The many roads to Rome: family resemblance concepts in the social sciences

What are the different ways in which family resemblance conceptual structures can be used in the social sciences? When should they be used and what are the specific advantages and challenges they pose? What are the descriptive and causal implications derived from their use? This paper advances our understanding of these conceptual structures by answering these questions and illustrating them with examples from established scholarship in political science. The paper first breaks down the broad category of family resemblance concepts into three types of structures, using set theoretic logic: individual sufficiency, INUS, and mixed structures. Second, we discuss when these alternative structures are more useful for concept formation, proposing three different prototypical situations to use them: when we seek to disaggregate an abstract attribute, when our concept relies on the negation of classic concepts, and when our concept rests on what we label as cumulative signification. Third, the paper introduces the logic of subtype formation in family resemblance and compares it with classical subtypes, showing that unlike the latter, subtypes in family resemblance do not require additional attributes; sufficient combinations of the main concept can at the same time be subtypes. Forth, the paper analyzes the descriptive and causal implications of using family resemblance conceptual structures, showing a trade-off between empirical differentiation and potential causal heterogeneity. A concluding section evaluates some misuses of family resemblance, highlighting the importance of avoiding empirical and theoretical pitfalls.