Part I Chapter 1
On the Roles and Limits of Ideal Theory
In book 5 of the Republic, Glaucon replies to Socrates:
Take it that I agree that there would be all these things and countless others if this regime should come into being, and donât talk anymore about it; rather, letâs now only try to persuade ourselves that it is possible and how it is possible, dismissing all the rest. (471e)
Socrates presented his account of justice earlier in the dialogue, although without fully revealing it. Glaucon and Socrates âgot to this point while seeking what justice and injustice are likeâ (Plato: 472b). They had so far relied on the ideas of compliance of citizens and of motivation to comply with the law (471d). Glaucon was ready to agree with Socrates, for that was not the real issue Socrates should be addressing. The biggest and most difficult âwaveâ assaulting Socratesâ argument was to know whether his account of a just city is possible, and how. Glaucon notices the sharp contrast between the hypothetical acceptance of an ideal account of justice and the questions of feasibility that arise when seeking to transform a society such that it conforms with this account. At first, it seems that Socrates was claiming that an ideal would remain valid even if it could not be realised. He indeed claims that a role of his ideal city is to judge actual cities on how well they approximate the ideal (472bâ473b). In this passage of the Republic, Socratesâbelieving that his ideal city could be realisedârather shifts the focus of his argument in order to show that a theory wishing to provide guidance ought to compare ideal and non-ideal cities (Kraut 1992, Ypi 2010: 538). He first concentrated on the role of an ideal of justice when thinking about the virtues of the polity. At one point he began to think about the agents charged to carry out these duties and on the feasibility constraints the theory should consider. Whether Socratesâ arguments in favour of the practical role of his account of ideal justice are persuasive largely depends on the role of motivation as it relates to the knowledge of justice, which is an eminently complex problem in Platoâs Republic (Kraut 1992, Brown 2011). Nonetheless, it seems clear that for Plato the philosopherâs justice should be taken as a useful paradigm that can be approached by the nonphilosophers.
In a similar vein, Rousseau opens book I of The Social Contract with the famous quote: âI mean to inquire if, in the civil order, there can be any sure and legitimate rule of administration, men being taken as they are and laws as they might beâ (Rousseau 1762). For Rousseau, the reform of any social institutions should begin with people as they are and laws as they could be. Our goals in society should reflect our best accounts of what an ideally just society should look like. Yet our strategies for achieving it must take into account how society is, with its non-ideal agents and non-ideal political structures.
Following Plato and Rousseau, Lea Ypi observes that a theory of justice has two stages: the formulation of ideal principles that will shape the institutions under the assumption of full compliance, and the assignment of responsibilities to agents who will realise the project (Ypi 2010). She claims that contemporary theories, especially Rawlsian theories, follow a similar pattern. One stage of these theories is devoted to defining the standard of social interaction under favourable conditions and full compliance. Then they leave this ideal level of reasoning to deal with problems of partial compliance, lack of motivation, and historical failures. For Plato, everyone would endorse just principles after an appropriate assessment of the circumstances of justice (Republic, 472). This would support the methodology according to which we first need to know what is justice by assuming that everyone would be motivated to promote it, only after which can we start to think about the relations in the non-ideal world that could support its realisation.
Rawls is even more adamant than Plato with regard to the analytical priority of âideal theoryâ. He says,
The reason for beginning with ideal theory is that it provides, I believe, the only basis for a systematic grasp of these more pressing problems [i.e. of partial compliance]. (Rawls 1999b: 8â9)
In other words, not only are there two stages of reasoning, but the formulation of regulative principles necessary to guide the transformation of political institutions comes first; the reflection concerning their implementation comes after (ibid. 216).
Even if we admit that the ideal principles are the standards by which we will evaluate and change non-ideal circumstances, we will still need to know by whom and how these changes will take place. Concerning agents, Socrates would have appointed philosophers, Rousseau the legislators, and Kant the moral politicians to rule. The realisation of an ideal need not assume that everyone will effectively comply. Socrates replies to Glaucon that if we discovered how a city might be governed nearly as we proposed, we will have discovered the possibility sought by Glaucon (472).
Accepting for the moment that this structuring of political theories in two stages is plausible, we may note one other commonality in the views of Rawls and Plato. It concerns the stage of the theory at which facts intervene and gain normative relevance. In the construction of their theories, we observe that Plato and Rawls make assumptions, or idealisations. Both assume full compliance. Plato also assumed motivation to obey the laws,1 while Rawls assumes favourable circumstances. That does not mean that they rule out the possibility that facts could have a role to play in political theorising. As Ypi points out, specifying the stage at which facts come into play is a weaker thesis than one that denies the role of factual assumptions altogether or asserts the validity of fact-insensitive principles (Ypi 2010: 539). The latter stronger thesis is held by Cohen, who denies the role of facts in the formulation of first-order, fundamental principles. Yet he admits a role for facts in second-order principles he names ârules of regulationâ (Cohen 2003).
The division of our theorising about politics in a more ideal dimension and a more practical dimension is thus not uncommon. It has sources in prominent figures of ancient, Enlightenment, and contemporary political philosophy. It also echoes, as mentioned in the introduction, important dichotomies in our common vocabulary addressing social and political issues, such as realism and idealism, the moral and the political, the descriptive and the normative, and the theoretical and the practical. This debate is very much alive today. Explicit defences and critiques of ideal theory are numerous in contemporary political theory.2 Implicit endorsements of one or another approach are even more numerous. I will talk in terms of ideal and non-ideal theory, as these terms will become clearer in the course of this book, in a debate that appears at times to become increasingly murkier.
We should see that, depending on how we understand this division of theories of justice into two stages, we face considerable problems. I will explore these problems in the next sections and chapters. I mention two here. First, it is not clear that the distinction between one ideal level of formulation and one non-ideal level of implementation is defensible. Second, it is unclear that there is a clear-cut sequence between these levels. I challenge the view according to which ideal theory has priority.
Rawls was the first to formulate the debate in terms of âidealâ and ânon-idealâ theory. He saw this as a good way to characterise the relation between political philosophical theory and political practice. And much of the debate proceeds from (and interrogates) Rawlsâs definitions, assumptions, and questions. It is thus important to review his characterisation of the problem.
1.1 Rawls and Ideal Theory
John Rawls described his theory of justice as an ideal theory. From Rawls, in A Theory of Justice, we get a detailed exposition of the function and features of ideal theory. Many of these functions and features reflect intuitions of the past about ideal theorising. The defences and objections to ideal theory as coined by Rawls provide fruitful ground to examine important aspects of the question. The coverage of this rich literature in the first chapters of this book should open the door to the examination of underexplored aspects of non-ideal theorising. And, more importantly, a wider view of this debate should reveal that a new way of approaching political theorising has been neglected in the literature, or at least has not received a systematic treatment. We may distinguish between the features, functions, and spirit of Rawlsian ideal theory.
Two idealised assumptions are central to the Rawlsian notion of ideal theory. First, ideal theory defines the principles for an idealised society where almost everyone complies with the principles (Rawls 1999b: 7â8, 2001: 13). In contrast, non-ideal theory assumes that not everyone complies with principles and deals with this by addressing issues such as civil disobedience (1999b: 8). The extent of noncompliance can vary depending as much on the problem as on the responses to it. A second feature of ideal theory is the assumption of a well-ordered society under favourable circumstances (1999b: 216). Favourable circumstances indicate the conditions that make a constitutional democracy possible (Rawls 2001: 101). A well-ordered society is a constitutional democracy. The favourable conditions include the social and economic means that are necessary to sustain this regime, such as education and a sufficient level of economic development. It is worth noting that while Rawlsâs idea of favourable circumstances can be observed in a good number of contemporary societies, full compliance is a feature of almost no society (Stemplowska and Swift 2012). They remain the two central features of ideal theories for Rawls.
Another feature of Rawlsâs ideal theory is the notion that ideal theory is a theory of âperfect justiceâ. For Rawls, the full-compliance assumption is necessary for us to find out what perfect justice requires. That is, the features of full compliance and favourable conditions are meant to lead us not to any theory of justice, but to a state of perfect justice. The exact definition of the idea of perfect justice seems in fact less important than the role it accomplishes (I will return to the definition, and to the full-compliance assumption, in section 1.6 below). The argument proceeds from the notion that we cannot know what perfect justice is if we have to think about whether agents will or will not actually comply with its principles. If we factor in noncompliance, the principles will not hold. Still, the real question is, why do we have to know what perfect justice requires (Stemplowska and Swift 2012: 376)?
The notion of perfect justice is associated with a central function of ideal theory: perfect justice provides the only basis to systematically grasp the problems that non-ideal theory seeks to investigate (Rawls 1999b). This claim about the âanalytic primacy of ideal theoryâ is a central object of study of this book. Ideal theory must be undertaken prior to non-ideal theory, so the argument goes, for it is the only way to achieve the basis from which we can systematically grasp the problems of non-ideal theory. In fact, it is paramount to distinguish between two functions of ideal theory which are constitutive of the claim for the analytic primacy of ideal theory. Setting the basis to systematically grasp the problems of ideal theory is done in two ways. First, ideal theory provides a target for non-ideal theory: âuntil the ideal is identified . . . non-ideal theory lacks an objective, an aim, by reference to which its queries can be answeredâ (1999b: 90). The function of the target is explicit in Rawlsâs treatment of non-ideal theory in the Law of Peoples: âNon-ideal theory asks how this long-term goal [of ideal theory] might be achieved, or worked toward, usually in gradual steps. It looks for courses of action that are morally permissible and politically possible as well as likely to be effectiveâ (1999a: 89). Call this ideal theory âas targetâ.
Second, ideal theory provides the measure to assess injustices in the real world âidentified by the extent of the deviation from perfect justiceâ (1999b: 216). The function of measure is explicit in Rawlsâs explanation of the connection of the two theories:
Viewing the theory of justice as a whole, the ideal part presents a conception of a just society that we are to achieve if we can. Existing institutions are to be judged in the light of this conception and held to be unjust to the extent that they depart from it without sufficient reason. (1999b: 246)
That is, we can only know what situation is unjust, and how unjust it is, by determining how it departs from the pre-identified ideal (although Rawls admits that part of this exercise is left to intuition).3 This is an explicit claim to the effect that ideal theory has an action-guiding function. Call this ideal theory âas measureâ.
Finally, besides its features and the functions, there is something particular to Rawlsâs understanding of ideal theory, which qualifies the spirit in which his ideal theory is conceived. For Rawls, ideal theory must remain a ârealistic utopiaâ (1999b: 11â12). It must be kept within the realm of what is realistically practicable (2001: 13). This can be clarified further on the basis of an account of what is expected of agents, which comes from Rawlsâs take on moral psychology (Rawls 1999b: 126): theories that are too difficult to expect agents to comply with should be ruled out. Agents should be able to comply with principles given merely their sense of justice. That is, no agent is required to be a saint or a moral hero (1999b: 419). In a similar vein, favourable circumstances do not imply a society of abundance, but rather âmoderate scarcityâ.
In a nutshell, for Rawls, ideal theory defines principles of perfect justice for an idealised society characterised by full compliance and favourable conditions, but must remain a realistic utopia. This exercise has to be undertaken prior to non-ideal theorising in order to provide the latter with its target and the measure through which injustices can be assessed. Non-ideal theory deals with more urgent and pressing questions and asks what justice demands of us in non-ideal circumstances. For Rawls, ideal and non-ideal theories aim at identifying the principles that should regulate societiesâ basic institutions. These are the main functions and characteristics of ideal theory according to Rawls and many commentators. Objections regarding the relevance of these features and functions have been raised and will be discussed below.
Two minor clarifications are warranted before concluding this section. We may rightly ask whether it is Rawlsâs principles that are ideal, or the method he uses to obtain these principles which is idealised (Stemplowska and Swift 2012). It seems clear that a method that consists in utilising idealisations is idealised. The assumptions of full compliance and favourable conditions are idealisations. This is even more clear in that the principles are chosen from the conditions set forth by the âoriginal positionâ under the âveil of ignoranceâ. Claiming that he formulates principles of perfect justice also denotes that the principles as such are ideal âas targetâ and âas measureâ.
More importantly, the relation between the epistemic and the practical nature of ideal theory will have to be examined. We may agree that Rawlsâs ideal theory aims at providing the values and the corresponding duties that apply to the political sphere. But again, should this be understood as if the goal of ideal theory is to inform us about the necessary actions to realise these values or rather as if this is only a question about knowledge of the values in question?
1.2 Idealisations
Prominent debates in political philosophy have been centred around questions such as âwhat is the right theory of justice?â, âwhat is the right âcurrencyâ, âmetricâ or âfocal variableâ of justice?â, and âwhat would a perfectly just society look like?â (Nozick 1974, Dworkin 1981, Arneson 1989, and Rawls 1999b are examples). These theories are ideal theories, at least with respect to their outputs: they put forward one framework that presents what should be distributed and what an ideal distribution should look like.4 Ideal theories are very present among contemporary liberal egalitarian theories. Also, these theories are ideal because of the idealisations that they make when justifying these principles to arrive at the description of the just society. Robert Nozickâs conditions for the basic appropriation and distribution of libertarian rights or Ronald Dworkinâs auction system for the distribution of resources are developed under idealised conditions, which allow the modelling of an ideal distribution. Whatever technical limitations in society to effect this distribution or lack of motivation of the citizens of that society to follow the chosen principles are assumed away at this level of theorising (Robeyns 2008, Valentini 2009). That is, ideal theories rely on idealisations.
It is not disputed that ideal t...