Is political realism barren?: normativity and story-telling

ABSTRACT Political realism has been criticised for its methodological claims about normativity and for its criticisms of moralism. Realists themselves should be more concerned that for all its methodological wrangling, realists have struggled to produce much positive theorising, rendering realism barren. I argue that realism, in both its liberal and radical forms, is currently barren in the sense of being unproductive, and show how the two dominant forms of realism are barren in different ways. Bernard Williams’s liberal realism exclusively derives its normative resources from existing political practices, values, and institutions, which leads it to replicate the status quo and precludes external criticism. Radical realism, represented here by Raymond Geuss, is barren because the fear of succumbing to moralism or ‘normativism’ leads radical realists to wrongly abjure normativity altogether. Having shown that both forms of realism are barren I argue that radical realists should adopt a method of political theory as story-telling as a way to avoid both moralism and barrenness. While realists are right to be wary of normative prescription, I argue that realists need to tell a story about which political transformations are desirable and why or else risk succumbing barrenness. I give a brief illustration of story-telling in practice then address some objections to story-telling as a productive method.


Introduction
Political realism irritates a lot of political theorists. Political theory also irritates a lot of political realists, but more on that later. This mutual animosity has its roots in a disagreement about normativity and the relationship between normativity and political theory. This disagreement manifests in different ways. One source of irritation towards realists is reflected in Lorna Finlayson's question: 'With Radicals Like These, Who Needs Conservatives?' (Finlayson, 2015). The irritation here lies in the tension between realist claims to methodological unorthodoxy, and their often orthodox substantive conclusions about political legitimacy, human behaviour, the limits of the state, and the fixity of politics. Realism's most significant and controversial claim is that political theorists should draw on a 'distinctively political normativity' when theorising (Jubb, 2019, p. 360). This form of normativity is contrasted with the moral or ethical normativity that realists decry in much political theory since Rawls. This points to the second major source of irritation towards realists: the target (usually Rawls) is misidentified, uncharitably mischaracterised, and his sins overstated. The faults that realists find in mainstream political theory are less pronounced and less widespread than they claim, and so the force of their criticism is less significant than they imagine. This irritation leads to one final grouse (I'm sure fervent anti-realists could find additional things to vex themselves over): even if the realist critique holds, realists seem unable to turn their methodological insurgency into positive theorising. 1 Or, to paraphrase E.H. Carr, realism is barren. Realism's barrenness corresponds to the three irritations already mentioned: its substantive conclusions are bleak, its critique is stillborn, and its attempts to move beyond a technical dispute about normativity are both theoretically and practically unproductive. It is the final irritation about productiveness that I address here, in order to show why realism in its current state is indeed barren, why realism should avoid barrenness, and how it might overcome that barrenness.
Realism, as everyone writing on the subject notes, is a broad church. This diversity might suggest incoherence to non-adherents, but a loose alliance of realists has been held together by an opposition to what they term 'moralism'. Within this alliance are two factions: the liberal realists inspired by Bernard Williams and the radical realists inspired by Raymond Geuss. 2 Despite differences in historical touchstones and differences in substance, both are united by a rejection of 'ethics first' or 'applied ethics' approaches to political theory (represented for both factions by Rawls). Instead, realists look to articulate and defend forms of political normativity that are 'autonomous' from ethics. Realism therefore rejects the equivalence of 'ethics' and 'normativity'. As a consequence, realists reject the view that all political questions of what ought to be done will ultimately rest on one moral claim or another. So when Kant demands that 'all politics must bend its knee before right', realists are apt to respond that the right must sometimes genuflect to politics. While there are other common realist themes of disagreement, conflict, and contingency, it is some version of the autonomy of the political claim which unifies the realist coalition.
The idea of political normativity is both a rallying-cry and a stumbling block for realists. This is, in part, because the autonomy of the political claim requires some tricky arguments about the purposes of political philosophy, the nature of politics, and the nature of normativity. These arguments are unlikely to be decisive or to persuade self-identified moralists who share Isaiah Berlin's view that political philosophy is 'but ethics applied to society' (Berlin, 1990, p. 2). What follows here is an argument internal to political realism and I have no direct ambition to initiate those who sympathise with Berlin. Instead, I hope to shed a little light on how realists should think about normativity, their methods, and political theory.
Just as the communitarian critique of liberalism came unstuck when its proponents looked for substantive and attractive alternatives to liberalism, so too has realism floundered for the best part of a decade in what one of its advocates, Enzo Rossi, calls a methodenstreit, or a dispute about methods (Rossi, 2016, p. 411). Derek Edyvane, in a scathing attack on realism's contribution to political theory, asks what we would lose by forgetting about realism altogether. His answer is that 'we lose, I suppose, the cottage industry that has built up around realism in recent years, but which is anyway descending by now into navel-gazing irrelevance' (Edyvane, 2020, p. 285). Even if realists judge this assessment to be unfair (and Edyvane goes on to suggest some ways in which realism might still have value), it should still be a cause for concern. Philosophers, even comparatively practically-minded political philosophers, cannot shy away from methodological arguments, but realists should be worried that for all their wrangling there is not much substantive meat on realism's methodological bones. The barrenness of realism captures this worry. I argue that realism's barrenness is caused by its refusal or inability to do normative political theory, but that by embracing political theory as story-telling realism might yet overcome barrenness. 3 I make three connected arguments in the following sections. The first argument is that there are two strands of realism which should not be considered together. I argue that Williams-inspired liberal realism is barren because its account of normativity is derived from, limited by, and legitimises a contingent status quo. Geuss-inspired radical realism also risks barrenness because it wrongly denies itself the resources to engage in the full range of normative theorising. I argue that radical realism should, by its own lights, be committed to doing normative political theory and outline why normative prescription is not incompatible with Geuss's rejection of 'normativism'. In the positive section of the argument, I suggest that story-telling offers realists a source of normativity which avoids both moralism and barrenness.

Two realisms, two types of barrenness
Before making that positive argument, I first address the charge: is realism barren? There are many reasons that realism's critics have thought it barren, not least that its insight about the distinctiveness of political normativity is false and nothing can spring from the dead land of error. But the idea of barrenness that I use here is to suggest something that is unproductive. An approach or method is unproductive when it fails to produce determinate prescriptions about political action, fails to allow concrete evaluations of political practices and institutions, or fails to offer any means of ordering and weighting political values or courses of action. By productive, then, I mean capable of furnishing anyone engaged in politics (citizens, academics, activists, politicians, protestors, disobedients, anyone with a passing interest in the news) with ways to think about and engage in politics. Why then is realism unproductive? My argument is that realism is unproductive because it is either unable or unwilling to engage in normative theorising.
Substantiating the charge of barrenness is complicated by the diversity of realists. Talking about 'political realism' as if it were a coherent and unified programme misrepresents the variety of realists and disguises some deep schisms. While realism can fracture along several fault lines, the most important of these concerns the status quo. Attitudes towards the status quo split contemporary realists into liberal and radical camps. These two groups are insufficiently well-distinguished in most of the literature, even though everyone writing on realism is well-aware of the division. Although both have common cause against Rawlsian political theory, the two groups diverge on substantive issues about legitimacy, obedience, justice, the state, and revolution. These differences mean that the two cannot be usefully considered together, despite their shared hostility towards moralism. It is also helpful to distinguish these two types of realism because they succumb to different kinds of barrenness. As Benjamin McKean puts it, liberal realism's 'insistence on seeing things as they are can easily curdle into an insistence that things are as they must be, especially because new political possibilities are both risky and difficult to perceive' (McKean, 2016, p. 877). Radical realists, on the other hand, deny themselves the normative resources they need for the political transformations they aspire to. My focus here is primarily on the radical realism of Raymond Geuss, but in order to reckon with the barrenness of Geuss's realism it is helpful to grasp how Williams's liberal realism is differently barren.

The barrenness of liberal realism
Of the two forms of realism, liberal realism is less troubled by talk of normativity. Liberal realist normativity is far removed from the normativity advanced by moralists, but realists such as Robert Jubb, Matt Sleat, and Edward Hall see political theory as essentially normative, and are committed to making normative prescriptions despite their opposition to moralism. The disagreement between liberal and radical realists about the status quo reflects a deeper disagreement about normativity. While both factions agree that political philosophy should avoid the kind of 'ethics first' approach that they identify in Rawls and Nozick, they do not agree about what realist political theory should look like. Liberal realists oppose moralism, but they are happy with, and dedicate their project to, identifying and articulating the features and consequences of a distinctively political normativity.
For liberal realists, realism arises in response to the impossibility of a stable moral consensus and consists principally in re-setting the ambitions of political philosophy to better attune it to 'what is platitudinously politics' (Williams, 2005, p. 13). For Williams, this entails a shift from the primacy of justice in Rawls to a renewed focus on a reimagined and de-moralised conception of legitimacy. Williamsian realism is characterised by its insistence that the moralist ambition of finding universal moral grounds for the legitimacy of liberal states is impossible. Williams's political theory grows out of his scepticism about attempts to systematise (and inevitably simplify) ethics. He claimed that his moral philosophy centred on reminding 'moral philosophers of truths about human life which are very well known to virtually all adult human beings except moral philosophers' (Williams, 2005, p. 13). This kind of common sense-ism is a prevalent strand in liberal realism. Instead of systematising the irreducibly complex, realists should focus on what all sensible people know about politics and prioritise 'order, security, and the conditions of cooperation' as a means of answering 'the first political question'. Answering that question, which Edward Hall has argued is the question of whether there can be politics at all (Hall, 2015), will rely on telling an historical story about what justifies the power of the state to those subject to its power. A similar idea was present in Williams's earlier work, Truth and Truthfulness, where he claimed that we 'have to appeal to a historical story about our situation, about the origins, development, and character of modernity. As with all largescale historical interpretations, we could not properly convince anyone of this story, or of the need for this story rather than another, without telling it' (Williams, 2002, pp. 263-4). Williams's realism, like John Gray's modus vivendi liberalism, tells a story about liberalism as a theory of legitimacy, which contradicts the dominant Rawlsian tale about liberalism as a theory of justice. From Williams's project two stories about liberalism can be discerned: the moralist story which sees liberalism as the political creed that grew out of secular morality, and the liberal realist story which sees liberalism as a way of answering the political question of legitimacy in terms of 'order, security, and the conditions of cooperation'. There are certainly other facets to both stories and more overlap than my binary admits, but whatever the liberal realist story is, it must contain the element of order for it to be distinctive from the moralist story. Realists of all stripes will prefer the realist story of liberalism to the moralist story of liberalism, but it is not clear that those are not the only two stories available. For liberalism's critics, including radical realists, there is a different story of liberalism about imperial conquest, environmental destruction, domination of the weak, the exploitation of the working class, the ossification of social hierarchies, and its imbrication with capitalism. It is not that one of these stories is necessarily better or more veridical than the others, but rather that Williams's account of politics and legitimacy is dependent on his story about liberalism being convincing. But his approach is not to convince us of his story's superiority, but instead to assume that we are convinced and then to see what follows. This assumption precludes liberal realists from envisioning deviations from the status quo because they must insist that the only story that can be told is one derived from the platitudes that realists can find in the history of liberalism. This view is expressed in Williams's slogan 'LEG [legitimacy] + Modernity = Liberalism', which implies that liberalism is uniquely able to tell legitimation stories under the conditions of modernity (Williams, 2005, p. 9). The charge that realism legitimises the status quo can be mistaken for the claim that the status quo is objectionable simply in virtue of being the status quo. But the former methodological judgment does not require the latter normative judgment. It is troubling to think that normative standards should be derived from a contingent present regardless of the normative judgments made about that present. Hobbes's story, which Williams follows, about the need for a powerful state, the capacity of fear to motivate obedience, and the constant threat of disorder and anarchy is crucial to the history of liberalism, but it also prevents liberal realists from finding any normative resources that cannot be found within this story. 4 This limit to liberal realism also indicates where radical realism can find its normative resources. As Rossi describes it, the 'radical approach . . . acquires its normativity by contesting what one may call legitimation stories' (Rossi, 2019, p. 642). I return to this in the final sections of the article.
It is telling that both Williams's liberal realism and Rawls's political liberalism bottom-out in the same appeal to values that are present or latent in liberal society already. For Rawls this is found in his claim that equality, liberty, and reasonableness are 'implicit in the public political culture of a democratic society' (Rawls, 2005, p. 223). Similarly, Williams's account of legitimacy is derived from 'standards internal to politics'. The values, commitments, and practices that one can find in these different sources will depend on where one looks, but both forms of grounding imply certain things about what politics is and what normative resources can be found in politics. When Williams invokes 'standards internal to politics' what he actually invokes are the idealised standards internal to postwar liberal democracy. What are these idealised standards? That those 'now and around here' will not submit to arbitrary political authority, that illiberal regimes would not be acceptable to those subject to their power, and as Michael Freeden notes, that the principles 'necessary to the political sphere . . . emerge from a liberal state' (Freeden, 2018, p. 357). It is not that they do emerge from a liberal state, it is that they can only emerge from a liberal state. If the answer to the question 'how can political power be legitimate?' is derived from the contingent social conditions of power and the beliefs about power produced by history then the status quo will inevitably fall out of that method because history has led us here (wherever that might be). The risk of attending too much to reality is that theory becomes trapped by a resignation to how things already are. This cannot help but lead to the theoretical reproduction of existing political, economic, and social relationships and structures. Resignation cannot yield genuine normative prescription or orientation because the foundations of those prescriptions and orientations must be derived from something other than the political status quo.
One response to my portrayal of liberal realism is that it underestimates the contingent character of Williams's analysis. After all, the internal standards 'now and around here' are liable to transform into 'then and around there', where different standards will apply. On this view, Williams's affirmation of the liberal status quo is not a necessary consequence of his method, but rather it is a result of applying his method to current contingencies. If the contingencies collapse then so too does his affirmation. Stears and Honig pursue something like this line of argument in contrasting Geuss's 'thoroughgoing pessimism' with Williams's desire to 'maintain a role for aspiration in his politics' (Honig & Stears, 2011, p. 189). They point to Williams's defence of abandoning foundationalist (hence moralist) approaches in support of their interpretation. Williams writes that contextualism does not foreclose social criticism because it instead 'provides a possibility of deploying some parts of it against others, and of reinterpreting what is ethically significant, so as to give a critique of existing institutions, conceptions, prejudices, and powers' (Williams, 2005, p. 37). But this receptiveness to social criticism is only present within the status quo affirming framework of liberalism. This is evident in the quotation Honig and Stears cite. Social criticism can only 'deploy some parts of it against others', but it can never deploy something alien against the whole. Just as Rawls's conception of reasonableness excludes everyone except for liberals, so too does Williams's method preclude criticism beyond liberal criticism. There can be no wholesale rejection of the contingent internal standards because no standards can be derived from the present contingent circumstances which would justify it. Liberal realism is parasitic on the status quo for its normative standards, but like all parasites it needs its host to survive. 5 In short, liberal realism entails a legitimisation of the status quo and is barren in its inability to theorise beyond the limits of currently existing practices, institutions, and values. If, as Glen Newey claimed, 'the major project of modern liberalism is to use ethics to contain the political', then the major project of liberal realism is to use the status quo to contain the political (Newey, 1998). In the interest of allowing some space to make a positive argument this critique of liberal realism is all too brief, but if the essence of the criticism is right then this would go some way to explaining why liberal realism is so bereft of the politically normative prescriptions that it hopes to produce. 6

The barrenness of radical realism
Raymond Geuss's radical realism seems to offer an antidote to the status quo affirming barrenness of liberal realism. Where liberal realism rests on standards internal to politics 'now and around here', Geuss's critical focus demands an evaluation of these standards with the aim of altering them. Where Williams seeks to manage the actual, Geuss outlines the possibilities of its sundering. Where liberal realists ask questions that arise from politics as we find it, and arrive at answers that draw from that same well (or 'tepid slimy puddle' in Geuss's acid phrase), Geuss has the opposite ambition. Geuss wants to change the question, to alter the terms of the debate, to reveal new and ignored facets, and to bring all these re-evaluations to bear on the practice of politics and political theory. But he also aims to do all of this without moralism, or what he elsewhere calls 'normativism' (Geuss, 2016a). Radical realism's relationship to normativity is more troubled than liberal realism's relationship to normativity because Geuss seems to eschew normativity altogether, rather than advancing a distinctively political normativity. Radical realism faces a problem: if the aim is to alter political standards (and so it cannot merely reflect pre-existing standards, as liberal realism does) and realism cannot draw on the ethics-first prescriptions of moralism, how can radical realism be normative at all? I will attempt to answer that question in the next section, but first it needs to be established that radical realists are in the business of making normative judgments and prescriptions at all. One obvious rejoinder to my question is that radical realists can and should abandon normative talk altogether. In certain moods Geuss defends the overwhelmingly negative nature of his writing as 'getting rid of pernicious illusions' (Geuss, 2016a, p. 62). On this view, radical realism is an essentially destructive enterprise which looks to free us from the moralistic wishful thinking that blights political theory. Implied in this programme are two features of radical realism: antinormativism and anti-positivity. These features are related, but it is useful to separate them for the sake of clarity and because Geuss responds to them in different ways.
Geuss is opposed to what he calls 'normativism', the view that there is a 'fully autonomous, closed, fully rationally grounded doctrine that prescribes for us in all important cases how we ought to act' (Geuss, 2016b, p. 17). This is a wider rejection than the liberal realist rejection of moralism because it entails a rejection not just of the priority of morality to politics, but also a rejection of ethical systems altogether. Although Geuss grudgingly admits that we might acknowledge that a 'certain "normative dimension" runs through all of human life', you 'cannot isolate it . . . and make it . . . the object(s) of a coherent, unitary, separate rational study'. This leads Geuss to conclude that 'the "normative turn" in political philosophy which started with Rawls is a mistake. It is a turn toward something that does not exist, the purely normative' (Geuss, 2016b, p. 17). Instead, Geuss claims 'that it does not matter that there is no unitary and separate "normative ethics" because political philosophy is always in itself already praxis-orienting and potentially interventive. Even political philosophy that takes itself to be simply descriptive and enjoins abstention from action is already intervening, because taking any position at all is intervening' (Geuss, 2016b, p. 18). But Geuss's description here sidesteps the problem of whether political philosophy is normative because the distinction between 'interventive' and 'normative' does not resolve the question at hand. Many kinds of intervention, and frequently the most important kinds of political intervention, are normative interventions. Some examples: A campaigner makes a normative intervention that the underfunding of the criminal justice system has detrimental effects for all involved; a politician makes a normative intervention against a government policy on trade that would further worsen the material conditions of those in the developing world; a political philosopher makes a normative intervention against imminent military intervention in a distant country. None of these interventions, despite being normative, are disbarred by Geuss's opposition to normativism. While theorists should avoid normativism as a systematising and fully comprehensive approach, one can still be a radical realist and engage in normative theorising.
The second feature of Geuss's programme is anti-positivity. By this I mean a rejection of the demand that political theory should provide positive or 'constructive' criticism of whatever practice, theory, ideology, or argument is being subjected to criticism. This demand requires that the critic not only tell us why something is bad but also what better alternatives could be offered in its stead. Geuss is alive to this demand and strongly rejects the view that 'you can't criticise the police system, the system of labour law, the organisation of the health services, etc., unless you have a completely elaborated, positive alternative to propose' (Geuss, 2008, p. 96). To accept this demand is to 'allow the existing social formation to dictate the terms on which it can be criticised, and to allow it to impose a theoretically unwarranted burden of positive proof on any potential critic' (Geuss, 2008, p. 96). He elaborates on this defence in A World Without Why. The worry about a purely negative form of theorising is that 'it is a concomitant of an anarchic abdication of responsibility on the part of the critic' (Geuss, 2014, p. 68). But that worry is mostly overridden because while, 'appeals to the need for "constructive" criticism can in principle represent a (generally laudable) attempt to remind those involved in some evaluation of human action of the need to remain aware of a kind of internal demand under which such criticism operates, namely of the need to keep Tscheryschevsky's (and later Lenin's) central question "What is to be done?" firmly in mind; in fact, however, the demand for "constructive criticism" in general functions as a repressive attempt to shift the onus probandi and divert attention from the possibility of radical criticism.' (Geuss, 2014, p. 90) Geuss's objection to the demand for constructive criticism is not then a principled objection, but is rather an objection to how this demand tends to function. But this means that it is, in principle, possible to offer constructive and radical normative criticism, even if the demand to do so tends to distract from genuine radical criticism. Would it not be a good thing to have both criticism and a positive alternative? And further, when Lenin's question is 'firmly in mind', does that not itself produce the demand for an alternative? Geuss's anti-positivity does not seem to preclude normative theorising. Geuss says as much when he acknowledges that there is not a 'special problem with the use of evaluative language'. A rejection of normativism 'in no way means we can't say it is better to drink this glass of clean water than that glass of water from downstream of the huge petrochemical complex, or that we would be better advised to support the development of solar power than to permit fracking' (Geuss, 2015, p. 66). Even though Geuss presents these examples as self-evident, they are still normative evaluations. As normative evaluations they stand in need of normative argument and justification because the question of 'what is to be done?' necessarily entails the question 'what should be done?' and that question ends up with the permanent question of 'why should that be done and not something else?'. Geuss attempts to resolve this difficulty through his advocacy of diagnosis and orientation over prescription. Rather than looking 'for general principles, one should rather engage in the diagnosis of one's current context . . . What are the pressing problems, what are the most acute dangers of the present?' (Geuss, 2012, p. 10). But diagnosis and orientation rest on evaluative criteria of some sort, some of which will be normative. It is only apparent that a problem is worthy of attention and remedy if we think that the problem is a) not ineluctable and b) objectionable in terms of some evaluative criteria. The alternative is mere description, which Geuss also rejects: 'we do not simply want to understand how the apartheid system worked in South Africa in the 1970s; we wish to judge it as being better or worse (in some respect) than other systems' (Geuss, 2008, p. 39). Geuss is right here, but if radical realists are comfortable with evaluation, then I cannot see how they are not also comfortable (in principle) with normative evaluation, even while rejecting normativism.
Despite Geuss's vilification of normativism, his roots in critical theory make his grudging acceptance of normativity unsurprising. Iris Marion Young defines critical theory as, 'a normative reflection that is historically and socially contextualized' but which seeks to 'evaluate the given in normative terms' (Young, 1990, p. 5). Prinz and Rossi make a similar claim in their defence of realism as 'ideology critique' when they say that 'ideology critique is normative but not normativist'. The 'normative element of ideology critique is already present therein through the concern with the inner normativity of the context in question' (Prinz & Rossi, 2017, p. 361). Critical theory and radical realism both take their methodological cues from Marx. Following Marx, prescribing the new is a task for social movements and revolutions to decide, not for theorists to preordain. This is not a methodological point about what theorists ought to do, but rather the much more practical point that any attempt to define the contours and boundaries of a way of organising and imagining society which hasn't yet been realised will be a fruitless exercise. This points to a crucial element of Geuss's realism: normativity is created by politics, rather than constraining politics. This is the positive re-telling of Geuss's line that 'ethics is usually dead politics' (Geuss, 2008, p. 48). My belaboured point here is that despite appearances, Geuss's radical realism does not eschew normativity altogether. Radical realism, by its own lights, is not opposed to normative theorising, but it should also be clear that it does not succumb to the status quo objection I levelled against liberal realism. Why then is radical realism barren?
The first reason is contingent: radical realists have not, to date, offered any substantive or sustained normative prescriptions, or in Geuss's terms, evaluative interventions. Geuss refuses to play any version of the normativist game, which he considers 'morally and politically reprehensible' (Geuss, 2016c: ix). Elsewhere, Geuss admits that 'political theory and philosophy are connected to practical interventions' but 'that does not mean that one must expound them explicitly in every possible discussion' (Geuss, 2008, p. 95). Geuss is right and one can do diagnostic or orientating normative theorising without engaging in prescription, but if my argument that radical realists should not be afraid of normativity holds, even if they are rightly wary of normativism, then radical realists have no principled objection to normative theorising and should be willing to engage in normative prescription. This is the point at which the charge of barrenness resurfaces: realists struggle to develop and defend ways in which the world could be made better. Put more philosophically, there is no meliorism in radical realism.
The second reason is that Geuss's radical realism fails to live up to Marx's ambition to transform the world. This transformational aspect is all but absent in Geuss's work. I cannot identify passages of any length where Geuss or any self-identified radical realist sets out what political transformations are desirable, why they are desirable, or how these transformations could be achieved in practice. If it is part of your methodological schema that the transformation of politics is desirable, then talking about what ought to be done is inescapable. That does not mean repeating the sins of 'normativism', but it does mean proposing ways of existing and relating to others, ways of thinking about politics, ways of thinking about and relating to the state, and ways of opposing domination and injustice that move us beyond where we currently stand. My plea here is for methodological pluralism. Even if Geuss is himself unwilling to engage in normative prescription, radical realism should not join him in that refusal. If one of the purposes of political theory is not just to rid us of 'pernicious illusions', but also to mitigate disillusionment then prescription cannot be avoided. Without normative judgments of some sort, how are we supposed to assess whether the cure is better or worse than the disease, or whether the cause is deserving of personal sacrifice, or whether there is any hope of improving the conditions of the dispossessed? Answering these questions at the very least permits, if not demands, prescriptive answers. No one denies the validity of critical methods, but unless we have a rough idea of what the alternatives could look like, it is hard to see why the present state of affairs is so unsatisfactory.
It is necessary to make one small note before moving onto the positive section of the article. I have taken Geuss and his followers as the exclusive representatives of radical realism. This is, undeniably, a narrow way of interpreting radical realism. There are other thinkers and schools of thought, including the agonism of Chantal Mouffe and Bonnie Honig or the pragmatism of Jeffrey Stout, which share realism's hostility to moralism but nevertheless envisage the kinds of transformation which are absent in Geuss. If this is right, then my criticism only applies to Geuss and has minimal wider relevance to political theory, which has long been practising radical realism under other names. While I wish I could do justice to those thinkers, who in some respects do meet the joint criteria of 'radical realism', none of them are primarily motivated by the rejection of 'ethics first' approaches, which I claim is central to realism as a distinctive school of thought in normative political theory. Further, Geuss's brand of radical realism, for better or worse, determines the dominant understanding of radical realism in the debate. Realism, as I claimed above, is so amorphous that it is almost impossible to identify its boundaries in a satisfactory way. But the argument I make here requires the imposition of some boundaries in order to proceed and this is the unsatisfactory place I have drawn them. But that is not to say that the wider point does not stand: if one is both a realist and a radical one may find more useful resources in agonism, pragmatism, or critical theory than one can find in contemporary political realism and my argument here might give further reasons to think that the case. 7

Sources of normativity: story-telling and political theory
Having argued that radical realists should engage in normative theory, a problem persists: if the aim is to alter political standards and realism cannot draw on the ethics-first prescriptions of moralism, how can radical realism say anything normative at all? In the spirit of 'doing realist theory', I have a methodological suggestion and an example to illustrate it. Realists are right that political normativity cannot draw exclusively or primarily from the universalising, non-contextual, and non-political well of moral philosophy. But what realists often neglect is a methodological commitment to storytelling. My suggestion is that realists should turn to story-telling as a way to make normative judgments, prescriptions, and evaluations without resorting to moral theory building and excessive abstraction. 8 Recently, Prinz and Raekstad have argued for the use of genealogy as a 'way of bridging the empirical-normative divide in order to say something concrete about what real political agents should value and be guided by' (Prinz & Raekstad, 2020, p. 11). This function of genealogy shares the kind of normative ambition I argued that radical realism should embrace in the previous section. Prinz and Raekstad identify three forms of genealogy: Geuss's 'genealogy as critique', Quentin Skinner's 'genealogy as conceptual mapping and reconstruction', and Williams's 'imaginary genealogy' as a thought-experiment. They argue that each form can serve a different purpose, but only describe Williams's 'imaginary genealogy' as a story. While there are affinities between genealogical approaches and storytelling, one important difference is that story-telling is necessarily forwardlooking. Whereas genealogies look to unearth and expose, story-telling serves a different purpose. Story-telling tries to connect the past with the present and the future by encouraging revised judgments about the limits of political possibility. The risk of genealogy without story-telling is that it identifies a formal method for the purely negative theorising that I argued Geuss should avoid. If successful, political theory as story-telling serves both the orientating and diagnostic functions identified as desirable by Geuss, but goes further by charting a course between purely negative genealogies and vindicatory genealogies. Geuss's mode of theorising is deliberately corrosive. It intends to corrode commitments to liberalism, or to corrode dominant legitimation stories about the state or about private property. As I argued in the previous section, this is a legitimate realist method to pursue, but it is not the only method one can pursue. Where genealogy is corrosive, storytelling is generative and creative. Again, this does not entail a rejection of the view that purely negative theory cannot be generative, but merely to note that neither realists nor radicals should limit themselves to purely negative theory.
The second source for this strategy is Richard Rorty. While Rorty is occasionally mentioned as a kind of realist (or at least realist-adjacent) he does not fit neatly into either realist faction. His jokey endorsement of 'postmodern bourgeois liberalism' alienates him from the radical realists, while his disagreements with Williams about truth alienate him from liberal realists. 9 Geuss both admired and despaired of Rorty, reflecting on their relationship that Rorty was 'both extraordinarily perceptive and, at times, intensely irritating' (Geuss, 2010, p. 100). But it is Rorty's understanding of how people actually come to act and think that radical realists should turn to. If 'you find yourself a slave', Rorty counsels in his Tanner Lectures, 'do not accept your masters' descriptions of the real; do not work within the boundaries of their moral universe. Instead, try to invent a reality of your own by selecting aspects of the world that lend themselves to the support of your judgment of the worthwhile life' (Rorty, 1990, p. 21, emphasis in the original). While talk of the 'worthwhile life' sounds ethical rather than distinctively political, it is not by accident that he chooses the essentially political slave-master relationship to make his point. Rorty, who eschewed foundationalist philosophical projects (like Williams and Geuss), was not interested in whether your judgment of the worthwhile life conformed to reason, but rather whether your inventions allowed you to make such judgments. Rorty's creative account of storytelling chimes with the transformational ambitions of radical realism. Rorty's call to invent a reality against the one that binds you is echoed by Geuss when he writes that philosophy should not try 'to sharpen, restrict, and discipline the language but to enrich it, change it, reconfigure it so that it is able to . . . permit us to see the point of new forms of human experience' (Geuss, 2014, p. 88). This builds on the status quo division between liberal realists and radical realists by indicating how liberal realist and radical realist story-telling will differ. Where liberal realist story-telling aims to 'restrict and discipline' what politics is and can be, radical realist story-telling aims to enrich, change, and reconfigure the stories we tell ourselves about why things are the way they are, and why they might not need to be that way.
It is worth offering an example at this point to illustrate my suggestion before considering some objections to the idea of political theory as storytelling. The Paris Commune, which existed for 72 days in 1871, transformed the city into a self-governing socialist association which rejected the authority of the French state, and which sought to alter the economic, social, and political relationships of Parisians and the Communards. This example is pertinent for realists because the Commune posed a direct challenge to the legitimation story of the newly founded French Third Republic. Given the priority accorded to legitimacy by realists, the story of the Commune offers a way to ask whether the stories we tell ourselves about the state and its authority are stories that we can make sense of, or which can help us to make sense of us. The Commune, while fleeting, had a deep influence on Marx's thought. For Marx 'the circumstance of the Commune proved enormously generative: creating ways of framing or reading or taking part in the moment of its intervention that then alter the frame of perception and open up the field of the possible' (Ross, 2016, p. 77). In particular, the Commune's transformation of exploitative wage labour into 'emancipated' labour, which served the needs of Parisians and not the demands of capital, led to Marx's revised view of history as the history of the relations of production (Ross, 2016, p. 81). As Kristin Ross argues, the Commune led Marx to believe that, 'actions create dreams and not the reverse' (Ross, 2016, p. 25).
There are many aspects of the experience of the Commune that could furnish realists with normative resources: Communards called (in their 72 days of resistance) for 'enrolling women on more equal terms in various groups such as trade unions; the call for equal primary and secondary schooling for boys and girls; a decree giving equal pay to women and men teachers' (Schulkind, 1985, p. 137). Needless to say, these were radical demands that grew out of political practice and were buttressed by the solidarity of the Commune. They were also normative prescriptions (e.g. 'pay should be equal regardless of gender') underpinned by normative commitments. The real commitments held by real people and briefly realised indicate that political reality suddenly and momentarily became pliable. The Commune is often mythologised, but its existence showed that actual political practices and associations which deviated dramatically from what anyone believed possible can be realised. The opening up of this possibility can furnish contemporary realists with non-moralist and non-status quo affirming normative precepts. While political moralists look for principles, realists should first look for stories. How did the Commune realise the socialist values of the Communards? What commitments did the Communards express and why were these expressions capable of securing the support of many Parisians? What motivated the political actors to act and why did they act in the way that they did? These are descriptive and empirical questions, but they demand normative descriptions, which might in turn permit If realists want to identify 'what really does move human beings to act in given circumstances' (Geuss, 2008, p. 9) then realists can look to stories like the Paris Commune to discover how new norms of human interaction developed, why they were desired, what needs they fulfilled, and what prescriptions about political action we might derive from them now. If radical realists hope for political transformation, then they must be able to tell a story about which transformations are desirable and why, and from that story develop normative prescriptions which are sensitive to the realities of politics. Telling these alternative stories can also undermine the dominant stories which legitimise existing forms of power and so enable realists to follow Rorty and 'invent a reality of your own'. This story is, inevitably, partial and incomplete and it emphasises certain aspects and neglects others. Any aspiration that a story could be all-encompassing or unified is misplaced. There should be no expectation in telling stories that they form a unified and harmonious narrative that places all of the relevant pieces in exact relation to one another. It is possible that one story can contain incompatible normative resources. The story of the Commune might suggest how and why the authority of the state should be resisted and also how such resistance is likely to fail. That realist theory is messier, uglier, and more discordant than moralism is both predictable and nonregrettable; so be it.
There are three challenges to political theory as story-telling that I now address: how is story-telling normative, how is it realist, and what nonarbitrary and non-moralist criteria can guide the choice of story? 10 In the previous section on Geuss's relationship to normativity I argued that radical realists can and should engage in all forms of normative theorising, including prescription. Story-telling is normatively orientating because it provides us with a better sense of the options available and an account of why some of those options are more desirable than others. But it is less obvious that story-telling is normatively prescriptive. This impression depends on what makes a theoretical method normatively prescriptive. The standard way of thinking about prescription is in terms of action-guidance. A theory is normatively prescriptive if it tells some relevant agent how to act in a certain situation. How does telling a story about what happened in the Paris Commune in 1871 tell any agent how to act in any situation now? Storytelling can be action-guiding in two ways. Firstly, it can allow a division of labour between the teller and the reader. Simply by the story being told, normatively justified forms of action can be inferred. Pulling down the statue of Napoleon in the Place Vendôme is not only an historical event, but also a precedent that might be followed. The motivations and aims of those who toppled the statue, the consequences of the act, and the meaning and significance attached to the action might all allow a reader to infer that such acts of public destruction are not only permitted, but are also normatively desirable. This might seem like thin prescriptive gruel because all the responsibility lies with the reader. But this is no different to any theoretical prescription. No one is compelled upon reading a prescription to follow it regardless of how determinate or exacting it is. Just as a reader might be persuaded by the content of an analytical argument to accept the conclusion, so too might a reader infer a conclusion from a story. Second, story-telling does not have to operate alone. I could tell the story of the Communards and also draw out the normative commitments they held, rationally reconstruct their values, and translate those reconstructions directly into prescriptions. While these are likely to be hypothetical prescriptions ('If one wishes to challenge the legitimation story of the state, one should . . . '), hypothetical prescriptions are still prescriptions. The risk here is that my account of storytelling looks less and less realist as it looks more prescriptive. This worry is addressed in the following two challenges.
What, then, is distinctively realist about political theory as story-telling? Bluntly: nothing. There can be moralist story-telling and realist story-telling and it is not a method that is unique to realists. Just as there can be moralist conceptual analysis and realist conceptual analysis, so too can story-telling be used to support theories committed to the continuity of ethics and politics or conclusions committed to their discontinuity. But like any story, what matters is how you tell it. Having said that, realists are likely to be drawn to real history rather than the 'imaginary genealogy' that Prinz and Raekstad identify with Williams. Realist story-telling will be unlikely to heed Rorty's advice to meet cold-hearted libertarians not with rational argument but with the 'sob stories about what happens to the poor in nonredistributivist societies' (Rorty, 1998, p. 121). The risk of imaginary story-telling is that such stories are less likely to address what actually moves people to act because no real person has been moved to act. In the same way that evolutionary psychology is plagued by the wholly-justified criticism that it postulates untestable 'just-so' stories, so too is imaginary story-telling prone to the kinds of wishful thinking that realists see in moralism. Some historical veracity might then be useful for realists, but little depends on this point because veracity is less important than whether the story is productive in terms of its orientating, diagnostic, and prescriptive effects.
The final and crucial challenge to story-telling is that I have not specified a criterion of choice for which stories to tell. The absence of criteria means that the choice of story is arbitrary because any story could be told to make any point, so story-telling is covertly moralist (or perhaps just rationalising) because the teller will choose stories which conform to their pre-political moral commitments. After all, I chose the Paris Commune as my illustrative story and not the story of Marshal Pétain or the story of the Restoration of the Stuart monarchy because the Commune tells a story about resistance to the state which I am sympathetic to. This problem then begins to resemble the charge of 'cryptonormativism' levelled against Foucault's genealogical method by Habermas (Habermas, 1985, p. 284). Rather than honestly announcing my moral commitments at the front door, I instead smuggle them in the back. There are two responses to this challenge. The first is that realists can also find normativity in the stories of political associations and movements which are repugnant. The capacity of nationalist politicians to attract millions of voters or of misogynistic online communities to create a brotherhood of imaginary persecution are also possible sites of norm creation. That radical realists find these causes repugnant does not mean that nothing can be gained by seeing how people come to relate to one another in new ways to take action. Solidarity among bigots is just as real as honour among thieves. If realists care about theorising how and why people act then that enquiry cannot be limited to only those stories which lend support to a theorist's predetermined conclusions. Stories can be parables or they can be cautionary tales and nothing in what I have said suggests that realists must choose only one kind of story. Good realist story-telling will not necessarily be morally-edifying nor will it always end in success. After all, the story of the Communards ends in their bloody defeat. Stories, like history, cannot prove anything and it would be folly to try to use them to that end. The shorter response to the challenge is that radical realists should again follow Rorty and 'claim no authority for our premises save the assent we hope they will gain from our audience' (Rorty, 1999, p. 173). The choice of the story can only be justified in terms of how persuasive and productive it is to those who encounter it. It is worth pointing out that such a method is only as underdetermined as any moralist method. Peter Singer and Robert Nozick can both be moralists, albeit of different kinds, and share few methodological commitments or substantive conclusions. It should not be surprising that realism will contain similar multitudes. So when I claim that story-telling can furnish realists with resources to specify which political transformations are desirable and why, the authority of 'why' rests entirely with the assent of the audience. There is one criteria when reading theory, which is: are you persuaded?

Conclusion
I have argued that political realism, both in its liberal and radical forms, is barren because it is unproductive. Liberal realism is barren because its normativity can only legitimise the status quo, while radical realism is barren because its refusal to engage in normative theorising forestalls its transformational ambitions. Radical realists must be willing to venture out into the parlous waters of which transformations are desirable and why. I argued that story-telling can furnish realists with a non-moralist method to theorise these transformations, using the Paris Commune as an illustration of this method. I then addressed three challenges to story-telling as a method for realists. Realism always runs the risk of becoming barren by attending too much to reality, but the imaginative alternatives to reality presented through stories can enrich that reality and provide realists with the normative resources they need. Notes 7. Thanks to an anonymous reviewer for raising this point and suggesting some thinkers I hadn't encountered before. 8. There is scant research on story-telling as a method in political theory, with the exceptions of Abbott (1991) and, more recently, Stevens (2019). 9. See Bacon and Rutherford, 'Rorty, Habermas, and Radical Social Criticism' (2021) for an account of Rorty as a more politically radical thinker than his adherence to liberalism suggests. 10. Thanks to the editors for raising these challenges.