Version classiqueVersion mobile

Wellbeing, Freedom and Social Justice

Ingrid Robeyns

2. Core Ideas and the Framework

Texte intégral

2.1 Introduction

  • 1 In developing the account of the capability approach as presented in this book, I have started in a (...)

1The previous chapter listed a range of fields in which the capability approach has been taken up, and in chapters 3 and 4 we focus in more detail on how the capability approach can (or cannot) make a difference for thinking about wellbeing, social and distributive justice, human rights, welfare economics and other topics. This broad uptake of the capability approach across disciplines and across different types of knowledge production (from theoretical and abstract to applied or policy oriented) is testimony to its success. But how is the capability approach understood in these different fields, and is it possible to give a coherent and clarifying account of how we can understand the capability approach across those fields?1 In other words, how should we understand the capability approach as an overarching framework, that unites its more specific uses in different fields and disciplines?

2This chapter gives an account of that general framework. It provides explanations and insights into what the capability approach is, what its core claims are, and what additional claims we should pay attention to. This chapter will also answer the question: what do we mean exactly when we say that the capability approach is a framework or an approach?

3In other words, in this chapter, I will give an account (or a description) of the capability approach at the most general level. I will bracket additional details and questions about which disagreement or confusion exists; chapter 3 will offer more detailed clarifications and chapter 4 will discuss issues of debate and dispute. Taken together, these three chapters provide my understanding of the capability approach.

  • 2 To the reader who finds this ultra-brief summary of the modular account of the capability approach (...)

4Chapter 2 is structured as follows. In the next section, we look at a preliminary definition of the capability approach. Section 2.3 proposes to make a distinction between ‘the capability approach’ and ‘a capability theory’. This distinction is crucial — it will help us to clarify various issues that we will look at in this book, and it also provides some answers to the sceptics of the capability approach whom we encountered in the previous chapter. Section 2.4 describes, from a bottom-up perspective, the many ways in which the capability approach has been developed within particular theories, and argues why it is important to acknowledge the great diversity within capabilitarian scholarship. Sections 2.5 to 2.11 present an analytic account of the capability approach, and show what is needed to develop a particular capability theory, application or analysis. In 2.5, I propose the modular view of the capability approach, which allows us to distinguish between three different types of modules that make up a capability theory. The first, the A-module, consists of those properties that each capability theory must have (section 2.6). This is the non-optional core of each capability theory. The B-modules are a set of modules in which the module itself is non-optional, but there are different possible choices regarding the content of the module (section 2.7). For example, we cannot have a capability theory or application without having chosen a purpose for that theory or application — yet there are many different purposes among which we can choose. The third group of modules, the C-modules, are either non-optional but dependent on a choice made in the B-modules, or else are completely optional. For example, if the purpose you choose is to measure poverty, then you need to decide on some empirical methods in the C-modules; but if your aim is to make a theory of justice, you don’t need to choose empirical methods and hence the C-module for empirical methods (module C3) is not relevant for your capability application (section 2.8).2 In the next section, 2.9, I discuss the possibility of hybrid theories — theories that give a central role to functionings and capabilities yet violate some other core proposition(s). Section 2.11 rounds up the discussion of the modular view by discussing the relevance and advantages of seeing the capability approach from this perspective.

5Section 2.12 summarises the conceptual aspects that have been explained by presenting a visualisation of the conceptual framework of the capability approach. Section 2.13 uses the modular view to illuminate the observation, which has been made by several capability scholars, that the capability approach has been used in a narrow and in a broad sense, and explains what difference lies behind this distinction. In the broader use of the capability approach, supporting or complementary theories or additional normative principles are added to the core of the approach — yet none of them is itself essential to the capability approach. These are choices made in B-modules or the C-modules. The modular view of the capability approach that will be central to this chapter can thus help to formulate in a sharper manner some observations that have already been made in the capability literature. The chapter closes by looking ahead to the next chapter.

2.2 A preliminary definition of the capability approach

  • 3 Amartya Sen often uses the term “social arrangement”, which is widely used in the social choice lit (...)

6The capability approach has in recent decades emerged as a new theoretical framework about wellbeing, freedom to achieve wellbeing, and all the public values in which either of these can play a role, such as development and social justice. Although there is some scholarly disagreement on the best description of the capability approach (which will be addressed in this chapter), it is generally understood as a conceptual framework for a range of evaluative exercises, including most prominently the following: (1) the assessment of individual levels of achieved wellbeing and wellbeing freedom; (2) the evaluation and assessment of social arrangements or institutions;3 and (3) the design of policies and other forms of social change in society.

  • 4 See Nussbaum (1988, 1992), Sen (1993a, 1999a, 14, 24); Walsh (2000); Qizilbash (2016); Basu and Lóp (...)
  • 5 A partial theory of justice is a theory that gives us an account of some aspects of what justice re (...)
  • 6 On the relationship between capabilities and human rights, see section 3.14.
  • 7 For earlier attempts to describe the capability approach, see amongst others Deneulin (2014); Gaspe (...)

7We can trace some aspects of the capability approach back to, among others, Aristotle, Adam Smith, Karl Marx and John Stuart Mill,4 yet it is Sen who pioneered the approach and a growing number of other scholars across the humanities and the social sciences who have significantly developed it — most significantly Martha Nussbaum, who has developed the capability approach into a partial theory of social justice.5 Nussbaum also understands her own capabilities account as a version of a theory of human rights.6 The capability approach purports that freedom to achieve wellbeing is a matter of what people are able to do and to be, and thus the kind of life they are effectively able to lead. The capability approach is generally conceived as a flexible and multipurpose framework, rather than a precise theory (Sen 1992a, 48; Alkire 2005; Robeyns 2005b, 2016b; Qizilbash 2012; Hick and Burchardt 2016, 78). The open-ended and underspecified nature of the capability approach is crucial, but it has not made it easier for its students to understand what kind of theoretical endeavour the capability approach exactly is. How should we understand it? Isn’t there a better account possible than the somewhat limited description of the capability approach as ‘an open, flexible and multi-purpose framework’? Answering these questions will be the task of this chapter.7

8The open and underspecified nature of the capability approach also explains why the term ‘capability approach’ was adopted and is now widely used rather than ‘capability theory’. Yet as I will argue in section 2.3, we could use the terms ‘capability theory’ and ‘capability approach’ in a more illuminating way to signify a more substantive difference, which will help us to get a better grip on the capability literature.

  • 8 A ‘technical term’ is a term which is used in a specialist debate, and has a meaning that is define (...)
  • 9 To ‘remain agnostic’ means that, for the purpose of that analysis, one does not make a choice betwe (...)

9It may be helpful to introduce the term ‘advantage here, which is a technical term used in academic debates about interpersonal comparisons and in debates about distributive justice.8 A person’s advantage is those aspects of that person’s interests that matter (generally, or in a specific context). Hence ‘advantage’ could refer to a person’s achieved wellbeing, or it could refer to her opportunity to achieve wellbeing, or it could refer to her negative freedoms, or to her positive freedoms, or to some other aspect of her interests. By using the very general term ‘advantage’, we allow ourselves to remain agnostic between the more particular specifications of that term;9 our analysis will apply to all the different ways in which ‘advantage’ could be used. This technical term ‘advantage’ thus allows us to move the arguments to a higher level of generality or abstraction, since we can focus, for example, on which conditions interpersonal comparisons of advantage need to meet, without having to decide on the exact content of ‘advantage’.

  • 10 Following Amartya Sen (1985c), some would say there are four different ideas of advantage in the ca (...)
  • 11 Of course, it doesn’t follow that mental categories or the material means play no role at all; but (...)

10Within the capability approach, there are two different specifications of ‘advantage’: achieved wellbeing, and the freedom to achieve wellbeing. The notions of ‘functionings’ and ‘capabilities’ — which will be explained in detail in section 2.6.1 — are used to flesh out the account of achieved wellbeing and the freedom to achieve wellbeing.10 Whether the capability approach is used to analyse distributive injustice, or measure poverty, or develop curriculum design — in all these projects the capability approach prioritises certain people’s beings and doings and their opportunities to realize those beings and doings (such as their genuine opportunities to be educated, their ability to move around or to enjoy supportive social relationships). This stands in contrast to other accounts of advantage, which focus exclusively on mental categories (such as happiness) or on the material means to wellbeing (such as resources like income or wealth).11

  • 12 On whether the capability approach can be used for explanatory purposes, see section 3.10.

11Thus, the capability approach is a conceptual framework, which is in most cases used as a normative framework for the evaluation and assessment of individual wellbeing and that of institutions, in addition to its much more infrequent use for non-normative purposes.12 It can be used to evaluate a range of values that draw on an assessment of people’s wellbeing, such as inequality, poverty, changes in the wellbeing of persons or the average wellbeing of the members of a group. It can also be used as an evaluative tool providing an alternative for social cost-benefit analysis, or as a framework within which to design and evaluate policies and institutions, such as welfare state design in affluent societies, or poverty reduction strategies by governments and non-governmental organisations in developing countries.

12What does it mean, exactly, if we say that something is a normative analysis? Unfortunately, social scientists and philosophers use these terms slightly differently. My estimate is that, given their numerical dominance, the terminology that social scientists use is dominant within the capability literature. Yet the terminology of philosophers is more refined and hence I will start by explaining the philosophers’ use of those terms, and then lay out how social scientists use them.

  • 13 Alkire (2008) calls these normative applications “prospective analysis”, and argues that we need to (...)

13What might a rough typology of research in this area look like? By drawing on some discussions on methods in ethics and political philosophy (O. O’Neill 2009; List and Valentini 2016), I would like to propose the following typology for use within the capability literature. There are (at least) five types of research that are relevant for the capability approach. The first type of scholarship is conceptual research, which conducts conceptual analysis — the investigation of how we should use and understand certain concepts such as ‘freedom’, ‘democracy’, ‘wellbeing’, and so forth. An example of such conceptual analysis is provided in section 3.3, where I offer a (relatively simple) conceptual analysis of the question of what kind of freedoms (if any) capabilities could be. The second strand of research is descriptive. Here, research and analyses provide us with an empirical understanding of a phenomenon by describing it. This could be done with different methods, from the thick descriptions provided by ethnographic methods to the quantitative methods that are widely used in mainstream social sciences. The third type of research is explanatory analysis. This research provides an explanation of a phenomenon — what the mechanisms are that cause a phenomenon, or what the determinants of a phenomenon are. For example, the social determinants of health: the parameters or factors that determine the distribution of health outcomes over the population. A fourth type of research is evaluative, and consists of analyses in which values are used to evaluate a state of affairs. A claim is evaluative if it relies on evaluative terms, such as good or bad, better or worse, or desirable or undesirable. Finally, an analysis is normative if it is prescriptive — it entails a moral norm that tells us what we ought to do.13 Evaluative analyses and prescriptive analyses are closely intertwined, and often we first conduct an evaluative analysis, which is followed by a prescriptive analysis, e.g. by policy recommendations, as is done, for example, in the evaluative analysis of India’s development conducted by Jean Drèze and Amartya Sen (2013). However, one could also make an evaluative analysis while leaving the prescriptive analysis for someone else to make, perhaps leaving it to the agents who need to make the change themselves. For example, one can use the capability approach to make an evaluation or assessment of inequalities between men and women, without drawing prescriptive conclusions (Robeyns 2003, 2006a). Or one can make a prescriptive analysis that is not based on an evaluation, because it is based on universal moral rules. Examples are the capability theories of justice by Nussbaum (2000; 2006b) and Claassen (2016).

14The difference with the dominant terminology used by economists (and other social scientists) is that they only distinguish between two types of analysis: ‘positive’ versus ‘normative’ economics, whereby ‘positive’ economics is seen as relying only on ‘facts’, whereas ‘normative economics’ also relies on values (e.g. Reiss 2013, 3). Hence economists do not distinguish between what philosophers call ‘evaluative analysis’ and ‘normative analysis’ but rather lump them both together under the heading ‘normative analysis’. The main take-home message is that the capability approach is used predominantly in the field of ethical analysis (philosophers’terminology) or normative analysis (economists’ terminology), somewhat less often in the fields of descriptive analysis and conceptual analysis, and least in the field of explanatory analysis. We will revisit this in section 3.10, where we address whether the capability approach can be an explanatory theory.

2.3 The capability approach versus capability theories

15The above preliminary definition highlights that the capability approach is an open-ended and underspecified framework, which can be used for multiple purposes. It is open-ended because the general capability approach can be developed in a range of different directions, with different purposes, and it is underspecified because additional specifications are needed before the capability approach can become effective for a particular purpose — especially if we want it to be normative (whether evaluative or prescriptive). As a consequence, ‘the capability approach’ itself is an open, general idea, but there are many different ways to ‘close’ or ‘specify’ this notion. What is needed for this specifying or closing of the capability approach will depend on the aim of using the approach, e.g. whether we want to develop it into a (partial) theory of justice, or use it to assess inequality, or conceptualise development, or use it for some other purpose.

  • 14 I kindly request readers who are primarily interested in the capability approach for policy design (...)

16This distinction between the general, open, underspecified capability approach, and its particular use for specific purposes is absolutely crucial if we want to understand it properly. In order to highlight that distinction, but also to make it easier for us to be clear when we are talking about the general, open, underspecified capability approach, and when we are talking about a particular use for specific purposes, I propose that we use two different terms (Robeyns 2016b, 398). Let us use the term ‘the capability approach for the general, open, underspecified approach, and let us employ the term ‘a capability theory’ or ‘a capability analysis, capability account or capability application’ for a specific use of the capability approach, that is, for a use that has a specific goal, such as measuring poverty and deriving some policy prescriptions, or developing a capabilitarian cost-benefit analysis, or theorising about human rights, or developing a theory of social justice. In order to improve readability, I will speak in what follows of ‘a capability theory’ as a short-hand for ‘a capability account, or capability application, or capability theory’.14

17One reason why this distinction between ‘capability approach’ and ‘capability theory’ is so important, is that many theories with which the capability approach has been compared over time are specific theories, not general open frameworks. For example, John Rawls’s famous theory of justice is not a general approach but rather a specific theory of institutional justice (Rawls 2009), and this has made the comparison with the capability approach at best difficult (Robeyns 2008b). The appropriate comparison would be Rawls’s theory of justice with a properly developed capability theory of justice, such as Nussbaum’s Frontiers of Justice (Nussbaum 2006b), but not Rawls’s theory of justice with the (general) capability approach.

  • 15 Yet even for capability theories, it is unlikely that Sen would agree that he has to draw up a list(...)

18Another reason why the distinction between ‘capability approach’ and ‘capability theory’ is important, is that it can help provide an answer to the “number of authors [who] ‘complain’ that the capability approach does not address questions they put to it” (Alkire 2005, 123). That complaint is misguided, since the capability approach cannot, by its very nature, answer all the questions that should instead be put to particular capability theories. For example, it is a mistake to criticise Amartya Sen because he has not drawn up a specific list of relevant functionings in his capability approach; that critique would only have bite if Sen were to develop a particular capability theory or capability application where the selection of functionings is a requirement.15

19In short, there is one capability approach and there are many capability theories, and keeping that distinction sharply in mind should clear up many misunderstandings in the literature.

20However, if we accept the distinction between capability theories and the capability approach, it raises the question of what these different capability theories have in common. Before addressing that issue, I first want to present a bottom-up description of the many modes in which capability analyses have been conducted. This will give us a better sense of what the capability approach has been used for, and what it can do for us.

2.4 The many modes of capability analysis

21If the capability approach is an open framework, then what are the ways in which it has been closed to form more specific and powerful analyses? Scholars use the capability approach for different types of analysis, with different goals, relying on different methodologies, with different corresponding roles for functionings and capabilities. Not all of these are capability theories; some are capability applications, both empirical as well as theoretical. We can observe that there is a rich diversity of ways in which the capability approach has been used. Table 2.1 gives an overview of these different usages, by listing the different types of capability analyses.

22Normative theorising within the capability approach is often done by moral and political philosophers. The capability approach is then used as one element of a normative theory, such as a theory of justice or a theory of disadvantage. For example, Elizabeth Anderson (1999) has proposed the outlines of a theory of social justice (which she calls “democratic equality”) in which certain basic levels of capabilities that are needed to function as equal citizens should be guaranteed to all. Martha Nussbaum (2006b) has developed a minimal theory of social justice in which she defends a list of basic capabilities that everyone should be entitled to, as a matter of human dignity.

23While most normative theorising within the capability approach has related to justice, other values have also been developed and analysed using the capability approach. Some theorists of freedom have developed accounts of freedom or rights using the capability approach (van Hees 2013). Another important value that has been studied from the perspective of the capability approach is ecological sustainability (e.g. Anand and Sen 1994, 2000; Robeyns and Van der Veen 2007; Lessmann and Rauschmayer 2013; Crabtree 2013; Sen 2013). Efficiency is a value about which very limited conceptual work is done, but which nevertheless is inescapably normative, and it can be theorised in many different ways (Le Grand 1990; Heath 2006). If we ask what efficiency is, we could answer by referring to Pareto optimality or x-efficiency, but we could also develop a notion of efficiency from a capability perspective (Sen 1993b). Such a notion would answer the question ‘efficiency of what?’ with ‘efficiency in the space of capabilities (or functionings, or a mixture)’.

Table 2.1. The main modes of capability analysis

Table 2.1. The main modes of capability analysis

Source: Robeyns (2005a), expanded and updated

24Quantitative social scientists, especially economists, are mostly interested in measurement. This quantitative work could serve different purposes, e.g. the measurement of multidimensional poverty analysis (Alkire and Foster 2011; Alkire et al. 2015), or the measurement of the disadvantages faced by disabled people (Kuklys 2005; Zaidi and Burchardt 2005). Moreover, some quantitative social scientists, mathematicians, and econometricians have been working on investigating the methods that could be used for quantitative capability analyses (Kuklys 2005; Di Tommaso 2007; Krishnakumar 2007; Krishnakumar and Ballon 2008; Krishnakumar and Nagar 2008).

25Thick description or descriptive analysis is another mode of capability analysis. For example, it can be used to describe the realities of schoolgirls in countries that may have formal access to school for both girls and boys, but where other hurdles (such as high risk of rape on the way to school, or the lack of sanitary provisions at school) mean that this formal right is not enough to guarantee these girls the corresponding capability (Unterhalter 2003b).

26Finally, the capability approach can be used for conceptual work beyond the conceptualisation of values, as is done within normative philosophy. Sometimes the capability approach lends itself well to providing a better understanding of a certain phenomenon. For example, we could understand education as a legal right or as an investment in human capital, but we could also conceptualise it as the expansion of a capability, or develop an account of education that draws on both the capability approach and human rights theory. This would not only help us to look differently at what education is; a different conceptualisation would also have normative implications, for example related to the curriculum design, or to answer the question of what is needed to ensure that capability, or of how much education should be guaranteed to children with low potential market-related human capital (McCowan 2011; Nussbaum 2012; Robeyns 2006c; Walker 2012a; Walker and Unterhalter 2007; Wigley and Akkoyunlu-Wigley 2006).

27Of course, texts and research projects often have multiple goals, and therefore particular studies often mix these different goals and methods. Jean Drèze and Amartya Sen’s (1996, 2002, 2013) comprehensive analyses of India’s human development achievements are in part an evaluative analysis based on various social indicators, but also in part a prescriptive analysis. Similarly, Nussbaum’s (2000) book Women and Human Development is primarily normative and philosophical, but also includes thick descriptions of how institutions enable or hamper people’s capabilities, by focussing on the lives of particular women.

28What is the value of distinguishing between different uses of the capability approach? It is important because functionings and capabilities — the core concepts in the capability approach — play different roles in each type of analysis. In quality of life measurement, the functionings and capabilities are the social indicators that reflect a person’s quality of life. In thick descriptions and descriptive analysis, the functionings and capabilities form part of the narrative. This narrative can aim to reflect the quality of life, but it can also aim to understand some other aspect of people’s lives, such as by explaining behaviour that might appear irrational according to traditional economic analysis, or revealing layers of complexities that a quantitative analysis can rarely capture. In philosophical reasoning, the functionings and capabilities play yet another role, as they are often part of the foundations of a utopian account of a just society or of the goals that morally sound policies should pursue.

29The flexibility of functionings and capabilities, which can be applied in different ways within different types of capability analysis, means that there are no hard and fast rules that govern how to select the relevant capabilities. Each type of analysis, with its particular goals, will require its own answer to this question. The different roles that functionings and capabilities can play in different types of capability analyses have important implications for the question of how to select the relevant capabilities: each type of analysis, with its particular goals, will require its own answer to this question. The selection of capabilities as social indicators of the quality of life is a very different undertaking from the selection of capabilities for a utopian theory of justice: the quality standards for research and scholarship are different, the epistemic constraints of the research are different, the best available practices in the field are different. Moral philosophers, quantitative social scientists, and qualitative social scientists have each signed up to a different set of meta-theoretical assumptions, and find different academic practices acceptable and unacceptable. For example, many ethnographers tend to reject normative theorising and also often object to what they consider the reductionist nature of quantitative empirical analysis, whereas many economists tend to discard the thick descriptions by ethnographers, claiming they are merely anecdotal and hence not scientific.

30Two remarks before closing this section. First, providing a typology of the work on the capability approach, as this section attempts to do, remains work in progress. In 2004, I could only discern three main modes of capability analysis: quality of life analysis; thick description/descriptive analysis; and normative theories — though I left open the possibility that the capability approach could be used for other goals too (Robeyns 2005a). In her book Creating Capabilities, Nussbaum (2011) writes that the capability approach comes in only two modes: comparative qualify of life assessment, and as a theory of justice. I don’t think that is correct: not all modes of capability analysis can be reduced to these two modes, as I have argued elsewhere in detail (Robeyns 2011, 2016b). The different modes of capability analysis described in table 2.1 provide a more comprehensive overview, but we should not assume that this overview is complete. It is quite likely that table 2.1 will, in due course, have to be updated to reflect new types of work that uses the capability approach. Moreover, one may also prefer another way to categorise the different types of work done within the capability literature, and hence other typologies are possible and may be more illuminating.

31Second, it is important that we fully acknowledge the diversity of disciplines, the diversity of goals we have for the creation of knowledge, and the diversity of methods used within the capability approach. At the same time, we need not forget that some aspects of its development might need to be discipline-specific, or specific for one’s goals. As a result, the capability approach is at the same time multidisciplinary, interdisciplinary, but also forms part of developments within disciplines and methods. These different ‘faces’ of the capability approach all need to be fully acknowledged if we want to understand it in a nuanced and complete way.

2.5 The modular view of the capability approach

32It is time to take stock. What do we already know about the capability approach, and what questions are raised by the analysis so far? The capability approach is an open approach, and depending on its purpose can be developed into a range of capability theories or capabilitarian applications. It is focused on what people can do and be (their capabilities) and on what they are actually achieving in terms of beings and doings (their functionings). However, this still does not answer the question of what kind of framework the capability approach is. Can we give an account of a capability theory that is more enlightening regarding what exactly makes a theory a capabilitarian theory, and what doesn’t?

33In this book, I present an account of the capability approach that, on the one hand, makes clear what all capability theories share, yet on the other hand allows us to better understand the many forms that a capability theory or capability account can take — hence to appreciate the diversity within the capability approach more fully. The modular view that I present here is a modified (and, I hope, improved) version of the cartwheel model that I have developed elsewhere (Robeyns 2016b). The modular view shifts the focus a little bit from the question of how to understand the capability approach in general, to the question of how the various capability accounts, applications and theories should be understood and how they should be constructed. After all, students, scholars, policy makers and activists are often not concerned with the capability approach in general, but rather want to know whether it would be a smart idea to use the capability approach to construct a particular capability theory, application or account for the problem or question they are trying to analyse. In order to answer the question of whether, for their purposes, the capability approach is a helpful framework to consider, they need to know what is needed for a capability theory, application or account. The modular view that will follow will give those who want to develop a capability theory, application or analysis a list of properties their theory has to meet, a list of choices that need to be made but in which several options regarding content are possible, and a list of modules that they could take into account, but which will not always be necessary.

  • 16 To say that the insertion of those theories is fully optional is not the same as saying that capabi (...)

34Recall that in order to improve readability, I use the term ‘a capability theory’ as a short-hand for ‘a capability account, or capability analysis, or capability application, or capability theory’. A capability theory is constructed based on three different types of module, which (in order to facilitate discussion) will be called the A-module, B-modules and C-modules. The A-module is a single module which is compulsory for all capability theories. The A-module consists of a number of propositions (definitions and claims) which a capability theory should not violate. This is the core of the capability approach, and hence entails those properties that all capability theories share. The B-modules consist of a range of non-optional modules with optional content. That is, if we construct a capability theory, we have to consider the issue that the module addresses, but there are several different options to choose from in considering that particular issue. For example, module B1 concerns the ‘purpose’ of the theory: do we want to make a theory of justice, or a more comprehensive evaluative framework for societal institutions, or do we want to measure poverty or inequality, or design a curriculum, or do we want to use the capability approach to conceptualise ‘social progress’ or ‘efficiency’ or to rethink the role of universities in the twenty-first century? All these purposes are possible within the capability approach. The point of seeing these as B-modules is that one has to be clear about one’s purpose, but there are many different purposes possible. The C-modules are either contingent on a particular choice made in a B-module, or they can be fully optional. For example, one can offer a comprehensive evaluation of a country’s development path, and decide that as part of this evaluation, one wants to include particular accounts of the history and culture of that country, since this may make more comprehensible the reasons why that country has taken this particular development path rather than another. The particular historical account that would be part of one’s capability theory would then be optional.16

35Given this three-fold structure of the modular view of capability theories, let us now investigate what the content of the compulsory module A is, as well as what the options are within the B-modules and C-modules.

2.6 The A-module: the non-optional core of all capability theories

36What, then, is the content of the A-module, which all capability theories should share? Table 2.2 presents the keywords for the eight elements of the A-module.

Table 2.2. The content of the compulsory module A


Functionings and capabilities as core concepts


Functionings and capabilities are value-neutral categories


Conversion factors


The distinction between means and ends


Functionings and/or capabilities form the evaluative space


Other dimensions of ultimate value


Value pluralism


Valuing each person as an end

2.6.1 A1: Functionings and capabilities

  • 17 See section 2.2 for an explanation of the technical term ‘advantage’.
  • 18 For some core differences in the way Martha Nussbaum and Amartya Sen use the terms ‘functionings’ a (...)

37Functionings and capabilities are the core concepts in the capability approach. They are also the dimensions in which interpersonal comparisons of ‘advantage’ are made (this is what property A5 entails).17 They are the most important distinctive features of all capabilitarian theories. There are some differences in the usage of these notions between different capability theorists,18 but these differences do not affect the essence of these notions: capabilities are what people are able to be and to do, and functionings point to the corresponding achievements.

38Capabilities are real freedoms or real opportunities, which do not refer to access to resources or opportunities for certain levels of satisfaction. Examples of ‘beings’ are being well-nourished, being undernourished, being sheltered and housed in a decent house, being educated, being illiterate, being part of a supportive social network; these also include very different beings such as being part of a criminal network and being depressed. Examples of the ‘doings’ are travelling, caring for a child, voting in an election, taking part in a debate, taking drugs, killing animals, eating animals, consuming great amounts of fuel in order to heat one’s house, and donating money to charity.

  • 19 See also section 3.3 which discusses in more depth the kind of freedoms or opportunities that capab (...)

39Capabilities are a person’s real freedoms or opportunities to achieve functionings.19 Thus, while travelling is a functioning, the real opportunity to travel is the corresponding capability. A person who does not travel may or may not be free and able to travel; the notion of capability seeks to capture precisely the fact of whether the person could travel if she wanted to. The distinction between functionings and capabilities is between the realized and the effectively possible, in other words, between achievements, on the one hand, and freedoms or opportunities from which one can choose, on the other.

  • 20 The exceptions are those philosophers who want to develop normative theories while steering away fr (...)

40Functionings are constitutive of human life. At least, this is a widespread view, certainly in the social sciences, policy studies, and in a significant part of philosophy — and I think it is a view that is helpful for the interdisciplinary, practical orientation that the vast majority of capability research has.20 That means one cannot be a human being without having at least a range of functionings; they make the lives of human beings both lives (as opposed to the existence of innate objects) and human (in contrast to the lives of trees or animals). Human functionings are those beings and doings that constitute human life and that are central to our understandings of ourselves as human beings. It is hard to think of any phenomenological account of the lives of humans — either an account given by a human being herself, or an account from a third-person perspective — which does not include a description of a range of human functionings. Yet, not all beings and doings are functionings; for example, flying like a bird or living for two hundred years like an oak tree are not human functionings.

41In addition, some human beings or doings may not be constitutive but rather contingent upon our social institutions; these, arguably, should not qualify as ‘universal functionings’ — that is, functionings no matter the social circumstances in which one lives — but are rather ‘context-dependent functionings’, functionings that are to a significant extent dependent on the existing social structures. For example, ‘owning a house’ is not a universal functioning, yet ‘being sheltered in a safe way and protected from the elements’ is a universal functioning. One can also include the capability of being sheltered in government-funded housing or by a rental market for family houses, which is regulated in such a way that it does not endanger important aspects of that capability.

42Note that many features of a person could be described either as a being or as a doing: we can say that a person is housed in a pleasantly warm dwelling, or that this person does consume lots of energy to keep her house warm. Yet other functionings are much more straightforwardly described as either a being or a doing, for example ‘being healthy’ (a being) or ‘killing animals’ (a doing).

  • 21 Arguably, some of that work is being done by social choice theorists and others working with axioma (...)

43A final remark. Acknowledging that functionings and capabilities are the core concepts of the capability approach generates some further conceptual questions, which have not all been sufficiently addressed in the literature. An important question is whether additional structural requirements that apply to the relations between various capabilities should be imposed on the capability approach in general (not merely as a particular choice for a specific capability theory). Relatively little work has been done on the question of the conceptual properties of capabilities understood as freedoms or opportunities and on the question of the minimum requirements of the opportunity set that make up these various capabilities. But it is clear that more needs to be said about which properties we want functionings, capabilities, and capability sets to meet. One important property has been pointed out by Kaushik Basu (1987), who argued that the moral relevance lies not in the various capabilities each taken by themselves and only considering the choices made by one person. Rather, the moral relevance lies in whether capabilities are truly available to us given the choices made by others, since that is the real freedom to live our lives in various ways, as it is truly open to us. For example, if a teenager lives in a family in which there are only enough resources for one of the children to pursue higher education, then he only truly has the capability to pursue higher education if none of his older siblings has made that choice before him.21

2.6.2 A2: Functionings and capabilities are value-neutral categories

44Functionings and capabilities are defined in a value-neutral way. Many functionings are valuable, but not all functionings necessarily have a positive value. Instead, some functionings have no value or even have a negative value, e.g. the functioning of being affected by a painful, debilitating and ultimately incurable illness, suffering from excessive levels of stress, or engaging in acts of unjustifiable violence. In those latter cases, we are better off without that functionings outcome, and the functionings outcome has a negative value. Functionings are constitutive elements of human life, which consist of both wellbeing and ill-being. The notion of functionings should, therefore, be valueneutral in the sense that we should conceptually allow for the idea of ‘bad functionings’ or functionings with a negative value (Deneulin and Stewart 2002, 67; Nussbaum 2003a, 45; Stewart 2005, 190; Carter 2014, 79–81).

45There are many beings and doings that have negative value, but they are still ‘a being’ or ‘a doing’ and, hence, a functioning. Nussbaum made that point forcefully when she argued that the capability to rape should not be a capability that we have reason to protect (Nussbaum 2003: 44–45). A country could effectively enable people to rape, for example, either when rape is not illegal (as it is not between husband and wife in many countries), or when rape is illegal, but de facto never leads to any punishment of the aggressor. If there is a set of social norms justifying rape, and would-be rapists help each other to be able to rape, then would-be rapists in that country effectively enjoy the capability to rape. But clearly, rape is a moral bad, and a huge harm to its victims; it is thus not a capability that a country should want to protect. This example illustrates that functionings as well as capabilities can be harmful or have a negative value, as well as be positive or valuable. At an abstract and general level, ‘functionings’ and ‘capabilities’ are thus in themselves neutral concepts, and hence we cannot escape the imperative to decide which ones we want to support and enable, and which ones we want to fight or eliminate. Frances Stewart and Séverine Deneulin (2002, 67) put it as follows:

[…] some capabilities have negative values (e.g. committing murder), while others may be trivial (riding a one-wheeled bicycle). Hence there is a need to differentiate between ‘valuable’ and non-valuable capabilities, and indeed, within the latter, between those that are positive but of lesser importance and those that actually have negative value.

  • 22 On the complex nature of ‘care’, and what the need to care and be cared for requires from a just so (...)

46The above examples show that some functionings can be unequivocally good (e.g. being in good health) or unequivocally bad (e.g. being raped or being murdered). In those cases, there will be unanimity on whether the functionings outcome is bad or good. But now we need to add a layer of complexity. Sometimes, it will be a matter of doubt, or of dispute, whether a functioning will be good or bad — or the goodness or badness may depend on the context and/or the normative theory we endorse. An interesting example is giving care, or ‘care work’.22 Clearly being able to care for someone could be considered a valuable capability. For example, in the case of child care, there is much joy to be gained, and many parents would like to work less so as to spend more time with their children. But care work has a very ambiguous character if we try to answer whether it should be considered to be a valuable functioning from the perspective of the person who does the care. Lots of care is performed primarily because there is familial or social pressure put on someone (generally women) to do so, or because no-one else is doing it (Lewis and Giullari 2005). There is also the hypothesis that care work can be a positive functioning if done for a limited amount of time, but becomes a negative functioning if it is done for many hours. Hence, from the functionings outcome in itself, we cannot conclude whether this is a positive element of that person’s quality of life, or rather a negative element; in fact, sometimes it will be an ambiguous situation, which cannot easily be judged (Robeyns 2003).

47One could wonder, though, whether this ambiguity cannot simply be resolved by reformulating the corresponding capability slightly differently. In theory, this may be true. If there are no empirical constraints related to the observations we can make, then one could in many cases rephrase such functionings that are ambiguously valued into another capability where its valuation is clearly positive. For example, one could say that the functioning of providing care in itself can be ambiguous (since some people do too much, thereby harming their own longer-term wellbeing, or do it because no-one else is providing the care and social norms require them to do it). Yet there is a closely related capability that is clearly valuable: the capability to provide hands-on care, which takes into account that one has a robust choice not to care if one does not want to, and that one does not find oneself in a situation in which the care is of insufficient quantity and/or quality if one does not deliver the care oneself. If such a robust capability to care is available, it would be genuinely valuable, since one would have a real option not to choose the functioning without paying an unacceptable price (e.g. that the person in need of care is not properly cared for). However, the problem with this solution, of reformulating functionings that are ambiguously valuable into capabilities that are unequivocally valuable, is the constraints it places on empirical information. We may be able to use these layers of filters and conditions in first-person analyses, or in ethnographic analyses, but in most cases not in large-scale empirical analyses.

48Many specific capability theories make the mistake of defining functionings as those beings and doings that one has reason to value. But the problem with this value-laden definition is that it collapses two aspects of the development of a capability theory into one: the definition of the relevant space (e.g. income, or happiness, or functionings) and, once we have chosen our functionings and capabilities, the normative decision regarding which of those capabilities will be the focus of our theory. We may agree on the first issue and not on the second and still both rightly believe that we endorse a capability theory — yet this is only possible if we analytically separate the normative choice for functionings and capabilities from the additional normative decision of which functionings we will regard as valuable and which we will not regard as valuable. Collapsing these two normative moments into one is not a good idea; instead, we need to acknowledge that there are two normative moves being made when we use functionings and capabilities as our evaluative space, and we need to justify each of those two normative moves separately.

  • 23 The Multidimensional Poverty Index is developed by the Oxford Poverty and Human Development Initiat (...)

49Note that the value-laden definition of functionings and capabilities, which defines them as always good and valuable, may be less problematic when one develops a capability theory of severe poverty or destitution. We all agree that poor health, poor housing, poor sanitation, poor nutrition and social exclusion are dimensions of destitution. So, for example, the dimensions chosen for the Multidimensional Poverty Index developed by Sabina Alkire and her colleagues — health, education and living standard — may not elicit much disagreement.23 But for many other capability theories, it is disputed whether a particular functionings outcome is valuable or not. The entire field of applied ethics is filled with questions and cases in which these disputes are debated. Is sex work bad for adult sex workers, or should it be seen as a valuable capability? Is the capability of parents not to vaccinate their children against polio or measles a valuable freedom? If employees in highly competitive organisations are not allowed to read their emails after working hours, is that then a valuable capability that is taken away from them, or are we protecting them from becoming workaholics and protecting them from the pressure to work all the time, including at evenings and weekends? As these examples show, we need to allow for the conceptual possibility that there are functionings that are always valuable, never valuable, valuable or non-valuable in some contexts but not in others, or where we simply are not sure. This requires that functionings and capabilities are conceptualised in a value-neutral way, and hence this should be a core requirement of the capability approach.

2.6.3 A3: Conversion factors

50A third core idea of the capability approach is that persons have different abilities to convert resources into functionings. These are called conversion factors: the factors which determine the degree to which a person can transform a resource into a functioning. This has been an important idea in Amartya Sen’s version of the capability approach (Sen 1992a, 19–21, 26–30, 37–38) and for those scholars influenced by his writings. Resources, such as marketable goods and services, but also goods and services emerging from the non-market economy (including household production) have certain characteristics that make them of interest to people. In Sen’s work in welfare economics, the notion of ‘resources’ was limited to material and/or measurable resources (in particular: money or consumer goods) but one could also apply the notion of conversion factors to a broader understanding of resources, including, for example, the educational degrees that one has.

51The example of a bike is often used to illustrate the idea of conversion factors. We are interested in a bike not primarily because it is an object made from certain materials with a specific shape and colour, but because it can take us to places where we want to go, and in a faster way than if we were walking. These characteristics of a good or commodity enable or contribute to a functioning. A bike enables the functioning of mobility, to be able to move oneself freely and more rapidly than walking. But a person might be able to turn that resource into a valuable functioning to a different degree than other persons, depending on the relevant conversion factors. For example, an able-bodied person who was taught to ride a bicycle when he was a child has a high conversion factor enabling him to turn the bicycle into the ability to move around efficiently, whereas a person with a physical impairment or someone who never learnt to ride a bike has a very low conversion factor. The conversion factors thus represent how much functioning one can get out of a resource; in our example, how much mobility the person can get out of a bicycle.

52There are several different types of conversion factors, and the conversion factors discussed are often categorized into three groups (Robeyns 2005b, 99; Crocker and Robeyns 2009, 68). All conversion factors influence how a person can be or is free to convert the characteristics of the resources into a functioning, yet the sources of these factors may differ. Personal conversion factors are internal to the person, such as metabolism, physical condition, sex, reading skills, or intelligence. If a person is disabled, or if she is in a bad physical condition, or has never learned to cycle, then the bike will be of limited help in enabling the functioning of mobility. Social conversion factors are factors stemming from the society in which one lives, such as public policies, social norms, practices that unfairly discriminate, societal hierarchies, or power relations related to class, gender, race, or caste. Environmental conversion factors emerge from the physical or built environment in which a person lives. Among aspects of one’s geographical location are climate, pollution, the likelihood of earthquakes, and the presence or absence of seas and oceans. Among aspects of the built environment are the stability of buildings, roads, and bridges, and the means of transportation and communication. Take again the example of the bicycle. How much a bicycle contributes to a person’s mobility depends on that person’s physical condition (a personal conversion factor), the social mores including whether women are generally allowed to ride a bicycle (a social conversion factor), and the availability of decent roads or bike paths (an environmental conversion factor). Once we start to be aware of the existence of conversion factors, it becomes clear that they are a very pervasive phenomenon. For example, a pregnant or lactating woman needs more of the same food than another woman in order to be well-nourished. Or people living in delta regions need protection from flooding if they want to enjoy the same capability of being safely sheltered as people living in the mountains. There are an infinite number of other examples illustrating the importance of conversion factors. The three types of conversion factor all push us to acknowledge that it is not sufficient to know the resources a person owns or can use in order to be able to assess the wellbeing that he or she has achieved or could achieve; rather, we need to know much more about the person and the circumstances in which he or she is living. Differences in conversion factors are one important source of human diversity, which is a central concern in the capability approach, and will be discussed in more detail in section 3.5.

53Note that many conversion factors are not fixed or given, but can be altered by policies and choices that we make. And the effects of a particular conversion factor can also depend on the social and personal resources that a person has, as well as on the other conversion factors. For example, having a physical impairment that doesn’t allow one to walk severely restricts one’s capability to be mobile if one finds oneself in a situation in which one doesn’t have access to a wheelchair, and in which the state of the roads is bad and vehicles used for public transport are not wheelchair-accessible. But suppose now the built environment is different: all walking-impaired people have a right to a wheelchair, roads are wheelchair-friendly, public transport is wheelchair-accessible and society is characterised by a set of social norms whereby people consider it nothing but self-evident to provide help to fellow travellers who can’t walk. In such an alternative social state, with a different set of resources and social and environmental conversion factors, the same personal conversion factor (not being able to walk) plays out very differently. In sum, in order to know what people are able to do and be, we need to analyse the full picture of their resources, and the various conversion factors, or else analyse the functionings and capabilities directly. The advantage of having a clear picture of the resources needed, and the particular conversion factors needed, is that it also gives those aiming to expand capability sets information on where interventions can be made.

2.6.4 A4: The means-ends distinction

54The fourth core characteristic of the capability approach is the meansends distinction. The approach stresses that we should always be clear, when valuing something, whether we value it as an end in itself, or as a means to a valuable end. For the capability approach, when considering interpersonal comparisons of advantage, the ultimate ends are people’s valuable capabilities (there could be other ends as well; see 2.6.6). This implies that the capability approach requires us to evaluate policies and other changes according to their impact on people’s capabilities as well as their actual functionings; yet at the same time we need to ask whether the preconditions — the means and the enabling circumstances — for those capabilities are in place. We must ask whether people are able to be healthy, and whether the means or resources necessary for this capability, such as clean water, adequate sanitation, access to doctors, protection from infections and diseases and basic knowledge on health issues are present. We must ask whether people are well-nourished, and whether the means or conditions for the realization of this capability, such as having sufficient food supplies and food entitlements, are being met. We must ask whether people have access to a high-quality education system, to real political participation, and to community activities that support them, that enable them to cope with struggles in daily life, and that foster caring friendships. Hence we do need to take the means into account, but we can only do so if we first know what the ends are.

  • 24 This is a critique that the capability approach shares with the happiness approach, which also focu (...)

55Many of the arguments that capability theorists have advanced against alternative normative frameworks can be traced back to the objection that alternative approaches focus on particular means to wellbeing rather than the ends.24 There are two important reasons why the capability approach dictates that we have to start our analysis from the ends rather than the means. Firstly, people differ in their ability to convert means into valuable opportunities (capabilities) or outcomes (functionings) (Sen 1992a, 26–28, 36–38). Since ends are what ultimately matter when thinking about wellbeing and the quality of life, means can only work as fully reliable proxies of people’s opportunities to achieve those ends if all people have the same capacities or powers to convert those means into equal capability sets. This is an assumption that goes against a core characteristic of the capability approach, namely claim A3 — the inter-individual differences in the conversion of resources into functionings and capabilities. Capability scholars believe that these inter-individual differences are far-reaching and significant, and hence this also explains why the idea of conversion factors is a compulsory option in the capability approach (see 2.6.3). Theories that focus on means run the risk of downplaying the normative relevance of not only these conversion factors, but also the differences in structural constraints that people face (see 2.7.5).

56The second reason why the capability approach requires us to start from ends rather than means is that there are some vitally important ends that do not depend very much on material means, and hence would not be picked up in our analysis if we were to focus on means only. For example, self-respect, supportive relationships in school or in the workplace, or friendship are all very important ends that people may want; yet there are no crucial means to those ends that one could use as a readily measurable proxy. We need to focus on ends directly if we want to capture what is important.

  • 25 However, while a focus on available options rather than realised choices is the default normative f (...)
  • 26 On the distinction between general capabilities and specific capabilities, see 3.2.4.

57One could argue, however, that the capability approach does not focus entirely on ends, but rather on the question of whether a person is being put in the conditions in which she can pursue her ultimate ends. For example, being able to read could be seen as a means rather than an end in itself, since people’s ultimate ends will be more specific, such as reading street signs, the newspaper, or the Bible or Koran. It is therefore somewhat more precise to say that the capability approach focuses on people’s ends in terms of beings and doings expressed in general terms: being literate, being mobile, being able to hold a decent job. Whether a particular person then decides to translate these general capabilities into the more specific capabilities A, B or C (e.g. reading street signs, reading the newspaper or reading the Bible) is up to them. Whether that person decides to stay put, travel to the US or rather to China, is in principle not important for a capability analysis: the question is rather whether a person has these capabilities in more general terms.25 Another way of framing this is to say that the end of policy making and institutional design is to provide people with general capabilities, whereas the ends of persons are more specific capabilities.26

  • 27 The relationship between means and capabilities is analysed in more depth in section 3.12.

58Of course, the normative focus on ends does not imply that the capability approach does not at all value means such as material or financial resources. Instead, a capability analysis will typically focus on resources and other means. For example, in their evaluation of development in India, Jean Drèze and Amartya Sen (2002, 3) have stressed that working within the capability approach in no way excludes the integration of an analysis of resources such as food. In sum, all the means of wellbeing, like the availability of commodities, legal entitlements to them, other social institutions, and so forth, are important, but the capability approach presses the point that they are not the ends of wellbeing, only their means. Food may be abundant in the village, but a starving person may have nothing to exchange for it, no legal claim on it, or no way of preventing intestinal parasites from consuming it before he or she does. In all these cases, at least some resources will be available, but that person will remain hungry and, after a while, undernourished.27

59Nevertheless, one could wonder: wouldn’t it be better to focus on means only, rather than making the normative analysis more complicated and more informationally demanding by also focusing on functionings and capabilities? Capability scholars would respond that starting a normative analysis from the ends rather than means has at least two advantages, in addition to the fundamental reason mentioned earlier that a focus on ends is needed to appropriately capture interindividual differences.

60First, if we start from being explicit about our ends, the valuation of means will retain the status of an instrumental valuation rather than risk taking on the nature of a valuation of ends. For example, money or economic growth will not be valued for their own sake, but only in so far as they contribute to an expansion of people’s capabilities. For those who have been working within the capability framework, this has become a deeply ingrained practice — but one only needs to read the newspapers for a few days to see how often policies are justified or discussed without a clear distinction being made between means and ends.

61Second, by starting from ends, we do not a priori assume that there is only one overriding important means to those ends (such as income), but rather explicitly ask the question: which types of means are important for the fostering and nurturing of a particular capability, or set of capabilities? For some capabilities, the most important means will indeed be financial resources and economic production, but for others it may be a change in political practices and institutions, such as effective guarantees and protections of freedom of thought, political participation, social or cultural practices, social structures, social institutions, public goods, social norms, and traditions and habits. As a consequence, an effective capability-enhancing policy may not be increasing disposable income, but rather fighting a homophobic, ethnophobic, racist or sexist social climate.

2.6.5 A5: Functionings and capabilities as the evaluative space

  • 28 I am using the term ‘normative’ here in the way it is used by social scientists, hence encompassing (...)

62If a capability theory is a normative theory (as is often the case), then functionings and capabilities form the entire evaluative space, or are part of the evaluative space.28 A normative theory is a theory that entails a value judgement: something is better than or worse than something else. This value judgement can be used to compare the position of different persons or states of affairs (as in inequality analysis) or it can be used to judge one course of action as ‘better’ than another course of action (as in policy design). For all these types of normative theories, we need normative claims, since concepts alone cannot ground normativity.

  • 29 Moreover, further extensions of this proposition may be needed. One issue is that we should not onl (...)

63The first normative claim which each capability theory should respect is thus that functionings and capabilities form the ‘evaluative space’. According to the capability approach, the ends of wellbeing freedom, justice, and development should be conceptualized in terms of people’s functionings and/or capabilities. This claim is not contested among scholars of the capability approach; for example, Sabina Alkire (2005, 122) described the capability approach as the proposition “that social arrangements should be evaluated according to the extent of freedom people have to promote or achieve functionings they value”. However, if we fully take into account that functionings can be positive but also negative (see 2.6.2), we should also acknowledge that our lives are better if they contain fewer of the functionings that are negative, such as physical violence or stress. Alkire’s proposition should therefore minimally be extended by adding “and to promote the weakening of those functionings that have a negative value”.29

64However, what is relevant is not only which opportunities are open to us individually, hence in a piecemeal way, but rather which combinations or sets of potential functionings are open to us. For example, suppose you are a low-skilled poor single parent who lives in a society without decent social provisions. Take the following functionings: (1) to hold a job, which will require you to spend many hours on working and commuting, but will generate the income needed to properly feed yourself and your family; (2) to care for your children at home and give them all the attention, care and supervision they need. In a piecemeal analysis, both (1) and (2) are opportunities open to that parent, but they are not both together open to her. The point about the capability approach is precisely that it is comprehensive; we must ask which sets of capabilities are open to us, that is: can you simultaneously provide for your family and properly care for and supervise your children? Or are you rather forced to make some hard, perhaps even tragic choices between two functionings which are both central and valuable?

65Note that while most types of capability analysis require interpersonal comparisons, one could also use the capability approach to evaluate the wellbeing or wellbeing freedom of one person at one point in time (e.g. evaluate her situation against a capability yardstick) or to evaluate the changes in her wellbeing or wellbeing freedom over time. The capability approach could thus also be used by a single person in her deliberate decision-making or evaluation processes, but these uses of the capability approach are much less prevalent in the scholarly literature. Yet all these normative exercises share the property that they use functionings and capabilities as the evaluative space — the space in which personal evaluations or interpersonal comparisons are made.

2.6.6 A6: Other dimensions of ultimate value

  • 30 For example, if Henry Richardson (2007) is right in arguing that the idea of capabilities cannot ca (...)
  • 31 This distinction, and its relevance, will be discussed in more detail in section 3.3.

66However, this brings us straight to another core property of module A, namely that functionings and/or capabilities are not necessarily the only elements of ultimate value. Capabilitarian theories might endorse functionings and/or capabilities as their account of ultimate value but may add other elements of ultimate value, such as procedural fairness. Other factors may also matter normatively, and in most capability theories these other principles or objects of evaluation will play a role. This implies that the capability approach is, in itself, incomplete as an account of the good since it may have to be supplemented with other values or principles.30 Sen has been a strong defender of this claim, for example, in his argument that capabilities capture the opportunity aspect of freedom but not the process aspect of freedom, which is also important (e.g. Sen 2002a, 583–622).31

67At this point, it may be useful to reflect on a suggestion made by Henry Richardson (2015) to drop the use of the word ‘intrinsic’ when describing the value of functionings and capabilities — as is often done in the capability literature. For non-philosophers, saying that something has ‘intrinsic value’ is a way to say that something is much more important than something else, or it is used to say that we don’t need to investigate what the effects of this object are on another object. If we think that something doesn’t have intrinsic value, we would hold that it is desirable if it expands functionings and capabilities; economic growth is a prominent example in both the capability literature and in the human development literature. Yet in philosophy, there is a longstanding debate about what it means to say of something that it has intrinsic value, and it has increasingly been contested that it is helpful to speak of ‘intrinsic values’ given what philosophers generally would like to say when they use that word (Kagan 1998; Rabinowicz and Rønnow-Rasmussen 2000).

  • 32 “[…] the relevant values can be said to be ‘end-point values’, insofar as they are not simply condu (...)

68In philosophy, the term ‘intrinsic’ refers to a metaphysical claim; something we claim to be intrinsically valuable only derives its value from some internal properties. Yet in the capability approach, this is not really what we want to say about functionings (or capabilities). Rather, as Richardson rightly argues, we should be thinking about what we take to be worth seeking for its own sake. Richardson prefers to call this ‘thinking in terms of final ends’; in addition, one could also use the terminology ‘that which has ultimate value’ (see also Rabinowicz and Rønnow-Rasmussen 2000, 48).32 This has the advantage that we do not need to drop the widely used, and in my view very useful, distinction between instrumental value and ultimate value. Those things that have ultimate value are the things we seek because they are an end (of policy making, decision making, evaluations); those things that do not have ultimate value, hence that are not ends, will be valued to the extent that they have instrumental value for those ends.

69Of course, non-philosophers may object and argue they are using ‘intrinsic value’ and ‘ultimate value’ as synonyms. But if we want to develop the capability approach in a way that draws on the insights from all disciplines, we should try to accommodate this insight from philosophy into the interdisciplinary language of the capability approach, especially if there is a very easy-to-use alternative available to us. We can either, as Richardson proposes, speak of the selected functionings and capabilities as final ends, or we can say that the selected functionings and capabilities have ultimate value — that is, they have value as ends in themselves and not because they are useful for some further end. It is of course possible for a capability to have ultimate value and for the corresponding functioning to have instrumental value. For example, being knowledgeable and educated can very plausibly be seen as of ultimate value, but is also of instrumental value for various other capabilities, such as the capability of being healthy, being able to pursue projects, being able to hold a job, and so forth.

70However, the question is whether it is possible to change the use of a term that is so widespread in some disciplines yet regarded as wrong from the point of view of another discipline. It may be that the best we can hope for is to become aware of the different usages of the term ‘intrinsic value’, which in the social sciences is used in a much looser way than in philosophy, and doesn’t have the metaphysical implications that philosophers attribute to it.

2.6.7 A7: Value pluralism

71There are at least two types of value pluralism within the capability approach. One type is the other objects of ultimate value, which was briefly addressed in the previous section. This is what Sen called in his Dewey lectures principle pluralism (Sen 1985c, 176). Expanding capabilities and functionings is not all that matters; there are other moral principles and goals with ultimate value that are also important when evaluating social states, or when deciding what we ought to do (whether as individuals or policy makers). Examples are deontic norms and principles that apply to the processes that lead to the expansion of capability sets. This value pluralism plays a very important role in understanding the need to have the C-module C4, which will be discussed in section 2.8.4.

72It is interesting to note that at some stage in Sen’s development of the capability approach, his readers lost this principle-pluralism and thought that the capability approach could stand on its own. But a reading of Sen’s earlier work on the capability approach shows that all along, Sen felt that capabilities can and need to be supplemented with other principles and values. For example, in his 1982 article ‘Rights and Agency’, Sen argues that “goal rights, including capability rights, and other goals, can be combined with deontological values […], along with other agent-relative considerations, in an integrated system” (Sen 1982, 4). Luckily, the more recent publications in the secondary literature on the capability approach increasingly acknowledge this principle pluralism; the modules A7 and C4 of the modular view presented in this book suggest that it is no longer possible not to acknowledge this possibility.

73The second type of value-pluralism relates to what is often called the multidimensional nature of the capability approach. Functionings and capabilities are not ‘values’ in the sense of ‘public values’ (justice, efficiency, solidarity, ecological sustainability, etc.) but they are objects of ultimate value — things that we value as ends in themselves. Given some very minimal assumptions about human nature, it is obvious that these dimensions are multiple: human beings value the opportunity to be in good health, to engage in social interactions, to have meaningful activities, to be sheltered and safe, not be subjected to excessive levels of stress, and so forth. Of course, it is logically conceivable to say that for a particular normative exercise, we only look at one dimension. But while it may be consistent and logical, it nevertheless makes no sense — for at least two reasons.

74First, the very reason why the capability approach has been offered as an alternative to other normative approaches is to add informational riches — to show which dimensions have been left out of the other types of analysis, and why adding them matters. It also makes many evaluations much more nuanced, allowing them to reflect the complexities of life as it is. For example, an African-American lawyer may be successful in her professional life in terms of her professional achievements and the material rewards she receives for her work, but she may also encounter disrespect and humiliation in a society that is sexist and racist. Being materially well-off doesn’t mean that one is living a life with all the capabilities to which one should be entitled in a just society. Only multidimensional metrics of evaluation can capture those ambiguities and informational riches.

75Second, without value pluralism, it would follow that the happiness approach is a special case of the capability approach — namely a capability theory in which only one functioning matters, namely being happy. Again, while this is strictly speaking a consistent and logical possibility, it makes no sense given that the capability approach was conceived to form an alternative to both the income metric and other resourcist approaches on the one hand, and the happiness approach and other mental metric approaches on the other. Thus, in order to make the capability approach a genuine alternative to other approaches, we need to acknowledge several functionings and capabilities, rather than just one.

2.6.8 A8: The principle of each person as an end

76A final core property of each capability theory or application is that each person counts as a moral equal. Martha Nussbaum calls this principle “the principle of each person as an end”. Throughout her work she has offered strong arguments in defence of this principle (Nussbaum 2000, 56):

The account we strive for [i.e. the capability approach] should preserve liberties and opportunities for each and every person, taken one by one, respecting each of them as an end, rather than simply as the agent or supporter of the ends of others. […] We need only notice that there is a type of focus on the individual person as such that requires no particular metaphysical position, and no bias against love or care. It arises naturally from the recognition that each person has just one life to live, not more than one. […] If we combine this observation with the thought […] that each person is valuable and worthy of respect as an end, we must conclude that we should look not just to the total or the average, but to the functioning of each and every person.

77Nussbaum’s principle of each person as an end is the same as what is also known as ethical or normative individualism in debates in philosophy of science. Ethical individualism, or normative individualism, makes a claim about who or what should count in our evaluative exercises and decisions. It postulates that individual persons, and only individual persons are the units of ultimate moral concern. In other words, when evaluating different social arrangements, we are only interested in the (direct and indirect) effects of those arrangements on individuals.

78As will be explained in more detail in section 4.6, the idea of ethical individualism is often conflated with other notions of individualism, such as the ontological idea that human beings are individuals who can live and flourish independently of others. However, there is no such claim in the principle of ethical individualism. The claim is rather one about whose interests should count. And ethical individualism claims that only the interests of persons should count. Ultimately, we care about each individual person. Ethical individualism forces us to make sure we ask questions about how the interests of each and every person are served or protected, rather than assuming that because, for example, all the other family members are doing fine, the daughter-in-law will be doing fine too. If, as all defensible moral theories do, we argue that every human being has equal moral worth, then we must attach value to the interests of each and every one of the affected persons. Thus, my first conclusion is that ethical individualism is a desirable property, since it is necessary to treat people as moral equals.

  • 33 As Richardson (2016, 5) puts it, “all capabilities […] are dependent on the body. Without relying o (...)
  • 34 Some have argued in favour of what they call ‘collective capabilities’, which I will discuss in sec (...)
  • 35 However, the question remains whether the capability approach is fully compatible with indigenous w (...)

79But ethical individualism is not only a desirable property, it is also an unavoidable property. By its very nature the evaluation of functionings and capabilities is an evaluation of the wellbeing and freedom to achieve wellbeing of individual persons. Functionings are ‘beings’ and ‘doings’: these are dimensions of a human being, which is an embodied being, not merely a mind or a soul. And with the exception of the conjoined twins, and the case of the unborn child and the pregnant mother, bodies are physically separated from each other.33 We are born as a human being with a body and future of her own, and we will die as a human being with a body and a past life narrative that is unique. This human being, that lives her life in an embodied way, thus has functionings that are related to her person, which is embodied. It is with the functionings and capabilities of these persons that the capability approach is concerned with.34 However, as I will explain in detail in section 4.6, from this it does not follow that the capability approach conceptualises people in an atomistic fashion, and thus that the capability approach is ‘individualistic’ — meant in a negative, pejorative way. And it also does not imply that a capabilitarian evaluation could not also evaluate the means (including social institutions, structures, and norms) as well as conversion factors, as well as non-capabilitarian elements of value — as long as we are clear what the role or status of each of those elements is.35

80Note that the use of the term ‘normative individualism’ is deeply disputed. Some scholars see no problem at all in using that term, since they use it in a technical sense that they believe should not be conflated with any pejorative use of the term ‘individualism’ in daily life. Other scholars resist the term ‘ethical individualism’, since they cannot separate it from (a) the notions of ontological and explanatory individualism, and/or (b) from the pejorative meaning that the term ‘individualism’ has in daily life, which is probably close to a term such as ‘egoism’. While the first group is, in my view, right, the second group conveys important information about how the capability approach will be perceived in a broader setting, including outside academia. It may therefore be recommendable to replace the term ‘ethical or normative individualism’ with the term ‘the principle of each person as an end’ whenever possible.

2.7 The B-modules: non-optional modules with optional content

81I believe that the best way to understand the capability approach is by taking the content of the A-module as non-optional. All capability theories need to endorse the content of the A-module (ideally in an explicit way) or at a very minimum should not have properties that violate the content of the A-module. But there are also properties of a capability theory where the module is non-optional, yet there is choice involved in the content of the module. This doesn’t mean that ‘anything goes’ in terms of the choice of the content, but it does mean that within each module, there is a range of options to choose from. These are the B-modules, each of which contains a range of possible content, from which the capability theorist can decide what content to adopt. However, the range of content of the B-modules must not contradict the A-module. The following table lists the non-optional B-modules with optional content.

Table 2.3. The B-modules: non-optional modules with optional content


The purpose of the capability theory


The selection of dimensions


An account of human diversity


An account of agency


An account of structural constraints


The choice between functionings, capabilities, or both


Meta-theoretical commitments

2.7.1 B1: The purpose of the capability theory

82The first module, which is itself non-optional, but where the content can be chosen, is the purpose of the theory. For example, one could use the capability approach to construct a theory of justice, to develop an international empirical comparison, to reform an educational curriculum, to develop alternative welfare economics, or to evaluate the effects of laws on people’s capabilities. Questions of scope and reach also need to be addressed in this module. For example, is a theory of justice a political or a comprehensive theory? Is such a theory domestic or global? Other questions that need to be addressed involve the intended audience. Is one constructing an academic theory where great attention is given to detail and even the smallest distinctions are taken as relevant, or is one addressing policy makers or societal organisations for whom every detail does not matter and the time to think and read may be much more constrained, while the accessibility of the ideas is much more important?

83Of course, one could argue that B1 is not specific to capability theories, and also holds for, say, deontological theories, or utilitarian theories, or theories that use care ethics as their basic normative foundation.

84While that is true, there are two reasons to highlight B1, the purpose of the capability theory, in the account that I am developing. The first is that it will help us to be explicit about the purpose. There are plenty of pieces published in the capability literature in which the purpose of the application or theory is not made explicit, and as a consequence it leads to people based in different disciplines or fields talking alongside each other. Second, it seems that the need to be explicit about the purpose (including the audience) of one’s capability theory or application is stronger in the capability literature than in other approaches, because in comparison to those other approaches it has a much more radically multidisciplinary uptake.

2.7.2 B2: The selection of dimensions

85The second B-module is the selection of capabilities and/or functionings. We need to specify which capabilities matter for our particular capability theory. This is a deeply normative question, and touches the core of the difference that the capability approach can make. After all, the dimensions that one selects to analyse will determine what we will observe — and also, equally importantly, what we will not observe since the dimensions are not selected.

86There is, by now, a large body of literature discussing the various ways in which one can make that selection, including some overview articles that survey the different methods for particular purposes (e.g. Alkire 2002; Robeyns 2005a; Byskov forthcoming). These methods explicitly include various participatory, deliberative and/or democratic approaches, which are widely used in capability applications.

87There are two crucial factors determining which selection procedure is suitable. The first is the purpose of the capability theory (hence the choice made in B1). If we develop an account of wellbeing for thinking about how our lives are going, we are not constrained by questions such as the legitimate scope of government intervention, whereas a theory of political justice would need to take that element into account. Another example is if we would like to use the capability approach to think about what is universally demanded by moral principles, hence to develop the capability approach into a theory of morality: there the selection may be constrained by a method of moral justification for categorically binding principles, which is much more demanding than a method that justifies principles we offer to each other as rationally defensible proposals in the public realm. At the empirical and policy level, similar questions arise. For example, one could take the international human rights treaties as reflecting a given political consensus, and use those to select capabilities (Vizard 2007). Or, one’s main goal may be to analyse what difference the capability approach makes for poverty or inequality analysis in comparison with income metrics, in which case one may opt for a method that makes the normativity explicit but nevertheless stays close to existing practices in the social sciences, assuming the epistemic validity of those practices (Robeyns 2003).

88The second factor determining which selection procedure is suitable is the set of constraints one takes as given in the normative analysis one is making. In an ideal world, there would always be cooperation between scholars with different disciplinary expertise, who would understand each other well, and who would be able to speak the language of the other disciplines involved in developing the capability theory. In an ideal world, there would also be no time constraints on the amount of time one has to develop a capability theory, and no financial constraints on the data gathering, or social, psychological or political constraints on the types of question one can ask when conducting a survey. One would be able always to conduct one’s own fieldwork if one wanted, one would have access to all the empirical knowledge one needed, and one would not be constrained in gathering the information one wanted to gather. Clearly, the methods for such an ideal world would be very different from the methods that are used in practice — where database-driven selection may be the best one can do.

89Still, whichever method one uses, what always remains important, and very much in the spirit of the capability approach, is not to act in a mechanical way, or to see the question of the selection of dimensions as a technocratic exercise. Even if one cannot, for example, collect certain data, one could nevertheless still mention the dimensions that one would have wanted to include if it had been possible, and perhaps provide some reasonable informed guess of what difference the inclusion of that dimension would have made.

2.7.3 B3: Human diversity

  • 36 For an introduction to libertarianism, see Vallentyne and Van der Vossen (2014).

90Within the capability approach human diversity is a core characteristic, and indeed a core motivation for developing the capability approach in the first place. Yet the account of human diversity that one endorses can differ. For example, scholars with a background in structuralist sociology or Marxism often believe that the social class to which one belongs is a very important factor in human diversity, which has great influence on which options lie open to a person, but also on how a person’s character and aspirations are formed. For those scholars, class interacts with, and in some cases even outweighs, all other identity aspects. For others, such as libertarians, these differences are not so important.36 They would not attach much (normative or explanatory) importance to one’s gender, ethnicity, race, social class, and so forth: everyone is, first and foremost, an individual whose personal ambitions and projects matter. Yet, whether one is a Marxist or a libertarian or one of the many other positions one can take, one always, either implicitly or explicitly, endorses a view on human nature and on human diversity. That choice should be made in capability theories, since the capability approach rejects the use of an implicit, unacknowledged account of human diversity. Hence such an account belongs to the B-modules: one has to have an account of human diversity, but, as long as one is willing to defend one’s account and it survives critical analysis, there are several accounts that one can opt for.

91Note that if one puts all the modules A, B and C together, a picture will emerge about the great importance attached to human diversity in the capability approach; this will be analysed in more detail in section 3.5.

2.7.4 B4: Agency

92Another B-module is the acknowledgement of agency. As a working definition, we can use Sen’s definition of an agent as “someone who acts and brings about change, and whose achievements can be judged in terms of her own values and objectives, whether or not we assess them in terms of some external criteria as well” (Sen 1999a, 19).

93Applications of the capability approach should endorse some account of agency, except if there are good reasons why agency should be taken to be absent, or why in a particular capability application agency is simply not relevant (for example, when one wants to investigate the correlation between an income metric and some achieved functionings).

  • 37 Martha Nussbaum explicitly refrains from integrating the notion of ‘agency’ in her capability theor (...)

94But clearly, as with other key ethical concepts such as ‘wellbeing’ or ‘freedom’, the concept of ‘agency’ can be fleshed out in many different ways. The capability approach is not committed to one particular account of agency. Similar to the acknowledgement of structural constraints, there is no agreed-upon or standard claim about how much agency, or what particular type, should be assumed; the claim is minimalistic in the sense that, as with the structural constraints which will be discussed in the next section, agency cannot simply be ignored and must be accounted for. One can give agency a key role in a capability theory (e.g. Crocker 2008; Claassen 2016) or a more restricted role, perhaps also using different terminology. One can also develop the account of agency by spelling out some of its preconditions, which may include capabilities. For example, Tom de Herdt (2008) analysed the capability of not having to be subjected to public shame as a precondition of agency, and showed how this may be relevant for social policymaking by illustrating its importance in a food relief programme in Kinshasa. For empirical scholars and policy scholars, an empirically sound account of agency will be crucial; for moral philosophers, a more theoretical account of what conceptualisation of agency is morally relevant will be needed. Thus, the precise content of this B-module will differ significantly between different capability theories and applications — but, in all cases, some acknowledgement of agency will be needed.37

2.7.5 B5: Structural constraints

95The fifth B-module is the account of structural constraints: the institutions, policies, laws, social norms and so forth, that people in different social positions face. Those differences in the structural constraints that people face can have a great influence on their conversion factors, and hence on their capability sets. For example, if relationships between people of the same sex are criminalised, then gay people may have all the means and resources they would wish, but they will still not be able to enjoy a happy family life. Or if people of colour face explicit or implicit discrimination on the labour market, then they will not be able to use the same labour-market resources (their degrees, training, experience) to generate the same levels of capabilities in the professional sphere of life, compared with groups that face no (or less) discrimination.

96In addition, structural constraints also play a role in the shaping of people’s capabilities that are not heavily dependent on material resources. If one group of people is, for cultural, historical or religious reasons, stigmatized as outcasts, then they will be treated with disrespect by other groups in society. The same holds for all groups that suffer from stigma, such as, for example, people with psychiatric disorders or other mental health issues. These structural constraints will also affect the capabilities that do not rely on resources directly, such as opportunities for friendships or for a healthy sense of self-confidence.

97Which of those structural constraints will be important for a particular capability analysis will depend on the context. For example, in her study of the living standards of waste pickers, scavengers, and plastic recycling and scrap trading entrepreneurs in Delhi, Kaveri Gill (2010) showed that caste plays a very important role in the capability sets of different castes. For example, those at the very bottom of the hierarchical ladder of waste workers — the waste pickers — have no opportunities for upward mobility due to social norms and societal discrimination related to their caste. In this study, social norms related to caste are key as a structural constraint; in other studies, it may be the anatomy of twenty-first century capitalism, or gender norms in gender-stratified societies, or some other set of structural constraints.

98In sum, structural constraints can have a very important role in shaping people’s capability sets, and therefore have to be part of capability theories. Structural constraints vary depending on one’s caste, class, ethnicity, age, gender, sexual orientation, (dis) abilities, and the economic system in which one lives. These structural constraints are very likely to have an influence on a person’s capability set (and in most cases also do have that influence). Having an account of structural constraints is therefore non-optional: every capability theory has one, although sometimes this account will be very implicit. For example, I will argue in section 4.10 that part of the critique of mainstream welfare economics is that it has a very weak or minimal account of structural constraints. Heterodox welfare economists who are embracing the capability approach are not only doing so because they think the endorsement of the capability account of wellbeing is better than the preferences-based accounts that are dominant in mainstream economics, but often also because they hope that the minimal account of structural constraints in welfare economics can be replaced by a richer account that is better informed by insights from the other social sciences and from the humanities.

2.7.6 B6: The choice between functionings, capabilities, or both

99In developing a capability theory, we need to decide whether we think that what matters are capabilities, functionings, or a combination of both. The core proposition that functionings and capabilities form the evaluative space (A5), was not decisive regarding the question of whether it is only functionings, or only capabilities, or a mixture of both, that form this space.

100There are various arguments given in the literature defending a range of views that only capabilities matter; or that primarily secured functionings matter; or that for particular capability theories it is impossible only to focus on capabilities; or that we sometimes have good reasons to focus on functionings. These various claims and arguments will be reviewed in section 3.4; as will be argued in that section, there are good reasons why people could reasonably disagree on whether the capability analysis they are conducting should focus on functionings or capabilities or a mixture. It follows that a choice must be made, but that there are various options to choose from.

2.7.7 B7: Meta-theoretical commitments

101Finally, each capability theory will embrace some meta-theoretical commitments. Yet often, these meta-theoretical commitments are shared commitments within one’s discipline or one’s school within that discipline, and as a graduate student one has become socialised in accepting these meta-theoretical commitments as given. As a consequence, it often happens that scholars are not even aware that there are such things as meta-theoretical commitments. For example, if one wants to conduct a measurement exercise (a choice made in the module B1) then one may be committed to the methodological principle of parsimony (to build a model with as few assumptions and as elegantly as possible) or, instead, to providing a measurement that is embedded into a rich narrative description aimed at a better understanding. Or, if one wants to construct a theory of justice (again, a choice made in the module B1), then one may aim for an ideal or non-ideal theory of justice, or for a partial or a comprehensive account of justice. Or one may espouse certain views about the status of theories of justice or meta-ethical claims related to, for example, the role that intuitions are permitted to play as a source of normativity. Some debates within the capability approach, but also between capability scholars and those working in other paradigms, would be truly enlightened if we made the meta-theoretical commitments of our theories, accounts and applications more explicit.

2.8 The C-modules: contingent modules

102In addition to the compulsory content of the core A-module, and the optional content of the non-optional B-modules, a capability theory could also add a third type of module, which I will call the contingent modules. These are either modules that need to be taken on board due to some choices that have been made in a B-module, or else they are entirely optional, independent of what one has chosen in the B-modules. The following table gives an overview of the contingent modules.

Table 2.4. The C-modules: contingent modules


Additional ontological and explanatory theories


Weighing dimensions


Methods for empirical analysis


Additional normative principles and concerns

2.8.1 C1: Additional ontological and explanatory theories

103Two capabilitarian thinkers could each aspire to make a theory of justice, yet embrace very different views on human nature and on the degree to which certain outcomes can be explained solely by people’s choices or are also affected by structural constraints. This can matter a lot for the particular capability theories that one develops.

104For example, in earlier work, I showed that the capability approach’s answer to whether there is anything wrong with the traditional gender division of labour depends a lot on the social ontological claims related to gender that are (implicitly) endorsed as well as the explanatory views of how that division of labour came about (Robeyns 2008c). If one believes that the fact that women end up doing most of the unpaid and care work, while men end up doing most of the paid labour market work, is a result of differences in talents, dispositions and preferences, then one would judge that the different functionings outcomes that result for men and women within households provide them with maximal levels of wellbeing given the formal institutional background that they face. But if one endorses a feminist explanation of this division of labour between men and women within households, then one is likely to stress power differences, the role of societal expectations and social norms in decision making, and so forth (e.g. Okin 1989; Folbre 1994). The same observed functionings outcomes in households with a traditional gender division of labour would then be evaluated differently.

105Similarly, Miriam Teschl and Laurent Derobert argue that a range of different accounts of social and personal identity are possible, and this may also impact on how we interpret a person forfeiting a capability that we would all deem valuable (Teschl and Derobert 2008). If we believe that our religious identities are a matter of rational deliberation and decisionmaking, then we will judge the choice to physically self-harm because of one’s religion differently than if we have an account of identity where there is much less scope for choice and rational deliberation regarding our religious affiliation or other group memberships.

106In short, different ontological and explanatory options are available in module C1, and they may have effects on various other elements or dimensions of the capability theory that are being constructed. However, we should be careful and not mistakenly conclude that ‘anything goes’ when we add additional ontological theories, since there should not be any conflicts with the propositions of the A-module — and, in addition, some ontological and explanatory accounts are much better supported by critical analysis and empirical knowledge.

2.8.2 C2: Weighing dimensions

107For some capability theories, the prioritising, weighing or aggregating of dimensions (functionings and capabilities) may not be needed. For example, one may simply want to describe how a country has developed over time in terms of a number of important functionings, as a way of giving information about the evolution of the quality of life that may give different insights than the evolution of GDP (e.g. Van Zanden et al. 2014). Weighing dimensions is therefore not required for each capability theory or capability application, in contrast to the selection of dimensions, which is inevitable.

108However, for some other choices that one can make in B1, the capabilitarian scholar or practitioner needs to make choices related to the weighing of the different dimensions. If that is the case, then there are different methods for how one could weigh. When considering which weighing method to use, the same factors are relevant as in the case of selecting the dimensions: the purposes of one’s capability theory, and the constraints one has to take into account when choosing a method.

  • 38 In the case of democratic theory, the discussion is often about which laws to implement, but the sa (...)

109In contrast to the overview works that have been written on how to select dimensions (e.g. Alkire 2002; Robeyns 2005a; Byskov forthcoming), capability scholars have written much less about which methods one could use to decide on the weights given to each dimension, specifically focussing on functionings or capabilities as the dimensions. What lessons and insights can we learn from what has so far been argued in this literature on the weighing of dimensions? (Alkire 2016; Alkire et al. 2015, chapter 6; Robeyns 2006b, 356–58) First, the selection of weights for the capability approach is structurally similar to other multidimensional metrics (in the case of evaluations) or decision-making procedures (in case one needs to decide to which capabilities to give priority in policies or collective decision making). Hence one should consult existing discussions in other debates where multidimensionality plays an important role. Let us first look at the group of applications in which the capability approach is used to make decisions about what we, collectively, ought to do. That may be in an organisation; or at the level of a community that needs to decide whether to spend tax revenues on investing more in public green spaces, or in social services for particular groups, or in taking measures to prevent crime, or in anything else that can likely be understood as leading to positive effects on our capabilities. In those cases, we can learn from social choice theory, and from theories of democratic decision making, how we could proceed.38 Decisions could be made by voting, or by deliberation, or by deliberation and/or voting among those who are the representatives of the relevant population.

110Second, the applications of the capability approach that involve a multidimensional metric of wellbeing or wellbeing freedom could use (most of) the weighing methods that have been discussed for multidimensional metrics in general. Koen Decancq and María Ana Lugo (2013) have reviewed eight different approaches to set weights for multidimensional metrics, which they categorize in three classes: data-driven weights in which the weights are a function of the distribution of the various dimensions in the population surveyed; normative approaches in which either experts decide on the weights, or the weights are equal or arbitrary; and hybrid weights that are in part data-driven but in addition depend on some normative decision. Note that in the data-driven and hybrid approaches, the selection of dimensions and the weights tends to be done through a process in which the selection of dimensions and the determination of the weights go hand in hand. One example is the proposal by Erik Schokkaert (2007) of using happiness as the master-value by which we weigh the various capabilities that together form the multidimensional account of wellbeing. In this proposal, if the functionings do not contribute to one’s happiness, they are given a zero weight and hence no longer count in the wellbeing index. In methods such as this one, there are two rounds of the selections of the dimensions: the first before one collects the data, and the second when one uses econometric techniques to determine the contribution that the various functionings make to the master-value (here: life-satisfaction) and uses those as weights; those functionings that will make no contribution will receive a weight of zero, which is the same as being deleted as a dimension in the wellbeing index.

111Third, for non-empirical applications, we can categorize methods to determine weights in the same way as we could categorise methods for the selection of dimensions. Morten Fibieger Byskov (forthcoming) distinguishes between ad-hoc methods (such as the data-driven methods discussed by Decancq and Lugo), procedural methods, or foundational methods. A theoretical capability application could include answers to all B-modules (including the selection of dimensions) yet decide that the weighing of those dimensions should be done in a procedural way, e.g. via a democratic decision-making process. Alternatively, one could introduce one master-value that will determine which capabilities are relevant, and also what weights they should be given. One example is the empirical work done by Erik Schokkaert (2007), which was discussed above. Another example, which is theoretical, is Rutger Claassen’s capabilitarian theory of justice, in which the selection and weighing of capabilities is done based on their contribution to that person’s “navigational agency” (Claassen 2016).

112Note that in the case in which one has essentially a monistic theory in which there is a master-value, one may doubt whether this doesn’t violate property A7 from the A-module. At first value, it seems that it does. But proponents of a monistic theory may respond that all theories or measures ultimately must choose one principle or value that tells us something about the relative weight of the different dimensions. In Nussbaum’s work, they argue, there is also an implicit master-value, namely human dignity. It seems to me that this issue is not sufficiently analysed and the dispute not settled. One question one could raise is whether all master-values have the same function. It seems to be different whether the capabilities constitute the dimensions of a good life (as in the case of flourishing), or whether they contribute to the mastervalue. For the time being, we should in any case flag this as an issue to which more attention should be paid in the further development of our understanding of the capability approach.

2.8.3 C3: Methods for empirical analysis

113If in B1 one chooses an empirical study, one needs to know which methods to use. This is the task of the module C3. For example, the study could contain choices about which multivariate analysis tools to use or whether certain existing data sets are capturing functionings, capabilities, or merely rough indicators. In C3, we also make methodological choices related to empirical analysis: does a particular capability issue require quantitative analysis, qualitative analysis, or a combination? In part, the contours of the empirical analysis will be influenced by one’s ambitions and goals: is one trying to measure functionings and/or capabilities directly, or is one measuring resources and conversion factors in order to infer the capability set?

114For empirical capability applications, these are of course huge methodological questions that need to be answered. These empirical methods questions may be particularly challenging for the capability approach for two reasons. First, because it is a radically multidimensional approach, and multidimensional analysis is by its very nature more complicated than a one-dimensional analysis. Second, in many cases, the relevant dimensions will include dimensions on which the collection of data is difficult, or on which no data are available — such as the quality of our social networks, the degree to which we do not suffer from excessive levels of stress, or our mental health. Nevertheless, as Alkire (2005, 129) rightly points out in her discussion on what is needed for the empirical operationalisation of the capability approach, one has to adopt the best existing empirical research (and its methods) that exists, and either master those new techniques that have been developed in other fields, or else engage in collaborations. Hick and Burchardt (2016, 88) raise the related point that there is a need for capability scholars to reach out and engage with related fields where similar themes and problems are faced. Only after that route has been travelled can we know the limits of empirical analyses of the capability approach.

2.8.4 C4: Additional normative principles and concerns

115Finally, module C4 provides room for additional normative concerns or moral principles that capability scholars aim to add to their capability theory. For example, in a particular capability theory, a principle of non-discrimination may play a role or, alternatively, one may want to work out a capabilitarian theory that subscribes to the non-domination principle as it has been defended by Republican political theory (Pettit 2001, 2009). Or, if one ascribes to a rich account of empowerment that stresses the relevance of ‘power’ and hence strongly incorporates relational aspects (e.g. Drydyk 2013; Koggel 2013), then one may add a principle related to enhancing people’s empowerment, or prioritising the empowerment of the worst-off, as an additional normative principle to be added in module C4. Again, there are several elements belonging to module C4 that could be added to a capability theory.

2.9 The modular view of the capability account: a summary

116In the previous sections, we have looked at the different elements of the modular account of the capability approach, consisting of the core A-module, the compulsory B-modules with optional content, and the contingent C-modules. For an easily accessible overview, the different elements of the modular view of the capability approach are summarised in table 2.5.

Table 2.5. The modular view of the capability approach

Table 2.5. The modular view of the capability approach

117What, exactly, is the status of this characterisation of the capability approach? Is this list of modules and the core properties exhaustive, and is this a proposal to change the current definitions on offer in the literature?

118The answers have been given throughout the sections so far, but now that we have gone through the different modules and know their content, it is worthwhile to repeat and summarize this in a very explicit manner. The modular view is an attempt at understanding the plurality of capability theories on offer in the literature, doing justice to this plurality, yet at the same time avoiding the idea that ‘anything goes’. By distinguishing between three types of modules — the A-module, the content of which one must adopt, the B-modules, which are nonoptional but have optional content, and the C-modules, which are contingent — we can get a better grasp of the peculiar nature of the capability approach: not exactly a precise theory, but also not something that can be anything one likes it to be. I hope that this way of looking at the anatomy of the capability approach will help us to understand what the approach is, but also provide more guidance to those who want to use the general capability approach as a guiding theoretical framework to work on particular theoretical or empirical issues and problems.

119The content of the A-module, the B-modules and C-modules is, as with everything in scholarship, a proposal that can be modified to accommodate new insights. If someone has convincing arguments why one element or module should be deleted, modified, or added, then that should be done. Given what we know from the history of scholarship, it is rather unlikely that no further modifications will be proposed in the future.

2.10 Hybrid theories

120In the previous sections, we have seen which modules are core in a capability theory, which ones need to be addressed but have optional content, and which ones may or may not be necessary to add to a particular capability theory. One question that this modular view raises is what we should think of a theory or an application that uses the addition of normative principles that are in contradiction with a property of the A-module. For example, suppose one would want to add the normative principle that institutions and personal behaviour should honour the traditions of one’s local community. There may be aspects of those traditions that are in tension with the principle of treating each person as an end, for example, because women are not given the same moral status in those traditions as men. What should we then say? Would such a theory no longer be a capability theory, even if the bulk of the theory is trying to think about the quality of life and desirable institutions in terms of the enhancement of functionings and capabilities?

121I propose that we introduce the notion of a hybrid theory — theories or applications that use the notions of ‘functionings’ and ‘capabilities’ yet do not endorse all propositions in the A-module. Let me stress that categorizing these theories as ‘hybrid’ entails no value judgement, whether positive or negative; rather, it is only a matter of clarifying the possibilities of having capability theories but in addition also hybrid theories which use part of the A-module yet also insert elements from other ethical frameworks that go against some propositions in the A-module. Thus, appreciating the possibility of hybrid views enlarges the diversity of theories that are possible.

  • 39 Note that for other capability applications or capability theories, such as making quality of life (...)

122Can we give an example of such a hybrid theory? Perhaps surprisingly, an example may be Amartya Sen’s theorising about justice. According to the interpretation by Antoinette Baujard and Muriel Gilardone (2017), Sen’s (2006, 2009c) recent work on justice does not endorse functionings and/or capabilities as the metric of justice, but should rather be seen as a procedural or democratic account of justice, in which the idea of having functionings and capabilities as the evaluative space is merely a suggestion, which should be put to the public who eventually, in a process of public reasoning, have to decide what justice is about. If that interpretation is correct, then Sen is unwilling to commit to proposition A5 (‘functionings and capabilities form the evaluative space’) when theorising about justice, since that is something to be decided by a process of public reasoning.39 Clearly, when Sen theorises about justice, he has certain meta-theoretical commitments (module B7) that make it inconsistent for him to endorse A5, namely the meta-theoretical commitment that the nature of justice will be decided by a democratic process. Whether that is a plausible meta-theoretical position, has been subject of debate in the capability literature (e.g. Claassen 2011; Byskov 2017) but need not concern us here. The point that is relevant for us is that Sen’s theorizing about justice could be seen as a public reasoningcapability theory of justice.

123Other potential candidates for hybrid theories are the theories that we discussed in section 2.8.2, in which functionings and capabilities play an important role, yet in the theory or measurement construction those functionings and capabilities turn out not to be of ultimate value, but rather to be instrumental for some further end that is normatively prior to the functionings and capabilities themselves. There is, in those cases, a master-value that determines how important (if at all) those capabilities are: capabilities that we could value, but which do not contribute to the master-value, will then not be given any ultimate value. As we discussed there, it is unclear whether those theories that endorse a master-value violate module A7 (value pluralism) or not. If we conclude they do, then the best way to understand these views is to regard them as hybrid theories too.

2.11 The relevance and implications of the modular view

124Understanding the capability approach as having a modular structure leads to a number of insights. Let me highlight three important ones: countering the risk of inflation, whereby we have no criteria for deciding when a theory is or is not a capability theory; appreciating the diversity of capability theories that are possible; and getting a better sense of how Sen’s and Nussbaum’s writings relate to each other.

125First, the modular view can help us to contain the risk of inflation: too many things being labelled as belonging to the ‘capability approach’, whereas they do not meet the essential characteristics of the A-module. The modular view of the capability approach which I presented gives us a description that includes all the work in the capability approach that should legitimately be included. There are, of course, other descriptions of the capability approach available in the literature. Yet to my mind most of these descriptions (including my own previous attempts at describing the approach) were insufficiently detailed and illuminating. If a description is too vague, we run the risk of inflation.

126For example, one could aim to work on multi-dimensional poverty analysis and highlight the fact that we should be interested in the combination of achievements that people are able to have. This would point at two important insights in the capability approach — namely its multidimensional character, as well as focussing on opportunity sets rather than on outcomes. But if the opportunities one focusses on are not capabilities, but rather opportunities to access certain bundles of commodities, then it would be an unjustified inflation to call this a capability application; rather, it would be another type of opportunity-based multidimensional inequality measure.

  • 40 I believe that Martha Nussbaum makes a mistake when she argues that a commitment to the normative p (...)

127Second, to understand that capability theories have a modular structure is crucial in understanding the diversity of capability theories that are possible. Let me try to illustrate this. Module C4 states that additional normative principles may be part of a capability theory, and property A6 that functionings and capabilities are not necessarily all that matters in a capability theory. From this it does not follow that all capability scholars have to endorse each and every capability theory. Surely there will be capability theorists who will take issue with the normative principles that are added in module C4 by other capability theorists when they design their theory. That is perfectly fine, as long as both theorists recognise that (a) the capability approach entails the possibility to add such additional normative principles in module C4, and (b) the normative principles they have added in module C4 are not thereby required for each and every other capability theory.40

128Take the following example. One may defend a political theory of disadvantage which states that no-one should live in poverty, no matter whether people are partly causally responsible for having ended up in that situation. Such a theory would endorse a principle (in module C4) that there should be, at the level of outcomes (and hence not at the level of opportunities) institutionally enforced solidarity via redistribution. Let us call those who endorse this principle the S-theorists (S for solidarity). This is a strong normative claim: many other normative political theories rather defend that everyone should have a genuine opportunity to live a decent life, but still attribute some responsibility to all persons for realising that life. Let us call these theorists the O-theorists (O for opportunity). Both the S-theorists and the O-theorists can agree that we should understand people’s wellbeing in terms of functionings and capabilities. The S-theorists and the O-theorists are both capabilitarians. They have to acknowledge that the other group’s theory is a capability theory, without having to endorse the other theory. In other words, a capability theorist can agree that the normative position or theory that someone else is defending is a capability theory, without having to endorse that specific theory. There is absolutely no inconsistency in this situation.

129Thirdly, the modular view of the capability approach endorses the view that Martha Nussbaum’s work on the capability approach should be understood as a capability theory, that is, a theory in which specific choices are made regarding the modules. It is not, as Nussbaum (2011) suggests in her Creating Capabilities, a version of the capability approach structurally on a par with Sen’s more general capability approach. What Sen has tried to do in his work on the capability approach, is to carve out the general capability approach, as well as to give some more specific capability applications. Admittedly, Sen’s work on the capability approach (rather than his work on a variety of capability applications) would have benefited from a more systematic description of how he saw the anatomy of the capability approach. To my mind, that has been missing from his work, and that is what I have tried to develop here and in an earlier paper (Robeyns 2016b). Yet everything put together, I agree with the understanding of Mozaffar Qizilbash, who concludes an analysis of the difference between Nussbaum’s and Sen’s work on the capability approach by saying that “On this reading […] Nussbaum’s capabilities approach emerges as one particular application or development of Sen’s original formulation of the approach” (Qizilbash 2013, 38).

130It is a mistake to understand the capability literature as a field with two major thinkers who have each proposed one version of the capability approach, which have then inspired the work by many other scholars. Rather, there is only one capability approach which is a generalisation of the work by Sen together with further developments by many others. In addition, there are many dozens capability theories — about justice, human rights, social choice theory, welfare economics, poverty measurement, relational egalitarianism, curriculum design, development project assessment, technological design, and so forth. Clearly, Nussbaum has been one of the most prolific and important contributors; she has pushed the boundaries of capabilitarian theories and has rightly advanced the agenda to achieve more clarity on the essential characteristics that any capability theory should meet. However, she has offered us a more specific capability theory, rather than another version of the approach, even if it is the capability theory that is by far the most influential capability theory among philosophers. Establishing the anatomy of the capability approach and its relation to particular capability theories is very important, because it vastly expands the scope of the capability approach, and increases the potential types of capability applications and capabilitarian theories.

131In sum, there is much pluralism within the capability approach. Someone who considers herself a capabilitarian or capability thinker does not need to endorse all capability theories. In fact, it is impossible to endorse all capability theories, since different choices made in module C1 (ontological and explanatory theories that are endorsed) and module C4 (additional normative principles) can be in conflict with each other. It is presumably coherent to be a Marxist capabilitarian, and it is presumably also coherent to be a libertarian capabilitarian, but it is not coherent to endorse the views taken by those two positions, since they are incompatible.

2.12 A visualisation of the core conceptual elements

132We have now covered enough ground in understanding the core concepts of the capability approach to construct a visualisation of these concepts. Figure 2.1 below gives a graphical representation of the different elements of the capability approach, and how they relate. Note that the arrows do not indicate normative importance but rather indicate which parts of this conceptual system are determinants of, or have an influence on, other parts.

133Let us start our description where economists generally start (and often also end): with resources. In the capability approach, the term ‘resources’ is interpreted in a broader sense than the understanding of that term elsewhere in the social sciences. Economics and the quantitative empirical social sciences have traditionally focussed on material resources only: either income and wealth, or else on the consumption that these financial means (or unpaid production) generated. One important lesson learnt from feminist economics is that about half of economic production happens outside the market and the formal economy, which is the reason why the box at the far left in Figure 2.1 also includes resources created by non-market production (Folbre 2008; Folbre and Bittman 2004).

134Both the resources and the consumption could be conceptualised as capability inputs: they are the means to the opportunities to be the person one wants to be, and do what one has reason to value doing. The means do not all have the same power to generate capabilities; this depends on a person’s conversion factors, as well as the structural constraints that she faces. Those structural constraints can have a great influence on the conversion factors as well as on the capability sets directly.

135From this visualisation, we can also see the difference between the social conversion factors and the structural constraints. The structural constraints affect a person’s set of conversion factors, including the social conversion factors she faces. But recall that those conversion factors tell us something about the degree to which people can turn resources into capabilities. Conversion factors are thus, conceptually and empirically, closely related to the capability inputs — that is, the resources that are needed to generate capabilities. Structural constraints affect conversion factors, but can also affect a person’s capability set without impacting on the conversion of resources in capabilities. For example, if a certain set of social norms characterizes a group in society as not having the same moral status as others, then this affects the capabilities of the members of that group directly, not merely in terms of what they can get out of their resources. A good example is gay people, a significant percentage of whom are not worse off in financial terms than straight people, but they often cannot express their sexual orientation in public, at the risk of humiliation, aggression, or even risking their jobs or their lives.

136Another part of the visualisation to pay attention to is the choice that people make given the capability set they have. These choices are always constrained in some sense, and the question is which types of constraints a particular capability theory will take into account. Also, the term ‘choice’ is used here in a very thin (or, as philosophers say, ‘weak’) sense: it is not assumed that elaborate thinking and weighing is done before we decide which capabilities to use and realise into functionings. In fact, we have ample evidence from psychology that there are many other factors that influence the decisions we make, including how hungry or tired we are, the people in our company, or the amount of time we have to make a decision (Ariely 2010; Kahneman 2011). Most capability theories will have an (implicit) theory of choice: they will have views on the extent to which people’s past history (which includes the structural constraints they faced in their personal past) as well as societal processes, such as preference formation mechanisms, influence the choices that we make from the opportunities that are available to us.

137The last element of the visualisation is the level of satisfaction or dissatisfaction with one’s functionings and capability levels each person will have. Of course, this does not mean that we need to attribute ethical significance to those levels of satisfaction; rather, the point in the visualisation is that satisfaction with one’s functionings and capability levels is not the same thing as those capability sets and combination of functionings achievements themselves. For now, we will concentrate on deepening our understanding of the capability approach itself, but in section 3.8, we will engage in more depth with the question whether we should look at functionings and capabilities, rather than satisfaction, or some other mental metrics, such as happiness.

  • 41 In reality, almost everything is related to almost everything else in one way or another, but putti (...)

138Two remarks are important. First, this is a stylized visualisation, and also a simplification. It is meant to help us see the different elements of the capability approach and how they relate; it is not a fine-grained and exhaustive picture of all the elements that determine a person’s capability set. One important limitation is that this is not a dynamic visualisation and that many arrows, indicating relationships between various parts, are not present.41 In addition, the choices we make from our capability set at one point in time, will be determinants of our resources and our capability set in the future. Another important limitation of this visualisation is that it gives us only the resources, capabilities, functionings and satisfaction of one person, but, as was mentioned before, capability sets are interdependent; hence choices made from one person’s capability set will lead to changes in another person’s capability set.

Figure 2.1. A stylised visualisation of the core concepts of capability theories

Figure 2.1. A stylised visualisation of the core concepts of capability theories

Source: Based on Robeyns (2005b), updated and expanded

2.13 The narrow and broad uses of the capability approach

  • 42 Readers who have never come across that distinction between the narrow and broad use of the capabil (...)

139We have now reached the end of the discussion of the modular view of the capability approach. Yet before closing this chapter, let us pause to use this modular view to clarify something that has been noted by several capability scholars, namely that the capability approach has been used and can be used in narrower or more limited ways on the one hand, and broader or richer ways on the other (e.g. Alkire, Qizilbash and Comim 2008, 4–5; Crocker and Robeyns 2009; Qizilbash 2012).42 The distinction has been slightly differently presented by different authors, but the general gist of their analyses has been that the capability approach can either be seen as offering something limited, or else much more ambitious and wide-reaching:

[…] several interpretations of the scope of the capability approach are used in the wider literature […]. These can be charted between two poles: one narrow and broad, with the broad subsuming the narrow. […] The narrow interpretation sees the approach primarily as identifying capability and functionings as the primary informational space for certain exercises. The broad interpretation views the capability approach as providing a more extensive and demanding evaluative framework, for example by introducing human rights or plural principles beyond the expansion of capabilities — principles which embody other values of concerns such as equity, sustainability or responsibility. (Alkire, Qizilbash and Comim 2008, 4–5)

140In the narrow way, the capability approach tells us what information we should look at if we are to judge how well someone’s life is going or has gone; this kind of information is needed in any account of wellbeing or human development, or for any kind of interpersonal comparisons. Since the capability approach contends that the relevant kind of information concerns human functionings and capabilities, the approach provides part of what is needed for interpersonal comparisons of advantage.

141The modular view presented in this chapter can help to make sense of this observation that there is both a narrower and a wider use of the capability approach. In the narrow use of the capability approach, the focus is often strictly on the evaluation of individual functioning levels or on both functionings and capabilities. If we look at the narrow use of the capability approach through the lens of the modular understanding of the approach, we can see that the narrow view chooses interpersonal comparisons as the purpose of the capability theory (module B1); and that it will have to make a selection of dimensions (module B2) and make a choice between functionings or capabilities (module B6); its choice for human diversity (module B3) will be reflected in the choices it makes in B2, but also in which groups (if any) it will compare. Its meta-theoretical commitments (B7) are likely related to limiting research to those things that can be measured. The narrow use of the capability approach will most likely not have much to say about agency (B4) and structural constraints (B5) but adopt the implicit theories of agency and structural constraints that are used in the empirical literatures on interpersonal comparisons. Finally, the narrower view must decide on how to weigh the dimensions (module C2) and which methods for empirical analysis (module C3) to make.

142In its broad uses, the capability approach not only evaluates the lives of individuals (as in the narrow use), but also includes other considerations in its evaluations, which are ‘borrowed’ from other approaches or theories. For example, the broader use of the capability approach often pays attention to other normative considerations and other values than only wellbeing, such as efficiency, agency, or procedural fairness.

143The broad view would, in most cases, have a more ambitious purpose for its use of the capability approach, such as societal evaluation or policy design. It would also have (either implicit or explicit) richer theories of human diversity, agency and structural constraints, and — most importantly — add several additional ontological and explanatory theories (module C1) and additional normative principles (module C4). The narrow view does not include modules C1 and C4, and this can make a huge difference to the kind of capability theory that emerges.

144An example of the broad view is David Crocker’s (2008) book on development ethics, in which he has extended the capability approach with accounts of agency, democratic deliberation and participation into a more detailed account of development ethics. Yet Crocker acknowledges that not all versions of the capability approach embrace agency so explicitly. The capability approach proper need not endorse a strong account of agency, but there are several scholars who have developed particular capability theories and applications in which agency plays a central role (e.g. Claassen and Düwell 2013; Claassen 2016; Trommlerová, Klasen and Leßmann 2015).

145Why is this difference between the narrow and the broad uses of the capability approach relevant and important? There are several important reasons. First, to assess a critique of the capability approach, we need to know whether the critique addresses the capability approach in its narrow use, or rather a specific version of its broad use. Second, we need to be clear that many of the additional normative commitments in the broad use of the capability approach are not essential to the capability approach: rather, they are optional choices made in modules B and, especially, module C1 (additional ontological and explanatory theories) and module C4 (additional normative principles and concerns). This insight will also be important when we address the question, in section 4.9, of whether we can simply talk about ‘the capability approach’ and ‘the human development paradigm’ as the same thing.

2.14 Conclusion

146The purpose of this chapter has been to give a comprehensive explanation of the capability approach: what is it trying to do, what are the many ways in which it has been used, what are the properties that all capability theories share, and what is the structure that we can detect in the construction of capability theories and applications? In order to get a helicopter view, I have deliberately put aside a number of additional distinctions and details. They will be the focus of the next chapter, whereas critiques and areas of contestation and debate will be discussed in chapter 4.


1 In developing the account of the capability approach as presented in this book, I have started in an inductive way by trying to generalize from how the capability approach has been used in the literature. However, that literature has been critically scrutinized, and in some cases I have come to the conclusion that some ideas in this broad ‘capabilities literature’ do not survive careful analysis, and should be rejected. Put differently, my methodology has been to be as inclusive as possible, but not at the cost of endorsing (what I believe to be) confusions or errors.

2 To the reader who finds this ultra-brief summary of the modular account of the capability approach here unclear: please bear with me until we have reached the end of this chapter, when the different modules will have been unpacked and explained in detail.

3 Amartya Sen often uses the term “social arrangement”, which is widely used in the social choice literature and in some other parts of the literature on the capability approach. Yet this term is not very widely used in other disciplines, and many have wondered what “social arrangement” exactly means (e.g. Béteille 1993). Other scholars tend to use the term “institutions”, using a broad definition — understood as the formal and informal rules in society that structure, facilitate and delineate actions and interactions. “Institutions” are thus not merely laws and formal rules such as those related to the system of property rights or the social security system, but also informal rules and social norms, such as social norms that expect women to be responsible for raising the children and caring for the ill and elderly, or forbid members of different castes to work together or interact on an equal footing.

4 See Nussbaum (1988, 1992), Sen (1993a, 1999a, 14, 24); Walsh (2000); Qizilbash (2016); Basu and López-Calva (2011, 156–59).

5 A partial theory of justice is a theory that gives us an account of some aspects of what justice requires, but does not comment on what justice requires in other instances or areas.

6 On the relationship between capabilities and human rights, see section 3.14.

7 For earlier attempts to describe the capability approach, see amongst others Deneulin (2014); Gasper (2007, 1997); Alkire, Qizilbash and Comim (2008); Qizilbash (2012); Robeyns (2005b, 2016b).

8 A ‘technical term’ is a term which is used in a specialist debate, and has a meaning that is defined within that debate. In many cases, the term refers to something other than its referent in common-sense language (that is, a layperson’s use of language).

9 To ‘remain agnostic’ means that, for the purpose of that analysis, one does not make a choice between different options, and hence proceeds with an analysis that should be valid for all those options. This does not mean that one cannot make a choice, or really believes that all available options are equally good, but rather that one wants to present an analysis that is applicable to as wide a range of choices as possible.

10 Following Amartya Sen (1985c), some would say there are four different ideas of advantage in the capability approach: achieved wellbeing, freedom to achieve wellbeing, achieved agency, and freedom to achieve agency. Yet whether the capability approach should always and for all purposes consider agency freedom to be an end in itself is disputed, and depends in large measure on what one wants to use the capability approach for.

11 Of course, it doesn’t follow that mental categories or the material means play no role at all; but the normative priority lies with functionings and capabilities, and hence happiness or material resources play a more limited role (and, in the case of resources, a purely instrumental role). The relations between functionings/capabilities and resources will be elaborated in 3.12; the relationship between functionings/capabilities and happiness will be elaborated in more detail in section 3.8.

12 On whether the capability approach can be used for explanatory purposes, see section 3.10.

13 Alkire (2008) calls these normative applications “prospective analysis”, and argues that we need to distinguish the evaluative applications of the capability approach from the “prospective applications” of the capability approach. I agree, but since we should avoid introducing new terms when the terms needed are already available, it would be better to use the term ‘prescriptive applications’ or, as philosophers do, ‘normative analysis’, rather than introducing ‘prospective applications’ as a new term.

14 I kindly request readers who are primarily interested in the capability approach for policy design and (empirical) applications to read ‘capability application’ every time the term ‘capability theory’ is used.

15 Yet even for capability theories, it is unlikely that Sen would agree that he has to draw up a list of capabilities, since he is a proponent of a procedural method for selecting capabilities. At the beginning of this century, there was a fierce discussion in the capability literature about whether it was a valid critique of Sen’s work that it lacked a specific list (Nussbaum 2003a; Robeyns 2003; Sen 2004a; Qizilbash 2005). Luckily that debate seems to be settled now. For an overview of the different ways in which dimensions can be selected in the capability approach, see section 2.7.2.

16 To say that the insertion of those theories is fully optional is not the same as saying that capability theories that will be developed with different types of additional complementary theories will all be equally good. For example, historians are very likely to think that most theories, even normative theories, have to be historically informed, and hence the relevant historical knowledge will need to be added to a capability theory. But these are matters of dispute that have to be debated, and cannot be settled by narrowing down the definition of a capability theory.

17 See section 2.2 for an explanation of the technical term ‘advantage’.

18 For some core differences in the way Martha Nussbaum and Amartya Sen use the terms ‘functionings’ and ‘capabilities’, see section 3.2.

19 See also section 3.3 which discusses in more depth the kind of freedoms or opportunities that capabilities are.

20 The exceptions are those philosophers who want to develop normative theories while steering away from any metaphysical claims (that is, claims about how things are when we try to uncover their essential nature). I agree that the description of ‘functionings’ and ‘capabilities’ in this section makes metaphysical claims, but I think they are very ‘minimal’ (in the sense that they are not wildly implausible, and still leave open a wide variety of theories to be developed) and hence we should not be troubled by these metaphysical assumptions.

21 Arguably, some of that work is being done by social choice theorists and others working with axiomatic methods, but unfortunately almost none of the insights of that work have spread among the disciplines within the capability literature where axiomatic and other formal methods are not used (and, presumably, not well understood). See, for example, Pattanaik (2006); Xu (2002); Gotoh, Suzumura and Yoshihara (2005); Gaertner and Xu (2006, 2008).

22 On the complex nature of ‘care’, and what the need to care and be cared for requires from a just society, see e. g. Tronto (1987); Kittay (1999); Nussbaum (2006b); Folbre (1994); Folbre and Bittman (2004); Engster (2007); Gheaus (2011); Gheaus and Robeyns (2011).

23 The Multidimensional Poverty Index is developed by the Oxford Poverty and Human Development Initiative (OPHI), under the leadership of Sabina Alkire. See for a clear introduction of the Multidimensional Poverty Index. For scholarly papers on the Index, as well as other work done by the scholars in OPHI, see

24 This is a critique that the capability approach shares with the happiness approach, which also focusses on what it considers to be an end in itself — happiness. Still, capability scholars have reasons why they do not endorse the singular focus on happiness, as the happiness approach proposes. See section 3.8.

25 However, while a focus on available options rather than realised choices is the default normative focus of capability theories, there are some capability applications where, for good reasons, the focus is on achieved functionings rather than capabilities. This will be elaborated in the next section.

26 On the distinction between general capabilities and specific capabilities, see 3.2.4.

27 The relationship between means and capabilities is analysed in more depth in section 3.12.

28 I am using the term ‘normative’ here in the way it is used by social scientists, hence encompassing what philosophers call both ‘normative’ and ‘evaluative’. For these different uses of terminology, see section 2.2. It is also possible to use the notions of ‘functionings’ and ‘capabilities’ for non-normative purposes (see section 3.10). In that case, the basic notions from the core are all that one takes from the capability approach; one does not need this normative part of the core. I will suggest in the concluding chapter 5 that explanatory applications of the capability approach are part of how it could be fruitfully developed in the future.

29 Moreover, further extensions of this proposition may be needed. One issue is that we should not only focus on capabilities that people value, but also on capabilities that they do not, but should, value (see section 2.7.2). Another issue is that the evaluative space should not necessarily be restricted to capabilities only, but could also be functionings, or a combination of functionings and capabilities (see section 2.6.5).

30 For example, if Henry Richardson (2007) is right in arguing that the idea of capabilities cannot capture basic liberties, then one need not reject the capability approach, but instead could add an insistence on basic liberties to one’s capability theory, as Richardson (2007, 394) rightly points out.

31 This distinction, and its relevance, will be discussed in more detail in section 3.3.

32 “[…] the relevant values can be said to be ‘end-point values’, insofar as they are not simply conducive to or necessary for something else that is of value. They are ‘final’, then, in this sense of being ‘ultimate’” (Rabinowicz and Rønnow-Rasmussen 2000, 48).

33 As Richardson (2016, 5) puts it, “all capabilities […] are dependent on the body. Without relying on one’s body there is nothing one can do or be”.

34 Some have argued in favour of what they call ‘collective capabilities’, which I will discuss in section 3.6.

35 However, the question remains whether the capability approach is fully compatible with indigenous world views and normative frameworks, as well as thick forms of communitarianism. This is a question that doesn’t allow for a straightforward answer, and requires more analysis. For some first explorations of the compatibility of indigenous world views with the capability approach, see Binder and Binder (2016); Bockstael and Watene (2016); Watene (2016).

36 For an introduction to libertarianism, see Vallentyne and Van der Vossen (2014).

37 Martha Nussbaum explicitly refrains from integrating the notion of ‘agency’ in her capability theory (Nussbaum 2000, 14). However, this does not mean that there isn’t an account of agency in her theory, since the inclusion of the capability of practical reason on her list of central human capabilities can be understood as corresponding to one particular conceptualisation of agency.

38 In the case of democratic theory, the discussion is often about which laws to implement, but the same insights apply to policy making. Both the literature on democratic theory (e. g. Dryzek 2000; Gutmann and Thompson 2004) and social choice theory (e. g. Arrow, Sen and Suzumura 2002, 2010; Sen 1999c, 2017; Gaertner 2009) are vast and will not be further discussed here.

39 Note that for other capability applications or capability theories, such as making quality of life assessment studies, Sen has no problem endorsing proposition A5. Moreover, one could also ask whether regarding his earlier publications on justice it would be implausible to interpret Sen’s writings as an endorsement of A5. In my view (and pace Baujard and Gilardone’s interpretation), Sen has made several statements in earlier work that could be seen as an endorsement of all propositions of module A for the case of theorizing about justice (e. g. Sen 1980, 1990a, 2000).

40 I believe that Martha Nussbaum makes a mistake when she argues that a commitment to the normative principle of political liberalism is essential to the capability approach, hence to each and every capability theory. For my arguments why this is a mistake, see Robeyns (2016b). Political liberalism is an additional normative commitment that is not a property of the A-module, but rather a choice in module C4.

41 In reality, almost everything is related to almost everything else in one way or another, but putting all those arrows on a visualisation would make it completely uninformative. After all, the task of scholarship is to abstract away from distracting details to see more clearly.

42 Readers who have never come across that distinction between the narrow and broad use of the capability approach may simply ignore this section and move on, since the anatomy of the capability approach that has been presented in this book covers the same terrain. In essence, this section is written for those who came across this terminology in the literature, and wonder how it relates to what has been said so far in this book.

Table des illustrations

Titre Table 2.1. The main modes of capability analysis
Légende Source: Robeyns (2005a), expanded and updated
Fichier image/jpeg, 448k
Titre Table 2.5. The modular view of the capability approach
Fichier image/jpeg, 232k
Titre Figure 2.1. A stylised visualisation of the core concepts of capability theories
Légende Source: Based on Robeyns (2005b), updated and expanded
Fichier image/jpeg, 184k


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search