A theory of justice? P H I L I P P E T T I T A T H E O R Y O F J U S T I C E ? * ABSTRACT. This is a critical analysis o f J o h n Rawls's A Theory o f Justice. Rawls offers a theoretical justification o f social democratic principles o f justice. H e argues that they are the principles which rational men would choose, under defined constraints, in an original position o f social contract. The author criticises Rawls's assumption that men o f any background, o f any socialisation, would choose these principles in the original position. H e argues that the choice which Rawls imputes to his contractors reflects a specific s o c i a l i s a t i o n - one dominant in Western democracies. The theory is useful because it systematises a particular sense o f justice; it is in no sense however a universal theory. My intention is to show that the contractual theory o f justice defended by John Rawls does not have the status of a universal theory (John Rawls, A Theory o f Justice, Oxford 1972 - henceforth 'J'). By a universal theory, I mean a theory which people in different circumstances, particularly people in different cultures, would have equal reason to accept - granted that they could all understand the argument for it. I intend to show that the most Rawls can claim is that his theory explicates the sense of justice of people in a particular society. The paper has two subsidiary goals. The first is to suggest that the society for which Rawls provides a theory of justice is Western democracy, particularly in its twentieth century form - WD, for short. Rawls appeals to our intuition when W D nicely sums up what we have in common. Also he takes as natural attitudes which, if not exclusive to WD, are at least characteristic of it. The second subsidiary goal is to suggest that at the specific level o f Rawls's argument only a particular theory of justice is possible; a universal theory would be something quite distinct. Rawls is concerned mainly with the principles of social justice: "they provide a way o f assigning rights and duties in the basic institutions of society and they define the appropriate distribution of the benefits and burdens of social cooperation" (J4). The basic institutions are "the polit- ical constitution and the principal economic and social arrangements" (J7). The particular principles for which Rawls argues are: first, "Each Theory and Decision 4 (1974) 311-324. All Rights Reserved Copyright �9 1974 by D. Reidel Publishing Company, Dordrecht-Holland 3 1 2 P H I L I P P E T T I T person is to have an equal right to the most extensive total system of equal basic liberties compatible with a similar system of liberty for all" 0250); second, "Social and economic inequalities are to be arranged so that they are both (a) to the greatest benefit of the least advantaged and (b) attached to otfices and positions open to all under conditions of fair equality of opportunity" (J83; for a later reading see J302). These princi- ples are interpreted under the constraints of two priority rules: roughly, that the first principle may never be compromised out of consideration for the second and that, in the case of the second, fair equality of oppor- tunity may never be restricted out of consideration for the greatest benefit o f the least advantaged 0302-03). Rawls argues for these principles of justice by a version of the contrac- tual theory. "They are the 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" (J11; see also Jl18-19). Rawls does not resort to a quasi-historical myth of a state of nature to give substance to his idea of the original position. It is a purely hypothetical situation defined by certain constraints and, taking account o f the constraints, we are meant to be able to simulate the reflec- tions of the imaginary contractors 0120). The constraints which define the original position include constraints on the parties - they are to be rational, representative of possible social positions, mutually disinterested, reliable when it comes to complying with the principles and so on; these, and constraints on the task in hand - the principles to be chosen are principles of social justice, the society in which they are to apply is one o f moderate scarcity, the principles are to satisfy such formal constraints as generality and publicity, etc. 0146-47). The central constraint however, is that of the veil of ignorance. This requires that the parties to the contract be in ignorance o f their particular talents and fortunes in the society for which they are choosing principles, and indeed be in ignorance of the particular historical circumstances of that society - their knowledge extends only to general facts of politics, economics and psychology (J137). The original position, by the present argument, is not a device which enables the theorist to step outside the limits of his place and time in history. It does not give him a voice to speak for men o f cultures far removed from his own. This its inventor fails to appreciate: " t o see our A T H E O R Y OF J U S T I C E ? 3 1 3 place in society from the peispective of this position is to see it sub specie aeternitatis: it is to regard the h u m a n situation n o t only from all social but also from all temporal points o f view" (J587). I I There are two lines which my argument might take. The first I will mention but not develop. It is the argument t h a t the very idea o f the original position makes sense only against the background o f a certain social experience. It presupposes the experience o f a society where the distri- bution o f social and economic goods is regarded as something subject to h u m a n agency (1). More deeply perhaps, it presupposes the experience o f social mobility and the uncertainty t h a t this brings with it; otherwise the veil of ignorance constraint would seem quite outlandish. The point I a m making is reminiscent o f C. B. MacPherson's claim t h a t the state o f nature o f which Locke (or Hobbes) spoke reflected a society in which market relations are d o m i n a n t so that " t h e individual with which he starts has already been created in the image o f market m a n " (Possessive Individualism, Oxford 1962, p. 269). I am arguing t h a t the contractors with which Rawls starts also show signs o f socially specific modelling: they are limit cases o f socially mobile individuals considering a problem characteristic o f a society t h a t allows some economic inter- vention by the state. I f this is so then some d o u b t is cast on the universal status o f Rawlsian theory. I do no more t h a n mention this point because I do n o t think t h a t it can be pressed home. The contractual theorist can say t h a t t h o u g h the original position is modelled on a situation specific to a certain society, there is no logical reason why it should n o t make sense in other societies. It is true t h a t there is no reason in logic w h y it should n o t do so - but there is every psychological reason why it should not. The idea o f the original position is t o o m u c h the flower of one society - as I see it, W D - to have m u c h chance o f blooming in others, at least in some others. In the f o u r t h section I shall mention a further consideration which suggests this view. The line o f argument which I prefer to follow n o w concentrates on the details o f the choice attributed to the contractors. The choice is t h a t o f the two principles o f justice and it is presented as " t h e unique solution to the problem set by the original position" (J119). I wish to argue t h a t 314 P H I L I P P E T T I T Rawls makes a case for this conclusion which there is no reason to think would carry equal weight in all societies. Rawls's b o o k falls into three parts: on 'theory', 'institutions' and 'ends'. In each of these parts he provides a description which is meant to justify the choice he attributes to the original contractors. In the first part he describes the choice as the implementation of the maximin decision proce- dure, in the second as the adoption o f principles in reflective equilibrium with our considered judgments of justice and in the third as the rational choice in view o f the contractors' conception of goodness. The three descriptions recur in the book, but each is defended in its respective part. I I I The maximin procedure is defined by a conservative rule for choice under uncertainty. "The maximin rule tells us to rank alternatives by their worst possible outcomes: we are to adopt the alternative the worst outcome of which is superior to the worst outcomes of the others" (J152-53). What Rawls wishes to argue is (a) that the choice of his two principles o f justice by the parties in the original position is a maximin procedure and (b) that as such it is the appropriate procedure in the original position. He puts some faith in this argument: "if the original position has been described so that it is rational for the parties to adopt the conservative attitude expressed by this rule a conclusive argument can indeed be con- structed for these principles" (J153). A b o u t (a): Rawls can show that choice o f his principles would be a maximin procedure only by comparing those principles with alternatives. He finds the alternatives in " a short list of traditional conceptions of justice" (J122). The conceptions he considers are mainly variants of utilitarianism: in particular he considers the principle that average utility- which, unlike aggregate utility, is insensitive to size of population - should be maximised. Here it is already clear that the reflections with which the contractors are credited are not very radical anthropologically: they are reflections traditional in WD. But perhaps the difficulty of the task does excuse these "rough and ready methods" (J123). Rawls now has to show that choice of the two principles represents the maximin procedure, that the worst outcome o f these two principles is better than the worst outcome of alternative procedures. His case has A T H E O R Y OF J U S T I C E ? 315 some plausibility. A high average utility might m a r k a society which tolerated slavery and the worst outcome for an individual in such a society - being a slave - would seem to be worse t h a n anything allowed in a society ordered by the two principles o f justice; the criteria for this assessment o f slavery certainly reflect a specific social experience, b u t I a m willing to grant t h a t they might be accepted by people o f any back- ground. The straight principle of average utility however is n o t the only utilitarian alternative to Rawls' principles. J. E. J. A l t h a m has suggested this further alternative to me: "maximise average utility, subject to first having satisfied needs". It is n o t at all clear t h a t the worst outcome of this would be worse t h a n the worst outcome o f the two principles o f justice. I am n o t anxious however to get stuck on this point. Let us grant t h a t choice o f the two principles does represent the maximin procedure. The second point which Rawls has to establish is t h a t in the original position, the maximin procedure is indeed the rational one. He draws on economic decision theory a n d argues t h a t in any situation there are three features which we must obtain if 'maximining' is to be the rational course. They are, t h a t the situation be one o f uncertainty within which probabili- ties c a n n o t be assigned to the possible outcomes o f any plan; t h a t the m i n i m u m outcome promised by the maximin rule be satisfactory; and t h a t some o f the outcomes o f alternative plans be unacceptable 0154). The A l t h a m alternative does n o t involve any unacceptable outcomes b u t again I put this aside. Let us assume t h a t these three conditions are realised in the original position. This brings us to the central question. G r a n t e d t h a t maximinning does m e a n choosing the two principles o f justice, granted too t h a t the original position satisfies the conditions which normally make maximinning the rational course, would choice o f the two principles necessarily seem the rational choice for the contractors to m a k e ? H o w would it seem to someone o f an aristocratic society for example, someone w h o f o u n d duelling regrettable perhaps but on m a n y occasions the only reasonable course? N o n e o f us can say for sure. A n d t h a t is precisely the point. It is only i f we presuppose in the contractors a certain attitude to risk - one which contrasts the p r u d e n t a n d the imprudent, n o t the mean and the m a n l y - t h a t the maximin rule will seem the rational procedure for them to adopt. We readily make t h a t presupposition because in W D we are generally disposed to take a conservative attitude to risk. 316 P H I L I P P E T T I T Rawls claims, contrary to this objection, t h a t his argument does n o t rest directly on the assumption o f an aversion to risk a m o n g the contrac- tors. " W h a t must be shown is that choosing as i f one h a d such an aversion is rational given the unique features o f t h a t situation irrespective o f any special attitudes to risk" (J172). This begs the question as it supposes t h a t there is some independent test o f the rational. A t this stage, the only test is whether we can simulate the reflections o f the parties in adopting the maximin procedure and feel their weight. This we can do with ease but only because o f our characteristic attitude to risk. A little historical imagination enables us to see t h a t others might find a very different rule o f choice the natural one for the contractors to follow. There is only one situation - one version o f the original position - in which it is h a r d to imagine that people o f a n y society would find the conservative rule o f choice unnatural. This would arise i f each contractor h a d to assume, n o t t h a t he might, but t h a t he definitely would be in the least advantaged position in the society to come. In this case the principle 'I cut, y o u choose' would apply (2). It would leave the choice o f the maximin alternative contingent, n o t on a particular attitude to risk, but on an attitude to coming out worst. This would be an advantage insofar as aversion to coming out worst seems to be a more natural and universal h u m a n attitude t h a n aversion to risk-taking. It would undermine the idea o f a contract however: a contractor could not assume t h a t he would be in the least advantaged position i f he recognised t h a t every other contractor made the same assumption. I V The second description which Rawls gives o f the choice which he imputes to the contractors presents the principles chosen as principles in reflective equilibrium with our considered judgments o f justice. They are in equilib- rium because the conclusions which they yield a b o u t h o w to organise the basic institutions o f a society are in line with our intuitive j u d g m e n t s ; the equilibrium is described as reflective because it gives us a view o f the grounds and mode o f derivation o f those judgments (J20). Rawls does n o t say t h a t the principles must in every case generate con- clusions in line with our judgments. In central cases they will. In marginal cases where we are n o t sure o f our judgments they m a y not: here the con- clusions m a y shift our intuitive judgments somewhat, providing us with A T H E O R Y OF J U S T I C E 9. 3 1 7 independent theoretical guidance (J19-20). The theory of justice is Socratic: "we m a y want to change our present considered judgments once their regulative principles are brought to light" (J49). The i m p o r t a n t point however is t h a t in the main, the t h e o r y fits our considered judg- ments. This fit is invoked by Rawls to justify the two principles which he takes the contractors to choose. " O n e can work out their consequences for institutions and note their implications for fundamental social policy. In this way they are tested by a comparision with our considered judgments o f justice" 0152). I do n o t intend to go into the consequences which Rawls draws f r o m his principles: they are consequences which describe W D in fair outline. W h a t I want to point out is t h a t the test o f reflective equilibrium is inconsistent with a 'universalist' view o f the theory o f justice. This test ensures t h a t what Rawls' theory does is explicate the sense o f justice d o m i n a n t a m o n g those whose judgments are taken into account, system- atise the principles t h a t they find natural and attractive. There is no more reason to believe t h a t one theory o f justice would satisfy all cultures t h a n there is to believe t h a t all cultures share the same judgments o f justice. Rawls provides a theory which answers only to our judgments, a theory o f justice for WD. (3) Rawls does try to give scientific status to the test o f reflective equilib- rium, and this m a y seem to do something for the status of his theory o f justice. He points out - rightly, I think - t h a t such a test also appears in linguistics, at least as described by N o a m Chomsky. " I n this case the aim is to characterise the ability to recognise well-formed sentences by formulating clearly expressed principles which m a k e the same discrimina- tions as the native speaker" (J47). Here too the theory is Socratic, the principles m a y influence our intuitive j u d g m e n t s : "while we m a y n o t expect a substantial revision o f our sense o f correct g r a m m a r in view o f a linguistic theory the principles o f which seem especially natural to us, such a change is n o t inconceivable, and no d o u b t our sense o f grammati- calness m a y be affected to some degree anyway by this knowledge" (J49). The analogy with linguistics however is a give-away. The analogy Rawls draws is with a particular grammar, the g r a m m a r o f a particular language. It suggests t h a t what he offers himself is a particular t h e o r y o f justice, the theory which explicates a particular sense o f justice. He might have f o u n d a less embarrassing analogue in universal grammar, the theory o f 3 1 8 P H I L I P P E T T I T the constraints under which all particular grammars must work. The trouble is that his theory does not have the required similarities with that grammar. It just is not a universal theory of justice. A further interesting point is suggested by the linguistic analogy: it links up with the argument mentioned in the second section but not developed. In Chomsky's linguistics, the descriptive adequacy of a gram- mar is its capacity to generate all and only the grammatical sentences of a language - the fact that it is in reflective equilibrium with our sense of grammaticalness. The drawback with descriptive adequacy as a criterion of a grammar however is that there may be two grammars which are capable of generating the grammatical sentences of a language. Chomsky finds a further criterion, a procedure for evaluating two descriptively adequate grammars, in universal grammar: the one which fits in better with the universal theory of language is the better grammar. It now has explanatory adequacy as well as descriptive (Aspects of the Theory of Syntax, Cambridge, Mass. 1965, pp. 26-27). The question is, does Rawls have any corresponding procedure for choosing between two theories of justice? He needs such a procedure, for it is at least conceivable that another theory of justice should also meet the test of reflective equilibrium. Why should a theory not be possible which set up a divine legislator and gave him such attributes that his imagined judgments on justice would be in reflective equilibrium with our own? H o w would Rawls decide then be- tween such a theory and that which he defends himself? Clearly he would decide for his own; so, I suspect, would all of us. But why? The only reason I can see, in the absence of a universal theory of justice which would do the work of Chomsky's universal theory, is that Rawls' theory is a native flower in WD, the alternative an exotic growth; here I return to the argu- ment of the second section. A myth of rational contractors in an original position is a more plausible myth to our minds than the myth of a divine legislator. V The third description which Rawls gives of the choice of the two principles o f justice presents it as the rational choice for the contractors in view of their conception of goodness. Does this description do any more than the others to give universal status to the Rawlsian theory of justice? I am A T H E O R Y OF J U S T I C E ? 319 going to argue that it does not because I do not think that RaMs estab- lishes the validity of this description. First, what does he mean by 'rational'? He takes his cue from the economic theory of decision-making. "Thus, in the usual way, a rational person is thought to have a coherent set of preferences between the op- tions open to him. He ranks these options according to how well they further his purposes; he follows the plan which will satisfy more of his desires rather than less, and which has the greater chance of being success- fully executed" (J143). Here RaMs gives us the three main principles that he thinks of as guiding rational choice: the principles of effective means, inclusiveness and greater likelihood (J411-12). The principles are useful because they enable us to compare different plans; they may even be allowed a universal appeal and validity. What they do not do however is mark out one plan as the rational one in a situation of choice - at least not necessarily. Hence they cannot in themselves determine the choice of the original contractors. What more is necessary in a situation of choice if one plan is to stand out as the rational plan ? First, Rawls says, that the requirement of delib- erative rationality is met. A plan is chosen with deliberative rationality when all its consequences have been clearly foreseen and balanced against the consequence of alternative plans (J417). Even when this requirement is met, however, it is not possible to say what is the rational plan for an individual in a situation of choice; that plan has only been specified formally. What it is also necessary to know is the desires which weigh with that individual. Thus if we are to be convinced that the rational plan for the original contractors to choose is represented by the two principles of justice we must be told something about their desires, their conception of the good. The distinctive thing however about the choice which faces the contrac- tors is that they do not, and in simulating their reflections we do not, know anything of their particular desires; this, by the veil of ignorance condition. The contractors are to choose principles of social organisation, not out of concern for such individually variable desires, but out of con- cern for desires to which they are bound, regardless of who they are. These are described as "general desires" (J263) or desires for "primary goods" (J93): they are meant to be desires for conditions required for the pursuit of any particular goals. 320 P H I L I P P E T T I T Rawls maintains, that these conditions, the primary goods, are certain rights and liberties, opportunities and powers, income and wealth and - something supposed to be made possible in the pursuit of these - self- respect (J92). He also maintains that these are ordered so that liberty is prior to socio-economic advantage and equality of opportunity to econ- omic welfare (J302-03); other motivational assumptions, e.g. that in- centive requires inequality, I leave unexamined. If his claims are allowed, then it certainly follows that Rawls's third description is a fair account of the choice which he imputes to the contractors; this is indeed the rational choice for them to make in view of their desires, their conception o f the good. Does Rawls justify his claims, does he establish his list of the primary goods? He certainly sets out a strategy for doing so. They are meant to be the goods to which the contractors are directed in view of certain "general facts" about human beings 0424). These include facts about human desires and wants, capacities and abilities, and social interdepen- dence: they are not discussed - on the grounds that they are matters of "common sense knowledge" (J425). The only fact discussed is described as a "deep psychological fact" (J432). "It says only that we prefer, other things equal, activities that depend upon a larger repertoire of realized capacities and that are more complex" (J429-30). Rawls calls it the Aristotelian principle. It is worth noting one reason why Rawls may feel free to pass so quietly over facts that play a crucial role in his theory. It is that he shows a not unexpected faith in psychology, an optimism about the invariability of the facts which this discipline may be presumed to catalogue. Thus he speaks of general psychological principles of which he allows his contrac- tors to be aware (J24, 456), principles which include "laws of motivation" 026). In doing so he backs up his assumption that men of any society would attribute to the contractors in the original position the same psychology that he gives them. His positivism on this issue begs the question of whether the original position makes possible a theory to explicate a universal sense of justice. But even if we grant Rawls his general facts - and what they are is not at all clear - we find that he fails to clinch his argument that they would lead the contractors to recognise the primary goods that he lists. At the point where we expect to have the argument clinched he falls back on A T H E O R Y OF J U S T I C E . 9 321 self-evidence. " W e must assume, then, t h a t the list o f p r i m a r y goods can be accounted f o r b y the c o n c e p t i o n o f goodness as rationality in con- nection with the general facts a b o u t h u m a n wants an d abilities, their characteristic phases and requirements o f nurture, the Aristotelian prin- ciple, and the necessities o f social interdependence .... I shall n o t argue the case for the list o f p r i m a r y goods here, since their claims seem evident e n o u g h " 0434). Rawls's failure to argue this case means his failure t o establish the validity o f his third description o f the choice at t ri b u t ed t o the contractors. It means that t h a t description does n o m o r e t h a n the others to establish the universal status o f his t h e o r y o f justice. VI But in fairness: R a M s does offer an argument f o r the necessity o f one p r i m a r y g o o d - t h a t o f self-respect. H e defines self-respect as (a) having a sense o f the value o f one's plan in life and (b) having confidence in one's ability to carry it out. He appeals to intuition in defence o f the idea t h at this is a p r i m a r y good. " W i t h o u t it n o t h i n g m a y seem w o r t h doing, or i f some things have value for us, we lack the will t o strive f o r t h e m " 0440). One can see the force o f this appeal in some cases, t h a t o f a profes- sional m a n in W D for instance, b u t one asks whether it applies in all. R a M s thinks t h a t it does and argues that each p erso n must find something he is g o o d at - and, to satisfy the Aristotelian principle, it had b et t er be o f some complexity - and f o r which he earns the respect o f some g r o u p o f people, respect which encourages him in self-respect. " T h u s w h at is necessary is t h a t there should be for each perso n at least one c o m m u n i t y o f shared interests to which he belongs and where he finds his endeavours confirmed by his associates" (J442; see also J178-79). This definition o f self-respect is tied in one obvious respect t o a n o n - universal p a t t e r n o f social experience, a p a t t e r n characteristically exem- plified in W D . It is individualist. I suggest that fo r societies in which the category o f the individual received less emphasis, fo r societies where social facts were n o t presumed in the n o r m t o be the o u t c o m e o f free individual actions, this definition would n o t do. Th e definition does n o t m a k e it a condition o f self-respect that one should see oneself as p a r t o f a larger entity - the g r o u p - in which one believes. This might well a p p e a r to be a condition o f self-respect to s o m e b o d y socialised in a m o r e tightly 322 PHILIP P E T T I T integrated society t h a n WD. But Rawls leaves no r o o m for such a person in his original position. His very definition o f society - " a cooperative venture for m u t u a l a d v a n t a g e " (J84) - must alienate anyone of a collec- tivist mentality f r o m the t h o u g h t experiment o f the original position. To get contractors into the original position in the hope o f simulating their reflections, it appears t h a t we are to bring t h e m t h r o u g h the p u r g a t o r y of WD. Rawls' concept of self-respect is i m p o r t a n t because he tries to derive an a r g u m e n t f r o m it for the priority - and indeed the basic appeal - o f liberty. 'Liberty' includes a variety o f freedoms: the freedom to vote and stand for public office, the freedom of thought, speech and assembly, the freedom o f person and (personal) property, freedom f r o m arbitrary arrest (J61); these, and also apparently " t h e i m p o r t a n t liberty of free choice o f occupation" (:I274). Rawls does n o t comment on the heterogeneous appearance o f the list. In general, liberty is something generated by institutional arrangements - a " p a t t e r n o f social f o r m s " (:I63) - and it has the merit o f p r o m o t i n g self-respect. Rawls argues t h a t in the society which the contractors foresee equal liberty is a sure basis for self-respect, one which n o b o d y would want to compromise, even in the hope o f material advantage. It is a basis o f self- respect because o f what it allows: " t h e full and diverse internal life o f the m a n y free communities o f interests" (J554; see also J442). A n d also, apparently, by a shift in the concept o f self-respect, it is meant to provide such a basis in its own right. " T h e basis for self-esteem in a just society is n o t then one's income share but the publicly affirmed distribution o f f u n d a m e n t a l rights and liberties. A n d this distribution being equal, every- one has a similar a n d secure status when they meet to conduct the c o m m o n affairs o f the wider society" (:I544). Against this it must be said t h a t the connection between liberty and self-respect cannot be p u t up as one t h a t must c o m m a n d universal assent, even when self-respect is defined in an individualistic way. There can be m a n y communities o f interest, enough to give everybody a chance o f self- expression, in a society which we would n o t regard as free. A n d there m a y also be sufficient matter for self-respect at the general political level. One might argue t h a t in a society terrorised by a dictatorial power, one would n o t have much r o o m for self-respect o f a n y kind. But the argument can hardly go to show the need for full political liberty in the Rawlsian sense - A T H E O R Y O F J U S T I C E . 9 323 certainly n o t something like liberty o f occupation. Liberty f r o m arbitrary arrest m a y be p u t in the same list as other 'liberties', but this does n o t make it o f a kind with them. Rawls himself mentions the possibility o f a feudal or caste society in which individuals find a source o f self-respect in the role given them in the order o f things. He says that this is an acceptable guarantee o f self-respect only because o f the assumption t h a t the order o f things is n o t subject to h u m a n choice, an assumption which is ruled out by the nature o f the enterprise on h a n d (J547). On the contrary: the only assumption necessary is t h a t the order of things in question has the most to be said for it in view o f the primary goods recognised by the contractors. Only by begging the question can equal liberty be taken to be one o f those goods. V I I The conclusion I draw is that Rawls does n o t establish that universal status which he assumes belongs to his theory o f justice. The contractual device gives the principles o f justice as o u t p u t only because the input is a socially specific mentality - the mentality, I suggest, o f people in WD. It will n o t do to say that the device gives even the general outline o f a "finally adequate t h e o r y " (J581). It can only give the outline o f a t h e o r y adequate to a particular sense of justice. The reason for this should n o w be clear. It is not any dialectical inca- pacity in Rawls but a reason in principle. I f a theory o f justice is meant to generate judgments o f justice satisfying the test o f reflective equilibrium then it is going to be relative to the set o f judgments, the particular sense o f justice, with which it tries to achieve equilibrium. A theory o f this kind c a n n o t assume universal status. But there is no scandal here. I do n o t see what is wrong with saying t h a t the sense o f justice by which we order - or at least criticise - our society reflects the image o f m a n we have created and n o t an eternal h u m a n essence. I do n o t see w h y theory should give itself the task o f rescuing t h a t sense o f justice from history. The t h e o r y o f justice is the means by which we explicate and examine our sense o f justice, it is n o t a means o f providing it with metaphysical foundations. The linguistic analogy suggests t h a t we should have a universal t h e o r y o f justice as well as a particular; this would describe the constraints which a n y theory of justice must meet. The best candidate for the title o f uni- 324 P H I L I P P E T T I T v e r s a l t h e o r y is t h e t r a d i t i o n a l p h i l o s o p h y o f j u s t i c e - t h e m e t a t h e o r y o f m o r a l s a n d politics. W h a t prescriptivists a n d descriptivists d o f o r i n s t a n c e c a n b e d e s c r i b e d i n this w a y : t h e y f o r m u l a t e c o n d i t i o n s - o f f o r m a n d c o n t e n t - w h i c h e v e r y sense o f j u s t i c e - a n d , m o r e generally, e v e r y m o r a l sense - m u s t meet, c o n d i t i o n s w h i c h a n y p a r t i c u l a r t h e o r y o f justice m u s t respect. I n r e c e n t w o r k o n justice a n d m o r a l i t y , it is fair t o s a y t h a t w e h a v e h a d e n o u g h o f u n i v e r s a l t h e o r y . T h a t is w h y R a w l s ' s w o r k is so i n t e r e s t i n g a n d so c h a l l e n g i n g ; it offers a p a r t i c u l a r t h e o r y o f o u r sense o f justice. I t is o n l y a p i t y t h a t R a w l s h i m s e l f s h o u l d h a v e c o n f u s e d it w i t h t h e o t h e r s o r t o f t h e o r y a n d tried t o c r e d i t it w i t h u n i v e r s a l status. Trinity Hall, Cambridge N O T E S * My warm thanks go to Jimmy Altham, Harry Bracken, Stefan Collini, Martin Hollis, Steven Lukes, Quentin Skinner and Denys Turner. Without their comments this paper would certainly have been a poorer effort. I am also grateful to Professor Rawls for a useful correspondence about his position. 1 I owe this point to Jimmy Altham. 2 I owe this point to Elizabeth Anscombe. 3 In WD there is wide disagreement on particular judgments of justice between politi- cally defined groups. I take it, with Rawls, that there is equally wide agreement on more general judgments, when the judgments are considered out of a political context.