Party System Institutionalization: Do We Need a New Concept?

摘要

This paper revises the conceptualization of party system institutionalization, as defined in Mainwaring and Scully's (1995) groundbreaking work. Regarding the rules of aggregation involved in conceptual structure, my argument resonates with Goertz and Mahoney's (2012) identification of two different "cultures" in contemporary social science. They understand one culture as building what Goertz calls "family resemblance" concepts, usually measured through a latent variable approach. The other culture understands concepts as structured in terms of "necessary and sufficient" elements, which can then be measured using an ontological approach. I claim that party system institutionalization has an implicit "family resemblance" structure and show that conceptual structure to be empirically and theoretically inadequate. In its current form, the concept of party system institutionalization also suffers from deficiencies when it comes to indicator validity and aggregation, as per (Munck and Verkulien, Comp Polit Stud 35(1):5-34, 2002) criteria. Problems of validity are caused by analysts' frequent reliance on a single indicator for operationalizing the concept. Problems of aggregation and conceptual structure arise from inconsistencies between the implicit theoretical assumption that party system institutionalization is conceptually linear and non-linear patterns that are not only theoretically plausible but also empirically observable in a large set of cases. Therefore, this paper advocates revising the concept and the way it is currently applied in the comparative party system literature. Such revision will permit better understanding of both the characteristics and dynamic evolution of party systems.