current century. The Guardian has claimed that as many as 25% of Europeans votes for populist parties (1), and in several countries populist parties and politicians have become part of the government. However, while virtually nobody would deny that populism has become dramatically more visible and popular, some scholars consider that estimate greatly exaggerated, basing their disagreement on which political parties count as “true” populist parties.

Indeed, how to define and operationalize the notion of populism is a question that is far from settled in the literature. Even the Guardian acknowledged that populism “has different characteristics by region” (1) and that there is a “noisy dispute over the meaning of populism” (2). Many scholars have coalesced around the definition proposed by Cas Mudde in 2004, which centers on a conflict between “the pure people” and “the corrupt elite” (3), but no such consensus exists among journalists, pundits, or politicians.

This paper contributes to resolving the problem by analyzing systematically how the term “populism” is used in the print media of four different English-speaking countries: two where populism is widely considered to be politically powerful (the United States and the United Kingdom), and two where it is not (Canada and Ireland). Specifically, I apply computational methods developed in recent years to produce distributed representations of words and concepts. These methods — Google’s word2vec is the best known, but I will also use the GloVe method developed at Stanford — convert words and phrases to many-dimensional vectors that can be thought of as representing a location in semantic space. By applying these algorithms to a corpus containing thousands of newspaper articles from each country, I produce four different such spaces. In each, it is possible to examine the semantic neighbourhood of the word “populism”, to identify what the word means in each country’s media discourse.

Similarities and differences observed in this analysis will tell us a lot about how journalists and publics in each country understand the concept, and how this understanding may differ from the academic one mentioned above. They will also point to different causal factors that have produced cross-national variation in the shared public conceptualization of populism. And they will shed light on why some politicians are considered populist by observers in one country while they may not be considered so by observers from another country. Finally, the approach introduced can be applied to any number of other contested political concepts: terrorism, liberalism, welfare, etc. As such, the paper also makes an important broader methodological contribution

(1) John Henley. 20 Nov. 2018. “How populism emerged as an electoral force in Europe.” The Guardian.
(2) Peter Baker. 10 Jan. 2019. “’We the people’: the battle to define populism.” The Guardian
(3) Cas Mudde. 2004 “The Populist Zeitgeist.” Government and Opposition

Author

• A. Maurits van der Veen, College of William & Mary