Apologies for vanishing and temporarily interrupting the capability project! I’m resuming my series of posts on the capability approach, which I expect to continue till mid-July (and afterwards we’ll see where we are). I am now turning to the capability approach as a theory of justice (social or distributive justice). This may require more than one post, and in this first one I want to discuss two meta-theoretical problems with the capability approach to justice.
Problem #1: a single module, rather than an engine
A theory of distributive justice must always specify at least a metric of justice and a distributive rule. Metrics are about the good that is the subject of the just distribution. Influential metrics are resources, Rawlsian social primary goods, or subjective metrics such as welfare/utility/happiness (generally ‘laundered’ to take away unwanted influences, and to guarantee minimal conditions of autonomy and/or authenticity). Distributive rules could be straight equality, responsibility-sensitive equality, sufficiency, priority to the worst-off, amongst others.
The capability approach to justice defends functionings and/or capabilities as the metric of justice. That normative commitment could be combined with several different choices of the distributive rule. Two of the leading theorists defend sufficientarian accounts (Martha Nussbaum and Elizabeth Anderson), and the capability approach is therefore often interpreted as being a sufficientarian theory: but that need not be the case. Sen, as far as I can see, does not commit to one particular distributive rule, since he believes the quest for a full theory of justice is mistaken, and we should instead focus on injustices, and judge situations case by case, implying that sometimes equality is the right metric, sometimes sufficiency [I am not endorsing Sen’s views – in fact, I have argued here before that I think he is mistaken]. Thus, the capability approach to justice is only committed to the specification of a metric, not to a distributive rule.
Richard Arneson (2010, p. 103) rightly remarked that this makes it hard to assess the capability approach to justice, since we are asked to judge one module of a theory of justice, without knowing what the other modules will be. However, as Arneson rightly points out
in general one assesses a suggested module of a moral theory as one might assess a proposed part that is supposed to fulfill some function in an engine. One can see if the part works by seeing if the engine works better with that part inserted or with some substitute inserted instead or without anything of the sort deployed. You might be abel to tell by inspecting the part in isolation that it could not play its assigned role, but the definite test for success will be how the part functions in its place, alongside the rest of the engine. And so it would seem to be for proposed modules of moral theories. The unit of assessment is really a complete moral theory.”
So this is the first challenge/problem when studying/assessing the capability approach to justice. In my contribution in the capability approach to the Stanford Encyclopedia of Philosophy, I listed what I believe that is needed for a full capability theory of justice: (1) its grounding/justification; (2) whether functionings, capabilities, or a mixture is the appropriate metric; (3) the selection and aggregation of relevant dimensions; (4) a comparison with other metrics of justice; (5) a justified choice of a distributive rule; (6) the role (if any) for personal responsibility [this is related to the question of the distributive rule].
Did I miss anything?
Problem #2: the status of the theory
Another real challenge for the capability approach to justice has to do with its meta-theoretical status. Anyone who has been following the recent debate on ideal vs. non ideal theories of justice knows how confusing (and confused?) that debate is, but here follows the best I can make of it. In order to properly understand, let alone assess, a theory, we need to know what the theory aims to deliver. In theories of justice, this is seldomly spelled out, but it matters, in my view, quite strongly. The first question is whether the theory aims to merely be truth-seeking, or also ‘practical’, in the philosophical/technical sense of that word, namely that it gives us (as persons or in the design of our institutions) guidance on what to do. If the theory tries to be merely truth-seeking, it does not need to take into account that certain things are non-observable (e.g. whether one’s preferences are ‘authentic’), or take feasibility constraints into account. These theories can use assumptions that are not merely abstractions but also idealizations. If a theory tries to be practical (in the above-mentioned sense), there are more constraints on the theoretical tools it can employ. If it relies on idealizations, unobservable entities, or ignores those feasibility constraints that it should respect even in a maximally just society (such as: babies are born completely dependent and die if they are not properly cared for; people need oxygen to live; crops don’t grow if they don’t get water), then the theory becomes either biased or useless for action-guiding. Much of the work done in applied philosophy belongs to the action-guiding group, and most of the influential theories of distributive justice belong to merely truth-seeking group. And we know way too little, in my view, of how to move from a plausible set of principles in the merely truth-seeking group to recommendations on action-guidence (if that move can be made at all – a huge question I am not addressing now).
Assessing the capability approach to justice is not only tricky because of the issues related to it being only one module in an entire engine, but also because, in contrast to many better-worked-out theories it is not always clear what kind of theory it aspires to be. Hence it may need to be assessed again different meta-theoretical backgrounds. It could be a poor/strong theory in one of those meta-theoretical categories, but not in others.
My hunch is that, because of its strong commitment to acknowledging human heterogeneity/ human diversity, the strength and potential of the capability approach as a non-idealised theory of justice that aims to be action-guiding. There aren’t significantly strong idealizations in the capability approach, such as that representative agents are heads of families and their interests coincide with the interests of all family members, that people have no caring needs/obligations, or that they have no disabilities/long-term illnesses, or that they are all equally able to convert ‘all purpose means’ into valuable goods. But this ‘realistic human anthropology’ comes at a price, and that price is messiness. The capability approach to justice does not have the theoretical elegance that many other theories of justice have, and elegance and parsimony are seen by many (analytical) philosophers as desirable theoretical characteristics.
The upshot is that we should be very careful to compare apples with pears, and that this is a real risk in the literature on the capability approach to justice.
There is much more to be said, but I think I’ve said more than enough to start with.