4.3: Should we commit to a specific list of capabilities?
At an earlier stage of the development of the capability approach, a rather heated debate took place on whether or not it was necessary for Sen to list the capabilities he felt were relevant for the issue under consideration. This ‘question of the list’ debate wasn’t always very helpful, since participants were not making the distinction between capability theories and the capability approach, which, as I will show in this section, is crucial to answer this question. Several scholars have criticized Sen for not having specified which capabilities matter or for not giving us some guidelines on how the selection of capabilities could be conducted (e.g. Sugden 1993; Roemer 1996; Nussbaum 2003a). As is well known, Sen has explicitly refrained from committing himself to one particular list of capabilities. But should Sen (or anyone else) do so?
In order to answer that question, it is important to keep the distinction in mind between the general capability approach, and particular capability theories. As Mozaffar Qizilbash (2012) rightly points out, Sen has written on the capability approach in general and he has developed particular capability applications, critiques, and theories. When asking whether Sen (or anyone else) should commit to a particular list of capabilities, we have to keep that distinction firmly in mind — since it is relevant to our answer.
It is obvious that there cannot be one list that applies to all the different purposes for which the capability approach can be used — that is, one list that applies to more specific capability theories and applications. Hence insofar as it is argued that Sen (or any other capability scholar) should endorse a particular list of capabilities when discussing the capability approach , rather than more specific capability theories , this critique misfires. This is part of the answer that Sen has given to his critics. Each application or theory based on the capability approach will always require a selection of valuable functionings that fits the purpose of the theory or application. Hence the capability approach as such is deliberately too underspecified to endorse just one single list that could be used for all capability analyses (Sen 1993, 2004). It is quite likely that those who have criticised Sen, or the capability approach in general, for not entailing a specific list of capabilities, have not sufficiently appreciated the distinction between the capability approach in general and more specific capability theories.
But what then about specific capability theories, applications and analyses? Should these always commit to a particular list of capabilities? It is possible to distinguish between two types of critique addressing capability theories, which I labelled the weak and strong critiques (Robeyns 2005a). The strong critique entails that there must be a clear list of capabilities that we can use for all capability theories and their application.
The strong critique is most clearly voiced by Nussbaum, who has proposed a list of ten “central human capabilities” that specify the political principles that every person should be entitled to as a matter of justice. 1 Nussbaum’s capabilities theory differs in a number of ways from Sen’s version. Nussbaum (1988, 2003) not only argues that these ten capabilities are the relevant ones, but in addition claims that if Sen wants his version of the capability approach to have any bite for addressing issues of social justice, he has to endorse one specific and well-defined list of capabilities.
Sen does not accept the stronger critique as it applies to particular capability theories. The reason is the importance he attaches to agency, the process of choice, and the freedom to reason with respect to the selection of relevant capabilities. He argues that theory on its own is not capable of making such a final list of capabilities (Sen 2004). Instead, Sen argues that we must leave it to democratic processes and social choice procedures to define the distributive policies. In other words, when the capability approach is used for policy work, it is the people who will be affected by the policies who should decide on what will count as valuable capabilities for the policy in question. This immediately makes clear that in order to be operational for (small-scale) policy implementation, the capability approach needs to engage with theories of deliberative democracy and public deliberation and participation.
Sen’s response to the strong critique can be better understood by highlighting his meta-theoretical views on the construction of theories, and theories of justice in particular. One should not forget that Sen is predominantly a prominent scholar in social choice theory, which is the discipline that studies how individual preferences and interests can be combined to reach collective decisions, and how these processes affect the distribution and levels of welfare and freedom. Sen published ground-breaking work in social choice theory before he started working on the capability approach, and he has never ceased to be interested in and to contribute to social choice theory. 2 Sen’s passion for social choice theory is also a very likely explanation for his critique of the dominant forms of contemporary theories of justice, which, he argues, focus on describing a utopian situation of perfect justice, rather than giving us tools to detect injustices and decide how to move forward to a less unjust society (Sen 2006, 2009c).
According to my reading of Sen’s work on capability theories and applications, he is not against the selection of dimensions in general, but rather (a) against one list that would apply to all capability theories and applications, and (b) as far as those capability theories and applications are concerned, in favour of seriously considering procedural methods to decide which capabilities matter.
However, even if we all accept that view, it doesn’t settle all disputes. Even if we agree that a selection of capabilities for, say, a poverty evaluation should differ from the selection of capabilities for a theory of justice, this still allows for different views on how that selection should be made. Some scholars have argued that it should be based on normative grounds, in other words based on philosophical reasoning and argumentation (Nussbaum 2000; 2006b; Claassen 2016). Others have argued for a selection based on a procedural method (Byskov 2017). For empirical applications, it has been argued that the selection of dimensions should be made in a way that minimises biases in the selection (Robeyns 2003). For policy-relevant applications, it has been argued that the freedoms listed in the Universal Declaration of Human Rights could provide a good starting point, and should plausibly be playing a larger role in the selection of capabilities (Vizard 2007; Burchardt and Vizard 2011). There are by now various overviews published on how to select functionings and capabilities but, interestingly, they almost always are limited to a certain type of capability theory, such as wellbeing for policy making (Hick and Burchardt 2016; Alkire 2016), multidimensional poverty measurement (Alkire 2016; Alkire et al. 2015), human development projects and policies (Alkire 2002; Byskov forthcoming) and theories of justice (Robeyns 2016d). Thus, there are a range of arguments pointing out that the selection of capabilities for particular capability theories needs to be sensitive to the purpose of the theory in question, hence selection is a matter to be decided at the level of the individual capability theories , rather than at the more general and abstract level of the capability approach (see also Sen 2004a).
1 These ten capabilities are: Life; Bodily health; Bodily integrity; Senses, imagination and thought; Emotions; Practical reason; Affiliation; Other species; Play; and Control over one’s environment. For more details, see Nussbaum (2006b, 76–78).
2 Sen was also awarded the Nobel Memorial Prize for his contributions to social choice theory and welfare economics. For some of his work on social choice theory, see Sen (1970a, 1970b, 1976, 1977b, 1979, 1983, 1986, 1992c, 1999c, 2017).