A modern theodicy: John Rawls and The Law of Peoples

Abstract

John Rawls’ The Law of Peoples has typically been read as an intervention in the field of ‘global justice’. In this paper, I offer a different and widely overlooked interpretation. I argue that The Law of Peoples is a secular theodicy. Rawls wants to show that the 'great evils' of history do not condemn humankind by using a secularised form of moral faith to search for signs that the social world allows for the possibility of perfect justice. There are, I show, striking homologies between this argument and the Christian theodicy that Rawls wrote in 1942, A Brief Inquiry Into the Meaning of Sin and Faith. Perhaps more significantly, I draw out how there is, as Rawls himself appears to acknowledge, an intimate relationship between this redemptive project and Rawls' idealistic and moralistic approach to political philosophy.

Introduction

John Rawls’ The Law of Peoples (LP) has been overwhelmingly treated as an intervention in the field of ‘global justice’ since its publication in 1999. Its defence of a form of ethical partialism justifying only limited global redistribution was seen as a response to those ‘Rawlsians’ who had tried to globalise the difference principle (Beitz, 2000; Caney, 2002; Kuper, 2000; Martin and Reidy, 2006). In this essay, I offer a different and widely overlooked interpretation. I argue that LP is a secular theodicy.

Rawls wants to show that the world is good despite the ‘great evils’ of humankind by establishing grounds for hope in the possibility of perfect justice. He articulates a theodicy of two halves. First, that we are naturally good, and all evil is a result of our contingent disfigurement by injustice. Second, that there are forces within history that can help potentiate our latent moral nature and deliver us from evil. There are, I suggest, striking homologies between this argument and the Christian theodicy Rawls wrote in his youth, A Brief Inquiry into the Meaning of Sin and Faith (SF).

One catalyst for this paper was my discovery of a draft in which Rawls (1995a) speaks of LP as a ‘theodicy’. This is not an isolated remark. In a 1993 interview, Rawls judged that, ‘on the whole’, the aim of his philosophy had been to redeem the social world, and that this was a ‘quasi-religious question, or indeed it is religious’ (Bok, 2017: 30). Many of those closest to Rawls report that his life and work were animated by a ‘religious attitude’, while a recent wave of scholars have brought to light the lasting importance of theology to Rawls’ thought (Bok, 2017; McKean, 2017; Nelson, 2019; Reidy, 2010; Weithman, 2016). While he never managed to draw the connection to Rawls’ project of theodicy, Wellman (2012) has shown that the law of peoples is, curiously, functionally designed to eradicate what Rawls (1999a: 6–7) calls the ‘great evils’ of humankind: ‘unjust war and oppression, religious persecution and the denial of liberty of conscience, starvation and poverty, not to mention genocide’.

In one of his most forthright discussions of his redemptive ambitions, Rawls (2005: lx) admits that his ‘focus on these questions no doubt explains in part what seems to many readers the abstract and unworldly character’ of his philosophy. One of the most important debates in political theory over the last decade has been over the moralising and idealising character of Anglophone political theory, and Rawls as the summa of this outlook. ¹ I believe that, as Rawls implies, these views are intimately connected to his theodicy.

Rawls wants to redeem the world by showing that we are the kinds of beings for whom perfect justice is possible. In order to show that perfect justice is possible, Rawls must perforce focus upon what perfect justice would look like (Weithman, 2016: 29–40). It is the possibility of perfect justice that ‘reconciles’ us to the world, because what is fundamentally at issue is not the achievement of justice, but the moral status of humankind. Rawls also has to explain why the turpitude of humankind does not defeat the possibility of perfect justice by adopting hopeful beliefs about our moral potential on practical grounds. This helps to explain: why Rawls focuses upon ideal theory, the modality of this ideal (‘possibility’), and the grounds for his claim that ideal justice is possible (‘practical hope’). Given that elements of this argument run throughout Rawls’ work, elucidating how they hang together in LP could provide a starting point for a wider reappraisal of the ‘unworldly’ commitments of Rawls’ philosophy. I return to this subject in the conclusion.

I want to reconstruct the theodicy in The Law of Peoples as Rawls saw it, which requires that we draw together threads scattered across his writing, and follow them into and reread familiar parts of his philosophy, in order that we can gain an appreciation of this theodicy as a systematic current in his thought. It is therefore at once a historical and analytic task, and seeks to establish a groundwork of understanding on the basis of which further, more critical and historico-biographical work might proceed.

One way of highlighting the value of this reconstructive approach is by way of a discussion of Stefan Eich’s (2021) recent argument that A Theory of Justice (Theory) is a ‘theodicy of growth’. On Eich’s view, Rawls sought reconciliation with the iniquities of America’s present through an idealisation of its future, the transition to which hinged on an extrapolation of the roaring growth of post-war Keynesianism. When stagflationary crisis hit in the 1970s, however, Theory’s redemptive horizon snapped shut, moving Rawls to ‘discard’ the problem of transition altogether.

Eich has interesting things to say about the role of growth in Theory, but the fact is he does not try and explain how the key terms and influences of Rawls’ theodicy – realistic utopia, reasonable faith, reconciliation, Kant, Hegel and Rousseau – hang together. Eich therefore says little directly about the shape of the theodicy he claims to be present in Theory, and what comment he does venture is unavoidably allusive. Consider two examples. Eich (2021: 985; 990–991) tells us that Theory tries to answer the problem of evil using Hegel’s concept of reconciliation, which ‘can also be expressed’ in terms of a Kantian defence of ‘reasonable faith in the possibility of a just and stable order’. Yet, for Rawls, Hegelian reconciliation is a primarily backward-looking account of society’s development that shows how its institutions attained ‘their present, rational form’, while Kantian reconciliation is a primarily forward-looking justification of the future possibility of perfect justice. In turn, both are grounded in ‘reasonable faith’. Why does Eich bundle these three things together, and how do they figure in the theodicy he describes in Theory? We never find out. Central to Eich’s (2021: 1005) case is that, after the 1970s, Rawls abandoned the problem of transition, repudiating the exigencies of historical practicability in favour of ‘realistic utopias’. What Eich does not mention is that Rawls effectively treats ‘realistic utopia’ as shorthand for Kantian reconciliation. Eich appears, therefore, to use the very same idea to describe both the ‘theodicy of growth’, and Rawls’ supposed response to its failure.

Rawls only explicitly discusses secular theodicy decades after Theory, mostly in the 1990s. Eich relies on these miscellaneous passages, and his reading is so hazy because he does not know how to retroject them into Theory. Eich is simply looking at the wrong text. As I will explain, there is good reason to believe that Theory does not address the problem of evil, and therefore has no need for a theory of transition – growth or otherwise. ² I would suggest a different historical gloss to Eich. Rawls’ turn to the problem of evil is a product of the liberal humanitarianism of the 1990s, where ideas of democratic peace, conditional sovereignty and American unipolarism made a global moral community plausible to Rawls, and the coeval resurgence of Holocaust memorialisation in American culture – and its conjoining with human rights, and events in Bosnia and Rwanda – gave him the impetus to confront a question that had long haunted him. ³

What we need to do is, starting with the work where it is easily most explicitly developed, LP, reconstruct this theodicy as a systematically articulated current within Rawls’ thought, such that we detail where it is present in his argument and how it helps to structure his philosophy at large. It is to this task of making the theodicy in LP legible that the rest of this paper is dedicated.

Sin and faith

When Rawls died in 2002, a short essay was discovered, ‘On My Religion’, in which Rawls recounts a dramatic religious awakening while a sophomore at Princeton (1941–1942). The catalyst appears to have been the envelopment of Europe by Nazism, which precipitated within Rawls a deep, Spenglerian despair (Cheah, 2022). Soon after this discovery, Rawls’ senior thesis from Princeton was retrieved, SF. It contains a Christian theodicy.

Rawls begins (2009: 110–128) by asserting the truth of the Christian God and the division of all things into two realms. Personality is the spiritual dimension of human beings. Nature is everything extended in space. Persons relate to one another in a qualitatively different way than they relate to nature: they recognise they are dealing with a person, and not inert matter. This is the potential basis of community. Indeed, personality is ‘made for community’ in the ‘likeness of God’, for humans are capable of faith, responsibility and love (Rawls, 2009: 121–122). These relations of givingness and fellowship form true community, whereas pride and jealousy are sin for they pervert community.

Rawls (2009: 127) analyses these categories – nature, personality and community – to formulate a solution to the problem of evil: of ‘how personality and community can be achieved in the face of pervasive sin in the world’. This is the ‘chief problem of politics’. Rawls (2009: 190–192) begins by refuting the ‘Manichean heresy’, the attempt to blame evil on forces outside of human personality. If evil is intrinsic to nature, humankind is condemned to sin. But if it is we who are responsible for evil, then perhaps it is a contingent part of the world that can be overcome by the reordering of our personality. To defeat Manicheanism, Rawls has to tackle the age-old belief that our natural appetites misguide and debase us. Rawls (2009: 141–142, 147–148, 190–192, 200) turns to St. Augustine to counter that we are not corrupted by our flesh, and that the human personality is the locus of all sin. Augustine ‘knows full well that the order of nature is good’. Rawls (2009: 150–151) borrows a distinction made by Philip Leon to elaborate on this conception of the person. ‘Egoism’ is our natural self-centred drive for definite objects like food, sex and water. ‘Egotism’ is our personal drive for social superiority that ‘craves honour, distinction, glory and praise’. All sin has its foundation in egotism, in our free choices qua personal beings.

Rawls has shown that evil is not an intrinsic part of the world. But to solve the ‘chief problem of politics’, he also needs a positive account of how we can overcome the ‘pervasive’ sin of human life. The psychological ‘barriers’ erected by original sin mean that ‘man alone cannot save himself’ (Rawls, 2009: 222–233). Humans can only find salvation through a process of ‘conversion’, a dialectical reckoning with God in which one reveals one's sin and God manifests his love. The crux of this process is the eventual realisation that everything that one has is given to one by friends, parents and community – and ultimately, by God (Rawls, 2009: 238–251). This reveals the ‘lie of independence’, shattering the illusions of individual pride and merit. Personality then re-joins community, and, with recrudescent faith, humans come to serve God and help restore true community.

Rawls (2009: 261–266) describes in ‘On My Religion’ how several vertiginous wartime episodes scrambled his theodicy. Rawls was distressed, above all, by news of the Holocaust. How could he pray to a God who ‘would not save millions of Jews from Hitler?’ Rawls became incredulous towards divine providence, even disgusted with it. ‘To interpret history as expressing God's will, God's will must accord with the most basic ideas of justice’, yet ‘the Holocaust can’t be interpreted in that way, and all attempts to do so that I have read are hideous and evil’.

Rawls is clear that the fount of his faith was moral, that while theological proofs of God's existence showed ‘nothing of any religious significance’, the ‘ideas of right and justice expressed in Christian doctrine are another thing’. Equally, his difficulties with Christianity ‘were always moral ones’. Rawls recalls how he came to reject the ideas of original sin, heaven and hell, and priestly authority, not because ‘the evidence for them was weak’, but because ‘they depict God as a monster’. Rawls adds that this was not the end of his religion, only that from thereon he no longer considered himself orthodox. I think we can conclude from these remarks that Rawls adopted Christianity as a vision of universal redemption through moral community. But, when Rawls concluded that this vision no longer found a justifiable basis in Christianity, it was precisely his theodicy that broke through and shattered his doctrinal faith. Kenzie Bok (2017) characterises Rawls’ secularisation as a ‘staircase story’, and I think LP abides this pattern. Rawls charts a secular path back to the redemptive vistas of its youth.

Secularising theodicy

In the draft in which Rawls (1995a) discusses LP as a ‘theodicy’, he disavows Christian theodicy and turns instead to ‘Rousseau's answer’. Where theologians query whether the world ‘is good in the eyes of God’, Rawls wants to ask ‘whether the world is good in our eyes, politically speaking?’ The world is good politically, Rawls tells us, if it allows for the possibility of a ‘perfectly just society’.

Rawls (1995a) takes from Rousseau the idea that human nature is good in that it contains various tendencies, the expression of which depends on the basic structure of society, and that while it can be contingently malformed by unjust institutions, there exists a set of institutions that can potentiate our moral capacities and move us to support perfect justice. Responsibility for evil is diffused into a systemic socio-psychological process, whose solution is political justice. Unlike theologians, Rawls can treat the pain and suffering endemic to the natural world as a brute fact, his aim only being to show that ‘the social world is good’.

Rawls is attracted to ‘Rousseau's answer’ because it is functionally homologous to the first step of his Christian theodicy, but replaces its content with a naturalistic story that overcomes the faultline that fractured his orthodox faith. Its homologies owe to the fact that Rousseau modelled his Discourses on Augustine. ⁴ Rousseau conjectures that humans originally lived in solitude guided by instinct's benign desire for the means of self-preservation (amour de soi), but that as civilisation brought humans into ties of mutual and increasingly unequal dependence, the species developed a peculiarly social mode of self-love (amour-propre). Individuals began to esteem themselves in the eyes of others, and to vie for status in a corrosive dynamic of domination, avarice and jealousy. The parallel with Augustine is stark: pride leads to humankind's Fall from prelapsarian innocence into self-compounding vice. Like in SF, social self-love – Leon's egotism, Rousseau's amour-propre – is the wellspring of a psychic rot. Rousseau mirrors Augustine because he shares the desire to blame humankind for evil to exonerate God's creation. Rousseau, however, sees evil as a fundamentally naturalistic phenomenon arising from the volitional actions of humans in a process amenable to historical, sociological and psychological explanation. As Susan Neiman (2002: 55) writes, this was the first modern account of evil.

Rawls must go further, however. Rousseau remained a providential deist – at least in writing – meaning that his quasi-Augustinianism only pushes the problem back a step. As Bayle (1965: 148–150) reasoned, if God is omnipotent, then he must have known that the ‘gift’ of free will would lead to the ruination of the species, and could surely have modified the human temperament to forestall that fate. Rawls, similarly, did not see how free will could excuse God for creating a world in which the Holocaust took place. Rousseau's deism contains the seed of the answer. If nature is a self-sustaining moral order that functions independent of the intercession of God, then God can begin to appear not as an essential predicate of the world's goodness but as a complicating superaddition. Rawls is thus able to extricate Rousseau's naturalistic answer to the problem of evil intact while eliminating God. This is pivotal, as it releases Rawls from the expectation that all evil should be ultimately justifiable as a part of a benevolent God's creation. Rawls no longer has to try and justify enormities like the Holocaust.

But without God, how can Rawls possibly justify the belief that the world is a morally good domain? ‘Rousseau's answer’ envisions how our evil world might be consistent with the possibility of perfect justice, but it is less a credible and evidenced theory than a wish. Rawls cannot appeal to extant human behaviour, for that is the problem that he needs Rousseau to solve in the first place: we live in a world scarred by ‘great injustices and widespread evils’ (Rawls, 1999a: 89). It is therefore an essential assumption of LP that ‘the limits of the possible are not given by the actual’, and that we must therefore ‘rely on conjecture and speculation, arguing as best we can that the social world we envision is feasible and might actually exist’ (Rawls, 1999a: 12).

Rawls turns to Kant's moral-practical faith. Kant offered moral grounds for belief in Rousseau's beneficent natural order (Neiman, 2002: 72), and the very idea of a moral faith is equally Rousseaun. In the ‘Profession of Faith of a Savoyard Vicar’, Rousseau (1979: 295) discards all philosophical and theological pretensions to know His designs and, regressing to a position of Cartesian doubt, grounds religion in conscience. Kant famously took his Critical turn under the influence of Rousseau, and sought to tame the metaphysical excesses of philosophy and theology alike through the critique of the limits of reason (Beiser, 1992: 42–46). Blocking any straightforward theoretical knowledge of God, Kant's strongest case for religion is therefore – like Rousseau – moral. Rawls, recall, revealed a thoroughly moral conception of religion in ‘On My Religion’. Kant's doctrine systematises this Pietist creed.

Kant believes humans can only coherently act towards ends they consider possible (Wood, 1970). At the same time, he regarded the a priori end of morality as a condition of the ‘Highest Good’ wherein happiness is strictly proportional to virtue. Clearly, that is not a humanly achievable objective. Kant resolved the contradiction by arguing that morality gives us practical grounds for adopting the idea of God as a holy dispenser of happiness, independent of its theoretical verisimilitude. When Kant (1998: 24–26) criticised theodicy, he rejected philosophical theodicies that pretend to know His designs through theoretical reason – metaphysical or theological – concluding that the true believer does ‘not found his morality on faith, but his faith on morality’.

As with Rousseau, Rawls extracts Kant's doctrine from its vestigial theology. Rawls (2000: 313–317) discards the Highest Good as apologetics, a remnant of Kant's pre-Critical Leibnizian phase. He adopts an alternate, secular interpretation of Kant's doctrine that he names ‘reasonable faith’, or ‘reasonable hope’. On this view, the end of morality is given solely by the principles constructed by the categorical imperative. What is crucial is that these principles can be achieved within the order of nature by agents simply acting on their duties. But this does require a very powerful practical assumption, namely, that ‘the order of nature and social necessities are not unfriendly’ to morality, and ‘contain forces and tendencies that in the longer run tend to bring out, or at least support, such a realm and to educate mankind so as to further this end’ (Rawls, 2000: 319–322). We must look for a ‘plan of nature to force mankind … to form a federation of democratic states, which will then ensure perpetual peace’.

Can we read Rawls in the light of ‘reasonable faith’? His theodicy revolves around his stability argument, for it is here that he offers empirical grounds for the possibility of a community bound and perpetuated by morality. In the stability argument of Political Liberalism (PL), Rawls (2005: 172) explains that he is adopting one of the roles that ‘Kant gave to philosophy… the defence of reasonable faith’. Similarly, LP ‘shows us, in the tradition of the late writings of Kant’, the ‘conditions under which we can reasonably hope for’ for a just society of peoples (Rawls, 1999a: 126).

In the 1980s, Rawls elucidated the basis of his philosophy in Kantian constructivism, later remodelled as ‘political constructivism’. Constructivism is not an application of theoretical reason that tries to apprehend an independent order of moral facts, but an exercise of practical reason that constructs objects according to a conception of them (Rawls, 2005: 89–129). Rawls sets out a procedure of construction, the ‘original position’, modelled on principles of practical reason in union with conceptions of the person and society drawn from liberal political culture. It is essential for Rawls (2005: 108) that his conception of the person is modelled on what an agent must be like to engage in practical reason: they must be ‘reasonable’ and ‘rational’, possessing a capacity for a sense of justice and a conception of the good. Rawls uses this procedure to generate principles of justice.

Once principles are chosen in the original position, in a second step, it is checked whether they are stable (Rawls, 2005: 140–141). But it is the principles chosen in the first step that guide the selection and interpretation of facts in the second. Rawls (2005: 122) explains: ‘a constructivist procedure is framed to yield the principles and criteria that specify which facts about actions, institutions, persons, and the social world generally, are relevant in political deliberation’. What is significant about this is that Rawls starts with an ideal of the person as reasonable, constructs the principles of justice with a procedure modelled on this assumption, and then uses those principles to select and interpret facts.

We see this play out in PL. Rawls (2005: 86–88) draws his moral psychology from the ideal of the person as reasonable that he starts with, cautioning that it ‘is not a moral psychology originating in the science of human nature’. Rawls is clear that this psychology has to be consonant with natural psychology, but his declared aim is to identify the ‘most reasonable’ moral psychology that the ‘scope of the world allows’. Rawls admits that, in trying to gauge that scope, we are groping in the dark: ‘the difficulty is that beyond the lessons of historical experience and such bits of wisdom as not relying too much on scarce motives and abilities (say, altruism and high intelligence) there is not much to go on. History is full of surprises’. He is forced to go beyond what he can confirm, into the realm of hope. This is equally clear when Rawls (2005: 158–168) draws upon history to support his stability argument. It is a highly stylised, conjectural history, written for the pre-established purpose of vindicating the principles of justice.

Kant believed that when adopting the conditions of the possibility of a priori morality on practical faith, our only limit is that we cannot adopt assumptions known by science to be impossible. Rawls employs a more demanding standard in that he thinks that the principles selected within the original position can only be accepted if we can show that a society governed on their basis could emerge and sustain itself as a moral community. Rawls offers ‘conjecture and speculation’ to give us a reasonable basis of hope in this possibility, but he never spells out how much we have to accept on hope.

Rawls (1995a) recasts theodicy as the vindication of the world as good. Justifying the world with reference to an immanent standard of political justice inevitably demotes the significance of evil. Where for the theist evil is a problem of cosmological proportions, for the social theorist it may simply appear as a remediable injustice or brute fact. Rawls reimagines theodicy through Hegel, for whom ‘the full task of theodicy is … showing us that our absolute need for reconciliation with the world as a whole’ is satisfied (Geuss, 1999: 82–83). This helps explain the unity of Rawls’ corpus. Rawls is concerned throughout to vindicate the social world as good, a project for which the existence of evil is merely one possible objection. Only upon turning to the abject international scene does he systematically try and show why the ‘great evils’ of humankind do not condemn us all.

Rawls, borrowing from Hegel, takes ‘reconciliation’ to be one of the principal tasks of political philosophy. Rawls (2008: 10–11) describes two forms of reconciliation in his Lectures on the History of Political Philosophy, and two sources of our need for reconciliation in the private paper ‘Political Philosophy as Reconciliation’ (Rawls, 1993). I believe they dovetail to reveal two overlapping theodicies.

One conservative ‘form’ of reconciliation, drawn directly from Hegel, strives to ‘calm our frustration against our society’ by establishing how its ‘institutions, when properly understood … are rational’ (Rawls, 2008: 10). Philosophy actualises the objective order of society by bringing this fact to the subjective consciousness of citizens, who are then able to ‘affirm’ society as rational (Rawls, 2000: 331–336). A story of this kind is at the heart of PL. Rawls locates one source of our ‘need’ for reconciliation in the loss of the possibility of a community founded on shared comprehensive doctrines after the breakdown of Christendom. PL salves this loss by vindicating the political unity of liberal society in a primarily retrospective conjecture about the development of an ‘overlapping consensus’. While political unity may be inferior to a unity anchored in comprehensive values, Rawls admits, this loss is outweighed by the gains of liberal pluralism, liberty and toleration.

As for the second ‘need’, it is worth quoting Rawls (1993):

‘The other source of the need for reconciliation is the evident pervasiveness of grave injustice in our social world, and indeed throughout history, arising in part from the will to dominate and to lord it over others, the great enjoyments to which the exercise of a successful will to dominate gives rise, all this abetted by vanity and greed, and other petty vices. But it is not from these propensities alone that injustice arises, but from the collective fears and rivalries of peoples and nations, as these become collective hatreds and anxiety’.

The ‘form’ of reconciliation salving this need derives not from Hegel, but Rousseau and Kant. It shows that evil does not impugn humankind by extending the boundaries of practical possibility to give us a basis of hope that the social world allows for perfect justice (Rawls, 2008: 10–11). We see this clearly in the section of LP dedicated to the Holocaust. Knowing that humans are capable of this ‘demonic madness’, Rawls (1999a: 19–23) fears, may ‘undermine our hope for the future of our society’. To show that the social world is good despite this terror, we must ‘follow Kant's lead’ and ‘envision a realistic utopia’, so that ‘no longer simply longing’, our hope for humanity ‘becomes reasonable hope’. If we fail, the ‘great evils’ of others ‘destroys us too and seals their victory’, condemning humankind as a depraved species. Rawls (1995b) invokes this same phrase – ‘great evil’ – to describe the atomic bombing of Hiroshima, whose devastation he witnessed first-hand in 1945. Rawls, remember, canvasses the aim of LP as the eradication of the ‘great evils of history’. When a correspondent queried him about the relevance of the Holocaust, Rawls (1998) responded plaintively: ‘It's an important topic for me personally and I have spent (wasted) an inordinate amount of time wondering and reading about it’.

These are overlapping theodicies in that they both defend the possibility of a moral community by adopting practical interpretations of human psychology and history. The difference is how much weight their respective aims place on these constructs. PL vindicates the existing order through a primarily retrospective story whose object of faith, an overlapping consensus, is posited as a largely accomplished development. In LP, however, the disjuncture between the existing world and a globe-straddling moral community is so great that Rawls must canvas his vision as a realistic utopia challenging the boundaries of the possible. The differentia specifica is that, on balance, LP takes the international scene as a problem to be overcome, not a liberal success to be affirmed.

A natural question is where Theory fits into this picture. We cannot read it through these two categories of reconciliation because it long antedates them. A complete re-reading of the text is beyond the scope of this paper, but in the next section I detail how Theory lacks both a theory of evil, and a theory of transition. It therefore has no answer to the problem of evil, and I would tentatively suggest that it sits closer to PL.

What are the ultimate stakes of LP? In short, the moral status of humankind. Rawls poses the question, ‘Is perfect justice possible?’, because to vindicate humankind Rawls only needs to show that we are the kinds of beings, and this is the kind of world, for whom perfect justice could be possible. While the realisation of the law of peoples is ‘not unimportant’, the ‘very possibility of such a social order can itself reconcile us to the social world’ (Rawls, 1999a: 127–128). Showing that the ‘deep tendencies and inclinations of the social world’ allow for the possibility of justice does not commit Rawls to any claim about the likelihood or practicality of justice. It is true that Rawls’ theory of moral learning delineates a progressive moral sequence which he claims, historically, we have travelled a certain way along, and prospectively, we must complete to achieve perfect justice. But nowhere does he claim that progress has to happen.

Natural goodness

Rawls has always seen humans as moral personalities who manifest their true nature in a community based on the reciprocal recognition of that fact. The Augustinian conception of the person in SF was merely the first. We therefore need to disaggregate Rawls’ various conceptions of the person to isolate those elements that ultimately combine to yield ‘Rousseau's answer’.

Bok (2017) and Galisanka (2019: 115–135) have shown how Rawls scoured post-war philosophy for the materiel out of which to reforge his moral conception of the person in a secular cast. Following Philippa Foot, Stephen Toulmin and others, Rawls sought to derive morality from the analysis of the Wittgensteinian ‘form of life’. He proposed that the human form of life is encased in natural capacities, which engender moral affects and relations, such that one cannot but encounter others as moral beings.

Rawls first conceived of this as the ‘natural basis of justice’ (Bok, 2017: 175), but ultimately folded it into Part III of Theory to confirm the possibility of a morally self-stabilising community (Rawls, 1999b: 397–441). Rawls contends that when we develop attitudes of love, trust and fellow-feeling towards other humans, this makes us ‘liable’ to moral feelings, such that when we violate out duties of care and cooperation, we feel moral guilt. Our natural affections lead to and support – though do not cause – moral life.

To envision how this could reliably scale to support a moral community, Rawls relies on Rousseau and Jean Piaget, one-time Director of the Rousseau Institute. Key is the idea, credited to Rousseau, that we have an extra-rational tendency to ‘acquire attachments to persons and institutions according to how we perceive our good to be affected by them’ (Rawls, 1999b: 406, 433). Rawls then argues that as individuals grow up in a just society and come to recognise how it benefits them, they develop reciprocal attitudes of love and trust that engender successively more sophisticated moral feelings. This follows Piaget's developmental psychology: parents care for our rational self-love, associates our social self-love, and finally we recognise that we, our family and friends benefit from enduring institutions, and are thus moved to adopt society's principles as our regulative desire.

This is, broadly, a Rousseaun theory of natural goodness. Rawls (1999b: 401–403) implies as much when he contrasts the ‘empiricist’ and ‘rationalist’ traditions of moral psychology. The first views humans as born without the pro-social capacities for morality, and the role of learning as the fallible remedy of this deficiency through a behaviourist programme of rewards and punishments. The second sees humans as born with capacities for good, and learning as the creation of the conditions for their proper maturation. Rawls sides with the latter, which he associates with Rousseau, and two thinkers marked by his psychology, Kant and Piaget.

Later, in Lectures on the History of Political Philosophy, we can identify two problems of stability. One is the mechanism of stability, in which Rawls (2008: 78, 239–241) contraposes Hobbes’ coercive solution to Rousseau's idea that just institutions ‘generate in those who live under them the general will needed to maintain it’. The other is the problem of how just institutions emerge in the first place. Rousseau's (1997: 68–78) answer is the baldly voluntarist idea that a lawgiver might independently bring a well-ordered polity into being under certain, extremely specific circumstances. Rawls (2008: 240) dismisses this as a deux ex machina and, as Brooke (2015) points out, inserts his own theory of transition from PL.

Theory only addresses the first problem. Its theory of moral learning assumes the pre-existence of a just society, and Rawls gives only a smattering of pages to the question of how to move from ‘non-ideal’ circumstances towards justice. Rawls (1999b: 335–343) discusses how civil disobedience might redress injustice, but on his view protest should express a fidelity to the regnant polity, and appeal to the sense of justice of the majority, to bring society back into equilibrium with its own principles. This presupposes that society is ‘nearly just’, and suggests that, for Rawls, there is no condition of penury from which to transition in the first place.

It is therefore unsurprising that Theory also lacks a theory of evil. Essential to the idea that we develop affection for those tending to our interests is, clearly, some account of what those interests are. Rawls (1963: 287; 1999b: 406) classes the child's interest at the first stage of moral development as ‘rational self-love’, which appears inspired by Rousseau's amour de soi. But at the time, Rousseau's conception of social self-love, amour-propre, was almost unanimously interpreted as a vainglorious desire for status, following inevitably from society, and leading only to human ruin. In the second stage of his scheme, therefore, Rawls appears to rely upon Piaget's (1948: 83–95) parallel but distinct concept of ‘mutual respect’. This is no simple substitution, however, for amour-propre is a bivalent capacity whose corruptibility later allows Rawls to theorise evil.

Without a theory of evil or of transition, this is far from ‘Rousseau's answer’ to the problem of evil. We therefore need to turn to how it becomes that theodicy.

Let us start with Rawls’ (2000: 291–308) lectures on Kant, where he contends that Kant's Religion Within the Boundaries of Mere Reason contains an ‘Augustinian moral psychology’. Human nature comprises two forms of self-love (amour de soi and amour-propre) and an inalienable feeling for the moral law (moral feeling). Morality is a question of how these capacities are ordered, of whether we succumb to pride and give self-love free reign, or subordinate our self-love to the moral law. On this Augustinian view, ‘moral failures… must all arise, not from the desire of our physical and social nature, but solely from the exercise of our free power of choice’. Rawls contrasts this to a ‘Manichean’ psychology lingering in some of Kant's remarks, which divides humans into the ‘good self we have as intelligences belonging to the intelligible world; and that … bad self we have as natural beings belonging to the sensible world’. If this were true, our very nature would be inimical to morality. What does this tell us? Rawls can read Kant through Augustine because Kant is working from Rousseau, who mirrored Augustine. It follows, transitively, that Rawls is also reading Rousseau through the Augustinian argument in the first step of his Christian theodicy.

Written in the mid-1980s, these lectures suggest that Rawls still lacked a way of interpreting Rousseau's amour-propre consistent with the possibility of justice. Rawls (2000: 292) indicates that, for Kant, evil is ‘the inevitable outcome’ of the impact of society on amour-propre. It is therefore striking that Rawls’ lectures on Rousseau, written in the early 1990s, are organised around precisely this question. How can the ‘dark and pessimistic’ figures in the Discourses, where it appears ‘inevitable that the social evils and vices that Rousseau deplores will come about’, possibly support the ‘sunnier’ vision of justice in Emile and The Social Contract (Rawls, 2008: 193–194, 197)? Solving this exegetical problem is one and the same as finding an answer to the problem of evil.

The revelation is N.H. Dent's 1988 Rousseau, which reads amour-propre as a bivalent capacity (Brooke, 2015: 436–441; Rawls, 2008: 198–200). In its ‘unnatural’ form, amour-propre becomes a desire ‘to be superior and to dominate others, and to be admired by them’. But in its ‘natural’ form, it is merely a desire for equal standing, providing the affective basis for mutual respect – the recognition that your limits on my action, are equal to my limits on your action. If social institutions guarantee our equal standing, ‘moved by amour-propre we are ready to accept and act on a principle of reciprocity’ when it is made available to us (Rawls, 2008: 199, 219). But to do this, we require a further capacity: moral feeling. On this reading, the Discourses show how the iniquities of civilisation disfigured amour-propre, while Emile and The Social Contract show how amour-propre can be turned into the affective wellspring of mutual respect under egalitarian institutions.

It is this version of Rousseau's theory of natural goodness which undergirds LP. In the draft in which Rawls (1995a) reflects on theodicy, he explains that ‘following Rousseau, the world is good in our eyes’ when a ‘perfectly just society’ is possible given ‘men as they are and laws of nature as they might be’. Rawls recapitulates the reading of Rousseau from his lectures, insisting that this allows us to reconcile the Discourses to The Social Contract in an alternate understanding of evil to ‘the Christian doctrine of original sin’. Similarly, Rawls (1999a: 6–7) begins LP by declaring that the ‘great evils of history’ are attributable to political injustice, before continuing:

Following Rousseau's opening thought in The Social Contract, I shall assume that his phrase ‘men as they are’ refers to persons’ moral and psychological natures and how that nature works within a framework of political and social institutions; and that his phrase ‘laws as they might be’ refers to laws as they should, or ought to be … to say that human nature is good is to say that citizens who grow up under reasonable and just institutions … will affirm these institutions and act to make sure their social world endures.

Rawls transposes his moral psychology into a sharply-drawn theory of world politics, conceiving of liberal and decent ‘peoples’ as societies with a developed moral nature, in contrast to the maligned nature of ‘outlaw’ states, and undeveloped moral nature of ‘burdened’ societies.

Societies are moved by two ‘fundamental interests’. First, they want to secure their means of survival (amour de soi): their independence, territory, security, and the well-being of their citizens (Rawls, 1999a: 29, 34). Second, they desire to be recognised as equals among others. This ‘falls under what Rousseau called amour-propre’ (Rawls, 1999a: 34). But as we know, it has a twin nature. On the one hand, societies can become ‘inflamed by… arrogant or wounded pride or by lack of due self-respect’ (Rawls, 1999a: 47). On the other hand, when it comes to peoples, ‘the due respect they ask for is a due respect consistent with the equality of all peoples’ (Rawls, 1999a: 62, 35). While the latter is an essential precondition of justice, peoples require a further capacity to understand and act from principles of justice – moral feeling – which then regulates ‘their fundamental interests as permitted by their conceptions of right and justice’ (Rawls, 1999a: 29). In contrast, ‘states’ pursue their interests without these limits: their malformed amour-propre kindles a desire for expansion, wealth and domination, enacted through an unscrupulous policy of realpolitik (Rawls, 1999a: 28–29).

Rawls (1999a: 46–48) suggests that in the ideal of the law of peoples, liberal peoples are ‘satisfied peoples’, that because their ‘fundamental interests are fully compatible with those of other democratic peoples’, they ‘have nothing to go to war about’. What does this mean? First, their amour de soi is satisfied: trade allows the effective acquisition of goods and the Society of Peoples affords collective security. Second, their amour-propre is satisfied: peoples are not inflamed by ‘wounded pride’, or ‘swayed by the passion for power and glory’, and they ‘mutually respect one another’. For Rawls, it is because the law of peoples benefits peoples, and encourages their mutual care, that peoples are ultimately moved to adopt its principles.

Transition

Rawls has tried to show that evil arises from the contingent perversion of our nature by unjust social institutions and that we have latent capacities for good. But to complete this theodicy, he must provide a homologous but naturalised version of the second step of his Christian theodicy, of how we can transition from evil to good. While the first step of Rawls’ secular theodicy consciously reworks his Augustinian vindication of the natural cosmos, the same is not true of this second step. Rawls is not writing his theory of transition in the mould of the account of conversion in SF. But the homology persists, for the fact that Rawls begins from the same quasi-Augustinianism forces the corollary: how can a world of evil people lift themselves out of abjection?

We saw that Rousseau's lawgiver does not provide a workable answer. To whom then does Rawls turn? ‘The Law of Peoples is greatly indebted to Kant's idea of the foedus pacificum and to so much of his thought’, Rawls (1999a: 10, 86) tells us, and his ‘basic idea is to follow Kant's lead’. Kant's influence is, therefore, pervasive. Rawls (1999a: 36) thinks it enough to recite Kant to defeat the idea of world government, treats the derivatively Kantian ‘democratic peace thesis’ as the empirical anchor of his argument, and defines Raymond Aron's ‘peace by satisfaction’ as the ultimate end of the society of peoples, a condition which Aron (2003: 161) thought achievable by one of two means: universal empire or Kantian law. Rawls, as we already know, relies on Kant's doctrine of practical faith to ground this theodicy, the primary object of which in his lectures on Kant is a plan of nature tending towards a foedus pacificum.

LP adapts key features of this plan, which – schematically – runs as follows. ⁵ In the state of nature, the inclinations of uncoordinated wills collide, creating evils so great that self-interest compels its victims to erect a civil condition. This is how the first rudiments of both domestic and international lawfulness emerge, but from thereon they can only develop towards morality through an interactive, positive feedback loop. As states move closer to peace, societies have more room for moral learning, softening relations between states, and so on. Central to the idea that domestic moral learning strengthens international peace is the thesis that republics – Kant's ideal constitutional form – do not wage war because this requires the consent of the people, who are loathe to acquiesce because it is they who bear its costs. In time, ‘partly by an optimal internal arrangement of the civil constitution, and partly by common external agreement and legislation, a state of affairs is created which… can maintain itself automatically’ (Kant, 1991: 48). Rawls, whose two copies of Kant's writings of international politics are covered in a thicket of marginalia, understood this argument well.

Rawls (n.d.) takes from Kant that ‘evil tends to destroy itself’ through the ‘invisible hand of self-interest’, and again like Kant, pitches this story at both the domestic and international levels. In PL, Rawls (2005: xxxviii-xxxix, 158–168) confirms the possibility of an ‘overlapping consensus’ with a conjectural history of its emergence, modelled on the Reformation. Admitting no doctrinal compromise, Catholicism and Protestantism were locked in a ‘mortal combat’ whose only escape was ‘circumstance and exhaustion’. Relenting, they accept a modus vivendi for an entirely negative reason: as ‘the only workable alternative to endless and destructive civil strife’. We find a parallel argument scattered across LP: lawlessness suffocates and exhausts states, leading them to enter cooperation to escape geopolitical precarity and economic isolation (Rawls, 1999a: 34, 113, 123).

Rawls combines this with Rousseau's tripartite psychology to explain how moral learning gets off the ground: evil creates the conditions – social cooperation – of its own eradication. In PL, a self-interested modus vivendi slowly builds trust and cooperation, and a recognition of the worth of liberal toleration, leading to a principle-based constitution, and then a conception-based overlapping consensus (Rawls, 2005: 158–168). In a similar progression, self-interest brings societies together (amour de soi), who then ‘tend to develop mutual trust and confidence in one another’ (amour-propre), before coming to affirm the law of peoples ‘as advantageous for themselves and those they care for’ and accepting it as an ‘ideal of conduct’ (moral feeling) (Rawls, 1999a: 113).

Rawls also adapts Kant's theory of the interdependent development of domestic and international morality. On the one hand, Rawls (1999a: 48–51, 54) hypothesises that as societies approach the ideal of domestic justice, they become increasingly peaceful. Rawls appears to offer two justifications for this claim. First, the motives of the citizens of well-ordered peoples (Rawls, 1999a: n27): it is in their interests to secure a stable basis of trade and security, they are not corrupted by the pathologies of pride and envy because their societies enshrine the social basis of self-respect, and their conceptions of justice are compatible with the toleration of other societies. Second, their governments represent their people, such that the irenic motives of citizens guide policy (Rawls, 1999a: 24, 53, 50, 75). Other societies, Rawls decrees, lack one or both of these internal characteristics. ⁶ On the other hand, Rawls endorses ‘Kant's idea that a constitutional regime must establish an effective Law of Peoples in order to realize fully the freedom of its citizens’ (1999a: 10). He cites Kant's (1991: 47–49, 90) contention that the internal freedom of societies ‘cannot be solved’ until external peace is achieved. This is borne out at crucial junctures in the argument. Rawls (1999a: 62, 122) claims that a decisive reason for admitting decent peoples into the foedus pacificum is that, if this were denied, it would injure their self-respect and, therefore, their internal moral development. Rawls (1999a: 48, 81), like Kant, also suggests that peoples can only remain ‘satisfied’ if there is general peace.

If peace depends on internal justice, and internal justice on peace, then the two can only develop in lockstep. This is importantly disanalogous from the account of moral development in Theory in that it does not concern the maturation of individuals from birth within a just society, but pre-existing societies formed outside of an international civil condition. In this sense it represents a genuine transition from evil to good. It is difficult to model this transition on a theory of individual moral learning – like Theory – as this involves the fanciful scenario of autarkic states entering into a perfect law of peoples. Rawls thus relies on a theory of collective moral learning – like the steps to an ‘overlapping consensus’ in PL – that sets out how the society of peoples as a whole might emerge through a process of interdependent development.

Rawls (1999a: 45, 48–54) claims that his ‘conjecture’ that the law of peoples would be stable ‘needs to be confirmed by what actually happens historically’. He appeals to the democratic peace thesis: ‘since 1800 firmly established liberal democracies have not fought one another’. But this does little to support Rawls’ stability argument. It does not show that liberal peoples would not attack non-liberal societies, and says nothing of ‘decent’ peoples. While there is consensus on the absence of conventional war between well-established democracies, there is little agreement on its cause. Rawls can find no support for the idea that peoples have a moral nature that moves them to act on the law of peoples. Wenar and Milanovic (2009) have shown that Rawls argues far beyond the social science.

Read carefully, however, Rawls (1999a: 51–54) only suggests that the historical foreign policy of democracies is ‘not incompatible’ with the ideal of the law of peoples. He is cognizant of the ‘great shortcomings of actual, allegedly constitutional regimes’, that they have constructed globe-straddling empires, pursued war ‘for expansionist reasons’, and recurrently intervened ‘in weaker countries, including those exhibiting some aspects of democracy’. If Rawls knows there is a drastic shortfall between the actual behaviour of democracies, and what is required of the law of peoples, how can he invoke the former to justify the possibility of the latter? The answer is that he is merely claiming that whatever its shortcomings, the achievement of peace between well-established democracies gives us a reasonable basis of hope in the possibility of a global moral community. Rawls reveals just how much he is asking us to take on practical faith.

If the perfection of domestic justice requires general peace, and the only way to establish general peace without Hobbesian coercion is the law of peoples, then a condition of this entire project is the universalisation of justice. We see this in Rawls’ extensive discussion of non-ideal theory in LP. Ideal theory assumes: (a) conditions favourable to justice, and (b) full compliance with justice. Non-ideal theory collapses these assumptions to ask how, in ‘our world with its great injustices and widespread social evils’, we can gradually work towards ‘a world in which all peoples accept and follow the (ideal of the) Law of Peoples’ (Rawls, 1999a: 89). According to Rawls, we have to identify the ideal first, which then gives non-ideal theory its objective and orientation. Rawls discusses two kinds of non-ideal society, corresponding to (a) and (b).

‘Burdened societies’ endure unfavourable conditions. Rawls (1999a: 108) is adamant that there are no natural obstacles to justice, only contingent political obstacles. The social world must be constitutively ‘friendly’ to justice, remember. Rawls (1999a: 108–110, 116–117) touts Amartya Sen's finding that famines arise from political negligence, heralds the extension of human rights to women as a fortuitous solution to over-population, and chides Charles Beitz's argument that some societies are arbitrarily advantaged by their natural resources. For Rawls, then, while peoples have a duty to assist these societies, ‘assistance’ is not primarily economic, but help creating the right politics.

Rawls stipulates that ‘outlaw states’ violate the human rights of their own citizens and perpetuate the cycle of war. He suggests that peoples should be ready to levy sanctions and deny outlaw states entry to ‘mutually beneficial cooperative practices’, tipping the scales of their self-interest in favour of cooperation (Rawls, 1999a: 93). If this fails, ‘no peaceful solution exists except domination by one side or the peace of exhaustion’ (Rawls, 1999a: 123). According to Rawls (1999a: 80–81, 93–94), peoples have casus belli against ‘outlaws’ in two instances: to quell severe human rights violations, and in self-defence. The other option, ‘exhaustion’, expresses Rawls’ hope that life outside the shelter of the society of peoples will take its own toll, and force a rapprochement.

Conclusion

I have reconstructed Rawls’ theodicy on its own terms, studiously avoiding commentary or critique. I now break from that mould, and return to the animating question of this paper, the quasi-religious basis of the idealising and moralising character of Rawls’ thought.

It is helpful to begin with the ‘ethics-first’ approach to politics, which Raymond Geuss (2008) takes Rawls to exemplify. On this view, ethics is a self-contained domain, divorced from history, politics and economics, whose role is to formulate an ideal of society in abstracto, and then in a second step apply this antecedently given framework to society. In constructing a theory of ideal justice, it finds it necessary to idealise individuals, society and institutions, to ask how a polity would function were it fully moral. This creates a warped social ontology, where the motors of modern world history – states, empires, capitalism, ideologies, energy – are consigned to oblivion at the same time as highly selective, idealised scraps of information are used to create an etiolated picture of ‘how-the-ideal-world-works’. Why, as Charles Mills (2005) eloquently protested 20 years ago, would ignoring the sources, dynamics and context of injustices help us to identify and address them?

What I wish to add is that, once we appreciate this theodicy, we can see Rawls’ ethics-first approach, and how his quasi-religious ambitions constrain his wider philosophy, with new clarity, and the weakness of this construction when viewed from outside of Rawls’ unworldly commitments.

First, Rawls’ totalising theory of world society, anchored in Rousseau’s three-stage moral psychology, and set in motion with Kant’s ‘plan of nature’. Are states really corporate persons whose behaviour can be read off from the balance of their ‘moral psychology', and whose internal political development is not disadvantaged by anything non-moral? It is impossible to square this theory with history. Rawls wants to bridge the chasm with practical hope, but it is difficult to see how this matters as a question of action-guidance. If it is exceedingly unlikely that societies act as Rawls postulates, why act on prescriptions premised on the strength of that imaginary supposition?

Second, Rawls’ decent Muslim people, ‘Kazanistan’. Rawls confesses that whether societies of this kind exist or not is beside the point; it is a hypothetical ideal-type to show that if a non-liberal but decent society did exist, it would be found legitimate by, and be moved to support, the society of peoples (Rawls, 1995a: 21–24). Its point is to say something about the latent moral structure of the social world, not the empirical nature of the actual world. Or, as Rawls puts it, its focus is not practical but realistically utopian. It is therefore to be expected that, as Idris (2020) has scrupulously detailed, this idealisation of a fictional ‘Muslim people’ goes hand-in-hand with a carelessness as to the actual history and politics of the Muslim world, and its structural position in the world system, contact with which immediately calls into question LP’s coherence. This, like so much of LP, raises profound questions of ideology critique.

What is fundamentally at issue is whether a single philosophical system can coherently harbour Rawls’ redemptive ambitions and serve as an appropriate guide to action. If features of this theodicy run the gamut of Rawls’ secular thought, and mainstream Anglophone political philosophy emerged from Rawls’ shadow, what light does this cast on the discipline?