[Note for bibliographic reference: Melberg, Hans O. (1997), How vague may a theory
of justice be?, www.oocities.org/hmelberg/papers/971105.htm]
How vague may a theory of justice be?
Reflections on the criteria for a good theory of justice using
Rawls as a starting point
by Hans O. Melberg
1 Introduction
What do we want from a theory of justice? First of all we want it to give us clear and
unique answers to the question "what is just?" in as many situations as
possible. Second, we want the answers to be good. By good answers I mean answers that
correspond to our reflected intuition (but not necessarily our immediate intuition).
However, a good answer is not only intuitive, we could also demand that the answer be
practically feasible, and - more controversially - stable. In sum, I shall judge theories
of justice on two criteria: Does it give us answers (clear, unique) and are the answers
good (intuitive, feasible, stable)?
In this essay I shall apply the above criteria to evaluate Rawls' article Distributive
Justice. Before I do so, I need to present a short summary of his theory.
2 Rawls
Rawls' arguments proceeds through three stages. First he defines an original position,
that is a hypothetical situation in which we do not know our own position in society (the
veil of ignorance assumption). Second, he presents two principles that he thinks rational
people would choose in this original position. Third, he discusses the concrete
consequences of these principles. It is important to remember these three stages, since
when I criticise Rawls for being vague it is not because the principles are vague
(the second stage), but the last stage: the concrete implications.
The two principles Rawls think people would choose are as follows (quoted from Bojer p.
2-3, originally in Rawls, 1971, p. 60):
1. Maximize liberty : "Each person is to have an equal right to the most extensive
basic liberty compatible with similar liberty for others."
2. The difference principle : "... social and economic inequalities are to be
arranged so that they are both a) reasonably expected to be to everyone's advantage, and
b) attached to positions open to all"
2.1 Attack method
At this point, we can disagree with Rawls in two ways. First, we can agree with the
procedure but disagree with the principles i.e. one could argue that people behind the
veil of ignorance would choose different principles than those proposed by Rawls. For
instance, one might argue that Rawls assumes too much risk-aversion; people might be
willing to trade a very small risk of something bad against a large increase in improved
average life-prospects. I shall discuss this further in the section in which I evaluate
the intuitive appeal of Rawls principles.
2.2 Attack substance
The second line of attack, would be to disagree with the procedure. Communitarians like
Michael Sandel, think that the procedure presupposes individualism and eliminates too much
information about the individual (like race). Others, like R. Dworkin, have argued that
hypothetical consent does not generate legitimate obligation. The fact that I would have
agreed to something need not imply that I am bound by this if the actual situation arises;
Even if I were willing to sell you a painting for $100 on Wednesday (when I did not know
the real value of the paining), it does not follow that I am obliged to sell the painting
to you for $100 on Thursday (when I have received information that the painting is worth
$5000).
2.3 Most just society = best society?
I am not going to go further into these debates sine they are not central to may main
criticism: the vagueness of the specific implications of Rawls theory (note: not the
vagueness of the principles themselves). I do, however, want to make a general comment on
Rawls approach concerning the nature of his theory.
While reading Rawls I kept wondering whether what he was describing was a theory of
justice or a general theory of the morally most acceptable society. Of course, many people
would argue that the two are the same: the morally most acceptable society is that society
which is most just. As Rawls writes: "Justice is the first virtue of social
institutions, as truth is of systems of though." Except for the weak argument by
analogy (even if truth is the first virtue of a system of thoughts it does not follow that
justice is the first virtue of social institutions), my worry is that this focus on
justice leads us to ignore other potentially important values that we also want satisfied,
for instance liberty, material well being, peace and good health. Rawls reply would be
that these values are included in his theory - indeed, his first principle was to maximize
liberty. It may sound pedantic (and it may be!), but I do not know whether liberty should
be viewed as a sub-field of the theory of justice.
The ultimate question is this: Can all ideals be reduced to one deep ideal about
justice, or are there alternative deep values that cannot (or should not) be reduced to
each other. [It is easy to construct examples in which the "shallow" ideals of
justice conflict with efficiency e.g. there could be a conflict between justice (punishing
war criminals) and peace in Bosnia. However, the difficult question is whether the
conflict still exists when we use the "deep" principles]. I do not know. (See
the essay Anti-justice for some concrete examples). What I do know is that if all
ideals cannot be reduced to a question of justice, then we might be willing to sacrifice
some justice in order to satisfy some other ideal, and Rawls' lexicographic priority of
justice would be wrong: It need not be true that "a society is rightly ordered"
when it is just (p. 320).
3. Does Rawls give us answers?
Whether one agrees with the method or not, one criteria for a theory of justice is that it
is able to give specific advice to concrete questions. In the following I shall argue that
Rawls theory is both restricted and vague.
3.1 Comprehensiveness
First, Rawls explicitly restricts his theory to deal with "the basic structure of
society", that is the design of the social, political and economic institutions that
determine our life-prospects. Thus, Rawls' theory cannot, and is not intended to, answer
questions like "Is it just to allocate students homes on the basis of the number of
years studied, or should the student homes go to those who are new in town?" Of
course, one might try to use the same procedure on these question (what would you choose
if you did not know your own situation?), however this is not the topic of Rawls' book.
Does the restricted scope of the theory significantly reduce its appeal? And, is it
possible and desirable to create a theory of justice for "big" decisions which
is isolated from a theory of justice governing "small" decisions? I do not know
the answers. What I do know, however, is that most people most of the time face small or
medium size decisions, and hence need a theory for these issues . Exactly how should
scarce kidneys be allocated among needy dialysis patients? Exactly how large should the
minimum wage be (if there should be one)? These are the real questions facing us in
everyday life, and Rawls theory is not intended to answer them. Moreover, by
distinguishing between "small" and "big" decisions he ignores one
potential method of constructing a deep theory of justice. By studying how people decide
on small issues we might find common principles behind these decisions across different
issues. These common principles could then be said to constitute the deepest principles of
justice according to our intuition. Note that this procedure for finding our "deep
intuitions" is different from Rawls' method of finding justice through what we would
choose under uncertainty. [For more on this, see Jon Elster's book Local Justice]
3.2 Concrete answers?
Having noted as a weakness the restricted scope or Rawls theory, I now ask whether the
answers he gives on the issues which the theory is intended to cover (the basic structure
of society), are clear and well-defined. My short answer is no, and this, in my opinion,
is an important criticism of Rawls' theory. The answers are simply too vague to give me a
clear idea of what kind of society Rawls want. many of his statements are so general it is
difficult to disagree, and even more difficult to use as a guide for constructing a
society. To justify this argument, I now turn to some concrete examples.
Let me start with Rawls' general recommendation that "One has to set free market
arrangements within a framework of political and legal institutions that regulates the
overall trends of economic activity and preserves the social conditions necessary for
justice" (p. 336). Now, the argument is not totally empty (there are libertarians who
argue that the state has no right to interfere with the workings of the market,
particularly that the state has no right to alter the income distribution). Yet, for the
waste majority of people the real question is not whether the state should regulate the
economy, but the extent to which it should do so. There are no countries in the world
today that come close to Nozick's minimal state (even Hong Kong is not a good example!).
What kind of institutions are Rawls calling for? At one point he seems to recommend the
creation of four large agencies each dealing with the following issues: allocation,
stabilization, transfer, and distribution (p. 347ff). In the section about feasibility I
shall discuss this concrete proposal in more detail. On a smaller scale, Rawls discusses
education (p. 347), inheritance and gift tax (p. 349), social minimum (p. 351), and a
just-savings principle (p. 352). However, he has few concrete recommendations, often
admitting explicitly that "on this question the theory of justice has nothing to
say" (about the exact level of inheritance tax, p. 349), or that "All of this
is, unfortunately, terribly vague" (about the level of savings required by the
just-savings principle, p. 354). Although his honesty is admirable, it does not reduce the
force of the conclusion that a theory of justice which does not give concrete advice is of
little use.
In sum, Rawls' theory is both restricted and quite vague. By vague, I do not mean his
definition of the original position or the actual principles of justice (the first two
stages). Both are extensively discussed, for instance he explicitly discusses four
possible interpretations of his two principles (natural liberty, natural aristocracy,
liberal equality, democratic equality), and he clearly favour what he calls democratic
equality. but the actual practical implications - which is what I am really interested in.
4 Are Rawls' answers good?
Even if we assume that Rawls give clear and concrete advice on the basic structure of
society, we must ask whether this advice corresponds to what we think is just. We must
also ask whether the advice is feasible in the real world given unavoidable limitations ,
such as scarce information, resources and weaknesses in human nature. Finally, one might
ask whether the theory is stable, or what one might call "dynamically feasible"
i.e. Would a society based on Rawls' principles be stable over time?
4.1 Intuitive?
When we try to judge the intuitive appeal of Rawls, we once again encounter the problem
that he gives few clear and concrete advice. Hence, I shall have to focus on evaluating
the intuitive appeal of his principles, not the implications of these principles.
Rawls himself discusses whether his principles are compatible with our common sense
notions redress, fraternity, liberty, reciprocity, desert, rights and Kant´s principle of
not using people as means. I have little to add or criticise in this discussion, except
perhaps to note my weak belief that some of his interpretations are a bit strained (e.g.
Is Rawls' interpretation of fraternity the most intuitive interpretation?, see p. 340).
I do, however, have two problems with his principles: its extreme risk-aversion and the
problems created if people care about relative incomes as well as absolute incomes.
4.1.1 Risk-aversion
According to the difference principle inequalities are only justified if they work to the
advantage of those who are worst off (raise their life prospects). Is this really
plausible? Would not people in the original position be willing to trade a small risk of a
bad result in return for a large increase in average life-prospects? Real life experience
with choice under uncertainty, at least, suggests that people do not exhibit extreme
risk-aversion.
One might try to rescue Rawls by arguing that the small choices we face in real life
are not like the big choices he is dealing with, and that we cannot use evidence of
risk-taking in lotteries to infer that a person behind a veil of ignorance would accept
trade risk for a higher average. We could also point to a theorem by Arrow which states
that under radical uncertainty (as opposed to risk) - when the probabilities of the
various states are simply unknown - it is rational to make a decision based only on a
consideration of the best and the worst outcome. Still, I am unsure whether it is
intuitively plausible that people will be as risk-averse as Rawls suggests.
A third line of argument to defend Rawls, is to say that his theory does not focus on
the individual which is worst off, but the group that is worst off.
Sometimes this seems like the most intuitive interpretation of Rawls (e.g. when he speaks
of his individuals as representatives of groups), but he also writes that when
chain-connection and close-knitness do not hold "The stricter interpretation of the
difference principle should be followed, and all inequalities should be arranged for the
advantage of the most unfortunate even if some inequalities are not to the advantage of
those in the middle positions" (p. 330). The risk-aversion implied by this statement,
I think, is not intuitive.
4.1.2 Relative incomes
My first criticism was that Rawls places too much weight on those who are worst off. My
second criticism is that his principle permits those who are well off to be so well off
that it may be counterintuitive. The problem is this: Let us assume a change in the basic
structure would increase the life-prospects of those who are well off by 200 units, and
those who are least well off by 1 unit. Admittedly this is a bit theoretical, but it
brings out the following implication: According to Rawls individuals do not care about
relative differences. He thinks that the individuals in the original position would accept
this change to the basic structure, since it raises the life-prospects of those who are
worst off, although it raises the prospects for those who already are well off a lot more.
This, I think is counterintuitive. People do care about their relative positions, not only
their absolute level
One might argue that the reason people care is that they are envious, and that this is
an unethical preference that should not be counted. However, I do not think all preference
for greater equality are inspired by envy. One might, however save Rawls by arguing that
self-respect is a function of your relative income, and self-respect is one of Rawls' most
important primary goods. Hence, in practice the above change would not be accepted by
Rawls. Yet, this is a somewhat weak defense sine it smells a bit like the utilitarian
defence against unethical acts (torturing children, the survival lottery): that in
practice they are prohibited, but not in principle. Similarly, Rawls theory only
indirectly allow the very common concern about relative levels. This sounds
counterintuitive, although in practice it need not make a big difference.
4.2 Feasibility
Assume now, that Rawls answers are clear and intuitive. We are still left with one
problem: Are they feasible? In other words, can the principles be implemented in the real
world?
My first criticism would be that it is difficult to implement the principles because we
do not know enough about how society work (or indeed, what people actually would choose
behind a veil of ignorance). Even if we agree on the difference principle, we need to know
the causal mechanisms in society in order to work out what kind of system the principle
implies. Rawls simply assumes that we know all these facts in the original position, but
this is of little help to people who live in the real world and have to deal with real
uncertainty.
A second criticism of the feasibility of Rawls' proposal concerns his neglect of
government failure. For instance, consider the proposal that "the economy must be
surrounded with the appropriate framework of institutions ..." (p. 357), and the
somewhat more concrete (but still vague) to create four branches of government, each
dealing with, respectively: allocation, stabilization, transfer, and distribution. In
theory this may work fine, but given human weaknesses like the desire for power and money,
one also needs to discuss whether the relatively extensive regulation of society implied
by Rawls, is workable.
Once again Rawls theory can be saved simply by postulating that people in the original
position knows about this weakness, and hence will not create a state structure in which
individuals would be given the opportunity to abuse their power. This is a good defense,
but it also suggests that some of the implications commonly drawn from Rawls theory (also
by himself) may be wrong: They may decide that large four large agencies with large powers
intended to create a just society, is too risky a project; It creates a structure of power
that too easily can be abused by power hungry individuals.
Once again, the answer can only be found by empirical investigation: To what extent is
human nature guided by thirst for wealth and power, and to what extent does the various
structures allow themselves to be manipulated be these individuals. I do not know the
answer, but I do know that a discussion of these issues should be included in the
discussion of the best possible society.
4.3 Stability
Stability is really a sub-class of feasibility. My question is this (which is a little
different from Rawls' discussion of stability): Can we expect a society based on Rawls'
principle to be stable? My argument is that Rawls is wrong to argue that his system is
stable since ex-ante stability is different from ex-post stability.
If two people agree on something before we know our own position (ex-ante stability),
this agreement need not be stable after we have been given the positions (ex-post
instability). To be concrete, Assume that I am stronger than you, but that you know where
a large treasure is hidden. Assuming that you are not strong enough to dig up the treasure
yourself, we make an agreement that if you lead me to the treasure and I dig, we shall
split the sum. Obviously, after finding the treasure, I - being selfish - may decide not
to honour my agreement, and - being stronger than you, I simply take the whole treasure.
In sum, only credible ex-ante agreements will generate stable ex-post solutions (see game
theory for more on the concept of credibility).
At one point Rawls indirectly comments on the above problem when he writes that
"the more advantages representative individual cannot say, and being reasonable,
will not say, that he deserves ... to acquire benefits in ways that do not contribute to
the welfare of others" (p. 344, my emphasis). The problem is that claims in the real
world need not be based on reason; I may not give a good reason why I should get more than
you, but I may still demand more than you simply because I am selfish, and I may get away
with it, since the world is not governed by not by reason but also by power and force.
This makes the society constructed by Rawls principles unstable (though, the principles
themselves need not be instable. See Rawls, p. 345).
5 Conclusion
Some might argue that I have been too critical of Rawls. True, the theory is restricted
and vague, but it is not empty. True, the theory may not be intuitive on all accounts -
assuming too much risk-aversion and no concern for relative positions. However the general
idea that which is chosen behind a veil of ignorance seems intuitive. The theory may
require too much information to be feasible, and the specific system of government may be
too optimistic with respect to human nature. Yet, most people seems to agree that the
market has to be regulated in some fashion. Lastly, his theory may not be stable since
people are not motivated solely by reason, but by brute selfishness. However, reason is
also one motivating factor behind human behaviour since we have a need to justify what we
do in terms other than pure self-interest. Lastly, I have not presented a superior
alternative to Rawls.
These are all valid positions against my arguments, but it still remains that it is
difficult to find concrete guidance on concrete issues (even on the basic structure of
society) using Rawls theory. This must count as a serious weakness. To argue against this,
one might say that I am demanding too much. I should not insist on concrete implications
because it is too early to draw the these implications. Hence, a discussion of principles
is in itself valuable - since it gradually builds a theory which - at some later stage -
can be used to draw concrete implication. it is difficult to evaluate this "long
run" justification. Time will tell if the theory will every evolve into something
more concrete. Until then I reserve my judgment.
A note on sources
My main source for this essay is Rawls' chapter "Distributive Justice" in Edumnd
S. Phelps' book Economic Justice: Selected readings, Penguin education, 1973, pp.
319-362. Another source is Rawls (1971) A Theory of Justice (which I have only read
partially). I also used Bojer (1995), "Hva Rawls egentlig mente", Sosialøkonomen,
no. 11, pp. 2-6. Lastly, I base some of my comments on lectures by John Grey and A. Swift
(Oxford University), W. Kymlicka's excellent book Contemporary Political Philosophy,
and Jon Elster's Local Justice.
[Note for bibliographic reference: Melberg, Hans O. (1997), How vague may a theory of
justice be?, www.oocities.org/hmelberg/papers/971105.htm]