Beyond the nonideal: Why critical theory needs a utopian dimension

John Rawls famously argues that in order to arrive at a plausible conception of what justice requires politically, we ought to proceed in two steps (Rawls, 1971, pp. 245–246). First, we ought to develop an “ideal theory of justice” that lays out what principles of justice can be justified if we imagine them to govern a society that is unlike ours in that certain limitations are removed. Developing such a theory, Rawls argues, is necessary to allow us to grasp the correct conception of justice. In a second step, we can then use this conception to find out what we must do in our actual, nonideal circumstances.

The idea that we need to refer to a utopian, ideal state of affairs in order to understand which conception of our most basic political concepts we ought to endorse is one of the features of Rawls's thought that has attracted the most commentary (see Robeyns, 2008; Simmons, 2010; Valentini, 2012). Realist opponents of ideal theory in this sense sometimes argue that ideal theory pays too little attention to questions of feasibility, that it does not take the limitations of human nature seriously enough, and that it is ill-equipped to guide us when thinking about necessary trade-offs (Farrelly, 2007; Galston, 2010; Nagel, 1995). In this article, I will examine arguments of a different nature that have historically emerged from the tradition of radical social thought and critical theory. These arguments do not object to ideal theorizing on the grounds that it leads to unrealistic demands or that it is insufficiently constrained by a realistic conception of human nature. Rather, these arguments assume that, given the nonideal circumstances in which current political theorists find themselves, they face limitations to their epistemic, imaginative, and conceptual capacities that distort the ideals they formulate and thereby take them in a direction that accommodates the status quo too much.

In response to these arguments, this article pursues two goals, a critical and a constructive one. On the one hand, I will argue that most (but not all) of these arguments fail and that even the most successful versions do not require us to give up on the idea that our political projects ought to be guided by a conception of an ideal society. Critical theory need not be an exclusively nonideal enterprise. On the other hand, I will outline a methodological approach that integrates critical theory's defining commitment to the idea of immanent critique and the historical nature of reason, but preserves a role for the construction of idealized visions of the future in its normative parts.

In Section 2, I briefly survey Rawls's conception of ideal theory. In Sections 3, 6, I focus on four arguments against ideal theory that emerge from the critical theory tradition and that deny that a conception of what a genuinely ideal society would be like is accessible to us. I distinguish between an epistemic, an imaginative, a radical conceptual, and a moderate conceptual impossibility claim, and I conclude that only the moderate conceptual impossibility claim offers a plausible argument against ideal theorizing. In Section 7, I argue that rejecting ideal theorizing (in a broad sense) based on such arguments runs the risk of unnecessarily constraining critical theories and of condemning them to a form of conservatism. In Section 8, I develop a response to this: critical theories can be guided by conceptions of ideal states of affairs that derive from people's awareness of the contradictions that govern their societies. I conclude that this allows critical theories to take on certain commitments of ideal theory without falling victim to its pitfalls.

Rawls argues that we can only understand what the most plausible conception of justice is when we examine which principle of justice would govern an idealized state of affairs. Most importantly, Rawls argues that, for the purposes of this examination, we ought to imagine a well-ordered society under favorable circumstances in which agents display “strict compliance” (Rawls, 1971, p. 8) with the principles that we choose for such a society (Stemplowska & Swift, 2012). Just what Rawls means by “strict compliance” is disputed (see Sirsch, 2020, p. 30), but the idea behind the claim that we need to engage in such idealization is not difficult to understand. In politics, we have good reason to concede that others will frequently be imperfectly motivated to comply with what justice requires of them. Consequently, it is often the case that what we ought to do—taking such imperfect motivation into account—is to pursue realistic outcomes that are not fully just but that are closer to full justice than what we would achieve if we were to pursue an uncompromising program. However, this clearly should not lead us to identify justice with the outcomes that we ought to pursue under such nonideal conditions. Our conception of justice, ideal theorists argue, ought not to be determined this way; instead, it should be determined by thinking about what outcomes we would, counterfactually, be justified in pursuing in a situation without such limitations.

While Rawls clearly acknowledges that our on-the-ground political projects must be primarily informed by a nonideal theory that takes the fact of noncompliance into account, he argues that we can only develop such a nonideal theory of justice—which is justified, in terms of the compromises it makes, by its capacity to move us as closely as possible to an ideally just state of affairs—if we have previously determined the correct “non-concessive” conception of justice (Estlund, 2019, p. 6, 31).

Even though Rawls does not assume that ideal theory is immediately action guiding, it is clear that his theory depends in several ways on the project of imagining a utopian, idealized state of affairs in which—due to full compliance and favorable conditions—our chosen principles of justice can be fully realized. First, having a conception of which norms we would rationally choose to govern an imagined ideal society provides us with an understanding of the normative authority of these norms for actions in our actually existing society. The utopian reference to a society free of imperfections is thus something we must be able to appreciate in order to engage in political theorizing and justified political action. Second, this conception also gives content to nonideal theory, since making concessions is only justified when this cannot be avoided. Having a correct assessment of what we would aim for under ideal circumstances is thus a precondition of knowing what we ought to aim for under the actual circumstances.

In what follows, I examine objections to the idea that reference to a utopian state of affairs can serve the goal of orienting our political thinking and action. I argue that there are four such objections—relating to our epistemic, imaginative, and conceptual capacities—to be found in the tradition of critical theory that, even though different in terms of their claims, have the same structure. They do not (or do not primarily) deny that referring to an ideal state of affairs may be desirable in principle or perhaps necessary for political theory. Rather, they deny that political agents in current, nonideal societies are in a position to have any confidence in their ability to correctly determine what such a state of affairs would look like. This is because they take the nonideal circumstances in which we find ourselves to radically constrain the epistemic, imaginative, or conceptual abilities that are necessary for such a determination. Although not all the authors surveyed in the following explicitly turn against “ideal theory,” if alone for the reason that this concept did not yet exist when some of them developed their arguments, their arguments all entail that we cannot engage in such a project.

An epistemic argument against ideal theory is implicit in the classic theories of ideology, starting with Marx. Such theories assume, at least for modern capitalist societies, that the nature of the injustices that characterize these societies is structurally nontransparent to political agents, for two reasons. First, capitalist social structures have an essentially misleading appearance. In The German Ideology, for example, Marx argues that the (false) belief that it is people's ideas that determine the structure of their material forms of cooperation is more than a mere epistemic mistake. Rather, “this phenomenon arises just as much from their historical life-process as the inversion of objects on the retina does from their physical life-process” (Marx & Engels, 1970, p. 36). In other words, it is the social structures themselves that systematically produce false beliefs to that effect.

Second, people's positions in the social structure can also distort their epistemic capacities. In the critical theory tradition, this claim has been most explicitly defended by Lukács (1923/1971) and later by Marxist-feminist standpoint theorists (Hartsock, 1983). Both argue that members of some social groups are constitutively excluded from knowledge of the true nature of the social structures in which they find themselves (typically: members of dominant groups), whereas members of other (subordinated) groups can, in principle, gain correct knowledge as part of their social and political struggles.

This relates to ideal theory as follows: If it is true that social structures are systematically epistemically opaque (for some or all members of society), that means that theorists will typically lack some or all of the knowledge required to fully comprehend how these structures violate various political ideals. Not only will their conception of the utopian state of affairs that informs their ideal theory therefore likely be insufficiently utopian, since it will not contain adequate remedies for all relevant injustices, but ideal theorists will also lack any justification for their belief that they have developed a sufficiently comprehensive utopian conception. In this case, their belief that they have developed an appropriate conception of justice that can guide political activity will also lack sufficient justification.

Mills, I assume, endorses the conditional in (5) but rejects the claim in (7), and thus the idea that there is epistemic continuity between our current situation and a potential ideally just society [i.e., the consequent of the conditional in (4)]. This commits him to denying the possibility of an ideal theory that makes substantive claims. To avoid having to give up on political theory altogether, he therefore also rejects the basic commitment of ideal theory in (1).

The first objection is a systematic one. It is unclear whether the claim only affects ideal theory. From the assumption that unjust societies can distort our epistemic capacities it follows that we can never be certain about whether our claims about what an ideally just society would look like are true. However, the same problem applies to normative claims that appear in exclusively “nonideal” theories that do not rely on any concept of an ideal state of affairs—for example, claims about relative improvements, the justification of trade-offs, and so forth—at least unless we also have a story about why the epistemic limitations to which we are subject only impact our attempts to conceptualize ideal states of affairs but not otherwise (Stahl, 2022, p. 9).

However, the argument is vulnerable to the same objections as the epistemic argument. First, it is unclear why our capacity to imagine ideal states of affairs should be more affected than, say, our capacity to imagine incremental improvements. Again, it is hard to see why the argument entails a specific objection to ideal theory. The only form of ideal theory that would be affected is one that proceeds from the assumption that we should begin with a description of an ideal society because this description is safer against socially created epistemic distortions. However, not all defenders of ideal theory make this assumption. Rawls, in particular, seems to assume that we need to begin with a description of an ideal state of affairs in order to not constrain our conception of justice by considerations of feasibility. This does not entail that our capacity to imagine such a state of affairs is epistemically safer than our capacity to imagine incremental improvements. But then, such a theory is in no worse shape than an exclusively nonideal theory if we assume that social factors can affect our imaginative capacities across the board.

Second, the argument carries an enormous justificatory burden, since it relies not only on the hypothesis that there is nothing in the psychological household of human beings that allows them to resist the distortion of their imaginative capacities by social forces, but also on the claim that members of oppressed groups will never display internal resistance in the form of imagining a society without oppression. Of course, there is nothing strictly impossible about a situation in which this is true—and as negativists such as Freyenhagen (2013) have explored, building on Adorno's work, there are ways to construct critical theories under this assumption. But we ought to require considerable evidence before we accept such an impossibility claim, and the evidence for its being true is most likely evidence that would simultaneously support the epistemic impossibility claim. In other words, the imaginative argument is unlikely to independently succeed where the epistemic argument fails.

Another argument focuses not so much on subjective capacities but on the availability of social concepts and hermeneutical schemes. A radical version of this argument holds that, due to some feature of the conceptual resources that we currently have available to us, these resources will necessarily lead us to adopt misleading conceptualizations of ideal state of affairs. A more moderate version, by contrast, only holds that the conceptual resources we have available to us at any given point allow us to formulate only context-specific ideals that are always in need of further revision.

While this argument is widely shared by many forms of critical theory (for a feminist statement, see Harding, 1986, pp. 648–649), I will analyze it here in the way it has been worked out by Mills and Adorno, since their conceptual approaches are more closely tied to classical critical theory.

Mills argues that one major consideration in political theory must be “conceptual adequacy” (Mills, 2005, p. 174). Ideal theories, according to Mills, are not conceptually adequate, since their concern with describing ideal states of affairs will lead them to develop conceptual schemes that “tacitly [represent] the actual as a simple deviation from the ideal” (Mills, 2005, p. 168), as opposed to nonideal descriptive concepts, which are geared toward an informative analysis of oppression and injustice.

A political theory of this type must fail, Mills argues, because it attempts to describe an idealized state of affairs with the help of concepts which are constitutively unsuited to the purposes of political theory: They are unsuited because they are incapable of correctly describing those injustices that pose the most significant problems for normative political theory. The resulting descriptions of ideal states of affairs are not suited to fulfilling the orienting function that they need to have in ideal theory.

In particular, Mills assumes that the structure of unjust societies and the class-specific experiences of most political theorists, who belong to privileged groups, account for the fact that they frame their arguments in unsuitable concepts (Mills, 2005, p. 168). While this argument is powerful in offering a plausible explanation for some failures of particular ideal theories, it is unclear whether it supports Mills's objections to ideal theory.

Mills is right to argue that, if members of privileged groups share conceptual schemes that obscure certain aspects of reality, and if the concepts in question are then used to develop a model of an ideally just society, the resulting ideal theory will be misleading. This leaves open the possibility, however, that we might be able to construct better forms of ideal theory by consciously drawing on the conceptual resources of marginalized groups that will most likely be better suited to capturing the central forms of injustice. That Mills does not consider this possibility is likely due to the fact that he assumes that the “descriptive mapping concepts” (Mills, 2005, p. 175) that emerge from such groups are not suitable for the project of specifically conceptualizing ideal states of affairs.

If Mills argument is both that we should not formulate our descriptions of ideally just states with the help of concepts that are developed in isolation from an analysis of actual injustice, and that we cannot formulate these descriptions with the help of concepts that are informed by such an analysis, this rules out ideal theorizing a priori. However, it is not obvious that we can never use the fine-grained conceptual vocabularies we have developed for describing oppression and domination to describe an idealized society. Of course, concepts such as “exploitation” or “sexual harassment” are not suitable to be employed affirmatively to describe any features of an ideally just state of affairs. However, these concepts may well have semantic contents and inferential relations that can make them useful for the project of developing further conceptual tools in ideal theory. Indeed, Mills's explanation of the impoverished vocabulary of traditional ideal theory is precisely that this vocabulary has not been developed on the basis of an analysis of injustice which leaves the possibility at least open that we could develop better conceptual resources for ideal theory if we engaged in such an analysis. Radical negativists may claim, in response, that our best analyses of injustice and the concepts that are best suited for formulating them will never enable us to say anything positive about an ideal state of affairs, but this is at least not obviously true.

Adorno presents a more general, and perhaps more radical form of the same argument (Stahl, 2017, p. 511). Adorno argues that modern societies are integrated into a “totality” by an overarching system of social domination and the technical control of human nature (and nature more broadly). As part of the historical development of this system, the capacities that are involved in human reasoning have played an ambivalent role. On the one hand, humans have used their ability to distance themselves from their immediate reality through conceptual generalization in order to liberate themselves both from the constraints of their external environment and from their immediate impulses. On the other hand, this project of liberation from nature involves the subordination of outer nature, human impulses, and ultimately other humans (Adorno, 1973, p. 320). This authoritarian aspect of human reason is the ultimate explanation of the existence of particular forms of social domination in modern societies—the capitalist market and, subsequently, fascist authoritarianism. Because of these historical dynamics, human conceptual abilities have developed in response to the demands of an ultimately authoritarian project. The basic concepts we have available to make sense of the world are shaped by that function and thus incapable of allowing us to arrive at a complete critical distance from that project.

Being much broader than Mills's thesis, Adorno's argument for the rejection of ideal theorizing also depends on a background theory that will strike many as implausible, due to the massive theoretical burdens it carries. I will not review the debate on Adorno's social theory here. For the purposes of this article, it suffices to discuss one particular objection. This has to do with Adorno's conception of society as a totality. As Axel Honneth (1985/1993) has argued, Adorno assumes that social domination in modern societies is comparable to a process of technical control of nature and works mostly through the manipulation of the preferences and beliefs of the oppressed. This is unconvincing, since it denies the importance of processes of legitimation. That is, all forms of social power, including forms of oppression and domination, must to some extent be able to deal with the resistance of those who are subjected to them, which will find its expression in social conflicts. Social conflicts are not comparable to technical problems, however. They arise when normative expectations have been violated. As a consequence, powerful groups have to respond to some degree by offering legitimation stories, at least if they do not wish to rely on physical coercion alone but exercise some degree of social power that is based on the acceptance of the oppressed. If that is the case, however, we need to assume that such attempts are always contested and that subordinated groups offer counter-narratives and competing interpretive schemes. Consequently, we cannot just assume that societies are integrated to the degree necessary to assume that there is only one dominant conceptual framework available to people. Rather, concepts themselves will be contested and challenged, and often ambivalent.

While this objection is not conclusive—after all, it remains an empirical question to which degree there are actual social conflicts of this type in any given society—it offers strong reasons to remain skeptical of Adorno's radical version of the conceptual impossibility claim.

Next to Mills's and Adorno's versions of the radical conceptual impossibility claim, there is also a more moderate version that historically lies at the root of the critical theory tradition's skepticism towards utopianism. This version is to be found, in two different variants, in Hegel and Marx. Both thinkers assume that human beings' conceptual capacities are part of a historical dynamic and that they are therefore shaped by their social context. On that basis, we can use their arguments to reject the implicit assumption in ideal theorizing that there is continuity between agents situated in current, unjust societies and agents situated in a hypothetical ideal society.

As is widely agreed by contemporary readers of Hegel, he assumes that the conceptual structure of our grasp of natural and social reality is a matter of historical, social practices in which our concepts are rooted (Brandom, 2009, p. 66). This has immediate consequences for his political philosophy. In the preface to the Philosophy of Right, Hegel famously claims that the task of philosophy is not to design “a world as it ought to be” but rather to comprehend the rationality of the existing institutional order (Hegel, 1991, p. 21). This follows directly from the idea that our grasp of reality is mediated by concepts that are bound to historically specific forms of self-understanding. For that reason, individuals' normative beliefs and the forms of justification they have available to them are also always mediated by concepts made available to them through the practices of their social and institutional context.

If one is to avoid falsely believing that one has direct, institutionally, and socially unmediated access to a realm of moral facts but still wants to hold on to the idea that there is such a thing as justified social criticism, one must assume that the social environments in which critics historically find themselves are, at any given point, at least rational to the degree necessary to provide them with substantive standards of criticism that they can justifiably employ. We are entitled to this assumption, Hegel believes, because existing institutions always incorporate the results of a historical learning process into their normative constitution; that is, they answer to a demand for justification that emerges from the failure of earlier institutional arrangements. For this reason, Hegel can describe our conceptually structured self-understandings as being subject to a dialectical development (Pinkard, 1996; Stahl, 2021, pp. 26–28) and as the product of a history of revisions that constitutes a conceptual learning process (Brandom, 2005; Jaeggi, 2018).

While Hegel's commitment to the idea that social and political criticism always operates from within the conceptual frameworks made available by one's environment does not rule out radical criticism—after all, one may be able to use the normative concepts that govern a historical community's shared relationship with their institutions to radically critique those institutions—it rules out a form of social criticism that bases itself on a historically unsituated utopianism and that is exclusively concerned with designing an ideal society, without being able to reconstruct the social foundation of the normative premises of such an enterprise.

Marx takes up Hegel's anti-utopian commitments. He agrees with Hegel that the conceptual frameworks that people when thinking about politics are always tied to specific social–historical contexts. However, Marx clearly does not draw from this the conclusion that we should engage in a mere reconstruction of the rationality potentials of the existing institutional reality. Rather, he assumes that, in the historically normal case, the conceptually structured self-understanding that any given social context makes available to agents is to some extent ideological. Marx's theory of ideology is often read exclusively as a theory that explains the adoption of specifically ideological beliefs through analyzing their functionality for social domination. In The German Ideology, however, Marx and Engels develop at least a fragmentary philosophy of language that makes clear that ideology, on their view, is also a matter of conceptually structured self-understandings. They point out that, on their materialist view, language is to be understood as part of the broader, “material” social practices that it helps to coordinate (see Stahl, 2013, 2022). We should thus expect that the concepts that people use in language will reflect those distinctions that are relevant to the successful cooperative pursuit of their practical goals—at least in those cases where practices are indeed cooperatively structured. Marx and Engels then distinguish between this simple, primordial function of language and an “ideological” state of affairs that comes into being with the emergence of a hierarchy in which some people only perform the work of setting goals that others must then pursue (Marx & Engels, 1970, pp. 44–45). In such a situation, Marx and Engels argue, concepts no longer transparently express the distinctions of an underlying practice, and the dominant class will shape the conceptual resources underlying the socially shared self-understanding such that they respond to their need for social control rather than the needs of everyone who is engaged in that practice. As a result, conceptually structured self-understandings will begin to take on the appearance of independence from practical purposes.

On this Marxian version of the thesis regarding the historical nature of conceptually structured self-understandings, it becomes clear that this rules out a specific claim about conceptual continuity and thereby makes available an anti-utopian argument that Marx himself does not make explicit (for a more nuanced discussion of Marx's anti-utopianism, see Leopold, 2016; Paden, 2002). If our conceptually structured self-understandings respond to concerns that are relevant in a given form of material-reproductive social practice, then we cannot simply assume that the agents in a hypothetical ideal society will frame their political thought in the same concepts that we do. In particular, we must assume that their understanding of which concepts are most appropriate for making sense of the specific domain of the political will have developed through a more extensive learning process than the one we have gone through (if we are not committed to the implausible claim that the transition to an ideal society will not involve any historical learning process), or that these concepts will respond to the needs of material social practices that are deeply unlike ours in that they are liberated from various forms of domination and exclusion. But then, we cannot assume that the concepts in which we now frame our conception of what an ideal society would look like are the concepts that people in such a society would accept as the most appropriate and relevant. In other words, we must assume that there is most likely no conceptual continuity between our current society and any ideal state of affairs. But then it is unlikely that our present concepts will be suitable to formulate the principles which will govern an ideally just society—and if correctly stating those principles is a necessary part of describing it, our present concepts will be unsuitable to describe an ideal society.

I thus take both Hegel and Marx to accept a specific version of what has recently come to be known as a “practice-dependent conception” of justice (see also Stahl, 2022). According to most practice-dependence theorists, the nature of social practices determines the content of appropriate principles of justice (Sangiovanni, 2008, 2016). Authors like Michael Walzer, Thomas Nagel, and the late Rawls are often said to endorse practice-dependence in this sense. To distinguish their views from the conception that I discuss, I will call their conception substantive practice-dependence.

Marx clearly assumes that (at least) all normative concepts are semantically practice-dependent, including the concepts of political theory, and it follows from Hegel's account that he similarly rejects semantic practice-independence (although on different grounds). But if we accept this premise, it becomes clear that we also cannot take a specific kind of conceptual continuity between nonideal and ideal states of affairs for granted: We must acknowledge the possibility that the concepts that are currently available for conceptualizing an ideal state of affairs would be rejected by those who are actually in that situation.

This version of the impossibility claim has the advantage of being immune to the objections raised against Mills's and Adorno's more radical versions. It neither assumes that the concepts at issue are defective because they reflect a preoccupation with power-free states of affairs nor depends on a picture of a totally integrated society. It also has the advantage of being entailed by both Mills's and Adorno's views, and is thus less demanding, but still seems to support the desired conclusion: that ideal theory is a misguided enterprise. This more moderate (but still apparently destructive) form of the conceptual impossibility claim is the strongest objection—in the sense that it has the fewest implications that critical theorists might resist for other reasons—to be found in the repertoire of critical theorists against the project of ideal theory.

In the next two sections, I will examine what follows from this. In Section 7, I will argue that—even if we accept this claim—there are strong reasons for concluding that we should give up ideal theorizing in the broadest sense of the term. In Section 8, I will outline a form of ideal theorizing that is compatible with the moderate conceptual impossibility claim.

Having argued that, with the moderate conceptual impossibility claim, critical theorists can raise a plausible objection to one of the central premises of the very project of an ideal theory, I will now examine whether this entitles us to the conclusion that critical theorists ought to reject ideal theory altogether and merely engage in exclusively nonideal theorizing. In this section, I will briefly substantiate the claim that this leads critical theories into a dilemma.

The dilemma follows from two claims. The first is that critical theories are aiming not merely for explanatory success but also for practical impact. For Horkheimer, it is a defining feature of critical theories that they aim at “an emancipation and at an alteration of society as a whole” (Horkheimer, 1937/1975, p. 208) and that the question of the verification of such a theory cannot be divorced from the possible realization of that aim in a future society (Horkheimer, 1937/1975, pp. 220–221). Similarly, Geuss argues that it is a defining feature of critical theories that they aim to contribute, as theories, to the process of emancipation (Geuss, 1981, pp. 1–2).

The second claim is that political theories that are exclusively “nonideal” in the sense that they think that focusing on the possibility of a radically better society is a distraction are likely to have the effect of promoting a conservative attitude towards politics (Finlayson, 2017; Prinz & Rossi, 2017). Lorna Finlayson argues that a theory which takes “constraints on political possibility” as historical constants “will inevitably tend to nudge us towards a greater acceptance of the status quo, towards more modesty in the change that we are prepared to propose or demand” (Finlayson, 2017, p. 270). For present purposes, however, such a strong argument is not needed. It suffices to say that reference to an ideal state of affairs is likely to promote the internal goals of a critical theory in various ways and that eliminating the possibility of making such reference thus comes with certain costs.

This is because a positive conception of ideal states of affairs can serve the emancipatory goals of critical theories in more than one way (Böker, 2017; Weeks, 2011, p. 204). First, if we have a conception of an ideally just or free society that we are confident is correct, then this picture can serve an important justificatory purpose in the nonideal parts of our theory. We can show that we are correct to fight against certain forms of social inequality or oppression if we can show that they would not obtain in an ideal society. This gives us confidence that abolishing these undesirable features of our current society will not have inevitable unacceptable opportunity costs. Second, a positive conception of an ideal state of affairs has an important orienting function in critical theory (Adams, 2019, p. 2; Robeyns, 2008, p. 345). This is because we will know that any nonideal proposals for gradually improving social relations will be a step in the right direction if we can show that it will lead us more closely to an ideally just or free society. Third, a positive conception of an ideal state of affairs can support the critical force of the theory. We can assume that there is always a danger in political theory that some features of our current society will be falsely taken as unchangeable limitations of our political projects. If we know that these features would be absent in an ideal society, however, then we will be less likely to make this mistake. Fourth, a conception of an ideal state of affairs can play an important role in making a critical theory motivationally effective. We can assume that, in general, political agents will only be motivated to accept some temporary concessions to injustice—which many emancipatory projects will involve by necessity—as part of a political strategy if they are guided by an ideal toward which they are striving (Moellendorf, 2006, p. 426).

The dilemma that thus emerges is the following: On the one hand, if we accept Marx's and Hegel's account of the nature of our political concepts, this would seem to rule out not only ideal theorizing but, more broadly, the development of conceptions of ideal states of affairs for the purposes of guiding political agency. On the other hand, giving up on this project undermines the practical force of the theory, which is precisely what is supposed to set critical theories apart from other normative projects. It is imperative for critical theorists to examine whether there is a way out of this dilemma.

In this section, I argue that the strategy of immanent critique that follows from the conceptual impossibility claim endorsed by Hegel and Marx does, in fact, not rule out that conceptions of ideal states of affairs can play a guiding role in political agency, and actually preserves a place for the utopian content of ideal theory.

If we accept the moderate conceptual impossibility claim introduced in Section 6, which states that the distinctions that structure our conceptually structured self-understandings always reflect the practical purposes of agents embedded in a historically specific form of society and that, therefore, no moral concepts (or the ideals expressed through them) represent an Archimedean point outside the predominant form of society, it follows that the only justifiable form of social criticism is a form of immanent critique that proceeds from within these conceptually structured self-understandings.

As already discussed, such a commitment to immanent critique is not conservative in the sense that it would only allow for a piecemeal improvement of societies as they currently are. On the contrary, the belief that one's critique is rooted in the objective normative potential of an existing society—such as the potential of a society which is at some level already committed to the equal value of all persons—is completely compatible with the belief that the society in question needs to change radically to realize that potential. The commitment to immanent critique does suggest two further claims, however, which may be taken to lead into the conservatism dilemma sketched above.

First, it suggests a form of conceptual conservatism. Even if critical theories advocate for radical change, they seemingly must phrase their criticism in the vocabulary that is best suited to capturing the concerns that are constitutive of present practices in society and thus remain bound in their self-understanding to the conceptual form of those practices. This seems to preclude the possibility of radical conceptual innovation, which may very well be a precondition for formulating a genuinely utopian image of an ideal state of affairs that allows us to adequately respond to the shortcomings of our current society.

Second, as argued above, the commitment to immanent critique also rules out conceptual continuity between the ideals that capture our current normative concerns and the ideals that would capture the concerns of people in an ideally free or just society. If this is the case, this seemingly undermines the possibility that ideal theory can function as critique. If the concepts that political theorists presently use to develop a conception of an ideal state of affairs cannot be presumed to coincide with those that people in that ideal situation would use to formulate their political principles, then we cannot have confidence that we are justified in using any specific ideal theory that is currently accessible to us as a normative standard.

There is good reason, however, to resist the first claim. It does not follow from the claim that we can only engage in immanent critique that such critique cannot lead to radical conceptual innovation. To see why this is the case, we need to take into account that many forms of social critique emerge in response to social contradictions. Hegel and Marx argue that many of the incongruences between what people are rationally committed to as part of their conceptually structured self-understandings and the actual reality of their social practices are not contingent but necessary, or in other words, that the social practices at issue are structured by “social contradictions.” Social contradictions of this kind obtain whenever (i) a practice is structured by a norm N¹ which prescribes not only certain regularities of behavior but also a certain normative orientation towards a set of norms N², and (ii) it is true that the better the agents in the practice conform to N¹, the less possible it is for them to live up to the standards of N² (Stahl, 2021, pp. 251–256).

I take it that Marx sees liberal capitalism as a form of life that is contradictory in precisely this way. Liberal capitalism as a form of life involves a socially shared, normatively prescribed orientation towards individual liberty and freedom of choice and a set of institutionally secured strategies to act on this orientation by treating others as partners in contractual arrangements and voluntary exchanges. However, these forms of social relationship, once universalized, lead to the emergence of social structures that systematically undermine liberty and freedom of choice, at least for the majority of the population.

If an immanent critic (equipped with a sufficiently sophisticated social theory) realizes that this is the kind of situation they are in, they must realize that it would be insufficient to pursue a form of immanent critique that merely applies already available normative concepts. Rather, they will then be rationally compelled to engage in another form of immanent critique which takes the contradiction at issue as a sufficient reason to revise the conceptual commitments that structure their social self-understanding and the practices they reflect at the same time. This will still be immanent critique for, beyond the nonmoral norm that one ought not to pursue one's projects with strategies that cannot possibly succeed, such a critique is not rooted in any external moral principle.

Of course, this raises the question of which alternative set of conceptual commitments critics have reason to endorse in such a situation. The only possible answer to this question is that their search for a new set of conceptual commitments and practices must be guided by the idea that there ought to be a form of life that fulfills two criteria: First, it does not suffer from the contradictions at issue; second, it must allow its participants to continue to understand why the previous, contradictory form of life had seemed justified, in particular in comparison to earlier forms (MacIntyre, 1990; Taylor, 1993; Stahl, 2021, pp. 62–72, 247–251). In other words, the revision of the conceptual commitments that structure social self-understandings must be guided by an ideal of a society that is no longer structured by the contradictions that are constitutive of their current social situation. To develop such an ideal, the utopian project of imagining a society that fulfills these criteria will often be the most straightforward way.

If we add to this the standpoint-theoretical assumption that members of oppressed groups will be in an especially advantageous position to arrive at a diagnosis of the relevant social contradictions, in particular, if they are informed by a substantive social–-scientific theory that discloses the origins of these contradictions (i.e., the mechanisms by which the institutions that compel them to pursue an orientation toward social values systematically undermine that pursuit), we can also avoid the counterintuitive implications of the epistemic and imaginative impossibility claims. In particular, if such groups develop utopian conceptions of a better society, there is no particular reason why they should be unable to imagine a society in which the contradictions no longer obtain or why they would fail to correctly identify the right conditions under which this would be the case.

This leaves the second argument. This argument states that—if we are committed to the Marxian and Hegelian picture of concepts—we are not justified to assume conceptual continuity between the ideals that capture our current normative concerns and the ideals that would capture the concerns of people in an ideally free or just society and that we therefore cannot have confidence that any ideal theory currently available to us will ultimately turn out to be justified. However, even if we accept this argument, it does not entail that we ought not to be guided in our political agency by a conception of an ideal state of affairs. Rather, it only entails that the specific form of ideal theory as we find it in Rawls and other liberal theorists is misguided, since these theorists aim to use the specification of an ideal state of affairs not merely to help us orient ourselves politically, but to ultimately settle the question of what the right conception of justice is. The latter use of ideal theory indeed requires conceptual continuity, but the first use does not.

If Hegel's and Marx's assumptions about the relationship between conceptually structured self-understandings and broader social practices are correct, this means that any ideal that we can develop is always conceptually bound to the practical concerns of its context of origin. If this is the case, then it is unreasonable to assume that there will be continuity between the concepts that people currently use to frame their ideals (i.e., concepts aimed at capturing the specific contradictions of their society) and the concepts that people in a hypothetically ideal society would use to think about an ideal state of affairs.

As Böker (2017) argues, however, we can hold on to a utopian dimension in political thought while simultaneously accepting that our conceptions of what counts as ideal are subject to historical changes, some of which may be prompted precisely by our attempts to realize a given ideal. This kind of realism acknowledges the truth of the impossibility claim regarding the formulation of a utopian ideal that is justifiable to both agents in the current society and agents in the society in which it will actually be realized. However, as Böker convincingly shows, even if we acknowledge the impossibility of arriving at one substantive conception of an ideal state of affairs that remains static throughout the process of trying to realize it, this still allows for a second-order utopia, that is, a description of an ideal “meta-process of ongoing utopian visioning that comprises multiple, pluralistic utopias” (Böker, 2017, p. 94). In effect, a critical theory can formulate a vision of how the historical process by which agents are motivated by utopian conceptions of society could ideally unfold if it were as free as possible from the distorting influences of oppression.

It is unlikely, however, that such a proceduralized meta-utopia—that is, a reference to an ideally democratic process by which we will continuously revise our first-order utopian desires—will exhaust what a critical theory means by emancipation. Böker is right to argue that utopias are subject to historical change and that ideal theories should not promote a static conception of the ideal that would block such changes. What is missing in her account, however, is an explanation of why such changes amount to anything more than a succession of various ideals, where later ideals are not in principle more justified than earlier ones.

A critical theory that is supported by a model of immanent critique can fill this gap: When social movements lead to a revision of the conceptual commitments of a given society, this will typically be driven by a diagnosis of social contradictions that is in turn grounded in concrete demands for emancipation that aim to abolish historically specific forms of oppression. The need to formulate particular substantive utopian visions that drive conceptual change is thus rooted in the concrete shortcomings of a given society.

While we can never be assured that any such particular utopian vision will be ultimately justified, this does not mean that we are not justified to let our political agency be guided by such substantive utopian conceptions. They can be justified in comparison to other, currently available forms of utopia, first by their adequacy to the specific social contradictions to which they respond. The theory of social contradictions that I have sketched therefore adds a materialist dimension: The emergence of particular conceptions of an ideal society is a matter not of random changes but of rational responses to social irrationality.

They can also be justified according to a more proceduralist standard, by the degree to which they have emerged through a process that has overcome or avoided epistemic distortions by oppression. Thus, even if social movements themselves do not explicitly embrace the idea of a democratic meta-utopia to which Böker refers, we can understand the normative force of their specific first-level utopias by comparing their emergence to what would have happened in a more ideal society.

What does this mean for ideal theory? If we accept this line of argument, we can no longer assume that it is the task of political theory to first select one particular normative concept (e.g., justice) as basic, then conceptualize an ideally just society on that basis, and finally develop a conception of justice by establishing which principles would regulate such a society. But abandoning this assumption (and with it, the precise technical sense of “ideal theory” as used by Rawls) does not rule out a constructive role for ideal theorizing in a broader sense in critical theory—this is the project of seeing the development of a conception of an ideally just society as necessary for political guidance We can say something specific and informative about which concept, for example, of justice or freedom, it would be most rational to reason with at any given historical juncture if we examine which utopian vision social movements have reason to endorse in response to the social contradictions of their society. We can make headway in answering that question if we also consider which conceptions of an ideal society they would endorse if the relevant process of collective deliberation and knowledge production were not distorted by social oppression and domination.

In this article, I have argued that utopian thinking ought to be a part of critical social theory, despite the strong anti-utopian currents among its most important historical representatives. I have examined epistemic, imaginative, and conceptual arguments that object to the possibility of using a vision of what an ideal society would look like to guide critical theory, and I have argued that only a moderate version of the conceptual argument can be defended. This argument rules out the idea that we can use ideal theory to ultimately settle the question of which normative ideals and which conceptions of, for example, justice ought to inform political action. But it does not rule out the idea that conceptions of an ideal society—as they emerge from social movements that respond to contradictions in their society—can play a role in determining the meaning, at any given historical juncture, of central political concepts that we can use to orient our political agency. Nor does it rule out the development of a conception of a meta-utopia that refers to the democratic processes by which oppressed groups can collectively give content to concrete utopian visions.

I received no outside funding for the research leading to this article.

There are no conflicts of interest.