The label ‘ideal theory’ is applied in a number of different ways by different political theorists, so it is important to begin by clarifying my target. The term was first introduced by Rawls to describe a certain methodological move. To find out what justice requires of us here and now, we begin by constructing a theory of what justice would look like under ‘ideal’ circumstances. This involves abstracting away from a number of ‘nonideal’ features of the real world, such as non-compliance, in order to imagine what a fully just world would look like. Once we have this ideal theory in hand, we then proceed to ‘nonideal theory’, that is, theorizing about what ought to be done in our current context to move closer to the ideal. Some critics of ideal theory are simply criticizing Rawlsian philosophy for focusing too much on the first step to the (near-)exclusion of the second.11 The target of my critique is wider: it is the ideal-theoretic method as a whole. As such the targets of my critique also include nonideal theorizing insofar as this is understood as derivative of ideal theory in accordance with the Rawlsian method. I will define this ideal-theoretic method as any approach to political theory that derives the normative and/or practical authority of its prescriptions wholly or in part from what Mills calls ‘ideal-as-idealized-models’.

To clarify this definition, let me introduce a concrete example of this method in use. Tommie Shelby’s recent book The Idea of Prison Abolition applies an ideal-theoretic method to interrogate the arguments of prison abolitionists, primarily Angela Davis.12 I take Davis’s work to be a paradigm example of a materialist approach to political theory, and I will have more to say about her method later; for now, what matters is that she begins from the nonideal problem of mass imprisonment, analyses how the prison industrial complex is bound up with racism, capitalism, and patriarchy, and argues that attempts to reform prisons are insufficient, or even counterproductive. She does not rely on an ideal-theoretical account of justice to ground these arguments. Shelby recognizes the methodological differences between his approach and Davis’s; he self-consciously subjects her materialist treatment of prisons to an interrogation using the ‘tools of analytic philosophy’ to test whether ‘systematic moral arguments’ can be developed in support of prison abolition.13 By ‘systematic moral argumentation’ he means ‘ideal theory’.14 The assumption underlying the book is that the question of whether prisons should be abolished can be answered by asking whether ‘the practice of imprisonment [could] ever be justified in a just social order’.15 His answer is ‘yes’. Imagining a reasonably just society, he does not find that the moral principles governing such a society rule out any of the necessary, as opposed to merely contingent, features of the practice of imprisonment. Shelby thus constructs what Mills calls an ‘ideal-as-idealized-model’, that is, a model of what society ought to look like and the role of idealized prisons within that society, and from this he derives conclusions about what we should do here and now: reform, not abolition.

The details and cogency of Davis’s and Shelby’s arguments need not concern us here. I only want to point out their methodological differences, and key to these is the role of ideal models. Mills, importantly, distinguishes ‘ideal-as-idealized-models’ from other sense of ‘ideal’.16 Sometimes ‘idealization’ is used to mean something like ‘abstraction’, or what Mills terms an ‘ideal-as-descriptive-model’. This usage of the term ‘ideal’ owes its popularity in the social sciences to Max Weber’s concept of ‘ideal-typical constructs’.17 As Weber emphasizes, ‘the idea of an ethical imperative, of a “model” of what “ought” to exist is to be carefully distinguished from the analytical construct, which is “ideal” in the strictly logical sense of the term’.18 Davis, for example, employs idealized models of capitalism, racism, and patriarchy to explain the role prisons play within these interlocking systems. These models play an important diagnostic role, but they are, of course, not models of what ought to be.

This does not mean, however, that Davis’s arguments are not normative or that she does not appeal to normative concepts like ‘liberation’ and ‘oppression’. I take it to be uncontroversial that an argument about what should be done—e.g., that prisons should be abolished—is straightforwardly a normative one. As Mills acknowledges, all moral and political theory could be called ‘ideal theory’ in the ‘trivial sense’ of having to do ‘with normative/prescriptive/evaluative issues, as against factual/descriptive issues, and so involv[ing] the appeal to values and ideals’.19 What differentiates Shelby from Davis, and what Mills (and I) object to, then, is not to the mere invocation of ideals but a specific methodological role that ideal theory accords to ‘ideal-as-idealized-models’ (for brevity, I will henceforth refer to these simply as ‘ideal models’).

So, what does the objection consist in? Mills’s main criticism is that ideal theories of justice (or rights, or equality, etc.) are unhelpful and even ideological because they obscure the deep injustices, like patriarchy and white supremacy, that structure our societies, and which political theory ought to be concerned with countering. Now, sometimes, Mills recognizes, ideal models can be useful in helping us understand how a given object functions. When a mechanic deals with a car, it is helpful for them to have a model of how that car should function, and to treat any deviation from that ideal model as a shortfall which needs to be fixed. This is because the car was purposefully designed to function in a specific way, and its actual functioning, including any deviations from that ideal, are best understood in relation to how it was meant to function. If our societies had been designed by benevolent philosopher kings, the same method could plausibly be used to understand them and diagnose deviations from the ideal of justice. But Mills’s point is that this does not apply to our social world. The world we live in and its social and political institutions are not designed in accordance with ideals of justice, of which they may occasionally fall short; rather, they are shaped by historical and ongoing injustice through and through.20 To abstract away from the fact that existing societies function as systems of oppression and exploitation ‘is not to make a simplifying assumption for theoretical purposes, but to repudiate theorizing them altogether’.21 As such, using ideal models is unlikely to help us understand the political problems we face or to help us remedy them.

Now, there is a straightforward reply to this criticism that defenders of ideal theory often make. Mills simply misunderstands the methodological role that ideal models are supposed to play in normative theorizing, one might argue. For the point of ideal models plainly is not to help us understand the world as it is, but to help us understand the world as it ought to be. Ideal theorists can easily concede the need for more nonideal theorizing to understand the injustices we currently face but insist on the need for this endeavour to be guided by an understanding of what an ideal society would look like. One can admit that ‘political theorists have already made extensive use of ideal theory while neglecting nonideal theory’, but this is not the same as ‘questioning the purpose of ideal theory as such’, argue Stemplowska and Swift. After all, ‘[w]e need to know the goal before we can form an all-things-considered evaluation of transitional steps in its direction’.22 Shelby, for example, largely takes on board Davis’s nonideal analysis of the role prisons actually play within existing racist and capitalist social structures. But he adds to this an ideal-theoretic account of the role prisons should play in a just society; an account that supplies ‘goals to work toward’ and ‘normative standards for judging the overall justice of social arrangements’.23 What could be wrong with theorizing the ideal we want to achieve alongside theorizing the nonideal circumstances we are confronting?

It may seem that Mills and the ideal theorists are simply talking past each other; Mills appears to be criticizing ideal models for failing to do something they were never meant to do. But I would caution against this conclusion. While Mills (regrettably) does not make it explicit, I think his critique should be read as challenging the core methodological claim that ‘until the ideal is identified, at least in outline … nonideal theory lacks an objective’.24 Now, defendants of this claim sometimes make the argument from path-dependence: if we do not know what our aim is, then even a step that appears as an immediate improvement may make that aim less accessible.25 This is undeniably true.26 We should be clear about what we want to achieve both in the long and short term in order to make informed decisions about which actions to take here and now. But the suppressed premise in this argument is that the (long-term) aims of political action are supplied by ideal models—and this, I will argue, is the claim that materialists call into question.

To justify the claim that ideal models furnish long-term aims for social change, ideal theorists must explain why such models have normative and/or practical authority for us. I borrow these notions from Rachel Fraser. If you want to claim normative authority for the principles (or ideal models) you employ, she points out, ‘you must say what distinguishes principles [or models] with normative authority from those without’: why would it be wrong for society to deviate from your model of ideal justice?27 To claim normative authority for ideal models is to make a metanormative claim, a claim about where normativity comes from. It need not include any claims about whether the ideal serves any practically useful role. On the other hand, ‘[a] principle has practical authority for an agent or group just so long as that agent or group is disposed to take the mere fact that some course of action would be to act in accordance with the principle as a (pro tanto) reason to so act’.28 Thus, an ideal model has practical authority for us, if we take it as providing pro tanto reasons to make society more like that model. The claim to practical authority is logically independent of the metanormative claim,29 but to justify the ideal-theoretic method one needs to make at least one of these. To say that without an ideal model we ‘lack an objective’ requires explaining why ideal models authoritatively provide such objectives.

If we want to properly understand the core of the disagreement between ideal theorists and their materialist critics, we need to understand it as a dispute about the authority of ideal models. Hence, to fix the target of my critique, I here define ideal theory as any method that accords normative and/or practical authority to ideal models. My argument will be that, from a materialist point of view, theories that accord such authority to ideal models are guilty of a methodological self-understanding best described as fetishism. To be clear, while I think this materialist re-interpretation of the critique of ideal theory helps us make better sense of aspects of Mills’s classic critique, the arguments in this article are not meant as an exegetical account of his work; it is, rather, an independent attempt to take up his intriguing, but underdeveloped, suggestion that the critique of ideal theory is a version of the materialist critique of idealism.30

Before going on with my critique of ideal theory, let me explain what I mean by the term ‘fetishism’ here. The term has occasionally been mentioned casually in critiques of idealized theorizing.31 Onora O’Neill, for example, mentions that ethical theorizing based on abstract principles has been blamed for making a ‘fetish of rules’, by which she simply means a kind of ‘ethical rigorism’ or ‘superstitious rule worship’.32 While acknowledging that some philosophers advocate ‘over-rigid reliance on certain rules’, she does not take this to seriously challenge the claim that ‘any plausible view of reasoning about conduct must give principles and rules an important role’.33 My argument will be that the charge of fetishism, once understood more precisely, has more mileage than that.

In short, fetishism, in the sense I shall be using it, means to treat what is the result of human agency as having power independently of that agency. Marx’s famous analysis of commodity fetishism follows this structure: the commodity is really the result of human labour and has value qua embodiment of human labour;34 but it is treated fetishistically as having value in and of itself, while human labour appears valuable only because it produces commodities.35 The way we expend our productive capacities, under a system of commodity production, therefore falsely seems to us to be determined by the values of commodities set by the ‘natural’ laws of exchange; the latter are invested with power over the organization of human labour, rather than being recognized as results thereof.36 For commodity producers, Marx writes, ‘their own social movement has the form of the movement of things, which, rather than controlling, they are controlled by’.37 It is important to be clear about what exactly is false about commodity fetishism. It is not false, on the Marxist analysis, that we are all subject to the ‘mute compulsion’ of capital—in a sense, we really are subject to the laws of commodity production.38 These laws, though, are not immutable but themselves results of social activity; it is the treatment of these historically specific relations of production as ‘self-evident natural necessity’ that is fetishistic.39

Commodities are real physical things, but the objects of fetishism can also be ideas. Marx, in fact, largely based his criticism of commodity fetishism on an analogy with ‘the misty place that is the religious world, where things produced by the human mind seem endowed with lives of their own’.40 Here, he was referring mainly to Ludwig Feuerbach’s critique of Christianity. God, thus Feuerbach charged, was simply an idea created by humans which Christians fetishistically treated as if it had authority over us. In fact, in the image of an all-powerful, benevolent God, Christians were, unbeknownst to themselves, worshiping human qualities projected onto a supernatural being. The fetishistic inversion here lies in a misunderstanding of the source of the qualities we value: ‘The fact is not that a quality is divine because God has it, but that God has it because it is in itself divine’.41 In other words, we believe that we value, say, benevolence, because God has this quality; when, in fact, we have attributed this quality to God because we already value it.

To level the charge of fetishism, then, is to accuse someone of a self-misunderstanding. In both Marx and Feuerbach, the outcome of human activity, the commodity or God, is treated as having independent existence and power over our actions. In both cases there is also an inversion of the source of value: we fetishistically believe our actions to be valuable because they produce exchange-value, or because they accord with divine qualities; when, in fact, both commodities and God derive their value from human activity. In the same way, I will argue, the ideal-theoretic method treats the outcome of a process of problem-solving as if it were an independent source of authority; it proceeds as if actions are valuable because they bring us closer to an ideal when, in fact, the ideal is valuable only insofar as it helps us solve practical problems. Needless to say, one need not agree with Marx’s or Feuerbach’s specific claims about capitalism and religion to follow the rest of my argument—my appeal to these two authors has been purely to clarify the structure of the fetishism critique.

Let me now return to the question raised in Section I: are critics and defenders of ideal theory talking past each other, or is there a real methodological disagreement? I argue, in this section, that we can better make sense of the dispute between Mills and ideal theorists if we see that each side (implicitly or explicitly) relies on different metanormative assumptions. Mills is a materialist, and (most) defenders of ideal theory are idealists. From a materialist perspective, the mistake involved in ideal-theoretic methodology is not that it is necessarily ideological (though it may often be), but that it relies on metanormative fetishism when it accords normative authority to ideal models.

Let me begin by distinguishing this fetishism critique from some related criticisms of ideal theory. I have already mentioned above that some critics mainly object to the exclusive focus on ideal theory. Others, however, have gone further and argued that ideal theory is unnecessary, or even useless, for the practice of political theory.42 Another set of criticisms has to do with whether we can reliably know the ideal. This claim has been made in a Hayekian vein arguing that societies involve too much complexity for us to reliably know which institutional designs produce the best outcomes.43 It has also been made in an Adornian way by critical theorists who worry that existing capitalist societies distort our understanding of the Good such that we cannot reliably theorize it.44 Both critiques can plausibly be attributed to Mills too. Some have understood him to say that ideal theorizing is useless in countering injustice, and that the endeavour is ideological because it ‘generates opportunity costs by crowding out other theories’.45 Others have highlighted his claim that ‘hegemonic ideologies and group-specific experience … distort[] our perceptions and conceptions’, meaning that the ideals we come up with will be tainted.46

Neither version of the critique is, I believe, particularly strong on its own. To the criticism from opportunity costs, ideal theorists can, as we have seen, easily respond by conceding the need for more nonideal theory; their deeper methodological claims about the role of ideals in normative theory seem untouched by this critique. To the argument from epistemic limitations, the obvious retort is that this is not a problem for ideal theory in any special sense: if our thinking about ideal justice is seriously distorted, then why would our theorizing about injustice be any less distorted?47 If these distortions, as critical theorists usually claim, can be remedied by incorporating a reflexive metatheory about how one’s social position shapes one’s cognition,48 then why can the same method not be used to develop a kind of ‘standpoint-theoretically informed ideal theory’?49

The fetishism critique, on the other hand, denies that ideal theory can help us ‘discover what justice requires of us’,50 even in principle. This is not the same as saying that the ideal is epistemically inaccessible. Even if those epistemic worries run deep, they are contingent and could be overcome at least in theory. Nor am I merely saying that ideal-theoretic work is pointless. Let us grant, for the sake of argument, that the value of political theory does not depend on its being practically useful.51 My claim is that the ideal-theoretical method is misguided because it relies on wrong assumptions about the status of ideals in the making of normative judgments. What we have is a metanormative disagreement.

Recall the ideal-theoretic reply to Mills: Our normative judgments about the world require appeal to some principles and, if this is so, then it is surely useful to model what an ideal society looks like according to those principles. Stemplowska and Swift, for example, state that ‘one cannot develop concrete proposals without the philosophy—the theory of ideals—to identify the principles that should inform them’.52 Elsewhere Erik Olin Wright states that behind every critique, ‘there is an implicit theory of justice, some conception of what conditions would have to be met before the institutions of a society could be deemed just’.53 Neither author provides any argument in favour of these claims; they are simply stated as if they were obvious. But they are not, in fact, obvious. The claim that our normative judgments derive their authority from a set of principles that specify what a just world would look like is a specific metanormative view in need of justification.

Specifically, it is a Kantian understanding of moral norms which sharply distinguishes moral philosophy from practical activity and subordinates the latter to the former. For Kant, we have to seek ‘the ground of obligation … not in the nature of the human being or the circumstances of the world in which he is placed, but a priori solely in concepts of pure reason’. He makes a strict categorical distinction within practical cognition between moral philosophy, resting ‘entirely on its pure part’, and ‘everything else in which there is anything empirical’.54 While we inevitably have to act in a material world and take its constraints into account when designing our actions, the source of normativity cannot be found in this world; normative authority emanates from a moral law belonging to a non-material realm.55 From this metanormative view, the ideal-theoretic methodology follows quite straightforwardly: if the aim of political action is to make our world conform to the principles provided by pure reason, then it seems advisable to start by modelling the kinds of political institutions that would realize these principles. The ideal model has normative authority because it realizes authoritative moral principles. All else being equal, it would be right to move closer to that ideal and wrong to move further away from it.

Materialism denies this image of normativity.56 This is not just because moral theorizing is always ‘shaped by social structures, by “material” social privilege and disadvantage’ and therefore never ‘pure’.57 It is also, and more fundamentally, because all cognitive activity is part and parcel of our interaction with the material environment. In sharp contradistinction to the Kantian dualism of the material and the moral, Marx and Engels write in The German Ideology that ‘[t]he production of ideas, of conceptions, of consciousness, is at first directly interwoven with the material activity and the material intercourse of men—the language of real life’.58 Normative thought is no exception: it cannot be understood apart from our material, embodied interaction with our environment.

I do not here want to wade into the extensive debates about the sources of normativity in the (materialist) literature on social critique59 or the debates about Marx and normativity.60 For my purposes, let me just say that I think normativity is best thought of as arising from our experience of the world as problematic. Iris Marion Young writes:

Our ideas of a better society are not derived from rational principles but arise as potential solutions to overcome our dissatisfaction with the world. Appeals to ideals are, as Finlayson puts it, ‘shorthand for demands’ for change.62 The source of normativity is not these ideals themselves, but here in this material world in the pressing problems of subordination, exploitation, oppression that we face.

Or take Angela Davis’s thought as an example. As mentioned above, her arguments for prison abolition do not rely on an ideal model of justice, but this does not mean that she never appeals to ideals or values. Normative concepts like liberation and oppression are pervasive in her work. Yet, she does not derive the normative force of her arguments from an ideal model of liberation, but rather sees value in such normative concepts only insofar as they help us respond to the problems we face: ‘if the theory of freedom remains isolated from the practice of freedom or rather is contradicted in reality, then this means that something must be wrong with the concept’.63 We cannot, on her view, consult an ideal theory of justice to find out whether prisons are compatible with freedom. Rather, concepts like ‘freedom’ and ‘liberation’ are ways of articulating problems and demands for change.64

Note that seeing normative thought as part of our interaction with the world is not the same as instrumentally subordinating theory to action. David Estlund has defended the value of highly idealized and ‘hopeless’ political theory on the straightforward grounds that ‘[u]nderstanding justice is important whether or not that understanding promotes justice or anything else valuable, because it is the understanding of something which is, in itself, important’.65 But this works as a defence of ideal theory only on the assumption that ‘justice’ is the kind of thing we can understand by abstracting away from the real world until only pristine pure reason remains. If our normative concepts, like ‘justice’, are responses to problems in the real world, then understanding the concept of justice and its value for us, requires investigating the context that gave rise to it and its role in political struggles; in other words, genealogy rather than ideal theory. This inquiry—and here I agree with Estlund—does not depend for its value on being immediately useful to political action. Critical theorists, like Adorno, have notably been scathing about the demand ‘that every idea should immediately produce its own legitimating document explaining its own practical use’. Sometimes, instead, we need to ‘resist the call of practicality with all our might in order ruthlessly to follow through an idea and its logical implications so as to see where it may lead.’66 Nevertheless, Adorno does not believe that ‘following through an idea’ can be done in abstraction from understanding its role in practice, because ‘[i]n its origins thinking is no more than the form in which we have attempted to master our environment and come to terms with it’.67 Any attempt to unmoor the idea from practice means misunderstanding it.

If this is true, then the ideal-theoretic method is fetishistic. Recall Feuerbach: ‘The fact is not that a quality is divine because God has it, but that God has it because it is in itself divine’.68 We can say much the same of ideal theory: the fact is not that a certain action is worth pursuing because it brings us closer to the ideal, but that we have projected it onto that ideal because it is worth pursuing. The dissatisfaction with how things are leads people to articulate demands for change and sometimes even to come up with utopian ideas for radically different societies that will eradicate the sources of their unhappiness (I return to the difference between utopianism and ideal theory below). Ideal theorists then take ideals articulated as part of these struggles, like ‘justice’ or ‘equality’, and hypostatize them. They treat them in abstraction from the problems they were originally meant to solve and develop ideal models which are meant to tell us what justice ‘requires of us’—even, according to some, if those requirements cannot realistically be fulfilled by any human society.69 As such they have taken the products of our own creative efforts to come up with solutions to social problems and fetishistically invested them with normative authority over human action. Normative theorizing, no matter how specialized, abstract, and distant from real politics, is always still a part of the material world and it makes use of concepts which make sense only as part of our collective interaction with this world, part of ‘the language of real life’. When theorists deny this, they are guilty of self-misunderstanding.

Thus, the fundamental methodological challenge that materialism poses to ideal theory is best understood as fetishism critique, not ideology critique. At least in the way authors like Mills and Stahl have levelled it, the ideology critique of ideal theory entails accusing it of ‘stabilizing prevailing relations of power’ because it occludes pressing injustices, like patriarchy and racism.70 This is true of much ideal theory. But the same can be said for a lot of nonideal theory; for example, when it uncritically takes contingent features of our actual societies as immutable ‘feasibility constraints’ or accepts ‘whitewashed’ epistemological frameworks.71 If we want to criticize ideal theory as a method, we need to pinpoint the methodological mistake it makes, rather than just point to its unsatisfactory outcomes.

The fetishism critique is not incompatible with ideology critique, however. In fact, it helps us see more clearly why many versions of ideal theory may be ideological. The fetishistic treatment of ideals tends to make us treat contingent historical ideas, which are products of colonial, racist, and patriarchal societies, as if they were universally valid. But it is not clear that this is necessarily the case whenever we do ideal theory. A feminist theorist may well come up with a highly idealized theory which tells us, say, that, based on some abstract understanding of the ideal of autonomy, ideal justice requires the abolition of gender.72 Such a theory would rely on the same methodological self-misunderstanding as any other ideal theory; but it does not seem to stabilize existing relations of power.73

I have not here presented arguments in favour of a materialist approach to political theory—that is not my aim in this article. The point of this section has been to clarify what I see as the deep (and usually unarticulated) methodological differences between ideal-theoretic and materialist theorizing. The fact that both sides in the debate have generally left their metanormative commitments unarticulated has led to a situation where detractors and defenders of ideal theory seem to talk past each other. Stemplowska, for example, simply ‘find[s] it hard to imagine that anyone might object to such [ideal] theorizing’.74 Lorna Finlayson, meanwhile, laments the ‘profound and frustrating difficulty’ of expressing objections to that approach in ways that are not simply dismissed as missing the point by mainstream philosophers.75 In the first place, my argument is that being explicit about metanormative assumptions would clarify such frustrating misunderstandings. The fetishism critique presupposes a materialist metanormative theory, and, as such, ideal theorists who are willing to defend metanormative idealism need not be troubled by it.

But my sense is that many ideal theorists will be, or should be, somewhat troubled nevertheless. This is because they have often been reluctant to make the defence of their methods conditional on accepting controversial metanormative views. In fact, Rawls, at least in his later work, explicitly emphasized the metaethical neutrality of his theory of justice. The next section therefore asks if there is any way for someone like Rawls to defend an ideal-theoretic method without idealist metanormative commitments.

Rawls notably insisted that his method was compatible with a wide range of metaphysical views, including materialism.76 So, can we make space for ideal-theoretic methods within a materialist political theory? The fact that ideal models do not have independent normative authority does not establish that they have no authority. Many human artifacts, including governments and social norms, have practical authority for us. Thus, ideal theorists may argue that, whatever the metaethical status of ideal models, we inevitably act politically on the basis of our ideals, and hence political theory should be in the business of supplying coherent models of those ideals that can guide our actions. This, however, is idealism in the familiar sense of assuming that ideas determine our actions. While avoiding metanormative fetishism, I will argue that it is guilty of practical fetishism by according ideal models undue practical authority.

Rawls, at least in his later work, not only claims metaphysical neutrality for his theory, but also explicitly understands it as an attempt to solve a pressing political problem.77 His ideal theory, he claims, is from the very start determined by ‘the practical purposes and social role that a conception of social justice must fulfill’.78 The problem for members of a liberal society, according to Korsgaard’s interpretation of Rawls, ‘is what we might call the distribution problem: people join together in a cooperative scheme because it will be better for all of them, but they must decide how its benefits and burdens are to be distributed’.79 Solving this problem is complicated by the fact that, despite all participating in one cooperative scheme, we do not share any substantive conception of the Good which could ground agreement on principles of justice. Hence, for Rawls, the role that a liberal theory of justice is supposed to fulfil is ‘to serve as a shared point of view among citizens with opposing religions, philosophical and moral convictions, as well as diverse conceptions of the good’.80 In other words, it is a practical solution to what Rawls saw as the pressing political problem of the day: that citizens in modern liberal democracies hold a plurality of worldviews but nevertheless need to agree on the basic constitutional principles regulating the basic structure of their societies.

On this view, ideal models are not, in themselves, invested with any normative authority; rather, following Korsgaard’s interpretation again:

Justified in this way, ideal theory clearly does not fall foul of metanormative fetishism. Instead of having normative authority of their own, judgments about ideal justice can be ‘treated as modes of public contestation whose point is to reform, sustain, and replace specific normative orders and social structures’.82 Sangiovanni goes as far as claiming that, once understood in this way, we can ‘think about even the most ideal forms of social and political philosophy as critical theories’ which are ‘emancipatory, embedded, diagnostic, and self-reflexive’.83

I would caution against such a hasty reconciliation, however. For, once we accept that the validity of ideal theorizing depends on its ability to solve actual political problems, that ‘the ideal component of the theories would be assessed by their nonideal functions, rather than the other way around’,84 one actually needs to make empirically grounded arguments for ideal theory as a useful form of practical problem-solving. Defenders of ideal theory can no longer help themselves to the claim that ‘until the ideal is identified, … nonideal theory lacks an objective’,85 since the objective is now acknowledged to be provided by the nonideal function we want the theory to serve. And once this priority of the nonideal over the ideal is acknowledged, Mills’s original criticism returns with full force.

Recall that Mills criticized ideal models for not helping us understand present injustices—a criticism that puzzled some ideal theorists since, they argued, ideal models are not meant to help us understand what is but what ought to be. But if normative theorizing is meant to provide solutions to practical problems then it plainly does matter that we have a good understanding of the problems we want to solve. Mills’s point is that abstract reflection on ideals and principles simply is not an appropriate response to the political problems we face. In fact, ideal theory misconceives the problem: the idea that our societies are ‘cooperative ventures for mutual advantage’ which simply stand in need of mutually agreeable principles of justice is an assumption ‘so directly contrary to reality, so centrally distortional of the essential defining features of the phenomenon in question, that they guarantee that a theoretical grasp of it will never be achieved’.86 The belief that the remedy for injustice lies in perfecting our conception of ideal justice, arguably, seems plausible only from the perspective of those whose social position already make the world appear as ‘nearly just’.87 Once we take into account deeply entrenched power structures of white supremacy, patriarchy, and capitalist exploitation, it looks implausible to claim that the problem is that we just have not yet found a suitable conception of justice for everyone to agree on; that ‘it is just a matter of coming up with better arguments’.88

While Rawls’s justification of the ideal-theoretic method, then, was an eminently practical one, it is telling that later ideal theorists have tended to respond to critiques like Mills’s by conceding that ideal theory may serve no practical use but that it is still valuable theoretical work.89 If we read Mills as presupposing a materialist metanormative framework, then his critique has a lot more bite. And it has bite against Rawls, in particular, because he was explicitly committed to a pragmatic form of political theory that does not rely on metaphysical assumptions. Unfortunately, Mills did not make this explicit, which has allowed his opponents to dismiss his critique too easily.

These criticisms can, again, be redescribed as a form of fetishism critique—but this time at the level of social, rather than metanormative, theory. For Rawls’s ‘political’ conception, while not investing ideal principles of justice with independent normative authority, does bestow practical authority on them. In conceiving of the fundamental problem of political theory as one of disagreement over principles of justice, Rawls relies on an idealist social ontology where our ideas determine political practice, and what stands between us and justice is simply a better theory of justice.

Materialism, as a social theory, however, claims that this is precisely not how social change happens. The claim that social change is not driven by ideals has often been interpreted as crude economic determinism that assigns ideas a merely epiphenomenal role in the superstructure.90 On this view, however, not only ideal theory, but all theory, would be pointless. A more plausible version of historical materialism recognizes that ideas play an important role—‘theory, too, becomes a material force once it seizes the masses’, said Marx91—but denies that ideals have practical authority qua ideals. Recall that ‘[a] principle has practical authority for an agent or group just so long as that agent or group is disposed to take the mere fact that some course of action would be to act in accordance with the principle as a (pro tanto) reason to so act’.92 Rawls assumes that citizens act politically mainly on the basis of their conceptions of justice; hence, the role of political theory is to show them what ought to follow from the principles that already have practical authority for them. On a materialist view, though, ideal models have no such authority. Ideals, like justice and equality, may inspire us to act, they may help us imagine political alternatives, and they may embody practical solutions to political problems. Thus, they are not epiphenomenal. But it is not the mere fact that an action brings us closer to an ideal that motivates political actors. What motivates action is the need to overcome oppression, which is why materialist theories locate the agency for progressive social change among those who have a material interest in it, i.e., those currently oppressed.

Marx and Engels accused the utopian socialists of their time of wishful thinking in supposing that ‘[h]istorical action [would] yield to their personal inventive action, historically created conditions of emancipation to fantastic ones’.93 Rawls’s idea that political theory can motivate social change by providing the most reasonable theory of justice faces the same accusation. From a Marxist point of view, the notion that our ideals have such independent power is fetishistic. Even if it does not accord undue normative authority to ideals, it still accords a practical authority to them that they do not, in fact, have. This is, of course, a disputed social-theoretical claim; I do not pretend that it is obvious, nor can I defend it in detail here. My point is, again, just to clarify the nature of the materialist critique of ideal theory. The upshot is this: ideal theorists need to either admit that their method relies on controversial idealist metanormative commitments, or they need to provide empirically grounded social-theoretical arguments for why ideal theory in fact has a crucial practical role to play in politics.

Before concluding, let me anticipate one possible objection to or misunderstanding of my arguments. Opposition to ideal theory is, regrettably, often conflated with opposition to long-term, radical, political goals. Jordan Walters, for example, has argued that ‘nonideal theory is ideological in virtue of the fact that it rules out more radical utopian ways of theorizing by methodological fiat’.94 Simmons, similarly, claims that without ideal theory we ‘allow irrational free rein to the mere conviction of injustice and to eagerness for change of any sort’.95 However, the methodological critique of ideal theory as fetishism by no means implies that we should not think about long-term political goals, or that political theorists have no role to play in articulating such goals. Let me explain.

The methodological question of whether we begin from the ideal or from the problem we want to solve is perpendicular to the question of whether we want to be reformists or utopians. Clearly an ideal model can be radical and utopian to the point of being, in Estlund’s words, ‘hopeless’.96 But other ideal theories, especially practice-dependent ones, may build ideal models that (too) closely resemble the actual world and therefore not be radical at all. Rawls’s theory of justice has been accused of this fault in assuming (and naturalizing), inter alia, the existence of nation states,97 the bourgeois family,98 and market incentives.99 Equally, non-idealizing approaches can be reformist by focusing on ‘realistic’ policy solutions that can be implemented here and now without fundamentally changing society.100 But they need not be. A materialist political theory starts with the problems to be overcome, and what determines whether the resulting theory is reformist or radical is whether the problems identified are seen as surface-level or deeply entrenched. Mills himself identified global white supremacy as a fundamental obstacle to emancipation101—hardly something that can be addressed with reformist policy fixes! Equally, Marx rejected any philosophy that imposes ideal blueprints on society, but in diagnosing the root cause of the proletariat’s suffering as a necessary outcome of our mode of production, he made it clear that nothing short of a revolution would solve the problem. It is deeply unfortunate that the idea of ‘problem solving’ has, in the literature on ideal and nonideal theory, become synonymous with ‘a relatively narrow and short-term perspective, aiming to identify “remediable injustices.”’102 In terms of philosophical methodology, there is no deep difference between proposing a policy fix to a technical problem or proposing a complete reorganization of society.103 Neither accords independent normative or practical authority to the proposed solution—whether we want to pursue it depends on whether we think it will solve our problem.

The example of Tommie Shelby and Angela Davis on prison abolition illustrates this nicely. The latter, in her activism as well as her philosophical writing, begins from an urgent practical problem: ‘how to prevent the further expansion of prison populations and how to bring as many imprisoned women and men as possible back into what prisoners call “the free world.”’104 This nonideal concern with improving the plight of prisoners here and now may be expected to lead to arguments for short-term criminal justice reform, but not so. Davis’s social-theoretical analysis of the prison industrial complex and its imbrication with capitalism, racism, and sexism leads her to argue that the way forward is total abolition. To make prison abolition work, though, Davis acknowledges that we need to do much more than simply subtract prisons from the world as it is. Abolitionism ‘require[s] us to imagine a constellation of alternative strategies and institutions, with the ultimate aim of removing the prison from the social and ideological landscapes of our society’.105 Thus, starting from the urgent problem of the plight of imprisoned people, Davis ends up arguing not just for the radical position of prison abolition, but for a thoroughgoing transformation of society as a whole.

Davis does imagine a radically different society, but the normative and practical authority of her arguments is not derived from an ideal model.106 The normativity of her arguments comes from the pressing need to address deep injustice, and her utopian proposals serve to denaturalize the existing prison system and as proposals for practical solutions to the problems prisons are supposed to solve.

It is true that Davis’s positive proposals for transformative justice do play a certain role in justifying her claim that we should abolish prisons. After all, we could agree that prisons are terrible but believe that there simply are no alternative options; in this way, providing an alternative solution strengthens the argument for abolition. But this is very different from the role that ideal models play in ideal-theoretic approaches. To see why, consider the methodological status of small-scale reformist policy solutions. Say that we find it problematic that prisoners are socially isolated, and therefore we propose a policy that bans restricting prisoners’ use of phones and internet. That policy is clearly not an ideal model of what justice requires of us; it is simply a practical proposal that is meant to bring us closer to justice. But now suppose, with Angela Davis, that we think such reformist policy proposals are insufficient at best and outright counterproductive at worst. Instead of reform, we propose a fundamental transformation not just of the prison system but of society at large. Naturally, the more wide-ranging such transformative proposals are the more difficulties we will face in assessing their plausibility and desirability. But, leaving those issues aside, in terms of normative methodology, the role of utopian proposals for reorganizing society is not fundamentally different from the role of small-scale policy fixes. Shelby, on the other hand, invests his ideal model with independent authority: through reflection on moral principles, he concludes that prisons have a place in a just society and therefore need not be abolished. Clearly, other ideal theorists could come to a different conclusion; other ‘problem solvers’ may argue for small-scale reforms. But this example should have demonstrated that it is not true that, without ideal theory, we somehow limit the radicalism of our thinking; in some cases, just the opposite may be true.

Ideal theorists, of course, will say that in order to assess whether any proposal, small-scale or utopian, does in fact promote justice, we will need a background ideal theory of justice. I have not, here, been able to present a comprehensive refutation of that counterargument. What I have shown is that it relies on the very premise that materialists contest. Thus, without further argumentation, it is question-begging in assuming the normative and/or practical authority of ideal models that materialists dispute. In fact, insofar as Kantian constructivists, like Rawls and Korsgaard, already agree that normative theory is ‘the use of reason to solve practical problems’,107 they really owe us an account of why ideal theory is a good solution to any of the pressing political problems we face.

On the materialist account, thinking about our political aims in terms of a set of ideal models which tell us what justice ‘requires of us’ is fetishistic. But there are other ways of thinking about our long-term aims. If we start from the need to overcome existing injustices, then the radicalism of our proposed solutions depends on how entrenched we think those injustices are. Our attempts to transform the world no doubt benefit from reflection on what a radically different social world would look like—and insofar as it is the best proposal we can think of, we may even term it ‘ideal’ in a loose sense—but such ideas would have the character of practical proposals for social change without carrying normative or practical authority of their own. We may well decide to write ‘recipes … for the eateries of the future’.108 If they look enticing enough, they may inspire us to try them out. But those recipes will not tell us what to eat; they are mere suggestions for how to satisfy our needs. And the ‘proof of the pudding’, Friedrich Engels once wrote, ‘is in the eating’.109

I have tried to show in this article that Mills’s suggestion to see debates about ideal theory as a dispute between materialism and idealism is a fruitful one. Where different authors appeared to be talking past one another, I have argued, we can better make sense of their disagreements once we see that they (often implicitly) rely on materialist and idealist assumptions, respectively.

Let me offer a schematic reconstruction of how the dispute looks once we take this into account. Rawls originally justified his ideal-theoretic method in an explicitly pragmatic way that was supposedly neutral between idealism and materialism. Mills then offered a critique from a materialist point of view, claiming that Rawls’s method is ideological because it cannot help us address entrenched injustices. The response from defenders of ideal theory was largely to abandon Rawls’s pragmatic justification for ideal theory, justifying it, instead, on epistemological grounds: to know what social changes will make our society more just, we need a theory of what a wholly just society would look like.110 But this argument relies on controversial idealist metanormative premises: namely that ideal models have normative authority of their own. In fact, it relies on just the kind of Kantian dualism—that moral philosophy happens in the strictly non-empirical realm of pure reason—that Rawls claimed to have purged from his own Kantianism.111 But if defenders of ideal theory cannot rely on such metanormative assumptions, they will have a much harder time defending their methods against the materialist critique. They would have to make arguments for ideal theory compatible with materialism. This is hard, because a core claim of materialism is that ideals lack the kind of normative or practical authority that would justify the method. In any case, the upshot is that both proponents and critics of ideal theory would do well to be much clearer about their metanormative and social-theoretical commitments.

I have also argued that the objection materialists have against ideal theory is most precisely articulated as a fetishism critique. I am very sympathetic to Mills’s ideology critique; it is certainly true that, by abstracting away from the nonideal, ideal-theoretic methods tend to draw our attention away from pressing injustices. Yet, this is not a necessary feature of the method. The example of Shelby’s work discussed above shows that it is possible to do ideal-theoretic work informed by careful engagement with real-world racial oppression. If we want to object to ideal theory as a method, and not just as it is most often practiced, we need to explain why it is methodologically misguided. I have argued that it is so because it accords independent authority to ideal models which, in fact, do not carry such authority. On a materialist image of the world, our normative thinking is nothing but a continuation of our practical engagements with the world we inhabit; our concepts cannot be understood apart from the practical problems they were intended to help us solve. The idea that political theorists are here to ‘discover what justice requires of us’112 is an inversion; the question we should ask is what we require of our social world. This question arises for us, only because the world appears to us as problematic; it leaves us dissatisfied. This ‘desiring negation that action brings to what is given’113 may well drive us to think in a highly utopian manner about the kind of society that would satisfy our desires. But these positive proposals should be understood as the outcomes of human creativity; not as ideal models standing apart from our agency and constraining it.