In a recent book by the political philosopher and my compatriot, Cristobal Bellolio, he advances the following thesis: the so-called crisis of democracy largely depends on how we define democracy.1 Therefore, if we understand democracy as liberal democracy, then we can speak of a democratic crisis. Conversely, if we believe that democracy, for example, should include populist elements, then what we are experiencing is a correction of democracy.

For some people, this might sound too soft or even meaningless. How, for instance, could Hungary be considered a corrected version of democracy, while Norway a democracy in need of correction? Bellolio, however, does not play word games. In fact, he is a supporter of liberal democracy. His thesis is purely descriptive, not normative (although working with concepts like democracy makes it difficult to separate the two, but that is a discussion for another time). And as such, I believe he is right.

To illustrate how challenging working with a concept can be, consider the following questions: Can a woman who didn't give birth to a child but raised him be considered his mother? What if this woman is officially his aunt or grandmother—could she be considered his mother? Chess, football, and boxing are all sports, but what do they have in common that groups them together? Perhaps they all have rules and involve some form of competition, but these features are common to all games, and not all games are sports. In the following, I will describe three common forms of concept formation that almost every political science student learns while studying political science (and probably any other social science).2 The purpose is to demonstrate one of the many difficulties that political scientists face when conducting research, and this is just at the beginning of it—conceptualization.

Let's start with the basics: why do we need to define a concept? At the initial stage of any research project, it is necessary to define the concepts that will be used and establish their respective indicators in order to develop an operational research design for implementation. Social and political sciences are no exception. Moreover, as these disciplines are largely comparative, it is essential to be very careful with our conceptualizations to ensure that the objects being compared are similar. In simple terms, we cannot just compare pears and apples.

The most popular conceptualization method is based on the rules of concept formation elaborated by the political scientist Giovanni Sartori. Here, the focus is on the concept's level of abstraction and the trade-off between the number of cases we want to study and the concept's number of attributes: the more abstract the concept is (or the fewer attributes it has), the wider the range of cases we can analyze, and vice versa. Therefore, when moving from a narrower to a wider range of cases, we must make our concepts more abstract to avoid conceptual stretching. The opposite is also true: when moving to a narrower range of cases, we must make our concepts less abstract by adding more attributes. In other words, there is a trade-off between the extension of the concept (the range of cases it covers) and its connotation (the number of attributes).

To illustrate this, consider the differences between democracies and non-democracies. It is not possible to compare them without stretching the concept of democracy or non-democracy. However, if we move higher in the ladder of abstraction from democracy (or non-democracy) to regime, we can properly compare them (although we will examine fewer attributes of democracies and non-democracies). For example, when comparing France (a democracy) and Russia (a non-democracy), it is necessary to consider the countries in very abstract terms (at the level of regime) while bearing in mind a few elements.

So far, so good. Sartori’s rules of concept formation follow a clear logic based on the assumption that the extension and connotation of a concept always vary inversely. However, to move in the ladder of abstraction, we must identify a minimum definition of the concept and a clear hierarchy of its attributes. This is where the problems start.

Again, consider democracy as an example. What is its most distinctive feature? Multiparty elections? Illiberal democracies have those, too. Or the guarantee of basic rights and freedoms? This disregards the importance of accountability in democratic processes. Should we consider the two elements as necessary and sufficient as presented in liberal democracies? In that case, only a few cases would belong to the category of democracy. Therefore, the minimum definition would not be very extensive, that is, the minimum definition, which is supposed to be abstract, would not include many cases.

To overcome these problems, one alternative is to use radial categories, the second type of conceptualization. The process of developing them involves the following steps: First, we identify a primary or ideal-type category that includes all the main attributes associated with the concept. Not all attributes need to be found together in a real case. This is because it would be a mistake to consider cases that lack certain attributes as not fitting the category. Therefore, secondary categories must be created that focus on a single core attribute while allowing for variations or omissions of other attributes.

A familiar example of this applies to the concept of "mother." A stepmother, a genetic mother, a nurturing mother, and more are all types of "mother," with "woman" being the core attribute of this concept3. In this case, the aunt or grandmother of the kid mentioned above could easily be his mother.

But let's illustrate this with the Liberal Democracy Index (LDI) of V-DEM. The LDI is based on a mix of electoral institutions, being this the core attribute of democracy, plus the liberal dimension. The latter is defined as "constraints on the executive by the legislature and the judiciary, and the rule of law ensuring respect for civil liberties".4 According to the LDI, both Norway and Romania are democracies. However, Norway is closer to the ideal, while Romania belongs to a larger group of democracies (electoral democracies) than Norway (liberal democracies). Terms with a prefix (like "liberal democracy" and "electoral democracy") used to be examples of radial categories.

Finally, there is another form of conceptualization that overcomes the shortcomings of Sartori's rules without identifying the concept's core attributes. It is the family resemblance method inspired by the Austrian philosopher Ludwig Wittgenstein. This form "is intended to solve the problem of dealing with objects that are clearly linked or associated with one another, much as family members are linked genetically, but where none is characterized by all of the relevant set of attributes."2 Sports is an example of this.

The first problem with this form of conceptualization is knowing when to stop adding attributes to a concept. Since there is no trade-off between extension and connotation or ideal types with core attributes, adding more attributes results in more cases. Thus, we miss the point of conceptualization entirely. In other words, it's difficult to conclude anything relevant by comparing chess, boxing, and football. However, what case can be made for not classifying one of them as a sport?

I wrote this brief review of conceptualization to demonstrate the difficulty of the work of social and political scientists. Perhaps we don't have to be as precise as natural scientists do, but our object of study is constantly slipping out of our hands. I included more colloquial examples, such as the one about sports or mothers, to make the reader more familiar with this challenge. Additionally, I wanted to show that social scientists are not idle and that if we can't provide the answers people are expecting, it's because we can't, not because of a lack of seriousness or malice. As I mentioned earlier, Bellolio is right to point out that the so-called crisis of democracy depends on how we define democracy. In another article, I will explain how discourse influences which definitions sound more convincing, but I will also explain that this does not mean these definitions are the only right ones.

References

1 Bellolio, C. (2024). La era del pesimismo democrático. Debate.
2 Mair, P. (2008). Concepts and concept formation. In D. Della Porta & M. Keating (Eds.), Approaches and Methodologies in the Social Sciences: A Pluralist Perspective (pp. 177–197).
3 Cambridge University Press.Collier, D., & Mahon, J. E. (1993). Conceptual “Stretching” Revisited: Adapting Categories in Comparative Analysis. American Political Science Review, 87(4), 845–855.
4 Nord, M., D. Altman, F. Angiolillo, T. Fernandes, A. Good, and S. Lindberg. (2025). Democracy Report 2025: 25 Years of Autocratization – Democracy Trumped? V-Dem Institute.