JOHN RAWLS: A THEORY OF JUSTICE

D. J. BENTLEYt

The object of this piece is to offer a few reflections, provisional

and far from rigorous in character, on Professor Rawls' A Theory of
Justice. In essence, I shall be saying no more than "this seems difficult,"
or "important" or (rarely) "unsatisfactory." Rawls is, of course, a
philosopher; I am trained as a lawyer, and I can only ask that the
following remarks be received de bene esse.

I
It may be helpful to those who are not familiar with Rawls to
plot his position against a better-known moral theory, namely utilitari-
anism. At any rate, on this side of the Atlantic, where no modern
charter contains the compact under which we consent to be governed,
the defence of truth and justice has most often in the last hundred
years appeared to be in the hands of the utilitarians. This raises two
immediate questions: first, is justice a "special department" of moral-
ity? And second, do we by this mean to commit ourselves further to
the position that the province of justice is not (whatever be the case
for the rest of morality) governed by the principles of utility? After
looking at these questions, we can go on to assess the particular solution
proposed by Rawls.
It is perfectly possible to be, broadly, an utilitarian and yet to
give the principles of justice a special status. For instance, in The Con-
cept of Law, Professor Hart carefully distinguishes "fairness" (which
roughly covers justice) from morality in general.' And Hart, I suppose,
is a modern utilitarian. Why should the two ever have been confused?
The explanation can be found by looking at Mill's chapter on the con-
nection between justice and utility.2 Mill rightly saw the idea of justice
as an obstacle to the simple test of utility in assessing right and wrong.
People, that is, went on referring to this rather old fashioned notion,
and seemed remarkably attached to it, despite being urged to look to
expediency instead. But to Mill, this notion of justice had to be either
irrational or the reflection of the operation of a perhaps concealed
principle of utility.
I Tutor and Lecturer in Jurisprudence, Christ Church, Oxford. B.C.L. 1957, MA.
1960, Oxford University.
I H.LA. HART, Tim CoNcEPT or LAw 153-55 (1961).
2 J.S. MiLL, U=TmaNixTr, LiBExR, w REPRESENTATIVE GovmanmNT 51-80
(1951).
 RAWLS' THEORY OF JUSTICE

To arrive at this position Mill took one sure and one false step.
As he rightly says, "[t] hat a feeling is bestowed on us by Nature, does
not necessarily legitimate all its promptings. '8 To feel that something
is just or unjust cannot be the test of its justice or injustice unless, as
Mill says, one believes that justice is some special quality ("intrinsically
peculiar") which can only be perceived by what amounts to a "special
revelation." This is a pretty familiar sort of argument of course. But
then Mill offers a false choice for us: either the feelings of justice are
sui generis perceptions like taste or smell, so that debates about justice
are to be resolved by asking "does it feel just?," as one might ask
whether a cup of tea has enough sugar, or justice is only a combination
of certain other qualities of a thing "presented under a peculiar aspect."
Mill goes for the latter, and his treatment of justice appears at times
more akin to social psychology than moral philosophy, for he regards
the powerful sentiments evoked in this area as what are really in need
of explanation. The resulting index of justice is simple: utility.
In fact, the choice Mill puts before us-utility or the special
quality-is, at least as he presents it, too restricted. There is a middle
way, which a theory like Rawls' exemplifies, though no doubt other
non-utilitarian but rational theories of ethics can be devised. This view
accepts readily that justice is "a thing intrinsically peculiar, and distinct
from all . . . other qualities"4 but denies that it is to be discovered
by the "peculiar revelation" which Mill rightly distrusts. Briefly, this
sort of theory would assert (a) that justice is a special department of
morality; and (b) that what makes it special is not simply the ground
it covers (by which I mean fair procedure, etc.) but also the kinds of
arguments that are deployed over that territory. It is one of the more
entertaining consequences of this divergence of approaches that when
Mill considers the thoroughly Rawlsian maxim "so act that thy rule of
conduct might be adopted as a law by all rational beings," he turns
Kant into a utilitarian:8 but what else could Mill have done?
It is now time to look a little more closely at the way Rawls treats
justice; after that, I shall take up some points which, I am sure, are
both important and yet by no means obvious.

II
Rawls sets out his main thesis at an early stage; much of the
rest of the book is concerned with refining the early broad statement
3Id. 51.
4 Id. 52.
GId. 64-65.
1072 UNIVERSITY OF PENNSYLVANIA LAW REVIEW [Vol. 121:1070

of principle, developing the philosophical arguments on which his

case rests (and knocking down objections) and applying his theory
to particular problems. There is for Rawls a preliminary distinction
to be drawn between the concept of and various conceptions of justice.
Men recognize the concept of justice when they "understand the need
for, and are prepared to affirm, a characteristic set of principles for
assigning basic rights and duties and for determining what they take
to be the proper distribution of the benefits and burdens of social
cooperation."" A conception of justice, then, consists of any set of
definite principles selected as being correct. I must confess to some
doubt here: can justice, and perhaps other fundamental notions in
political philosophy, be so easily and unequivocally distributed be-
tween concept and conception?' A lot perhaps depends on the force
and fineness of the term "characteristic."
In any case, what Rawls proposes is a conception of justice. What,
he asks, are the basic principles that "free and rational persons con-
cerned to further their own interests would accept in an initial position
of equality as defining the fundamental terms of their association"? 8
There are two: "the first requires equality in the assignment of basic
rights and duties, while the second holds that social and economic
inequalities . . . are just only if they result in compensating benefits
for everyone, and in particular for the least advantaged members of
society."
This conception, which Rawls calls "justice as fairness," may
appear to be obvious, but Rawls is after big game here, and his pre-
mises cannot be accepted lightly. Rawls is not simply engaged in a
dispute over definition with his academic opponents, nor is he con-
cerned to defend a generally acknowledged good like liberty of expres-
sion. Rather, he is trying to produce a set of principles which anyone
who has felt the pull of justice (that is, the concept of justice) would
acknowledge. Sometimes one genuinely does not know what would be
the right thing to do, and Mill and Rawls could well disagree. In such
a case, by ranking justice as the prime virtue, and in turn ranking
his two principles of justice, Rawls is arguing that alternative solutions
to conflicts of principle are incorrect, whether or not they are presented
in the name of justice.
We are being offered, then, a social contract theory in place of

61J. RAwILs, A THEoRY OF JVSTICE 5 (1971). See also id. 7-11.

7 Cf. Dworkin, The Jurisprudence of Richard Nixon, N.Y. R v. or Boo3Ks, May 4,
1972, at 27 (employing such a distinction).
8 J. RAwLs, supra note 6, at 11.
9 Id, 14-15.
19731 RAWLS' THEORY OF JUSTICE

the long-lived rationalist ethic of utilitarianism. It is of course a hypo-

thetical contract for this is an exercise in moral philosophy, in the
criticism of institutions, and not an attempt to explain how social
arrangements have actually come about. Thus we are not entitled to
ask, as did Hume, "what authority any moral reasoning can have,
which leads into opinions so wide of the general practice of mankind."
But there is a further question which also turns on the relationship"
between statements of fact and moral judgments: how does the mere
fact that certain principles of the moral contract may be selected,
insure that those principles are the correct ones? This goes to the
heart of "justice as fairness," and in light of this question we may note
immediately two features of Rawls' theory. First, it does not attempt
to embrace the whole field of morality. "Rightness as fairness" must
wait its turn, although Rawis seems to consider that such generalisa-
tion would be worth a try. Second, and this seems to me to be more
serious for those trying to find their bearings in the Rawls universe,
his concern is primarily to give an account of the kind of ideal social
arrangements that adherence to his principles would engender.
This leads Rawls to devote a great deal of attention to very
fundamental issues of social organization (hierarchies and savings in
the design of institutions, etc.) and rather less attention to what
could roughly speaking be called jurisprudential problems, which fall
into what he terms "partial compliance theory", that is, the adapta-
tion of his theory to the world we actually live in, where people are
not necessarily free, rational or just, where disobedience exists and
punishment may be necessary.'0 As a consequence of this concentra-
tion on "strict compliance theory," the follower of Rawls is left with
a significant difficulty in explaining how a theory derived from the
postulated choices of men in an imaginary state of affairs can have
any applicability to conduct in the real world. I shall return to this
difficult question later.

III
I want now to look a little more closely at certain aspects of this
theory. First, I should like to go back to the question of what makes
this conception a moral one: why are we entitled to say that what men
would choose is what they are right to choose? Briefly, the answer is
that it is not what they choose, but the way they choose, that insures
the morality of their choice. The terms of the contract are settled by
men in a situation in which they know very little about themselves
10Id. 8-9.
1074 UNIVERSITY OF PENNSYLVANIA LAW REVIEW [Vol. 121:1070

and their circumstances; in Rawls' terms they choose behind a veil

of ignorance. In this position they will settle only for a society in
which basic fair treatment is guaranteed to each individual. Rawls'
conception is thus deliberately prepared to meet a basic question
which can be asked about any society: why should a person be ex-
pected to assent to its arrangements-what do they offer him? In this
way the theory is sharply differentiated from any utilitarian theory,
according to which the pursuit of happiness for the greatest number
precludes the possibility of guaranteeing fairness of treatment to the
individual. Rawls can argue then that the utilitarian. claim to count
each person as one, no one to count for more, no one for less, is
thoroughly misleading. Prior to it should be a commitment to the
notion of individual personality, too basic to be at the mercy of a
felicific calculus.
Two points need consideration. First, how does Rawls know that
this is how rational beings would behave in the "original position"?
I think the answer must be that Rawls has loaded the odds fairly
heavily against any other choice. A gambler would be taking a fearful
risk, not only for himself but also on behalf of his children and his
children's children, if he turned out to be born on the wrong side of
the tracks in a fundamentally unjust society. As a way of eliminating
some rather extreme options, this approach has its attractions, but
in less drastic situations its role is less clear. Were the gambler assured
of some measured amount of fairness of treatment, he might be will-
ing to risk any additional fairness. This goes, of course, to the mini-
mum definition of fair treatment. But the point is that that standard
might be sufficiently low that Rawis' two subsidiary principles would
not follow as a matter of course.
The second comment that I would make about the alleged ethical
superiority of Rawlsian over utilitarian justice is perhaps an index of
the difficulty that I find in interpreting the utilitarians rather than a
point against Rawis. Obviously if Rawls and the utilitarians really
diverge, this ought to show up in their treatment of rights against
society. At first sight this appears to be the case. Mill, for instance,
considers that justice is closely bound up with the notion of rights,
and that "[t] o have a right [is] to have something which society ought
to defend me in the possession of. If the objector goes on to ask,
why it ought? I can give him no other reason than general utility."'1
Now I think it is at any rate arguable that this dictum should not be
treated as conclusive proof that Mill does not understand what rights

11 J.S. MiL, supra note 2, at 66.

 RAWLS' THEORY OF JUSTICE 1075

are or rate them highly enough. My reason for making this suggestion
is that Mill has also set out a list of situations in which the precepts
of justice apply, and identified their common factor in the notion of
a right, or rather, rights. For Mill, to act morally one must act accord-
ing to a universalisable principle, so that there would be no way of
subverting the rights of any individual that would not also justify the
invasion of the rights of the other members of society. And Mill is
clear that the disutility of this state of affairs would be overwhelming.
(Which comes very close to the position of one who relies on an appeal
to "the very notion of a right.")
In short, where Rawls relies on a hypothetical contract, Mill
says simply that the world would be a much nastier place to live in,
but for the principles of justice.' 2 At least this seems a possible way
for the utilitarian to argue, even if I have read too much into Mill.
I now wish to turn to some matters which I find very difficult, but
which are, I am sure, essential ingredients in any critical account of
Rawls' theory. They are connected to the notion of the "original
position," and are similar to the sort of criticism of Rawls that points
to the unreality of the choices made by wholly ignorant contracting
parties. But I am not so concerned with the difficulty of knowing what
such people-and hence such choices-would be like. My problem
arises from the distinction that Rawls draws between strict and partial
compliance theory, which I have already mentioned. Rawls deals for
the most part with a perfectly just society in which "[e]veryone is
presumed to act justly and to do his part in upholding just institu-
tions."' 3 Partial compliance theory, on the other hand, treats injustice.
"It comprises such topics as the theory of punishment, the doctrine
of just war, . . . the justification of the various ways of opposing un-
just regimes" as well as "questions of compensatory justice."' 4 Only
by considering the just society, however, does Rawls think we can
hope to find the answers to the problems of our own imperfect societies.
I confess that I find this hard to accept confidently. For one thing, it
is not clear how this perfectly just society would work in practice.
This is not as silly an objection, I hope, as it may at first sound. It is
one thing to criticise existing institutions in light of standards external
to those of the society in which the institutions are to be found 3 5
12 See Mill's account of the "useful accommodation of language" by which "the
character of indefeasibility attributed to justice is kept up, and we are saved from the
necessity of maintaining that there can be laudable injustice." Id. 79.
13 J. RAwLs, supra note 6, at S.
14Id.
15
H art does this when making out a case against Devlin. See H.L.A. HART, LAW,
Lma'v" Aam Moam= (1963).
1076 UNIVERSITY OF PENNSYLVANIA LAW REVIEW [Vol. 121:1070

There is also an understandable desire in social theory to deal some-

times in "ideal types'-after giving the reader adequate warning of
what is going on. In both these cases it is then possible to measure the
distance between reality and the ideal state of affairs. But Rawls seems
to be going further than either of these critical methods. He appears to
be trying to lay down a pattern for a possible society; not in the sense
of a possible amendment to an existing society that can plausibly be
defended on the basis of knowledge of existing societies (as one might
hypothesize that to relax the law on obscenity would not really harm
the balance of payments) but in a stronger sense altogether, Rawls
is saying: "this would be an ideal society." Yet surely societies are
always the product of actual events (and perhaps even the visible
witness of inevitable laws)? The enterprise has the same sort of odd
ring to it as the notion of the perfect man. Even astronauts are human.
At two other places in A Theory of Justice, I have a further difficulty
with the absence of a theory of how society works, and I again feel
that this leads to trouble when one turns to the question of how society
ought to be made to work.
The subject of justice, Rawls says, is "the basic structure of
society, or more exactly, the way in which the major social institutions
distribute fundamental rights and duties and determine the division of
advantages from social cooperation. By major institutions I under-
stand the political constitution and the principal economic and social
arrangements."' 6 The key word is "fundamental." Even under the re-
vealed reason of a written constitution, the kinds of interests that this
term covers seem subject to debate (though no less worth defending
because of that). All of this, of course, in Rawls' own system, must
involve partial compliance theory; though perhaps "the way" should
not be taken too literally. In any case, I find a sort of circularity here,
only partially concealed by the blend of the conceptual and the empiri-
cal in the passage in question. The trouble is surely not that this kind
of objection merely shows my inability to distinguish the role of the
moral philosopher from that of the political economist. It is rather
that there may be legitimate doubt as to the appropriateness of purely
speculative systems in this area.
That is one example of the difficulty that I have found in pinning
down the thesis of this work. It recurs when Raws discusses the
principle of the rule of law. "[T]he law," he says, "defines the basic
7
structure within which the pursuit of all other activities takes place.")

16 J. RAwLs, supra note 6, at 7.

17 Id. 236.
1973] RAWLS' THEORY OF JUSTICE

How does this fit the statement that I have just quoted of the basic
structure of society? Of course, if legal systems have no gaps, then
in some sense all our activities can be placed under some rule of the
system; liberty is no longer the silence of the laws. But any serious
account of the basic structure of a society would presumably take that
for granted and would be concerned with the actual norms, whether
legal or not, according to which activity is in practice directed, and
additionally, where the norms are legal, with the function of the norms
as distinct from their content (as, for instance, Karl Renner uses the
notion of the functional transformation of norms). 8 What makes laws
basic in the usual case is simply that they do as a matter of fact lie
at the root of certain societies. 9 But the law comes a bad second in
comparison to other factors that necessarily determine this quality of
institutions, though there is a whole possible area of study which could
trace in given cases just how bad.
A last instance of what I find to be an insecure foundation, some-
where between heaven and earth, can be found in the chapter on
Distributive Shares,2" in which Rawls attempts to set out an arrange-
ment of institutions-neutral as between private property systems and
socialism-that will fulfill the requirements of his theory of justice.
There is no doubt that this must be taken as being a very important
and explicit statement of Rawls' views. But what status does Rawls
intend his precepts for the institutions of the just society to enjoy?
Are they supposed to work? This is not at all straightforward. One is
surely in the realm of partial compliance theory here, yet Rawls is
defining standards rather than predicting contractual arrangements,
and this apparently on the basis of the original position, that is, strict
compliance theory. Nor am I clear as to the usefulness in practice of
the Archimedean viewpoint"- from which society is to be measured
against Rawlsian justice. On the one hand, one is estopped, as it were,
from arguing that the fundamental arrangements of society simply
will not do, since the theory is neutral at this point. But one is still
expected to be able to show (remind?) erring members of society
that their desires and aspirations are incompatible with justice. There-
fore I repeat: anyone who is going to hitch his wagon to Rawls will

I8K. RENNim, THE INSTITUTIONS Or PRIVATE LAW AND THEm SociAL FUNCTIONs
81-95, 205, 251-60 (1949).
19Rawls' comparison with the rules of a game is a giveaway in this regard. See
J. RAwLs, supra note 6, at 235-43. To talk of what is basic, unless it is treated in an
ad hoc way, can be misleading.
20 Id. 258-332.
21Id. 261.
1078 UNIVERSITY OF PENNSYLVANIA LAW REVIEW

have to think very hard about the way the concepts he is using are
related to each other and the world. 2

IV
One last point may be made before I try to sum up my comments.
This is a grand theory, and perhaps it tries to do a little too much at
times. It is true that Mill puts all questions of justice under the rule
of utility, and that this may lead to an unsatisfactory or less than
entirely clear account of the problem. But I feel that some of the
questions traditionally considered by theorists of justice fit even less
readily into a scheme based on an original contract. As an example,
I wonder whether it is only the fact that they belong to "partial
compliance theory" that makes such notions as mens rea appear to
fit awkwardly into Rawls' theory. Similarly, Rawls says that "[i]t
would be an intolerable burden on liberty if the liability to penalties
was not normally limited to actions within our power to do or not
to do." 23 But the objection to punishing an unfree man cannot be
that the criminal law thereby operates as a limitation on his freedom
of choice, at least if that is why Rawls here invokes the notion of
liberty, since the man was, by hypothesis, not free to choose. The real
offense to our notions of responsibility and desert is rather that the
connection between the legal and the moral wrong is broken.
CONCLUSION
As I said in my opening remarks, this piece does not aim very
high. In it I have attempted two things: to sketch the way in which
Rawls diverges from the utilitarians, and to record some of the diffi-
culties that I have experienced while following his argument. Briefly,
these latter can be placed under two heads: the problem, stemming
from the notion of the "original position," of relating strict compliance
theory to the real world, and the problem of whether it is desirable to
refer all questions of justice directly or mediately to a single grand
theory. Of one thing I am convinced: Rawls will be ill-served by any
follower who falls to realise that the argument of this book, though
of great power, is neither simple nor self-evident.
22 Cf. the account of the various "branches of government," which is reminiscent of
the eulogies of the working of the separation of powers. Id. 277-84.
23 Id. 237.