Four concepts of rules: A theory of rule egalitarianism

Social justice – A hollow incantation but still a reasonable concern

Classical liberals are not usually portrayed as being sympathetic to a social understanding of justice. Early classical liberals, such as Adam Smith, David Hume and John Locke, were adamant about the importance of the rule of law, insisting that rules must treat everyone as equally free, but, overall, they spent little if any time talking about social justice. Afterwards, modern classical liberals, like Friedrich Hayek, Milton Friedman and Ludwig von Mises, were actively hostile to the idea of social justice, as they were primarily concerned with defending market capitalism against socialism. Hayek (2013: xix), for example notoriously characterised social justice as a ‘mirage’, as well as a ‘fraudulent’, ‘meaningless’, ‘thoughtless’, ‘empty’ and ‘vacuous’ concept. Lest there be any doubt about his views, Hayek also wrote the following – ‘I believe indeed that the greatest service I can still render to my fellow men would be if it were in my power to make them ashamed of ever again using that hollow incantation’. Such provocative remarks have been the source of much criticism on the part of those who follow Rawls (1999: 3) in considering justice as the ‘first virtue of social institutions’. Classical liberalism, hence, is usually thought to be incompatible with the main intuitions of what is variously referred to as ‘high liberalism’, ‘liberal egalitarianism’ or ‘social justice liberalism’. This assumption, I argue, is incorrect. The main contention of this article is that the classical liberal tradition can be reconciled with social justice, but, for that to happen, we must first ensure that a concern for the rule of law is etched in the very core of our understanding of social justice. I accordingly propose an original liberal approach, which I term ‘rule egalitarianism’, that will explain how this seemingly contradictory merger of classical liberalism and social justice is both conceivable and necessary through a proper understanding of the generality and abstractness of the rules of market institutions.

The aforementioned idea that classical liberalism is hostile to social justice is now shared by high liberals like Freeman (2007: 45) as well as, perhaps more surprisingly, by neoclassical liberals such as Brennan and Tomasi (2012). Neoclassical liberals are the intellectual heirs of the classical liberal tradition, though they nonetheless take a stance against the supposed classical liberal disapproval of social justice – many of them preferring to call themselves ‘bleeding-heart libertarians’. Neoclassical liberals who want to find a place for social justice within classical liberalism, therefore, feel they have to look to the high liberal tradition to find the intellectual tools permitting them to carry out such a project. Hence, neoclassical liberals concerned with social justice adopt a high liberal justificatory strategy, like Tomasi did on his own admission (2012: 95), while, in fact, I think, they need not do so. Simply put, neoclassical liberals adopt the high liberal assumption that the social order must be justifiable to everyone, including the worst-off, and they think that market institutions can be so justified – that is classical liberal institutions can be derived from the high liberal understanding of social justice. Brennan (2007), for example interprets Rawls’s difference principle as requiring the long-term maximisation of the welfare of the worst-off, and then claims that we should thus favour market capitalism. Unfortunately, such a neoclassical liberal approach does not explain how social justice may be valuable solely from a classical liberal viewpoint, and why some classical liberals had indeed good reasons to defend some aspects of the modern welfare state. Overall, then, the classical liberal tradition is still generally thought to be incompatible with social justice.

This portrait of social justice in the classical liberal tradition, however, is inaccurate. Once we get past the scornful rhetoric used by many classical liberals to talk about social justice, we can in fact see that they have repeatedly endorsed anti-poverty and even some egalitarian policies. Consider, for instance, the following measures Hayek recommended – an assured minimum income to counter extreme poverty (2013: 249), a reform of property rights and inheritance to reduce inequality of opportunity (2007: 134), and a limited form of progressive taxation to compensate for some indirect taxes that place a proportionally heavier burden on the smaller incomes (1978: 307). Are these about social justice? I think so, and, if I may add, these measures are perfectly acceptable from a classical liberal viewpoint. One could then ask the following question – why are these measures acceptable for classical liberals, who, need we remind, have also called social justice a ‘humbug’ and a ‘will-o-the-wisp’? The answer can be found in Lister’s recent article on Hayek, inasmuch as we take his main conclusion backwards. Lister (2013: 435) concluded by saying that ‘what Hayek resists teaching us is that the legal framework of capitalism can be designed to reduce inequality in the name of justice’. I believe the opposite is true. The legal framework of capitalism can be designed to reduce inequality in the name of justice, and this is precisely what Hayek and other classical liberals have taught us, though perhaps not in so many words – that is social justice within a capitalist framework is quite possible from a classical liberal point of view.

This article accordingly proposes a distinctively classical liberal theory of social justice, which does not rest on any of the high liberal premises. I call such an approach ‘rule egalitarianism’ so to emphasise its dual focus on the rule of law and social justice. Such an approach differs from the neoclassical liberal theory, according to which social justice requires market institutions. I propose a more modest, but also more widely palatable view – classical liberal ideas on market capitalism permit and even require some egalitarianism. I will first (second section) examine the classical liberal understanding of formal equality so to afterwards explain why classical liberals should rather favour the concept of isonomy, which is about the generality and the abstractness of the rules. I will then (third section) distinguish four concepts of rules, only one of which will be fully satisfactory for egalitarian endeavours from a classical liberal viewpoint, (fourth section) and then see why social justice becomes acceptable once we embrace the liberal commitment to the generality of rules. I conclude (fifth section) by explaining how classical liberals have repeatedly promoted social justice within the limits of the rule of law, much like neoclassical liberals or bleeding-heart libertarians do today, though without using high liberal assumptions.

The ultimate goal of this article is rather ambitious – it proposes to rescue egalitarianism from a skewed classical liberal understanding of formal equality, which is so ingrained that it is almost unquestioned. A theory of rule egalitarianism will permit us to carry out such a rescue, since it will balance the classical liberal concern with freedom under the law with an egalitarian theory focused on the basic structure. The high liberal tradition has already achieved a similar goal through what is now commonly called ‘justice as impartiality’, as defended, for example by Nagel (1995), Barry (1995) and Tan (2004) – that is social justice as applying to institutional arrangements and not to personal choices within the rules of such arrangements. Though rule egalitarianism will accept this line of argument must famously defended by Rawls (1955) in his seminal article ‘Two Concepts of Rules’, it will also go further, arguing that not any kind of rules may be enacted by liberal institutions. Simply put, the main argument of this paper is as follows – the classical liberal tradition has always favoured general rules over particular rules, and such a commitment to rules reaching beyond particular events explains the inescapability of social justice from a liberal point of view, provided we take the problem of imperfect generalisation seriously. Such a conclusion, showing how social justice stems from the main classical liberal premises, will be a useful addition to the literature that is too often silent about the classical liberal advocacy for social justice.

Equal treatment and the generality of rules in the liberal tradition

Let us begin by examining the classical liberal understanding of equal treatment. Egalitarian policies, after all, are cases of differential treatment inasmuch as they redistribute economic resources by treating rich and poor people differently. This section shows that many classical liberals have used a misguided understanding of formal equality to criticise egalitarian policies treating people unequally. Indeed, differential treatment is in no way a real problem from a classical liberal point of view, as the idea of formal equal treatment is at best an illusion – that is people cannot be treated equally.

At its core, the wariness of classical liberalism towards social justice springs from an apparent opposition between the vague concepts of ‘legal’ and ‘material’ equality, as it was often thought that these two are incompatible with one another. Friedman (2002: 195), for example contended that liberals and egalitarians argue for two irreconcilable understandings of equality, namely ‘equality of rights’ and ‘material equality’. Friedman accordingly asserted that one cannot be both an egalitarian and a liberal – that is liberals must reject egalitarian policies that gut equality of rights by taking from some to give to others. More recently, from a neoclassical liberal point of view, Tomasi (2012: 35) said that classical liberal systems of formal equality are inimical to social justice endeavours, and therefore, Tomasi argued, high liberals have had to advocate for a substantive conception of equality that goes well beyond what classical liberals can accept. That is, while classical liberals will solely argue for formal equality of opportunity, high liberals will argue for social justice, thus addressing arbitrary inequalities of birth and family circumstances about which classical liberals putatively remain silent. Perhaps the clearest statement of this opposition between formal and substantive equality comes from Hayek (1978: 87), who said the following:

From the fact that people are very different it follows that, if we treat them equally, the result must be inequality in their actual position, and that the only way to place them in an equal position would be to treat them differently. Equality before the law and material equality are therefore not only different but are in conflict with each other; and we can achieve either the one or the other, but not both at the same time. The equality before the law which freedom requires leads to material inequality.

In his book The Free and Prosperous Commonwealth, Mises (1962: 28) wrote, ‘Men are not equal, and the demand for equality under the law can by no means be grounded in the contention that equal treatment is due to equals’. In this quote, Mises chaotically used or alluded to no less than five different meanings of equality, namely two characteristics of persons as well as three modes of treatment by others:

(C1) Innate inequality – i.e. the fact that physiologically, neurologically, and socially individuals are not born the same.

(C2) Achieved inequality – i.e. the fact that individuals will acquire different social positions based partly on their own merit or through sheer luck.

(M1) Equal treatment under the law – i.e. being treated through the application of some rules with due process.

(M2) Strict equal treatment – i.e. being treated the same as the relevant comparison group.

(M3) Treatment as an equal – i.e. being treated with the same level of concern and respect as the relevant comparison group.

In other words, any instance of treating some entities the same in some respect will necessarily involve treating other entities differently in this respect, or possibly treating the same entities differently in other respects. For example, if the set of entities is picked out by the property of rational agency, as in contract law, one will treat competent adults the same, but most children and people with some cognitive disabilities differently. Or if the specified form of treatment is provision of income, then in ensuring the relevant set has the same level of economic resources, one will inevitably be responding differently to people’s choices – equality of income will indeed mean inequality of wealth, given different choices. No treatment can be purely equal, as one will always be treated differently on at least some level. The question, then, is not whether we endorse equal treatment, as it is rather about finding a liberal way in which people can be treated unequally. By framing the debate in terms of equal and unequal treatment, classical liberals have simply created an illusion. We must recognise that every application of a rule qua rule will treat some people differently in some particular circumstances, and this is not something that is specific to social justice. Classical liberals imperatively need to acknowledge such a reality so that they may have a better understanding of both the rule of law and social justice – and how these two are not necessarily opposed to one another.

The real debate for the classical liberal tradition, I think, is between general and particular rules – and it is through a proper understanding of the generality of the rules of market societies that we may reconcile classical liberalism with social justice, leading us to a rule egalitarian theory that offers a real alternative to the neoclassical liberal approach. First, the analysis above shows that it is impossible to be treated the same in all respects with regard to all comparison groups, and, second, it is doubtful that we should even want to treat people as equally as can be. After all, there is nothing wrong from a classical liberal viewpoint with maternity programs that only benefit women, or state-funded homeless shelters that only benefit homeless people, though these are obvious cases of differential treatment. Consider now what Hayek wrote (1978: 154) on the generality of the law:

The requirement that the rules of true law be general does not mean that sometimes special rules may not apply to different classes of people if they refer to properties that only some people possess. There may be rules that can apply only to women or to the blind or to persons above a certain age. (In most such instances it would not even be necessary to name the class of people to whom the rule applies: only a woman, for example, can be […] got with child.)

Four concepts of rules – Social justice within the bounds of the law

Let me now explain what generality and abstractness entail in a rule egalitarian theory. These two features, I am well aware, are not sufficient to guarantee that rules will in fact be compatible with the rule-of-law tradition. One could add, for instance criteria of intelligibility, reasonability, non-contradiction, and non-discrimination. The generality and abstractness of the law are nonetheless particularly relevant to any egalitarian debate, as they will later permit us to establish a distinction between nomos-observing and nomos-neglecting egalitarianism, as well as they will explain why social justice is crucial from a classical liberal viewpoint. Let me also mention that Hayek defined the concepts of ‘generality’ and ‘abstractness’ in many different and often contradictory ways. For example he occasionally used the concept of generality as an attribute of abstractness, thus blurring the distinction between these two terms. Although in defining generality and abstractness I essentially follow Hayek’s intuitions, I will endeavour to introduce a more workable definition of these two concepts than he did, so to be able to introduce four clearly distinct concepts of rules. Hayek himself was inspired by the ancient Greek process of nomothesia – that is legislation – for the enactment of laws of general and permanent validity, which took precedence over simple decrees. This intuition remains central for the classical liberal tradition, committed to the rule of law, and is also, I believe, to be fundamental for the liberal study of social justice moving onwards.

The generality of the law means, as Fuller mentioned (1969: 47, 210), that the law must act impersonally, namely that it must apply to general classes of people, contain no proper names, and must not attempt to direct the conduct of specific individuals. In the eighteenth century, Paley (1860: 235) expressed it as follows – ‘general laws are made […] without forseeing whom they might affect; and, when made, […] let them affect whom they will’. Simply put, the classical definition of generality focuses on the relative size of the population affected by the rule, as well as on the impersonal character of the rule itself. The abstractness of the law means that the rules must apply to an unknown number of future instances so to permit the formulation of reasonable expectations (Hayek, 2013: 97, 203). It thus permits people to plan private arrangements, secure in the knowledge of their protection. As Lord Mansfield classically established, ‘it is of more consequence that a rule be certain, than whether the rule is established one way rather than the other’ (Vallejo v. Wheeler, 1 Cowp 143 (1774)). I will further detail these two notions in a moment, so to see the consequences they have for rule egalitarianism, but for now let us see how they may affect our understanding of rules.

Following these two features of isonomy, we may establish a distinction between four concepts of rules. That is, we may distinguish between different ways through which any state policy can be framed. Though the above table presents the four concepts discontinuously, I should mention that these concepts are to be understood as continuous – that is they are part of a spectrum. It is a matter of degree, approximated for purposes of representational convenience by dichotomies. The goal is to have general and abstract rules as much as possible, such that if generality and abstractness need to be reduced or sacrificed that should be seen as regrettable and as a cost.

Using these four concepts of rules, I now propose two major conclusions that will permit us to see the errors of the previous approach focused on equal treatment, as well as to offer an alternative to the neoclassical liberal understanding of social justice. First, it is possible to pursue social justice from a classical liberal point of view inasmuch as we do so in proper institutional terms, namely through general and abstract rules. Second, as the next sections will show, once we accept the liberal commitment to general rules, which will inevitably have imperfect consequences in particular circumstances, we must also accept social justice as a way to correct the errors of generalisation in a market society.

Classical liberals, of course, will agree with high liberals that social justice should apply to institutional arrangements and not to personal choices within such arrangements – that is justice as impartiality –, and yet they will probably want to have additional guarantees that social justice will not overly hinder the people’s freedoms within a capitalist framework. This is where general and abstract rules come into play, providing us with the first step towards a reconciliation between the welfare state and the classical liberal tradition. Let me now further illustrate these two features of isonomic treatment, so that we may understand the strength of the first conclusion for the classical liberal tradition – that is social justice can indeed be pursued through general and abstract rules.

First, one of the most classical debates on the generality of the law is to be found in the case of Cicero’s exile. In 58 B.C., the tribune of the plebs Publius Clodius Pulcher introduced a series of laws – that is the Leges Clodiae –, one of which – the Lex Clodia de Civibus Romanis Interemptis – was engineered to personally drive Cicero into exile since he was a known enemy of Julius Caesar. Without explicitly naming Cicero, the law was still ingeniously directed at him alone. It stated that whoever put to death a Roman citizen without trial should be exiled – that is forbidden fire and water. The wording of the bill, however, clearly singled out Cicero. It was directed at him alone, following his role in 63 B.C. in the second Catilinarian conspiracy, during which he indeed executed some conspirators to defend the late Republic, though without formal trial. Naturally, Cicero complained that such a law was not a law proper, but rather a ‘privilegium’ (De Legibus (The Laws): 3.44) – that is a privilege or a prerogative. It disregarded the Twelve Tables, the foundation of the Roman law, and more specifically the ninth table – ‘No privileges, or statutes, shall be enacted in favor of private persons, to the injury of others contrary to the law common to all citizens, and which individuals, no matter of what rank, have a right to make use of’. Therefore, following Roman law, a general rule is one that is not directed at private individuals, with or without naming them, as well as one that does not introduce unequal statuses for the people – for instance individual-specific rights, separated from the general body of rights available to all citizens.

In the field of distributive justice, an obvious example of a general set of rules can be found in Rawls’s theory of justice, which intended to distribute social primary goods across social positions. To frame distribution to benefit the least advantaged is not to frame distribution so that it applies to some particular individuals. It is rather to frame distribution so that whomever the least advantaged are identified to be, they will be the ones to benefit. Contrary to what G.A. Cohen suggested (1995: 87f), hence, the difference principle does not require benefiting the particular individuals who happen to be worst-off. It requires raising the worst-off position, so that whomever it is that occupies this position, they are as well-off as those in the worst-off position can be. Classical liberals share a similar understanding of social justice. ‘There is every reason to help the poor man who happens to be a farmer’, Friedman contends (2002: 191), ‘not because he is a farmer but because he is poor’. The distribution of burden or benefit is then impersonal – it does not matter why one is poor, whether one deserves to be poor, or what group one belongs to. The message is quite simple – if you are poor, you will receive some help. Today, as some classical liberals noted, many egalitarian regulations are non-general, as they target specific entities. Although they do not always name specific entities, such regulations are not intended to apply to general classes of people. They rather help people as members of occupational or age groups, if they do not directly target particular organisations and industries.

Let me pre-empt one important question. Why do classical liberals think that the category ‘farmers’ is not a general one for wealth redistribution, while ‘poor’ and the ‘least-well-off’ are indeed general categories for the same purpose? The classical definition of generality, after all, focuses on impersonal treatment and the relative size of the population – and using such a definition, we could indeed think that ‘farmers’ is a general category, much in the same way as ‘poor people’ was general according to Milton Friedman. The problem in this case is that a definition of generality is somewhat incomplete if it is only focused on the impersonal treatment of a sizable population. Generality is also about the relative size of the population. The question, then, is the following – relative to what? As Schauer explains (1995: 635), general rules must be understood through the act of ‘giving reasons’. Simply put, because of their impersonal character, general rules rest on reasons that are ‘broader than the outcome they are reasons for’. When we help poor people, the reason behind the rule is to ameliorate poverty. Giving money to a subset of the population at large, identifiable under the description ‘farmers’, however, fails to embrace such a reason. In this case, the relative size of the population is not general enough, or so classical liberals commonly argued.

The case of the farmers, as detailed by Friedman, is then a case of failed generalisation, that is of the ‘factually unfounded attribution of a characteristic to a group’, to use Schauer’s words (1997: 292). For other rules, of course, ‘farmers’ would indeed be a general population, as there are other perfectly acceptable reasons to subsidize farmers. Concerning Friedman’s example, tough, we may say that there is indeed a problem with a rule intending to help the poor that only helps farmers. I will come back to such a question in the next sections, using the reason-giving aspect of general rules to justify social justice. But, again, the concept of generality should not be understood as discontinuous – it is a matter of degree. If the goal is to ameliorate poverty, helping poor people is more general than helping poor farmers, which, in turn, is more general than helping poor potato farmers, or simply helping your poor neighbour, John the farmer.

Second, regarding the abstract character of the law, the classical liberal tradition has always been adamant – one must be able to know what one may count on in one’s endeavours inside a given basic structure (Hayek, 2013: 41ff). This is precisely what abstract rules permit, since once one knows the rules, one will be able to direct one’s conduct so to achieve one’s goals. Such a definition of abstractness would also encompass the traditional requirements of non-retroactivity and publicity – as retroactive or unpublicised rules would foil any attempt to formulate some reasonable expectations. The case of Cicero’s exile, once more, is a good example of non-abstractness. Cicero was exiled for undertakings he had done many years before the Lex Clodia de Civibus Romanis Interemptis. A scheme in which such non-abstract rules can be enacted at any moment is naturally nomos-neglecting, because the people cannot direct their conduct through their knowledge of the law. Notice incidentally that even if abstract rules have to be designed for the long term, they do not have to be carved in stone, as long as any change is preceded by due warnings and possibly includes a grandfather clause.

In the field of social justice, Rawls’s theory of justice provides us again with a good example of abstractness. The subject of justice according to Rawls (1999: 6) is what he called the ‘basic structure of society’, which includes the political constitution as well as the principal economic and social arrangements. This, in turn, permits the formulation of legitimate expectations, since the rules of the basic structure are designed to act continuously and durably. Rawls certainly did not advocate for social justice at the spur-of-the-moment – for example for egalitarian rules that would only be enacted and enforced when there would be too much inequality. Nor did he advocate for an egalitarian system that could be abolished by every new legislatures. This would not permit the formulation of legitimate expectations, which was for both Hayek (2013: 169ff) and Rawls a fundamental aspect of a liberal polity – that is people should be secure in their planning of private arrangements, and rules must accordingly permit reasonable reliance (Schauer, 1991: 137–145).

Reconciling freedom under the law and social justice

The question is now the following – can classical liberals embrace social justice, though some of them, like Hayek, also characterised social justice as a ‘hollow incantation’? Yes, I believe they can, as this section will now show, and, what is more, they should, as the next section will explain. One of the main worries of classical liberals in regard to social justice was that it would be incompatible with freedom under the law. Social justice was an ‘empty concept’, according to Hayek (1985: 58), because ‘no preconceived scheme of distribution could be effectively devised in a society whose individuals are free, in the sense of being allowed to use their own knowledge for their own purposes’. This passage is illuminating, since it is either utterly incorrect or straightforwardly correct, depending on how one understands market rules.

On the one hand, it would be correct to say that one may not impose a specific pattern of distribution in a free society, as Nozick once noted (1974: 160ff). To achieve any such pattern, one would have to limit individual freedom, using non-general and non-abstract rules. For instance, if one were to try to impose an order guided by an allocative principle of desert, one could not do so without putting a stop to market interactions and directing the conduct of specific individuals. One would have to control every single transaction so to make sure that everyone gets exactly what he or she deserves. Parents could not give money to their undeserving children anymore, nor could we buy things from disreputable individuals. By using regulations, statutes, and decrees, one could then come close to the desired allocation of goods in a given society – and yet this is obviously unacceptable from a classical liberal viewpoint. This first approach is what I call nomos-neglecting egalitarianism – it is not compatible with freedom under the law. We could think that imposing some specific patterns can be done using general and abstract rules – and perhaps this could indeed be done, depending on the complexity of the pattern we want to establish. But any attempt to establish a specific pattern through general and abstract rules will probably be foiled by the problem of imperfect generalisation, which we will examine momentarily. The problem is that there are limits to what general and abstract rule may do, since such rules may normally only direct behaviour as a more or less systematic whole (Postema, 2011: 168), and therefore they are not suitable for such a complex task as imposing a certain scheme of distribution.

On the other hand, though, it would be incorrect to say that one cannot devise some egalitarian rules while still leaving individuals free to use their knowledge to pursue their own purposes, as Hayek would certainly concede (2013: 172). After all, people are still free to engage in free market transactions even if their government institutes some redistributive rules – for example mild taxation and some welfare programs. One can act as one will in a market society, even if some welfare programs have been enacted, inasmuch as these programs were properly framed through general and abstract rules. For instance a general sales tax used for redistributive purposes would do very little in a market economy to prevent individuals from going about their business. This, in a way, would be a ‘preconceived scheme of distribution’, and yet intuitively it would be incorrect to say that such a scheme could not be devised in a free society. This second egalitarian approach, then, is nomos-observing – properly speaking, it is compatible with freedom under the law as well as with a theory of isonomic treatment.

General and abstract rules, therefore, permit us to care about the general consequences of market capitalism without compromising freedom under the law. Social justice can then be pursued by classical liberals, provided they do so carefully through rules known in advance designed to govern the conduct of a general population. A basic income guarantee is perhaps the most typical example of a transfer that can be accepted from a classical liberal point of view. In fact, both Hayek (2013: 249) and Friedman (2002: 191) made a case for such an assured minimum income, that is a non-market welfare safety net, or ‘a floor below which nobody need to descend’. They were followed, more recently, by neoclassical liberals like Zwolinski (2011). Such egalitarian endeavours appearing in the form of the application of general and abstract rules invalidate the classical liberal understanding of social justice being a ‘humbug’ or a ‘will-o-the-wisp’. Hence, and this is my first conclusion, it is indeed possible to pursue social justice from a classical liberal point of view inasmuch as we do so in proper institutional terms, namely through general and abstract rules.

Simply put, the problem is now the following. We can have social justice in a classical liberal polity, but not any kind of egalitarian policy will do – regulations, statutes, and decrees are problematic, as they do not permit the formulation of reasonable expectations or they are not general enough. As Hayek noted (1978: 258), talking about the possibility of government taking the initiative in such areas as social insurance and education, ‘Our problem here is not so much the aims as the methods of government action’. It is for such a problem that the rule egalitarian theory I put forward is particularly relevant, addressing the methods of government action. Rule egalitarianism, then, is not only interesting for the classical liberal tradition, but also for the high liberal one, and it could thus be regarded as articulating presuppositions which were not fully articulated by high liberals with their theory of justice as impartiality. Not only should liberal egalitarians care about institutional settings, since it is the only way they can consistently remain liberals, but they should also care about the characteristics of the rules they enact, as it is the only way they can make sure that egalitarianism is compatible with the rule-of-law tradition. One should not redistribute economic resources by targeting specific people, one should not redistribute only when it pleases one to do so, yet one may reform some aspects of the legal framework of market capitalism by enacting egalitarian laws of general and abstract validity.

Before we move on to reconcile social justice and classical liberalism, let me address one final problem. I just introduced a conception of generality that focuses on the compatibility with the reasons underlying a law’s enactment, which though it will lead to original conclusions for the classical liberal tradition is not a wholly novel conception of generality. Marmor (2007: 10ff) advanced it earlier, using the example of a law to regulate pollution that ought to regulate all and only those who can pollute, and yet, as Gowder argues (2013: 611 n.111), such a position may also lead to some undesirable conclusions. The problem is as follows – Marmor claims that the apartheid was general inasmuch as the laws discriminating against black people were indeed moved by bad reasons. Hence, provided the reasons for a law are non-general enough, it seems the distinction between generality and particularity becomes immaterial – in effect, generality becomes a requirement of good faith relative to some underlying public purposes, without saying anything about such purposes. However, this seems incorrect.

Can we say that a law is general in virtue of the fact that it corresponds to the manifestly non-general reasons for its enactment? As Gowder explains, we are caught in a dilemma. On the one hand, if the answer is no, we then need a principle determining which reasons are acceptable – and, we may add, from a classical liberal view such a principle may very well exclude social justice. On the other hand, if the answer is yes, it looks like there is a serious problem with the idea of generality in itself, since it no longer protects freedom under the law, or at least the classical liberal understanding of such freedom. It is also quite possible that no answer can be found, since, as Gowder notes (2014a: 1028f), ‘It is notoriously difficult to find the expressive content of a law, or, indeed, anything, and there are many who worry that it might be indeterminate, and the inquiry futile.’ In trying to justify social justice through generality, hence, we run the risk of making generality less attractive from a liberal point of view – that is generality becomes some kind of purely formal idea, not especially useful to think about justice or freedom in general.

One possible solution, championed by Gowder (2013: 611), is to make public reason the relevance criterion for our understanding of generality, which would develop ‘the core egalitarian idea that we must justify ourselves to those whom we treat differently’. Rule egalitarianism proposes an alternative solution, which may perhaps be more appealing to classical liberals who would not accept Rawls’s theory of public reason. A law is general inasmuch as, first, it acts impersonally, and, second, following Marmor (2007: 12), there is an appropriate connection between the reasons for prescribing the norm and the characterisation of its population. Impartiality, therefore, is the relevance criterion for our understanding of generality, which, in turn, exercises some constraints on the underlying reasons of rules – of course, it does not mean that a rule cannot focus on certain traits that only some people would possess, but, as Hayek argued (2013: 93), the idea of impartiality implies the following:

If a need arises to call in an impartial judge, it will be because such a person will be expected to decide the case as one of a kind which might occur anywhere and at any time, and therefore in a manner which will satisfy the expectations of any person placed in a similar position among persons not known to him individually.

To be truly impartial, however, we must not make any such assumption, which, therefore, permits us to arrive at a different conclusion than Marmor. Though laws discriminating against black people were indeed moved by bad reasons, they were not general because they were not impartial – they were moved by disdainful attitudes, at a particular time and place, and set to satisfy the expectations of some wicked persons. A criterion of impartiality, hence, exercises some constraints on the classical liberal understanding of rules, which, I should note, does not bar social justice, nor the differential treatment of rich and poor, and which, indeed, as we will now see, demands social justice.

Rule egalitarianism – On imperfect generalisations and social justice

We have now shown that it is possible to pursue social justice from a classical liberal point of view. This conclusion, however, will be of little use if classical liberals reject social justice altogether, as they seemingly often do. Therefore, we must now show that such a rejection of social justice is mistaken from a classical liberal viewpoint – that is provided we accept the main classical liberal premises, we should also embrace social justice. I propose to defend such a thesis using the reason-giving attribute of generality, following Frederick Schauer’s work, which, in the end, will show that having general rules of market capitalism imposes some constraints on any institutional framework so that social justice becomes necessary to correct the injustices of imperfect generalisations.

Let us start with a simple statement – classical liberals have championed general and abstract rules of market capitalism. There are many reasons why that it so, but one such reason is that markets will make everyone better-off and will lead to the best overall results for society as a whole. This empirical claim is now used by neoclassical liberals to argue that social justice requires market capitalism. Let me defend the opposite – market capitalism requires social justice. Though one should probably not bundle classical liberalism with the consequentialist libertarian tradition of, say, David Friedman (1989), classical liberals like Hayek have indeed used some consequentialist arguments to back market institutions – for example they claim that markets will indirectly produce ‘aggregate happiness’, as opposed to socialist regimes that will make everyone worse-off.

The classical liberal tradition, of course, also rests on some principled arguments, such as individual autonomy and the non-aggression principle. Indeed, said Hayek (1978: 13), markets also protect freedom so that someone ‘can expect to shape his course of action in accordance with his present intentions’, without having somebody else having ‘power so to manipulate the conditions as to make him act according to that person’s will rather than his own’. Whether one accepts the aggregate-happiness or the liberty-based argument for market institutions, the problem from a classical liberal point of view is then the following – although the general and abstract rules of market societies will indeed make most people vastly better-off, they will also make a few people worse-off, that is some people will fall through the cracks of market institutions (Zwolinski, 2015), living destitute lives in unforgivable conditions, thus also making them vulnerable to private coercion in market situations.

Why is that a problem? After all, we all know that rules often lead to some unwelcomed consequences in particular circumstances, if only because of our limited knowledge Hayek would say (1945) – unfortunately, we could add, some people will be left behind. ‘Our faith in freedom’, said Hayek (1978: 31), ‘does not rest on the foreseeable results in particular circumstances but on the belief that it will, on balance, release more forces for the good than for the bad’. Not so – particular circumstances are still important, as we will now see, inasmuch as they are anticipated consequences of general rules.

The problem is that the generality of some rules is defined by the reasons given for the enactment of the said rules – again, ‘to provide a reason for a decision’, said Schauer (1995: 641), ‘is to include that decision within a principle of greater generality than the decision itself’. In other words, the generality of the law is not only defined by the impersonal character of the rule itself, as well as by the relative size of the population it is meant to apply to, but also by the act of reason-giving. Unlike statutes saying, ‘Do it!’, Schauer explains (1995: 633–635), general rules are based on reasons ‘broader than the outcome they are reasons for’. Such a feature of general rules explains why classical liberals have indeed championed laws of general validity in market societies – it also explains how the generality and the abstractness of the law are self-reinforcing. Giving reasons, Schauer writes (1995: 645), ‘carry the same kinds of promissory commitments vis-à-vis future actions or decisions as do locutions more explicitly seen as promises’. Hence, the act of reason-giving, which is an act of generality, leads to reasonable reliance that is the antithesis of the authoritarian approach of, say, decrees. Since they argue for the domestication of coercion, of course, classical liberals will favour general rules over non-general ones.

The difficulty for classical liberals is then as follows – general rules of market capitalism may be in error, ‘where the notion of an error’, according to Schauer (1997: 286), ‘is defined in terms of serving the background purpose or rationale of the rule, or in terms of what the rule-maker would have done, in person, had she confronted this individual case in the absence of any rule’. One may not want to call such occurrences ‘errors’, since, as Schauer notes, it is a common feature of rules that they ignore particulars in favour of tendencies. But the fact remains that many general and abstract rules of market societies manifestly do not fully serve the background rationale of making everyone better-off, which, recall, is indeed part of the classical liberal argument for market capitalism. Nor do market rules always protect individual freedom, as showed by Bertram et al. (2012) in a study about the systemic denial of freedom in private regimes of power, particularly in the workplace, as well as Paul Gowder in his recent paper on market unfreedom, a response to neoclassical liberals, which explains how economic desperation forces people to engage in immiserating work (2014b).

Simply put, some people are left behind in market societies, which, we could say, is an ‘injustice of imperfect generalisation’. Because classical liberals endorse general rules of market capitalism, and because one reason why they endorse such rules is to make everyone better-off or to protect individual freedom, they have to find a way to mend instances where some people are systematically worse-off in market societies or denied some basic liberties. In fact, we know that general rules will undercut the background classical liberal rationale for market institutions, at least for some few people – that is by design, general and abstract rules of market institutions ignore particulars so to only consider some collection of particulars. The rule egalitarian argument is then the following – the justice of the background institutions must be evaluated in terms of the rationale behind these institutions. Without social justice, I argue, a classical liberal polity is accordingly unjust, since general rules will inevitably fail some people.

One possible solution to the problem of imperfect generalisation, which was favoured by both Plato (Statesman) and Aristotle, is to say that it is unjust to refuse to correct the general rules to take account of the particular, as Schauer explains (1997: 286). After all, said Aristotle (Nichomachean Ethics: 1137b), ‘concerning some matters it is not possible to speak correctly in a general way’. According to such an Aristotelian theory, social justice could be understood as a particular rectification of legal justice. ‘This is in fact the nature of the equitable’, stated Aristotle, ‘a correction of law in the respect in which it is deficient because of its being general’. In other words, when we know that some general rules are in error, leading to some undesirable consequences in particular circumstances, we have to enact and enforce particular rules to counter such outcomes.

The classical liberal tradition manifestly disagrees with such an approach – that is a lack of generality should be considered as a deficiency, not a lack of particularity. As Hayek repeatedly argued, the justice of general and abstract rules is not necessarily imperiled by their regrettable consequences in some particular circumstances. Likewise, Schauer convincingly said (1997: 289) that ‘it is odd to think of every inaccuracy with personal consequences as an injustice, especially once we recognise that every application of a rule qua rule entails the risk of just this kind of inaccuracy’. Yet that some people fall through the cracks of market institutions is still a problem – and classical liberals never had any difficulty to acknowledge that there is a problem with extreme poverty, and, say, homelessness.

Let me therefore propose the following conclusion – social justice for classical liberals must be understood at least as a necessary general rectification of the rules of market institutions. Simply put, general and abstract rules of social justice should accompany general and abstract rules of market capitalism, without which background market institutions will be unjust. If anything, classical liberals can and should accept social justice as a way to correct the errors of imperfect generalisation. That is not to say that general rules of market capitalism are unjust because general, but that market institutions are indeed unjust if they rely on such rules without having some mechanism to reach those who will not be reached by the general rules – social justice is such a mechanism.

Using such a rule-based approach to social justice, classical liberals do not need to accept the high liberal concern with equality, or with the difference principle, nor do they need to accept Aristotle’s theory of equity. They only need to be consistent with their own ideas, most important of which is an unequivocal commitment to general and abstract rules of market capitalism, which, we know, will disserve some people. This is an institutional problem. The question is then the following – what do we do with people who are left behind as a consequence of having general rules? The answer, I think, is social justice – that is as I announced in the introduction, contrary to what Lister argued, I believe that the legal framework of capitalism can be designed to reduce inequality and ameliorate poverty in the name of justice, and this is precisely what Hayek and other classical liberals have taught us. After all, Hayek endorsed an assured minimum income, a reform of inheritance, and a limited form of progressive taxation. Each of these are egalitarian rules, which, I think, are better understood through a rule egalitarian theory.

Classical liberals who argue for very limited or time-sensitive measures of poverty alleviation fail to embrace the consequences of their own preference for general rules (Epstein, 2011: 148). There is a cost involved with having general rules of market capitalism – by endorsing such rules, classical liberals commit themselves in a way that they rarely recognise. They then have to endorse some egalitarian rules to catch those who will inevitably fall through the cracks of market institutions. Of course, poverty is not a problem that can be solved entirely, at least provided we accept the core classical liberal premises. But that some people live destitute life in market societies is still a problem that needs to be addressed, as most classical liberals acknowledged. That many classical liberals like Hayek did not fancy the concept of ‘social justice’ is inconsequential. Addressing such a problem is indeed taking social justice seriously, which, I argue, is not only possible from a classical liberal viewpoint, but also necessary provided we have some general market rules.

That the state should protect the institutions of private property and contractual freedom as always been a staple of the classical liberal tradition. That the state should also care about social justice has been hinted repeatedly, though often timidly, by modern classical liberals – for example Epstein (2011: 141ff), Friedman (2002: 190ff), as well as Hayek (2013: 249) who quite unwaveringly argued that the state has a ‘clear moral duty’ to assist those in extreme poverty. With a rule egalitarian theory, I believe, classical liberals need not be so timid anymore about their advocacy for social justice. A rule egalitarian approach is compatible with the main pillars of the classical liberal tradition, namely, limited government, limited administrative discretion, rule of law, individual freedom and market capitalism. It may accordingly be endorsed by classical liberals concerned with social justice who would not follow neoclassical liberals in accepting the main high liberal intuitions about, say, the difference principle.

The rule egalitarian approach, I should add, may lead to some interesting conclusions not only for the classical liberal tradition, but also for neoclassical and high liberals alike. High liberals may now want to accept both market capitalism and social justice, which they can if they use the core arguments of rule egalitarianism. Neoclassical liberals, conversely, may want to embrace a slightly more substantive understanding of social justice, that is, a theory of social justice that is not simply about accepting market institutions for the long-run maximisation of the welfare of the worst-off. A theory of rule egalitarianism will be concerned with the present plight of the worst-off in the midst of plenty. Such an approach, as I have shown, is compatible with the classical liberal framework in some salient sense, and yet it is also compatible a stronger understanding of social justice than the one defended by neoclassical liberals today.

The conclusion I would propose for the classical liberal tradition is then simple – contrary to what is often thought, some egalitarian endeavours are acceptable from a classical liberal point of view, and, what is more, they may be necessary provided we accept the classical liberal argument in favour of general and abstract rules of market capitalism.

OPEN IN VIEWER

Table 1.

Four concepts of rules.

	General	Not general
Abstract	(R1) Law: i.e. a rule known in advance designed to govern the conduct of a general population. e.g. traffic laws.	(R2) Regulation: i.e. a rule known in advance designed to govern the conduct of a specific entity. e.g. drilling regulations for Chevron.
Not abstract	(R3) Statute: i.e. a rule designed to temporarily direct the conduct of a general population. e.g. conscription laws.	(R4) Decree: i.e. a command that can be uttered at any moment to direct anyone’s conduct. e.g. amnesty laws and pardons.