In the recent debate on political normativity in political philosophy, two positions have emerged among political realists. According to the first view, political normativity is understood as orthogonal to moral normativity, and moral considerations do not figure in the reasons given in support of a political principle or a course of action in the political domain. Instead, theorists in this camp have been drawing on instrumental, functional or epistemic normativity in theorizing political normativity. According to the second view, moral norms and political norms are not dichotomous in this sense, as moral considerations may figure in the justification of a political principle or theory. The distinctness rather has to do with how moral norms and prescriptions are ‘filtered through’ the realities of politics such that they are altered by politics’ constitutive features (Sleat 2022; Jubb 2019; Hall 2017). While the former ‘non-moral’ view of political normativity has been severely criticized (Erman and Möller 2015, 2022a, 2022b, 2023a; Maynard and Worsnip 2018), the latter ‘filter view’ has remained underdeveloped and vague. To take the debate on political normativity forward, the aim in this paper is to explore what it could reasonably mean to claim that moral norms are filtered through politics and are aligned with its constitutive features. More specifically, we explore the role of moral norms in political theory. What work do they do, and under what conditions?

Very little in the form of systematic analysis has been made of this, with an exception of Charles Larmore’s attempt to demarcate political from moral philosophy. We take our starting-point in Larmore’s work and make two claims, one critical and one constructive. First, we argue against Larmore’s claim – following political realists – that because political philosophy is concerned with the regulation of basic institutions and legal-political orders, it should primarily focus on political legitimacy rather than justice and always focus on legitimacy before justice. In our view, nothing in the constitutive features of politics supports such a conclusion. Second, we argue that any reasonable political theory relies on foundational principles (even if implicitly), and that the role of moral norms is determined by the fact that since political principles regulate basic institutions and legal-political orders, they cannot fully constitute foundational principles in political theory. Foundational principles (or values) are moral principles (or values) that are independent of a society or polity. As such, they are more basic than political principles in the sense that they put up the normative boundary conditions for such principles. Therefore, any reasonable political theory needs to rely on at least one moral premise.

The paper is structured as follows. We begin by exploring, in broad strokes, what is distinctive of political philosophy, which sets it apart from moral philosophy (I). Thereafter, we defend our first claim through a critical assessment of Larmore’s view that political philosophy should prioritize a focus on political legitimacy over justice (II). This paves the way for defending our second claim about the proper role of moral norms in political theory in the third section (III). The fourth section concludes (IV).

1 What is Distinctive of Political Philosophy?

What seems clear from the ongoing debate on political normativity is that it is a deadlock to understand political normativity as a non-moral (or arbitrarily moral) kind of normativity. But how do moral norms come into play in political normativity? Realists who do not reject moral normativity altogether, claim that moral norms are ‘filtered’ through politics (Sleat 2022; Jubb 2019), “in the sense that the weight, direction and relevance of different considerations would all systematically be altered by politics’ constitutive features” (Jubb 2019: 362). This sounds reasonable. But what does it mean more concretely?

A good starting-point for answering this question is to investigate what is distinctive of political philosophy generally. Few, if any, today would see political philosophy as a subcategory of moral philosophy or simply as ‘applied’ moral philosophy. But what research is done in political philosophy compared to moral philosophy? Historically, political philosophy has implicitly been seen as a part of the broader domain of moral philosophy. But, as stressed by Larmore, properly understood political philosophy differs much more deeply than is typically supposed (Larmore 2020: 3).

According to another view, what is a distinctive feature of political philosophy is that it regulates collective behaviour, such as collective decision-making or problem-solving, whereas moral philosophy regulates individual behaviour (Resnik 1998; Rachels 1999). However, while this is often true, a moral theory may also focus on collective behaviour (e.g. moral actions of a collective agent). Questions related to collective agency in moral philosophy typically intersect with discussions about responsibility, action, and ethics. The discussion centres around groups or collectives (such as corporations, nations, and organized groups) and their capacity to act morally or immorally, intentionally or unintentionally, and to bear moral responsibility for such actions (see, e.g. Gilbert 1992; Pettit and List 2011; Tollefsen 2015).

A better approach for identifying what is distinctive of political philosophy is to look at what kinds of questions or problems that are responded to in political philosophy. On this view, political philosophy is always concerned with matters related to the governance of a political structure, such as the state, or the basic institutions of society – a famous example being the Rawlsian approach (Rawls 1999). This is distinctive of political philosophy since moral philosophy need not be concerned with this. Typical questions addressed in political philosophy include: What is the best form or governance? What is justice in societal arrangements? What are the rights and duties of citizens? What is political freedom? What is the source of political authority?

When investigating matters related to governance and basic societal institutions, what is furthermore characteristic is that there is not only deep disagreement about the right and the good, but also deep disagreement about the very nature of the right and the good. Therefore, one central task for political philosophy is to determine how we can justifiably establish and shape a political order given that there might be reasonable disagreement not only about justice but about the nature of morality itself. Indeed, before the modern era, this was not at all acknowledged, since reasonable disagreement about moral questions where rarely if ever recognized. Instead, the dominant idea was that the use of reason will, eventually, lead to unanimity (Larmore 2020: 6).

2 What Should be the Priority of Political Philosophy?

Given this understanding of the distinctness of political philosophy, should political philosophers focus primarily on political legitimacy rather than justice? Yes, according to Larmore. In his view, legitimacy “ought to be the primary object of political philosophy” because “legitimacy has to do with the conditions under which enforceable rules may be justifiably imposed on the members of a society” (Larmore 2020: 5). Hence, legitimacy should be of primary concern, “justice figuring only derivatively” (2020: 15). In fact, Larmore takes it one step further, arguing that the question of legitimacy must be responded to first, before we can address questions of justice: “Only if a system of political rule is more or less legitimate should it make sense to ask what principles of justice it ought to establish” (2020: 5).

The reason why political philosophy should focus on legitimacy and address questions of legitimacy first, according to Larmore, is thus because the fundamental task for legitimacy is to determine which political order may justifiably impose authoritative rules for handling conflicts between people, including conflicts arising from reasonable disagreement about justice (2020: 5). Legitimacy entails a right on part of a state to exercise coercive power over people, who in turn have an obligation to comply, not insofar as these authoritative rules (laws) happen to be just and not merely because of fear of sanctions, but “because the state is entitled to make laws to govern their conduct in particular” (Larmore 2020: 45). It is one thing to hold on to our conception of justice despite reasonable disagreement, quite another to impose that conception upon those who disagree. In Larmore’s words,

the principles defining the conditions under which the state may in general institute laws and thus exercise coercion constitute, as it were, “second-order” principles of justice, determining whether any particular view of justice may enjoy the force of law amid reasonable disagreement about what justice involves (2020: 48).

It is easy to see Larmore’s kinship with political realism. The priority of legitimacy over justice is indeed in line with the political realist understanding of the task of political philosophy. For realists, the primary goal of a political system is to maintain order and stability, which in their view could only be achieved if there is a legitimate authority in place. In this sense, legitimacy provides the necessary foundation for any further political and moral goals, including the pursuit of justice.

An immediate reaction to this ‘priority view’ is that, rather than addressing legitimacy ‘before’ justice, it seems that we precisely need justice to theorize conditions of legitimacy in the first place. However, this objection is something that Larmore addresses. We might say that the “principles explaining with what right and in what respects a state may exercise coercive power over some particular group of people are principles of justice. After all, a regime that imposes its rule without possessing such a right is commonly said to be acting unjustly” (Larmore 2020: 44). But while it is true that theorizing legitimacy involves justice, he continues, it does not “belong to the sphere of distributive justice” (2020: 44). This is so because prior to questions having to do with the fair distribution of benefits and burdens of social cooperation, is the question of under which conditions any such distributive scheme of rights and duties “may legitimately be instituted through state action” (Larmore 2020: 44). Hence, Larmore argues, while principles of legitimacy can be said to involve an idea of justice, since they determine when coercion and the like may be justly employed, they make principles of distributive justice “subject to the terms they lay down for becoming socially binding” (2020:46).

Here, again, we see the priority view unfold. Principles of legitimacy are seen as more foundational in the sense that they set up the boundary conditions, ‘rules of the game’, as it were, for distributive schemes. However, two observations are in order. First, also principles of distributive justice are dependent on fundamental principles of justice. As Rainer Forst points out in a criticism of mainstream liberal theory, we cannot treat goods to be distributed as if they were pre-existing resources or ‘manna from heaven’, as they do not exist prior to, or apart from, the social practices that generate them. By looking at the social structures that produce inequalities, Forst pushes the question of justice beyond a narrow focus on the allocation of resources to a more comprehensive examination of the conditions necessary for a just society (Forst 2007). Hence, on this view, principles of legitimacy and principles of distributive justice are on par, equally dependent on more fundamental moral principles. And if Forst is right, the priority view misses what is the correct priority: it is not between legitimacy and distributive justice, but between all normative ideals for the political domain and fundamental moral principles.

Second, the priority view runs the risk of restricting political philosophy’s domain unjustifiably, depending on what kind of distribution we are alluding to. A distributive scheme could involve, not only primary goods, but also the distribution of equal influence, the equal acceptance of authority, or the equal right to justification. To this, Larmore would presumably reply that legitimacy must be prioritized since reasonable people disagree on distributive schemes. It is true that reasonable people disagree on some distributive schemes and that without agreement on principles of legitimacy it would be challenging to form any consensus on policies related to distributive justice. At the same time, people may reasonably disagree about principles of legitimacy too. For example, it might be more difficult to agree on demanding principles of legitimacy than on very non-demanding principles of distributive justice. In well-functioning Western liberal democracies, we have agreed on both principles of legitimacy and principles of distributive justice, the latter of which are typically enforced through a constitution securing the distribution of basic rights and duties. Moreover, whether we reasonably agree or disagree on legitimacy and distributive justice seems to be an empirical matter and should not therefore be treated as a pre-theoretical limitation on the normative political space.

More importantly, even if it turns out that agreeing on principles of distributive justice is impossible in all feasible worlds, a priority of legitimacy does not follow. This is so because the former claim is empirical, while the latter is normative. Hence, even if we empirically need agreement on principles of legitimacy to realize distributive schemes, this empirical priority does not entail a justificatory priority. Indeed, this echoes the problem encountered by realists who follow Bernard Williams in assuming that since a political system must secure stability and order to realize justice, normative questions of legitimacy must come first, prior to questions about justice (see Williams 2005: 3; Jubb 2015: 921). But again, the second, normative claim does not follow from the first, empirical one (see more on this in the next section where we further discuss priority claims).

Moreover, such a view would quickly dissolve political theory as we know it. If a political theorist who aims to theorize a normative political principle (of, for example, legitimacy or justice) is also required to respond to the question of how to realize this principle in a society, the theorist would end up trying to answer empirical questions she is not competent to answer. Because whatever the subject in question, there is always a long chain of empirical preconditions that must be fulfilled (Erman and Möller 2023b).

For example, Larmore insists that the fundamental task for legitimacy is to determine which political order may justifiably impose authoritative rules for handling conflicts, including reasonable disagreement about justice (Larmore 2020: 5). But why jump on the train there? Using the same reasoning, we should require of such a theory to also offer an account of the empirical conditions under which this is realizable, such as that it would demand a sufficient amount of food, drink, oxygen, social stability, and so on (Erman and Möller 2023b).

While it would indeed be reasonable to object against an infeasible normative political theory that it is ill-construed, this is something different from demanding that such a theory must offer a full story of its own realization. This would not only dislocate political philosophy, but also require of theorists to focus on empirical aspects they are not equipped to study.

3 What is the Proper Role of Moral Norms in Political Theory?

So far we have treated our first, critical claim. In this section, we move to our second, constructive, claim, which states that any reasonable political theory must rely on foundational principles (even if implicitly). To do so, we take a closer look at the relationship between moral and political norms. As we rehearsed in the beginning of the paper, the idea that political normativity – i.e. political norms and principles – is ‘orthogonal’ to moral normativity has been severely criticized in the literature, and many political realists have either abandoned the claim or clarified that they never intended to argue for any such radical position (i.e. they were either unclear or misunderstood). Matt Sleat, for example, states that we should acknowledge “the relevance of morality to our normative thinking about politics” (Sleat 2022: 472).

But how, more exactly, we are to understand this ‘relevance’ is less clear in the literature. Central to political realists is the view that politics should not be reduced to morality (Hall 2017: 284; Sleat 2016: 254, 2022: 470). Instead, moral norms and prescriptions should be “filtered through or aligned to the realities of politics” (Sleat 2022: 474; cf. Jubb 2019: 362) For political realists, these ‘realities’ typically entail taking “disagreement, conflict, and power as ineradicable and constitutive features of politics” (Sleat 2022: 474). But while the idea that there are a number of inescapable empirical conditions to which any political norm has to somehow conform is not void of substance, it is still not clear how this ‘filtering’ is supposed to work, and what the resulting relation between moral and political norms is supposed to be.

3.1 On the Relation between Moral and Political Norms

In order to get a clearer understanding of the relation between moral and political norms, let us start by rehearsing the problems with the non-moral view that moral and political principles are orthogonal, in the sense that moral considerations have no bearing on what is politically right, one way or the other. This ‘orthogonality’ is most often cashed out in instrumental terms.Footnote 1 The political domain has a number of fundamental goals, and what is politically right solely depends on what best fulfil these goals. Let us now assume that politics has two fundamental goals: stability and order. The political system must be resilient and robust in the face of internal or external challenges, and there must be a presence of established rules, norms, and institutions that regulate behaviour, resolve conflicts, and maintain coherence within the political structure. These two goals are obviously reasonable in themselves, and definitely put restrictions on political actions and policies. But could they possibly determine what is, and is not, a politically right action/policy?Footnote 2 To answer that, we should start by distinguishing between whether stability and order is necessary for political rightness, or whether it is both necessary and sufficient.

If it is merely necessary, it means that we need some other norm or principle to choose between two different policies that both generate stability and order. But unless those further norms or principles are also instrumental – in which case we just add them to the list and ask the same ‘necessary or also sufficient’-question for the lot – we seem to need moral norms after all. In other words, what we seem to have here is rather a version of the filter view than the non-moral view: the upshot is that stability and order are necessary conditions for a rightful or legitimate political order, and only principles or norms which have ‘passed’ that filter are relevant.

Let us therefore interpret the claim that stability and order are both necessary and sufficient conditions for political rightness in the following way: if and only if a policy establishes (or retains, if starting from an already stable and ordered state) order and stability, it is politically right. Now, there are two further interpretive alternatives to consider. Either stability and order are understood as threshold (or binary) notions, so that at a certain level of resilient institutions and effective rule enforcement in a society, it is ordered and stable. Or stability and order are understood as scalar notions: the more stable and orderly a society, the better it is, and the policy or action that most increases the level of stability and order is the preferred one.

How reasonable would a non-moral account on either of these two interpretations be? Not reasonable at all, we will conclude, and we think that the reasons why are instructive for understanding the proper role of moral considerations in the political domain (which, of course, is the point of this rehearsal, since the premise of the current paper is that the non-moral view is incorrect).

For the threshold understanding of the stability and order principle, a similar problem as with the necessity-interpretation returns, but with a vengeance: as soon as a society has reached a level or stability and order, anything goes. As soon as the threshold is reached, each and every political policy has equal value and there is no (political) reason to pick one before the other. The scalar notion, one the other hand, selects (at least in theory) a single policy as the preferred one, notably, the one generating the highest level of stability and order.Footnote 3

Despite the first interpretation being very allowing and the second being very demanding, they both share the same fundamental problem: none of the normative considerations we have taken as central to a good political order is part of the principle governing what policy is politically right (on our current assumption). Whereas the former governing principle allows for any policy above the threshold and the latter only the one maximizing the order and stability of society, neither of them care about how unequal the distribution of social goods may be on the policy, how few liberties the citizen come to enjoy, how little the autonomy of persons is respected, how totalitarian or despotic the system of government might be, etc. – the list here includes all normative concerns debated in political theory the last 150 plus years.

The logical upshot of a non-moral account of political normativity is thus that everything from freedom of speech and distributive justice to citizens’ equal say in the decision-making, accountability, rights, duties and responsibilities play no normative part whatsoever in the political decision-making. Of course, the normative upshot of whether, say, freedom of speech or the right to self-determination is respected in a political policy is up for debate. Perhaps the level of stability and order is greatly improved in a context where some liberties are more restricted. But to say that considerations that we typically take to be central to a good society are completely irrelevant, tout court, is simply too extreme a bullet to bite. Such an account of political normativity is just too unreasonable.

Of course, the above account was only one (crude) example, and one can add other goals, making the resulting non-moral account more nuanced. But the general problem persists, no matter which (non-moral) aims you select: even if we grant that there is a set of properties that must be in place in a well-functioning society – such as stability, order, or the ability to enforce binding decisionsFootnote 4 – they may at most mark out a set of necessary properties. At least some level of stability and order, say, is arguably necessary for a society to aspire to be a functioning political order. Absent such a level, any freedom or right cannot be upheld in any case, and become ‘irrelevant’ (in a practical sense at least). But that there are a number of necessary properties for a political order to be practically functioning is a far cry from arguing that they are sufficient. Rather, they are the political equivalent to oxygen and food to sustain our bodily functions. These are necessary empirical conditions needed to be in place in order for us humans to work property at all. But while the ability to breath air and get sufficient food to eat are important conditions for our actions, they are not very helpful for determining what we ought to do.

In a similar manner, any non-moral ‘goals of politics’ at best signifies what we may call conditions of realization, i.e. they determine a minimum that must be in place for a functioning polity (Erman and Möller 2023b). But apart from that, they do not determine what political order, policy or action is preferable compared to another. Politics is an essentially contested domain, both in relation to its scope and content. What counts as politics – such as the public/private distinction – as well as what determines just, legitimate or otherwise right political actions is intensely debated in contemporary political theory.

For sure, for practices regulated by instrumental norms such as chess, it makes sense to claim that the best chess player is simply the one who is most successful in winning, or that someone who is treating all pieces on the chess board like checkers pieces is actually not playing chess at all. In the political domain, however, there are many alternative viewpoints on how to ‘move the pieces’, and the question of which kind of consideration speaks in favour of a political action or institution is open-ended, a matter of normative argumentation. How much, and to what extent, questions of stability and order matters in relation to, say, human rights and welfare, are normative questions within the political domain. Consequently, any reasonable political argument must include non-instrumental normative considerations, one way or the other.

3.2 The Proper Role of Moral Norms in Political Theory

The above discussion helps us grasp the role of moral considerations in politics as well as how to make sense of the ‘filtered through politics’ metaphor. So far, we have argued for the rather commonsensical claim that non-instrumental normative considerations are essential in political philosophy. More precisely, the two types of non-instrumental norms that we are interested in here are political norms and moral norms. Now, the question of the role of moral norms in political theory naturally (indeed, tautologically) depends on how we draw the line between moral and political norms. On one end of the spectrum, we may count as moral all but the instrumental norms having to do with necessary properties for a political community to persist – such as a threshold level of order and stability – or other ‘goals of politics’. On such an account, however, virtually all of the values and norms debated in the last century of political theory count as moral norms, and virtually all principles suggested in political philosophy should be called moral principles rather than political principles. Rawls’ principles of justice, then, are not political principles but moral principles. Such usage may in itself be innocent (as always, it depends on what is supposed to follow from the semantic distinction), but apart from trivializing the debate, it also fits badly with common usage. Similarly, at another end of the spectrum, any value or principle that has normative force in a political context may be called a political value or principle. Hence, the value of autonomy when utilized in a political context, should be counted as a political value rather than a moral one. But that also trivializes the debate, and seems to be far from the ordinary usage of the terms in question.

As suggested above, we take the most interesting demarcation between moral and political philosophy to be that the latter (but not the former) is fundamentally concerned with matters related to the governance of a political structure, such as the state, or the basic institutions of a society (see, e.g. Rawls 1999). Principles relating to these properties, then, is what we suggest should be called political principles. This means that for a principle to be labelled as a political principle, there is some contextual factor involved having to do with political governance of some sort.

Now, there will arguably be big differences with regard to the level of ‘contextual dependence’ here between very applied political principles such as policies for a particular area of politics, and rather abstract political principles, such as Rawls’ principles of justice. Whereas real life policies will depend heavily on current circumstances and limitations, Rawls’ principles aspire to a rather high level of generality, demanding only that certain ‘circumstances of justice’ are in place. For our purposes in this paper, however, it suffices to note that they are all context dependent: they rely on a certain set of empirical facts to obtain in order for their principles to be valid (such as Rawls’ assumption of ‘favourable conditions’ in society). This should be contrasted with what we here call foundational principles, by which we refer to principles (or values) that are independent of a society and polity. Given that distinction, our positive thesis about how moral principles or norms figure in the justification of a political principle is this:Footnote 5

Any valid political principle gains justificatory support from at least one foundational moral premise.

In other words, moral norms or principles figure in the justificatory support – the set of arguments providing the justification – for any valid political norm or principle. In that sense, moral norms always play a fundamental justificatory role in political theory. We will take two argumentative routes to support this claim, the first more dialectic while the second is more technical.

The first, ‘dialectical’ route involves looking at the kind of justifications for political principles that theorists as a matter of fact have used in the literature, also when arguing against a too ‘moralized’ view of political theory. That ‘mainstream’ political philosophers – who typically argue for principles of justice or legitimacy with a strong focus on egalitarian values, or on respecting the liberty and freedom of each citizen – gain support from moral premises is hardly surprising. Personal autonomy, self-determination and the equal value of each person are paradigmatic moral values, and it is hard to see how arguments for their political counterparts would be justified without reference to them. But this is also the case for classical political realists, who have argued for a much ‘thinner’ notion of political legitimacy, based on achieving social order in light of deep and irreducible disagreement. As Williams argues, peace, stability and security must be ensured first, since they serve as a pre-condition for any other political objectives (Williams 2005). Still, Williams and virtually all political realists who have followed him, argue that stability and security in itself are not sufficient for a political order to be legitimate (in perfect alignment with our argument against instrumental accounts of political normativity above). ‘Might is not right’, as the slogan goes; not all social orders that achieve stability and security can be considered legitimate. For example, “a peaceful situation preserved entirely through suppression or tyranny, or at the extreme by the exclusion or even elimination of those who do not accept it”, is not legitimate (Williams 2005: 3).

Consequently, also theorists who stress the historical context of a polity rather than idealized moral assumptions (e.g. Galston 2010: 387; Philp 2010: 466; Sleat 2010: 496), rely on normative conditions which, at bottom, depend on foundational, non-contextual premises. In order for a polity to be legitimate, a certain level of acceptance or agreement from the subjects are needed, they argue. But the agreement (Horton 2010) or willing consent (Bellamy 2010) is not an empirical matter only: it has to be perceived as free (Newey 2010), and cannot be coerced (Williams 2005) or be too tyrannical (Horton 2010). Non-coercion, freedom and absence of tyranny might be seen as commonsensical premises, of course, but they are foundational all the same: the conditions of (in this case) legitimacy, if contextual and historically dependent in other respects, still rely on these non-contextual values for its justification.

The second, ‘technical’ route to support our claim that political principles rely on foundational moral premises is to make use of one insight from G.A. Cohen’s famous argument that facts cannot ultimately ground principles. Cohen argues that no fact of any kind can ground a normative principle, unless there is a further normative principle which explains why this is so. Ultimately, Cohen holds, there will always be a normative principle that is not grounded on facts (Cohen 2003, 2008).

Cohen distinguishes between two kinds of principles: those that are dependent on facts and those that are fact-independent. On his account, “a principle can reflect or respond to a fact only because it is also a response to a principle that is not a response to a fact” (Cohen 2003: 214). The idea is that if we have a principle P that is sensitive to facts F, there is another “more ultimate principle that explains why F supports P” (Cohen 2003: 218). Cohen exemplifies this with the principle ‘we should keep our promises’ (A), which we assume depends on the fact that only then can people pursue their projects. Principle (A), Cohen argues, is true only because there is a further principle (B), ‘we should help people pursue their projects’ (Cohen 2003: 216–17). This further is also true – if true it is – only because of a further fact, namely that people can achieve happiness only if they pursue their projects. But the latter fact is explanatory salient only if there is yet another principle (C) that states ‘we should promote people’s happiness’. For a utilitarian hedonist, we have now reached a principle which is true regardless of any further facts; it is a fact-free normative principle. If we are not utilitarians, however, the regress continues at least one more step, since there must be some fact which explains in virtue of what we should promote people’s happiness. But eventually, we reach some normative principle which holds regardless of any further fact.

As a whole, Cohen’s argument is of course not uncontroversial. Indeed, Larmore argues against it, claiming that facts do ground principles, all the way down (Larmore 2013: 301–305; 2020: 60–67). Larmore’s argument is based on the close connection between a principle for action and reasons for action. “Principles,” he argues, “are general rules of thought and action, asserting that certain lines of conduct are what we have reason to adopt in the sorts of circumstances they stipulate” (Larmore 2020: 61). Reasons, he further argues, “consist in the way certain facts in the world count in favor (a normative relation) of certain possibilities of conduct” (2020: 62). A reason to carry an umbrella consists, he exemplifies, “not simply in the fact that it is raining, but in this fact counting in favor of the option of taking an umbrella” (2020: 62). He concludes that reasons, and thus principles, “do not float free of the (nonnormative) facts, but depend, as I have noted, on the facts being as they are” (2020: 62).

Although it does not matter for the insight of which we are currently interested (as we will see in the next paragraph), for what it is worth, we think Larmore is wrong, or at least that he misunderstands or misconstrues the sense of fact and fact-free which is of interest for Cohen.Footnote 6 Larmore is right that on most accounts, normative reasons are understood as facts.Footnote 7 But typically, the notion of ‘fact’ adhered to in these accounts is very undemanding, for the most part simply entailing that it is a true proposition (e.g. Smith 1994), or “that which can be designated by the use of the operator ’the fact that …’” (Raz 1975: 18). Cohen, on the other hand, utilizes the much stronger sense of ‘fact’, where it typically refers to an empirical fact (more specifically, the fact/value distinction). When Cohen calls a principle fact-dependent – that it “can reflect or respond to a fact” – he has in mind a principle that is true (if it is true) in virtue of another (empirical) fact that obtains. Whether we frame this in terms of principles or in terms of reasons make no difference: if I have a reason to carry an umbrella, that reason is fact-dependent if I have it only in virtue of the fact that it is raining (or that it is likely to start raining soon, or the like). And in that case, this reason is dependent (in the relevant sense) on further reasons, which ultimately becomes fact free, arguably referring to the good of avoiding harm, or something similar.Footnote 8 In truth terms: whenever there is a principle P that is true (or reason that obtains) only when a certain set of facts F obtains, this is because there is another true principle (or valid reason) of the form ‘if F then P’. If also that principle (or reason) is dependent on further empirical facts, the process continues until we reach a principle (or reason) that is true regardless of any further empirical facts, and for which we have no further principles or reasons to give.

In the present paper, however, it is not really the fact dependent/independent distinction that is in focus. Cohen’s argument, whether or not it truly demonstrates that all fact-dependent principles depend on fact-independent ones, demonstrates another important insight: the way in which political principles ultimately need justificatory support from some foundational premise. Remember that political philosophy, as we understand it, is fundamentally concerned with matters related to the governance of a political structure or the basic institutions of society. Naturally, principles guiding political institutions and the like are dependent on a myriad of factual circumstances: economic realities, temporal constraints of all sorts, and of course the cultural-political traditions of the entity in question. Political principles, thus, are dependent on the society in question. What Cohen’s argument demonstrates, however, is that as a matter of mere ‘normative logics’, such political principles are justified only in virtue of other principles which are less dependent on a particular society – and, ultimately, on premises which hold regardless of any particular society or polity. In other words, on what we here have called foundational premises.

Before discussing what we take to be the upshot of our discussion for the filter view – and more generally for how to understand the relationship between moral and political norms – let us treat a potential objection to our second, technical route to our thesis.Footnote 9 It may be objected that while Cohen’s argument demonstrates that political principles depend on foundational premises, their status as moral principles may be questioned. Miriam Ronzoni and Laura Valentini (2008), for example, have argued that the foundational principles of Cohen’s argument need not be moral, but can also be methodological. Their aim is to defend constructivism, the family of theories claiming that the validity of normative principles derives from the way they are constructed (e.g. through Rawls’s original position). While Cohen argues that constructivism is ruled out by his argument, since constructivists, on his interpretation, hold that all principles are fact-dependent (Cohen 2003: 213), Ronzoni and Valentini argue that there is an alternative interpretation of constructivism that is safe from his argument. A constructivist who thinks that mind-independent moral facts are beyond the limits of what we can plausibly claim to know, and who believes that a certain constructive procedure is the best way to justify normative principles without appealing to such moral facts, can, they argue, rely on a foundational principle like Pmethod:

When theorizing, one ought not to start from assumptions whose validity or truth is beyond the limits of what we can plausibly claim to know (Ronzoni and Valentini 2008: 409).

This principle is methodological rather than moral, they argue. So if they are correct, fact-independent ultimate principles need not be moral.

There are two main ways of responding this challenge. One might of course argue directly against Ronzoni and Valentini, and claim that the argument for a foundational principle being methodological rather than moral is somehow misconstrued.Footnote 10 For the purpose of the present paper, however, we need not take a stand on that issue. Whether or not they are correct, our conclusion that political principles depend on – i.e. gains justificatory support from – at least one foundational moral premise, still holds.

The reason can be inferred already from Ronzoni and Valentini’s argumentative structure. According to their argument, their foundational principle Pmethod grounds (though a number of steps) a principle which says that “one ought to act on those principles which the constructive procedure X delivers” (Ronzoni and Valentini 2008: 409). Depending on what procedure X is – Kant’s categorical imperative being a common contextualist example – a set of foundational moral principles is then justified, and eventually also a set of (contextual or fact-dependent) political principles. Now, the central aspect to note is that the methodological principles under debate are upstream from the foundational moral principles, and that the political principles are downstream from the moral principles. That is, regardless of whether foundational methodological principles also play a role in justifying substantial political principles (a claim we doubt, but in this paper do not take a stand on),Footnote 11 it is still a fact that political principles gain justificatory support from foundational moral premises. Hence, our conclusion stands in any case.

3.3 Revisiting the ‘Filter View’

Now, let us return to the initial question of how to understand the relationship between morality and politics. Does the filter view proposed by some political realists (Sleat 2022; Jubb 2019; Hall 2017), which suggests that moral norms and prescriptions are ‘filtered through’ or ‘aligned to’ the realities of politics, provide an enlightening picture?

On the one hand, the filter view suggests a primacy of the ‘realities of politics’ (the filter) through which some, but not all, moral norms (the liquid) may pass. On the other, Larmore seems to argue for a primacy of morality over politics. “[C]laims to legitimacy”, he states, “must always rest on assumptions expressing a morality prior to politics” (Larmore 2020: 108). Although conceptions of legitimacy may differ, Larmore further claims, “all these various understandings of legitimacy have in common that they regard it as rooted in antecedent principles of an essentially moral nature. (2020: 70).

Hence, it seems that we have rather different pictures of the relation between morality and politics present also among theorists who endorse some form of political realism. However, we argue that although there is a tendency to over-interpret both of these pictures, rightly interpreted there is truth in both of them.

On the one hand, there is a clear sense in which the filter view is compatible with the view of the relation between morality and politics that we have argued for in this section. Although the threshold limits are unclear and highly debatable, it seems reasonable to presume that we need a certain level of stability and order for a political system to persist. Politics is an essentially contested domain, both in relation to its demarcation (‘what counts as politics’) and substance (‘which values and norms should govern’), and it is reasonable to assume that there will be persistent disagreements on these matters on various levels. A political system must be stable enough to handle such disagreements and the potential conflicts that they might give rise to. In that sense, at least some moral considerations have to be conditioned on ‘political realities’.

Moreover, in politics, moral considerations, although typically important, do not always ‘trump’ political considerations. A political decision may be legitimate while at the same time being morally questionable. A political party, for example, may come to power by perfectly legitimate means (such as through an open, democratic election without any traces of manipulation) although its political agenda is morally corrupt – examples of which, unfortunately, are not only historical in nature. As such, it has gained political legitimacy for pushing that agenda, even when it is morally problematic. In this way, there is a sense in which moral norms do not pass the ‘filter’ of legitimate political realities.

On the other hand, this does not mean that the question of what determines whether a political power is legitimate or just (or whichever normative concept is referred to) is independent of moral considerations. According to what we have argued in this paper – and which has become increasingly accepted also within the realist camp – moral premises are always part of the justificatory support for a political principle. What might look like two incompatible claims at first glance – i.e. that what is politically legitimate and what is morally right may differ and that what is legitimate depends on moral premises – are in fact compatible on closer scrutiny: it is just a matter of the general context-dependence of normative claims. While it may be morally wrong of me not to give my seat in the subway to an older person in much more need of it, it may also be wrong to force me to do so. Likewise, even if we assume that it is morally wrong to ban abortion or allow the death penalty, it might be politically legitimate to do so, given the right procedure.

The ‘morality prior to politics’ claim expressed by Larmore is perhaps even easier to over-interpret than the former, ‘filter’ metaphor. This seems to be the ‘ethics first’ view so criticized by political realists. And surely, the claim that one domain has priority over another is a provocative one. Many of Larmore’s claims seem to repeat variations of this theme, such that political principles are “rooted in antecedent principles of an essentially moral nature” (Larmore 2020: 70, our italics; cf. also 93).

We think such ‘priority’ and ‘antecedent’ claims are misleading and should be avoided. There are many ways in which one discourse may be prior or antecedent to another and we would deny that morality is prior or antecedent in most of them. The talk of ‘antecedent’ seems to imply some sort of temporal – or even causal – priority, but that is very hard to get off the ground: the idea that moral considerations come temporally first, and cause political considerations, is both unsupported and unlikely. Moreover, there seems to be no particular conceptual priority of morality in relation to politics: our political concepts do not seem to demand a prior knowledge of moral concepts. Also, in most senses of epistemological priority, the claim that morality is prior to politics seems unsupported: we seem to be able to gain knowledge about political concepts without a prior knowledge of moral concepts (Erman and Möller 2015: 224 − 25).

Hence, this priority claim only seems reasonable if ‘priority’ refers to the justificatory dependence expressed in our discussion above. Political concepts like justice, legitimacy and liberty are in a justificatory sense dependent of – or in Cohen’s terminology ‘a response to’ – moral premises. While we would insist that ‘priority’ is not a very good label on this justificatory relationship, for the reasons just rehearsed, on a closer inspection it seems that Larmore’s ‘priority’ and ‘antecedence’ claims in actuality should be given no stronger interpretation than the one implied by our discussion above. Indeed, when he argues that despite the need to respect the ‘circumstances of politics’ the justification provided for legitimate rule “must still rely on principles of a moral character” (Larmore 2020: 108), he continues: “A good example is the liberal conception of legitimacy I outlined earlier … which holds that the state’s use of coercion, its use and threat of force, is justified insofar as it honors a principle of respect for persons” (2020: 108). When elaborated, it thus seems as if also Larmore’s claims are not as radical (from the standpoint of political realism) as he sets them out to be at first sight.

4 Conclusion

In this paper, we have discussed the relationship between moral philosophy and political philosophy (and between moral norms and political norms), interrogating one common view among political realists, namely, the filter view, in an attempt to theorize the proper role of moral norms in political theory. We have defended two claims. On a critical note, we have argued against the view, defended by Larmore and political realists, that because political philosophy is concerned with the regulation of legal-political orders and the basic institutions of society, it should primarily focus on political legitimacy as well as on political legitimacy before focusing on justice. On a constructive note, we have argued that any reasonable political theory must rely on foundational moral premises that are independent of a society or polity. As such, they are more basic than political principles in the sense that they put up the normative boundary conditions for such principles.

We ended the paper by discussing two contrasting pictures of the relation between morality and politics, which both are somewhat misleading. On the one hand, the filter metaphor is misleading and potentially problematic, since it prioritizes politics over morality in the sense that the filter is made up of a fixed ‘reality of politics’ through which some moral norms (the liquid) may pass. Such a view is in need of explanation. In this sense, the filter view, as it has been articulated so far in the realist literature, is still in need of further development and refinement. Equally misleading is Larmore’s claim of a ‘morality prior to politics’, which seems to entail an ‘ethics first’ view of the kind that political realists reject. In order to avoid potential misunderstandings, we recommended that the term ‘priority’ should be avoided. The only reasonable interpretation of ‘priority’ of the moral to the political is more clearly described as a justificatory dependence of political theory on at least one moral premise; and the only reasonable interpretation of ‘priority’ of the political to the moral is a constitutive dependence on certain empirical conditions (e.g. a level of stability and order). In a debate that is accused of the different camps talking past each other, we think that directly addressing these particular relations when they are of interest, instead of making sweeping claims of the priority of X over Y, is a preferable strategy.