Reading Guide 11: Rawls, A Theory of Justice

John Kilcullen

Copyright (c) 1996, R.J. Kilcullen.

First we will read two articles by Rawls and then some extracts from his book, A Theory of Justice (1972).

The first article is from 1958, "Justice as Fairness", Readings, p.241. It is divided into eight sections numbered with Roman numerals. I have underlined some phrases and written in some headings. Read Section I. The first two paragraphs give the reader some preliminary idea of what he will do in the rest of the article, the 3rd and 4th fend off possible misunderstandings. Read now section II. Some comments. First, in these two principles there are in fact four points: (1) that there should be as much freedom as possible; (2) that people should be equally free; (3) that inequalities should be to everyone's advantage, and (4) that the better positions must be open to all (compare "equality of opportunity"). It is as if he supposes that freedom is a good thing and inequality a bad thing.

In later writings Rawls gives a name to the third point, that inequalities should be to everyone's advantage; he calls it the "difference principle". The difference principle allows an inequality provided it is to everyone's advantage. How can inequality ever be advantageous to the person who gets the smaller share? The answer is that inequalities may constitute incentives which increase the size of the cake to be shared, so that the smaller piece may be larger absolutely than an equal share of the smaller cake that would have existed if no such incentive had been offered. This is presumably why, at the top of p.134, he mentions the idea of reward for services contributed.

On the face of it there is some conflict among these principles. The first stresses equality--an equal right to the most extensive liberty--to a liberty which is, however, compatible with a like liberty for all, which again implies equality. Everyone has to have the same amount of liberty, and if someone's liberty is so extensive that others can't have as much, then that person's liberty has become excessive. So what are the inequalities that can be allowed, provided they are advantageous and open to all? In the second paragraph, of p.134 Rawls explains that "The first principle holds, of course, only if other things are equal". This makes it what some philosophers call a prima facie principle, others call a defeasible principle, other call a presumption. The initial presumption is that people are to be equally free, there is an onus of justification on anyone who proposes something that would make freedom unequal. Equal freedom is the starting point, but we can move from there to an unequal freedom if someone can produce a good enough reason. Re-read the first half of that paragraph.

The justification offered for some inequality of liberty might be that it is to everyone's advantage. Notice to everyone's advantage, not just to the majority's or to the advantage of some privileged minority. In later versions of the theory Rawls treats all this a bit differently. He says in his book (1970) that at least in a reasonably advanced society liberty must be as extensive as possible, and equal, no matter what advantages there might be in making liberty unequal. The idea of incentives then has to be applied to something other than liberty--e.g. to income or shares of consumer goods. In this version the "equal liberty" principle is not just a defeasible presumption but a principle that takes priority over advantages that might be got by making liberty unequal. In the book he speaks of "the priority of liberty". But in this 1958 version of the theory the idea of a "priority of liberty" is missing.

Look again at the last paragraph on p.135 ("It should be noted"): He will not allow, as a justification for departing from the starting point of equality, an argument to the effect that it would be advantageous to the majority, or to some elite. The argument can't be in terms of advantages to some outweighing advantages to others.

Re-read the last 9 lines of section II ("It will be a mistake") which makes a link to the next section, and read section III.

In this section Rawls asks us to imagine a certain process. He has described it in very general terms, so that it is possible to imagine it concretely in a variety of ways. Notice at the top of p.137 he reminds us of the ambiguous way he is using the word "person". (Cf. p.134, line 3 and following.) We can imagine these persons as representatives of smaller bodies -- "families", he says, or we might substitute electorates and imagine the body as a parliament making laws. Alternatively, we might think of the persons as individuals and their discussions as taking place among members of a household. If we imagine it the first way, we can interpret some of his remarks as being about "the rule of law"--i.e. the notion that governments and government agencies should act in accordance with stated general laws that apply to everyone alike over a reasonable period of time ("standing" laws). The laws are made by a legislature and afterwards applied by a court. With this interpretation in mind re-read p.138, the paragraph beginning "Since". Now read it with a household in mind. The household has certain ongoing practices--mostly a matter of a division of labour, who does the shopping, who pays the bills and so on, also about which television shows get watched, and whatever else needs regulating. From time to time they have a "complaints" discussion (maybe regularly, more likely casually). But unlike most households (more like the political system) they agree on principles before they get down to particular complaints. The principles are to apply to all alike and for some reasonable time into the future. Of course self interested individuals may propose, will propose, principles that suit them. By the requirement that a proposed principle is to be binding on future occasions and, in addition, the proposition that the circumstances that will obtain in the future are "as yet unknown", Rawls hopes to ensure that fair principles will emerge, principles not tailored to give an unfair advantage to one person over another. This is why he calls this scheme "justice as fairness". He explains this later. In this passage he seems to suggest that the sort of discussion he has in mind might take place in real life. He remarks "such discussions are perfectly natural in any normal society". But in real life it will never really be true that in all relevant respects the future circumstances in which the principles are applied are as yet unknown. Members of a household know that their sexes, their relative ages, and many other things, are not going to change. In developing the later version of the theory Rawls did some hard thinking about which circumstances must be unknown to make the rules fair, and came to realise that the discussion could not be a real-life one; in the later version this discussion becomes frankly an artificial device, a fiction, a hypothetical exercise. It takes place under what he calls a "veil of ignorance" which prevents the hypothetical discussants from knowing certain things about themselves that in real life people do know. We'll come to that later. For the moment he asks us to imagine a pretty realistic discussion in which the differences in bargaining strengths are reduced simply by the uncertainties of the future, so that what emerges is a fair agreement. (If you look back briefly to the end of the opening paragraph of the article, on p.132, notice the reference to the social contact: the participants in this household discussion come to an agreement about the principles by which their practices are to be judged.)

Returning to p.139, notice his remarks on "The two main parts". Underline 7 lines down this paragraph "first part", and 7 lines further down underline "second part". It is not altogether clear which two parts of the previous material he is referring to-- look back through pp.136-139 and see whether you can identify "two main parts". Anyway, the comments he wants to make here on the process are reasonably clear. Justice is a part of morality, a species within a genus. Some features of his story correspond to the generic features of all parts of morality, others are specific to justice. What is specific to justice is that it comes into play when people press conflicting claims upon one another. Elsewhere Rawls refers to Hume's remarks (POL167 Readings pp.344-5) on the irrelevance of justice to the mythical golden age when human beings never quarrelled and when there was no shortage of the things people want. Rawls uses the phrase "circumstances of justice" to refer to the circumstances that make justice relevant--the selfishness or confined generosity of mankind and the scarcity of the things human beings want.

Re-read the next paragraph, "Thus the two parts...". Here Rawls suggests that his two principles (or was it four), the ones laid down at the beginning of section II (p.133), would naturally be agreed on as a result of the process described on p.138.

Finally, re-read the first paragraph of Section III, and now re-read the last paragraph of this section, "These remarks are not offered as a rigorous proof", p.140. So what is the proof being sketched here? It is akin to arguments in a branch of economics or mathematics that was attracting the attention of moral and political philosophers in the 50s, namely the theory of games. This consists of arguments designed to establish which tactic or game plan will yield the greater "pay-off" for a player confronted with a certain clearly defined simple situation. You may have met with "The Prisoner's Dilemma". Game theory aims at rigorous proof: it is not just an empirical prediction of what a player is likely to do, it aims at a proof that this or that would be the best tactic. Rawls has specified certain motives and certain psychological traits, which determine what counts as a good "pay-off", and has put his hypothetical profile into a determinately structured (and somewhat artificial) situation, and he suggests that to accept his two principles would be the best tactic or game plan such people in such a situation could adopt. However, he doesn't present a strict proof in game-theoretical style. You might expect that Rawls or someone else might later publish a formal proof, but as far as I know no one ever has--as he acknowledges in A Theory of Justice, giving a rigorous proof turns out to be a very difficult matter. So we've got a sketch and we may agree that the two principles look intuitively like a good solution to the problem situation Rawls has put his hypothetical people into.

Now read section IV. In this section Rawls explains the relationship between his theory and earlier "social contract theories", such as the theories of Hobbes, Locke and Rousseau. The reference to the Greeks might remind you of Glaucon's account in Plato's Republic, 358e-359b (POL167 Readings p.53). Notice in the footnote on p.143 the reference to Braithwaite's book. Look up David Gauthier in the library catalogue sometime and you'll find more applications of game theory to moral and political philosophy. Notice in the last paragraph of this section Rawls's claim that there is nothing fictitious about Rawls's social contract. As I remarked earlier, later versions of the theory became more artificial, and as a matter of fact the compact or agreement also ceases to be really a compact. We'll come to that later.

Now read section V. The first paragraph is the explanation of why Rawls calls his theory justice as fairness. At the end of the paragraph he suggests that in ordinary speech justice and fairness are used in somewhat different contexts, but he wants to combine the two terms to suggest that justice involves something we are familiar with in the context of voluntary practices where the voluntarily participating parties do not want to be used or taken advantage of. Fairness involves a kind of equality--the interests of each participant are equally to be taken into account; and fairness also involves competition on conflict between people who are in some sense, at least in that context, self-interested.

The next paragraph is about what Rawls calls, perhaps inappropriately, the duty of fair play. He means that people who have participated in a practice and derived benefits from it must not refuse to play their part when the practice requires them to accept loss for the benefit of others. You are not supposed to pick up your marbles and go home when one of the other players turns out to be better or luckier than you are. A gambler who has been winning is supposed to keep on playing if he starts to lose. When it comes to your turn in some practice you've benefited from, it's too late to say, "Well, actually, I never thought this practice was a good idea".

On p.147 Rawls links his theory with another idea influential at the time, namely that of "respect for persons". Cf. R.W. Downie and E. Telfer Respect for Persons (1960); see library subject catalogue Respect for persons. Notice here the notion of a "criterion", a way of recognising whether a person has a concept, some observable action he will engage in on certain suitable occasions if he does really have the concept. The formulation of such criteria was a common project in 1950s philosophy, under the influence of Wittgenstein and of Gilbert Ryle's Concept of Mind. The idea of formulating behavioural criteria by which you can decide whether someone has some concept is also behind certain earlier passages; look back to p.139, the second half of the page, where he tries to specify the signs by which we can judge that someone has a morality. To return to p.147, where Rawls connects his theory of justice with the recognition of others as persons (not mere things): Re-read pp.147-148, the three paragraphs from "Now the acknowledgment". Look back to the bottom of p.144, when Rawls makes a link with a related notion, the idea of community. These are perhaps allusions to Immanuel Kant's ethical theory. According to Kant it is a basic principle of morality to treat other persons as ends in themselves and never merely as means. Another formulation of his principle is to the effect that we should regard ourselves as members of a certain community, the kingdom of ends, consisting of all rational beings. In fact Rawls's first principle of justice seems to come straight from Kant's Critique of Pure Reason (B373):

A constitution allowing the greatest possible human freedom in accordance with laws by which the freedom of each is made to be consistent with that of all others--I do not speak of the greatest happiness, for this will follow of itself--is at any rate a necessary idea, which must be taken as fundamental not only in first projecting a constitution but in all its laws.

Now read section VI, comparing Rawls's theory with Utilitarianism. Some of this develops the point made at the end of section V, but there is also another point: that utilitarians decide from the point of view of a superior administrator allocating pleasures and pains to the individuals (of whom he may himself just happen to be one), and from that superior and impartial standpoint it does not matter who gets what as long as the net aggregate is maximised. Rawls's decision makers are rather parties pressing claims on their equals, with no impartial superior adjudicating.

Now read section VII, which continues the comparison with Utilitarianism. This is further elaboration of the point made in the last paragraph of section V. The reference to Mill (p.153, 15 lines from the bottom) is a reference to the last part of Chapter V of Mill's Utilitarianism, Readings from the bottom of p.93 to p.96.

The reference to Bentham on p.155, paragraph beginning "It is worth remarking" is to Bentham's arguments in favour of equality, which you can find in J. Bentham, Theory of Legislation, "Principles of the Civil Code", "Objects of the Civil Law", Chapter V.

Put on one side a thousand farmers, having enough to live on and a little more. Put on the other side ... a prince, well portioned, himself as rich as all the farmers taken together. It is probable, I say, that his happiness is greater than the average happiness of the thousand farmers; but it is by no means probable that it is equal to the sum total of their happiness, or, what amounts to the same thing, a thousand times greater than the average happiness of one of them.

ought to maintain the distribution as it is actually established. It is this which, under the name of justice, is regarded as his first duty... When security and equality are in conflict, it will not do to hesitate for a moment. Equality must yield ... if property should be overturned with the direct intention of establishing an equality of possession, the evil would be irreparable. No more security, no more industry, no more abundance! Society would return to the savage state whence it emerged.

Now read section VIII. If you read the papers referred to in the footnote on p.157 you will find that they put forward something similar to Adam Smith's idea of the impartial spectator (Readings, p.7, p.110 of Smith's Moral Sentiments). Rawls is saying here, in effect, that his two principles of justice are principles whose function is similar to that of principles in Adam Smith's system, viz. to articulate in general form the concrete judgements we make in particular cases. See Adam Smith, Readings, p.10, RH side, paragraph 8 (p.159 of Smith's Moral Sentiments).

Well, when you've digested "Justice as Fairness" turn to the next article (Readings, p.254), "The Justification of Civil Disobedience". Ignore the underlining, which is excessive. This article was first published in 1966, eight years after "Justice as Fairness". Rawls was writing at a time when the black civil rights movement in the United States was conducting campaigns of civil disobedience, i.e. openly breaking discriminatory laws as a protest against discrimination, not fighting the police but letting themselves be arrested. Read the introductory paragraph and then section II. Notice some developments in the theory since "Justice as Fairness". In the first paragraph of section II there is a new distinction between "natural duty" and "obligation". Rawls has adopted a distinction some philosophers were making between a duty and an obligation. I think myself that these words are interchangeable. Those who distinguish them say that an obligation is something that arises from some action you have voluntarily done (e.g. accepting a benefit, or making a promise, or accepting office in an organisation). A duty is something you have irrespective of any voluntary act on your part. You undertake or incur obligations, you just have duties. The obligation to do our share when our turn comes corresponds to the "duty of fair play" in the earlier article (cf. Readings, pp.247-8). But prior to this there is a natural duty to support just institutions--a duty that we have even before we have accepted any benefits from such institutions (they may not yet exist). So the idea of the natural duty to support just institutions is that you ought to work for just institutions even if you have not yet enjoyed the benefit of such institutions.

In the second paragraph you encounter the notion of an "original position" and the frank acknowledgment that this original position is hypothetical (contrasting with the suggestion in "Justice and Fairness" that "Such discussions are perfectly natural in any normal society" (Readings , p.244)). In this article the principles of justice are those principles that would be agreed to if people were placed in an original position suitably defined.

The third paragraph goes some way toward the suitable definition of the original position. Notice that there are "very strong restrictions" on what the hypothetical contracting parties are supposed to know. The corresponding part of "Justice and Fairness" is in Readings p.244, the bottom half of p.138. The very strong restrictions now proposed go beyond ordinary natural ignorance of further circumstances. As I said in commenting on "Justice as Fairness" the members of a household know that whatever happens the males will still be male and the females female. In "The Justification of Civil Disobedience" he says that they do not know whether they are men or women. A "Veil of ignorance" covers all information relevant to bargaining strengths and weaknesses. It also conceals "their own particular interests and preferences or the system of ends which they wish to advance: they do not know their conception of the good". The point of this is to make sure that the principles they acknowledge will provide for tolerance, freedom of religion, etc. Bear in mind the point made at the end of "Justice as Fairness", that these principles are supposed to express in general terms the particular concrete judgments that are made by competent judges; so Rawls tailors his definition of the original position to make sure that the right principles emerge.

You might wonder how the parties in the original position, behind this heavy veil of ignorance, will decide on any principles, if they don't know their conception of the good. In A Theory of Justice Rawls ascribes to them a "thin" conception of the good, just enough to motivate a choice of principles. They are allowed to know that they may have ends and that certain sorts of things (e.g. money) are likely to be useful to them whatever their ends may be. He calls such things primary goods. These are things likely to be useful whatever your particular ends may be. The thin theory of the good is a list of primary goods. But as he says here people in the original position don't know their own particular interests, preference and ends.

You may feel that this veil of ignorance makes the original position so artificial that it no longer corresponds with any process of reasoning we can actually engage in; the household complaints discussion of "Justice is Fairness" was at least imaginable. But in fact I think that what Rawls is now describing is not too dissimilar to the way in which we actually argue about justice. If we say "you just have to submit, because I'm stronger/richer etc", that is not an argument in terms of justice. Arguments about justice do not make any reference to differences of bargaining power. We argue as if we don't know what the differences are.

Now read section III

Line 8 from the beginning of section III refers to the "fact that we are normally required to comply with unjust laws as well". The requirement here is not just legal but moral. Of course we are legally required to conform to a law, even if it is unjust. But Rawls is saying that normally we have a moral obligation to obey a law even though it is unjust. For my own part I deny that this is a fact. If I want to do something and the state makes a law forbidding me, in my view unjustly, from doing it, then if they can't persuade me that their law is after all just, and if they can't coerce me, then I'll do the thing without any qualms of conscience. The doctrine that we morally ought to obey the law even if it is unjust comes from Paul's letter to the Romans Chapter 13 ("obey the powers that be... for conscience sake") by way of the last chapter of Calvin's Institutes, through John Locke's doctrine of tacit consent; I think we ought to reject this tradition entirely. Not believing that the requirement of obedience is a fact I don't demand that a theory account for it. But Rawls believes in the fact and asks how can we explain it? Skip over to p.245, almost half way down. He says: "In agreeing to a democratic constitution ... one accepts ... the principle of majority rule". Yes, as a principle of legality, but not of morality. We accept that what the democratic process produces is the law, that those who agree with the law do nothing illegal or unconstitutional in acting on it or enforcing it; we may even accept that normally, if they are acting in good faith, in the honest belief that the law is just, they do nothing immoral in acting on it and enforcing it. But none of this implies that we are morally obliged ourselves to act on it and enforce it, believing as we do that it is unjust. Even if we don't believe that it is unjust but merely that it is inexpedient or unnecessary, we are not morally obliged to obey it and enforce it.

But to reject the so called fact that Rawls says his theory can explain is not an objection to the theory. Let's go back and look at p.244 and the top of p.245. Notice the stages, with the veil of ignorance lifting somewhat after the first stage. Laws have to be both constitutional and just, the constitution has to be just: the decisions reached in any stage are binding in all subsequent stages. Notice that Rawls demands of legislators (and by implication of voters) that they set aside their knowledge of their own bargaining powers and interests and vote impartially.

Notice the distinctions between pure, perfect, and imperfect procedural justice, illustrated respectively by fair gambling, the practice of giving the person who cuts the cake the last piece, and the criminal trial. In politics, as in a criminal trial, justice does not consist in whatever comes out of it (as in gambling), and the procedure does not guarantee that the outcome will be just, so there may be unjust outcomes, unjust laws.

At the beginning of the last paragraph he seems to say that majority rule is not essential to democracy. I don't know what definition of democracy he has in mind. On the usual definition majority rule is necessary though perhaps not sufficient to make a regime democratic.

Read now Section IV

This is clear enough. Notice the point made at the end--Rawls does not rule out on principle other forms of resistance besides civil disobedience. This seems to imply that in some cases people may be justified in disobeying secretly, in avoiding arrest, even in using violent means. In this section he is merely characterising civil disobedience as the term is usually understood, not saying that it is the only alternative to obedience.

Now read Section V

This is all clear enough. Compare the points about self-esteem and respect in the bottom third of p.250 and the top third of p.251 with "Justice as Fairness", bottom third of p.144 (Readings p.247) and p.147.

Now read Section VI

In the first paragraph notice the statement that "no increase in efficiency in the form of greater advantages for many justifies the loss of liberty of a few". This is the familiar point basic to Rawls's approach, explained in "Justice as Fairness" e.g. in the discussion of slavery.

Notice the discussion of the sense of justice. Although the parties in the original position are thought of as self-interested when they choose among possible rules, this choice takes place in a wider context in which they are not merely self-interested but also concerned to be just. In the bottom third of p.253 when he discusses how the sense of justice manifests itself he may be once again thinking of criteria in Wittgenstein's sense (e.g. "Justice as Fairness", p.147, Readings p.248), but it also makes sense at a non-technical level.

Finally the question "Who is to say?" is of course Locke's question "Who shall judge?", Pol 167 Readings, p.311. He answers as Locke does, that each of us must judge. But this does not mean that we can say whatever we like. We must judge in good faith in a principled way.

This brings us to Rawls's book, A Theory of Justice. In Articles, Chapters and Lectures there are three lectures on the theory presented in this book, so I won't make detailed comments on the extracts in the Readings.

The first few pages are about three different conceptions of equality of opportunity (required, you will remember by Rawls's second principle). The three conceptions he calls natural liberty, the liberal interpretation of equality of opportunity, and the democratic interpretation. So read Readings from p.262 to the end of the first paragraph of section 13 (to the word "omission").

The next extract is on some of the game-theoretical aspects of the argument. Re-read "Justice as Fairness", p.140 (Readings, p.245), and then Readings from p.264 to near the bottom of the left-hand side of p.266 (to the word "omission").

The next extract is an analysis of freedom, or as Americans prefer to say, "liberty". Read from p.266 to the RH side of p.267 (to the word "omission"). A couple of points here need comment. The terms "serial order" (near the end of p.203) and "lexical order" (10 lines or so later) refer to the idea prominent in the later version of the theory that the first of the two principles of justice (calling for the greater possible equal liberty) has priority over the second principle. Contrast "Justice as Fairness" p.134 (Readings p.242) which envisages that liberty may be traded for other advantages. In A Theory of Justice trading of liberty for something else is (normally) forbidden. But notice that one liberty may be traded for another (p.203, Readings p.266)--what has to be maximised is the system of liberty as a whole, not each and every distinguishable liberty.

The next three extracts illustrate the trading off of one liberty for another so as to enhance total liberty. Read from Readings p.267 to p.271, three quarters of the way down the RH side. The next short passage sums up the outcome of what you've just read. Read from the bottom of p.271 to the top of p.272 (to the word "omission").

The next extract deals with the problem of determining a just rate of savings. It seems that questions of justice arise in relations between different generations. For example, if the members of this generation consume non-renewable resources without restraint and unrestrainedly pollute and destroy the environment, they may improve their own standard of living, but at high cost to later generations, and this seems unjust. Similarly this generation should do a certain amount--or rather an uncertain amount--of saving and investing for later generations, just as earlier generations did things from which we benefit. Consumption, pollution, saving and investment are aspects of the same problem of justice between generations. Read now from p.272 to three quarters of the way though p.275.

A brief comment on the points Rawls makes on pp.291-2 (Readings p.275, LH side). "The present time of entry interpretation" implies that when we go in behind the veil of ignorance, though we then don't know what stage of history we're at, we do know that we are the same persons as we were when we went in--and therefore that the other parties we meet there are all contemporaries of ours. This is bizarre, and Rawls's reasons for adopting this interpretation (see section 24 of the book, not in the Readings) are unpersuasive. It gives rise to the problem he mentions here--that the parties, if they are self interested, have no motive to agree to any saving for future generations. So he changes the motivation assumption: they are to care about future generations.

Let me point out that Rawls is mistaken in thinking that his theory involves a contract between several parties. The theory amounts to this: the principles of justice are the principles that would be accepted by any individual motivated in a certain way and knowing and not knowing certain things, if that individual understood that the principles accepted were to apply to everyone alike. When you go behind the veil of ignorance you don't meet any other parties at all; there are no parties, there is no contract.

The next extract tabulates the principles and priority rules as they have been elaborated in this book. The two principles of "Justice and Fairness" have developed into something more complex. Read pp.302-3 (Readings, pp.275-6).

The content of the next extract is indicated by its heading. Read to Readings p.277, RH side.

The next extract is about the priority of liberty. In A Theory of Justice Rawls distinguishes between a general conception of justice and a special conception. The general conception was stated under that heading on p.303 (Readings p.276 LH side). The special conception is the one that includes the two priority rules on p.302. The special conception applies once a society has developed beyond the struggle for mere survival. Before then, liberty may be traded for material goods--slavery may be justified. Has he taken back what he said "Justice as Fairness" about slavery against utilitarians? No: he still does not say that slavery for some may be justified by advantages for others, but he admits that slavery for some may be justified by advantages for the some who are enslaved--but only in the primitive stages of history. Read the next section, to p.279 RH side. Recall that "primary goods" are the things included in the "thin" conception of the good, the things likely to be useful whatever your ends.

From the passage you have just read you will have a sense of what Rawls thinks the good life is. This is also the topic of the last extract. Read p.279-281.

Three lectures on Rawls's theory of justice:
The Original Position
Decisions in the original position
Liberty

Return to Teaching Materials on Political Thought
Return to Home Page