Professional Documents
Culture Documents
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
C 2016 Metaphilosophy LLC and John Wiley & Sons Ltd
V
METAPHILOSOPHY
Vol. 47, No. 1, January 2016
0026-1068
Abstract: When we make public policy choices, is it helpful to know how utopia
(that is, the perfectly just institutional design) would look? Amartya Sen argues
that it is neither necessary, nor sufficient, nor even contributory. He claims that
before making a policy choice one should compare several feasible institutional
designs to see which promotes justice most (a “comparative assessment” of
justice), and that it is misleading to use the perfect design as a standard in those
comparisons. Principles of justice are the proper standard. The present article
contends that the perfect design has nevertheless an important role to play in the
prior task of identifying and refining our principles of justice. It also shows that
the perfect design—in at least one sense of this term—may be a legitimate long-
term goal for present policy choices.
1. Introduction
Let us assume that before we make policy choices it is crucial to com-
pare alternative feasible institutional designs, in order to tell which
design would improve justice, or reduce injustice, most. This task may
be called, following Amartya Sen, a “comparative assessment” of jus-
tice (Sen 2009, 17). Identifying the perfectly just institutional design is
regarded by him as neither necessary nor sufficient, nor of any particu-
lar help, for that comparative task (Sen 2009, 15–16). In this article I
contend that having a picture of the perfect design plays an important
role in comparative assessments of justice.
The article is structured as follows. In section 2, I introduce a set of
preliminary definitions and considerations. In section 3, I assess the
claim that the perfect design is helpful for real-world policy choices,
since it serves as a comparison standard between imperfect designs. I
show that although using the perfect design itself as a comparison
standard may result in wrong policy choices, this design nevertheless
has an important role to play in the prior task of identifying the proper
comparison standard. In section 4, I show that the perfect design has
yet another role to play in comparative assessments as a legitimate
long-term goal for present policy choices. This is shown after distin-
guishing several senses of the idea of a perfect design.
the latter.2 We may then roughly define the idea of a perfectly just insti-
tutional design (or perfect design) as the institutional design that fully
instantiates all principles of justice (that is, the institutional design that
has the highest justice performance).
A comparative assessment of justice is a comparison between different
feasible institutional designs in order to choose one, according to several
comparative dimensions. The most important, or focal, comparative dimen-
sion is justice performance (but there are other dimensions, which will be
explained later). Although it is obviously possible to comparatively assess
infeasible institutional designs in terms of justice (for example, we may com-
pare different utopias to assess which of them implements principles to a
higher degree), in the present article the idea of comparative assessments of
justice will refer exclusively to assessments of currently feasible designs, since
these are the main subject of actual policy choices. (For the idea of feasibil-
ity, see section 4.1.)
Let us illustrate all previous definitions by the following example. For the
sake of simplicity, suppose that the only principle of justice is SP. Imagine
that our current institutional design has trouble securing everyones access
to basic goods (nourishment, health care, and housing), and 50 percent of
the population lacks access to at least one of them. Imagine that there is an
alternative institutional design that would secure access to 80 percent of the
population. If we made a comparative assessment between both designs, we
would conclude that the alternative design is better than the current one in
terms of justice performance. If there were a third alternative design that
would secure access to everyone, that third design would then be the perfect
design, since it would perform highest in terms of justice.
In this article I avoid the common terminology of “ideal theory”
and “non-ideal theory,” since it may lead to confusion due to the many
diverse senses in which the terms “ideal” and “theory” may be under-
stood (Stemplowska and Swift 2012, 373–74). For similar reasons, I
also avoid the term “perfect justice” (used, for example, by Sen 2009),
since it is ambiguous; it could refer to the perfectly just institutional
design or to the set of principles that are fully implemented by it.
3. Comparison Standards
Consider the following simple argument, which aims to show that
knowing the perfect design is relevant for policy choices. Suppose we
2
It should be noted that certain social rules may be regarded as desirable states of affairs
themselves, and not as desirable due to the distributive consequences caused by their existence
(that is, “instrumentally” desirable). Procedimentalist conceptions of democracy, for instance,
do not claim that democracy is (only) valuable as a set of social rules capable of implementing
certain principles of justice (such as equality and liberty). These conceptions claim instead that
democracy is a normatively required state of affairs in itself.
Design A. This design would fully implement SP. Among other insti-
tutions, this design includes a system of public health care and edu-
cation. Design A is infeasible (assume this for the sake of the
argument).
Design B. This design is similar in most (but not all) aspects to design A.
It includes a system of public health care and education. Design B par-
tially implements SP (say, by 25 percent). Design B is feasible.
Design C. This design is different from design A in most aspects. It
includes a system of private health care and education, among other
differences. Design C partially implements SP (say, by 50 percent).
Design C is feasible.
8
I thank an anonymous referee for urging me to address these issues.
9
There may be cases in which a proposed set of principles of justice could be fully
implemented by more than one institutional design. And in some of those cases it may
happen that our considered judgments conflict with one of these designs but not with the
other(s). Should we then revise our principles, our judgments, or both? The answer (which
I cannot develop here) is that in each particular conflict we should choose the revision
that brings more overall coherence between theory and considered judgments. I thank
Ruth Zimmerling for helping me notice this problem.
justice in several other ways. This example shows that in some cases
choosing short-term justice-enhancing reforms could forever block even
higher justice-enhancing reforms in the future.
Path-dependence considerations are therefore relevant for compara-
tive assessments of justice. We should be mindful of the possibility of
long-term, indirect paths to institutions with better justice performance.
This does not, however, by itself imply that we should have in mind the
perfect design. Arriving at this conclusion requires that we differentiate
between several meanings of the term “perfect design” (section 4.2)
and the related notions of “feasibility” and “circumstances” (4.1).
10
Rawls claims that his proposal for national and international institutions attains a
balance between feasibility and justice performance (1999, 2; 2001, 13). He also claims
that his proposal is the best indirectly feasible design (2001, 5).
5. Conclusion
I have argued here that identifying the perfect design could contribute
to comparative assessments of justice. It is true that these assessments
11
I thank an anonymous referee for prompting me to clarify this point.
Acknowledgments
I thank Marcelo Alegre, Corrado Fumagalli, Florencia Luna, Pablo
Gilabert, Ignacio Mastroleo, Julio Montero, Eduardo Rivera L opez,
Andres Rosler, and attendees at the Colloquium at the Institut f€ur
Politikwissenschaft at Mainz University and the Political Philosophy
Group (GFP) Colloquium for helpful comments. Research leading to
this article was financed by the Argentine National Research Council
(CONICET).
References
BonJour, Laurence. 1985. The Structure of Empirical Knowledge. Cam-
bridge, Mass.: Harvard University Press.
Boot, Martijn. 2012. “The Aim of a Theory of Justice.” Ethical Theory
and Moral Practice 15:7–21.
Cohen, Gerald. 2001. Why Not Socialism? In Democratic Equality:
What Went Wrong? edited by Edward Broadbent, 58–78. Toronto:
University of Toronto Press.