Republicanism and moralised freedom

Abstract

A moralised conception of freedom is based on a normative theory. Understanding it therefore requires an analysis of this theory. In this paper, I show how republican freedom as non-domination is moralised, and why analysing this concept therefore involves identifying the basic components of the republican theory of justice. One of these components is the non-moralised pure negative conception of freedom as non-interference. Republicans therefore cannot keep insisting that their freedom concept conflicts with, and is superior to, this more basic concept. I demonstrate how we can use pure negative freedom to formulate the republican theory more precisely. This exercise is more fruitful than the common focus on the alleged conflict between the two freedom concepts.

Introduction

Republican freedom as non-domination has gained considerable popularity in recent decades. Its proponents consider it a more attractive political ideal than liberal freedom as non-interference. Freedom, republicans say, is not simply the absence of interference but rather the status people enjoy when they are protected against interference they have not themselves instructed. Liberal freedom, on the other hand, does not demand such protection, and is therefore criticised for implying, counter-intuitively, that people who are vulnerable under the arbitrary control of other agents are free as long as these agents do not actually interfere with them (Ingham and Lovett, 2019; Pettit, 1997: ch. 2, 2008a, 2011, 2012: ch. 1; Skinner, 2008).

We are therefore told that a slave can never be free on the republican account since he is subject to the master's power to interfere with him contrary to his instructions and must therefore act submissively so as not to provoke the master's interference. ¹ He is unfree whether the master exercises this power or not. On the liberal account, on the other hand, the slave will be free as long as the master does not interfere with him. Here republicans usually ignore that the master reduces the slave's overall freedom from interference by restricting the number of actions the slave can perform in conjunction with not acting submissively (Carter, 2008; Kramer, 2008).

Republican freedom is commonly regarded as a moralised conception of freedom. ² That is, it is defined with evaluative terms, as one kind of interference – instructed, or controlled, interference – is morally permissible and therefore not a source of unfreedom. And the prevention of impermissible acts, such as taking a slave, is not a source of unfreedom. Republican freedom therefore differs from liberal freedom, which, at least on the pure negative conception, is a strictly empirical, or descriptive, concept. On the pure negative view, an agent, A, is free to do x as long as no one makes it physically impossible for him to do x. And another agent, B, makes A unfree to do x by preventing A from doing x. ³ A remains free to do x if B merely makes it more difficult or costly for him to do x, but B will then reduce the number of actions A can perform in conjunction with x, thereby reducing his overall freedom. Pure negative freedom is an empirical concept, as we can observe whether or not an agent is physically prevented from performing an action. ⁴ We need no moral judgment for discriminating between types of prevention. This is indeed what makes this account of freedom purely negative. Positive freedom is to act on autonomous motivations. Pure negative freedom, on the other hand, is insensitive to whether a constraint is intentional, significant, or permissible (Steiner, 1994: 9–12, 17–21).

It is commonly observed that republican freedom is moralised while pure negative freedom is not. The issue of what this means for the alleged conflict between these freedom concepts, however, has been left largely unexplored. In this paper, I demonstrate its significance for the republican–liberal debate. I first show how republican freedom is moralised, and why Philip Pettit (2006: 278–280, 2008a: 117, 2012: 58), the main contemporary defender of republican freedom, fails in his attempts to deny that it is moralised. In fact, I show how republican freedom is moralised at two levels. As is commonly observed, republican freedom is moralised at the collective level, as it says government interference does not make people unfree insofar it tracks people's common interests. However, these common interests will frequently conflict with personal interests. Promoting these common interests might be attractive and a justified exercise of state power, but that means the distinction between constraints that make you unfree and those that do not is a distinction between justified and unjustified institutional constraints.

Pettit denies that this makes republican freedom moralised, since the distinction between controlled and uncontrolled interference is not moralising. ⁵ That is, controlled interference does not impose an alien moral doctrine on anyone. Pettit therefore understands republican freedom as ‘a neutral political ideal’ – it is neutral between conceptions of the good people may hold (Pettit, 1997: 97). ⁶ Controlled interference tracks the interests people share, not those they ought to share. And we can all agree, as a matter of fact, whether an act of interference is controlled or uncontrolled. This might all be correct, but it does not show that republican freedom is not moralised. We shall see that the republican distinction between controlled and uncontrolled interference is based on moral evaluation of people's interests. It is based on a theory of justice specifying which interests institutions should respect and not. ⁷

We become aware of moralisation also at the individual level when we examine the view that you are not made unfree when you are interfered with in accordance with your instructions. Citizens are said to be free when their government acts reliably in accordance with their common interests, and individuals are free when others interfere with them only in accordance with their instructions. Such interference occurs when the preference you expressed at the time you gave the instruction, t₁, conflicts with the preference you expressed at some later time, t₂. This priority to the instruction, or to the t₁ preference, is based on moral evaluation.

The observation that republican freedom is moralised means it is based on a normative theory and therefore cannot be a building block in the formulation of such a theory. It is because the concept is based on this theory that republicans can say the slave is always unfree and citizens remain free as long as their government is constrained to act in accordance with their common interests. Pure negative freedom, on the other hand, is not moralised, and can therefore serve a function in the formulation of this theory. Saying that the two concepts conflict is therefore to ignore that they lie at different levels of theorising.

By insisting on this conflict, republicans deprive themselves of a basic concept that would enable them to formulate their theory more precisely. As I show towards the end of the paper, republicans should not reject pure negative freedom, but instead apply it in the formulation of their theory of justice – the theory on which their freedom concept is based. In particular, they should clarify the value of freedom understood in this non-moralised sense. The value of freedom is here understood as the value in being free to do x, independently of the value in doing x. Following Ian Carter (1999), I distinguish between the specific and non-specific value of freedom. Freedom has specific value if only freedoms to do certain things are valuable. Freedom has non-specific value, on the other hand, if a freedom is valuable regardless of its particular instance; freedom as such is considered valuable. I show how the republican fundamental concern with individuals’ self-respect implies that freedom has non-specific value. This leads to a discussion of how freedom should be distributed, and I argue that Pettit follows Rawls in opting for a maximin distribution.

While these arguments concerning the value and distribution of freedom are debatable, the most important point is that the way forward for republicans is not to focus on the differences between these two concepts but rather on how the republican theory, on which republican freedom is based, promotes pure negative freedom.

Moralised freedom

Robert Nozick (1974: 160–164) considers taxation a source of unfreedom, as it prevents individuals from transferring their resources as they wish. He denies, however, that the government makes anyone unfree by protecting private property acquired in accordance with his theory of just appropriation. But as G. A. Cohen (1979) points out, both cases involve interference, since protecting private property means anyone not in possession of the property is prevented from using it. Nozick thus distinguishes between different types of interference and takes the type compatible with his libertarian theory not to make the individuals unfree, while interference conflicting with his theory does make people unfree. Unjustified interference therefore makes us unfree, while justified interference does not (Cohen, 1979: 12, 1988: 294–296). ⁸

Pettit's conception of republican freedom also distinguishes between a type of interference that makes the agent interfered with unfree and a type of interference that does not make the agent unfree. He calls the former uncontrolled interference and the latter controlled interference. In either case, an agent, A, imposes an obstacle on another agent, B's, ability to perform an action. But the interference is under B's control to the extent that A is required to act on B's instructions. A thus acts towards B in accordance with B's interests. Uncontrolled interference, on the other hand, does not accord with these instructions, or it accords with the instructions only because it happens to suit A. In such cases, B has no control over the way A acts towards him.

In Pettit's (2012) frequently used example, A gives the key to his alcohol cupboard to B with the instruction of giving it back only on 24 hours' notice when A later comes and asks for it. When A then wants a drink and B refuses to give him the key, B interferes with A – she prevents him from having a drink – but since B acts on A's instructions, she does not make A unfree. ⁹ Analogously, Pettit thinks citizens are not made unfree when the government acts towards them in a way they have instructed. The people's instructions are understood to express interests that people share and can therefore avow in public without embarrassment (Pettit, 2001: 156–160). Especially important is everyone's interest in developing a sense of dignity and equal standing in the society. This point will be important later in the paper.

These interests must be robustly protected. Freedom, Pettit says, requires not just the absence of interference conflicting with people's instructions or common interests, but also the robust protection against such interference. Such protection is provided by social institutions that constitute individuals’ freedom; freedom comes into existence simultaneously with these institutions (Pettit, 1997: 107). It is important to note that the protection against interference these institutions provide itself involves interference – they prevent or inhibit people from interfering contrary to their common interests. But since such interference is in the common interest, they do not make people unfree in the republican sense.

In the next two sections, I consider whether this distinction between controlled and uncontrolled interference makes for a moralised conception of freedom. I show that it does by identifying moralisation both when republican freedom is applied at the collective level and at the individual level.

Political decision-making

On the republican conception, a government under popular control does not make citizens unfree. Popular control is realised when the government is appropriately constrained from acting contrary to the citizens’ shared interests. These are interests individuals can avow in public because they are held by everyone who wants a government that treats people respectfully as full members of society (Pettit, 2001: 156–160). Pettit treats this citizen–government relation as analogous to the relation between individuals exemplified by the alcohol cupboard case. Citizens can instruct government interference in a way analogous to how an individual instructs other individuals’ interference with him.

Many critics, however, point out that the two cases are not analogous, since a government can never act on each citizen's interests in the way an individual can act on the instruction of another individual (Carter, 2008: 64–65; Christman, 1998: 205; Estlund, 2014: 799–802; List and Valentini, 2016: 1061–1063; Sharon, 2016). Individuals in a large society have conflicting interests, and it is therefore impossible for a government to track all of them. Many individuals will likely also have an interest in everyone but themselves being constrained in some manner. I might want everyone but myself to be taxed. And I might want the state to protect my private property but no one else's. And as Carter (2008: 65) points out, a thief convicted on the basis of common interests is very unlikely to have an interest in being imprisoned, even if he would like anyone else to go to prison for the same offence. Popular control therefore cannot mean the government acts in accordance with each individual's instruction. It instead appears to mean the government operates legitimately by serving common interests, and not personal interests conflicting with these common interests. And to say that government interference is no source of unfreedom as long as it is legitimate is to moralise freedom.

Pettit, however, has repeatedly denied that his freedom concept is moralised (Pettit, 2006: 278–280, 2008a: 117, 2012: 58). He is aware, of course, that individuals might want to free-ride on others’ tax-paying and to have laws that constrain everyone except themselves (Pettit, 1997: 55–56). But such interests are ‘irrelevant’, he says, with respect to the definition of the common interests that the government of free citizens promote (Pettit, 2004: 152). Living in a society is a ‘historical necessity’, and because we depend on our society, we cannot demand the satisfaction of interests that will undermine it (Pettit, 2012: 167). We must accept that we are bound by the same laws as everyone else, which means certain options will not be available to us (Pettit, 1997: 93, 2004: 152).

Pettit may be right that people are better off with than without a state imposing certain restrictions on their behaviour, but that just means these restrictions are justified for this reason. Only by moralising our definition of freedom – by basing it on a theory of justice specifying how institutions ought to operate – can we therefore say that interference in accordance with common interests does not make people unfree. On a non-moralised definition, by contrast, any restriction makes citizens unfree to perform certain actions independently of how collectively acceptable these restrictions might be. It makes no difference whether the interference is justified or not. A person is made unfree to perform various actions regardless of why he is imprisoned.

Pettit points out that republican freedom does not impose on people interests they do not already have, and says it is therefore not moralising. The common interests the government of free people promotes are acceptable to everyone. And he means ‘acceptable’ in a ‘non-normative sense, implying that the object or policy or whatever is such that people are disposed to accept it’ (Pettit, 2012: 170, fn. 34). The common interests are interests individuals in fact share, not interests they ought to have. The republican freedom ideal is neutral in the sense that it does not impose a particular conception of the good on people. ¹⁰ It consequently conflicts with paternalism (Pettit, 2008a: 127–128, n. 19, 2012: 58–59). And Pettit (2008a: 127–128, n. 19) thinks republican freedom therefore cannot be moralised, since a moralised conception of freedom ‘might justify a paternalistic concern for people's good, regardless of their perception of the good’.

This response to the moralisation charge tells us that Pettit is simply talking about a different kind of moralising than his critics do. Anti-paternalism and neutrality between conceptions of the good are perfectly compatible with a moralised conception of freedom, as the term is commonly used in the freedom literature. A moralised concept of freedom is based on a theory of justice, and if that theory is neutral between conceptions of the good, then so is the freedom concept. It may not be ‘moralising’ in the sense Pettit here has in mind, but it is moralised in the sense that it is based on a theory of justice. We moralise freedom by saying interference in line with this theory is not a source of unfreedom, while other interference is. Whether this just interference is, or can be, paternalistic is a separate issue dealt with in the theory.

A non-moralised conception of freedom, on the other hand, is ‘value-free’, to use Carter's (2015: 284) term, since it is defined without evaluative terms. It simply describes a state of affairs without reference to any moral evaluation. Pure negative freedom, or unfreedom, is descriptive in this sense, as it describes the social relation where B does not, or does, physically prevent A from doing x. It is defined without evaluation of the prevention. This non-moralised conception can have evaluative connotations insofar as we are inclined to think it is a good thing that A is not prevented from doing x, or, more generally, that people's interference with one another is minimal. And it is indeed because of these evaluative connotations that pure negative freedom can play a useful role in a liberal theory, as we shall see in a later section. But such evaluation is independent of the concept itself (Oppenheim, 1981). ¹¹

Pettit (2008a: 117) seems to suggest that also republican freedom is empirical in this way when he points out that people with conflicting views of whether an act of interference is justified or not might agree on whether it is controlled or uncontrolled. Whether an act of interference is controlled or uncontrolled is a factual question, he says, not a normative one (Pettit, 1997: 56, 2006: 279–280). The interference is controlled insofar as it accords with the interests of the agent interfered with and the interferer was constrained from interfering contrary to these interests. If the interference did not have these properties, it was uncontrolled and consequently an act of domination.

But this is not the issue when we consider whether a freedom concept is moralised. When Pettit has explained what he means by controlled and uncontrolled interference, we might agree on how to fit acts of interference into one or the other category. ¹² But this distinction is based on a theory of justice, and we can argue about the merits of this theory. The basis on which we consider whether an act of interference makes the agent interfered with unfree or not is therefore normatively controversial. Once the theory is explained, we might agree on whether the theory justifies the interferences or not, and therefore whether the interference is a source of unfreedom, in this moralised sense. But the freedom concept is still moralised, since we need this normative theory to distinguish the type of interference that does make you unfree from the type that does not.

Priority to instruction

So, republican freedom is moralised because the government promoting citizens’ common interests cannot promote each individual's personal interests. Individuals have conflicting interests, and those who are justly prevented from acting on interests not in line with common interests, such as the justly imprisoned criminal, are not made unfree. We have seen that Carter identifies this moralisation at the collective level, but he says ‘no independent moral considerations come into play’ at the individual level, where republicans say B does not make A unfree by constraining A in accordance with A's instruction (Carter, 2008: 65; see also Sharon, 2016: 137–138). In this section, however, we shall see how the moralisation of republican freedom appears also at the individual level. The same individual can have conflicting preferences, and republicans moralise freedom by saying the individual is made unfree only when he is prevented from satisfying one preference, but not the other. Since the interference here is not institutional, the evaluation is arguably not based on a theory of justice. It does, however, express a moral view of how individuals ought to treat each other.

Recall that in Pettit's example, A instructs B at t₁ not to give him the key to the alcohol cupboard when he later asks for it. When A then, at t₂, asks B for the key and B refuses to give it to him, B interferes with A, but not in a way that makes A unfree, since A instructed the interference. Republican freedom thus gives priority to the preference A expresses when he instructs B at t₁ over the preference A expresses later, at t₂. As Carter and others observe, this case differs significantly from the collective-level case in that A alone is the source of the instruction he gives to B. There is no need to account for moral considerations concerning everyone being equally bound by a collective decision, since A alone can determine how B is to behave towards him.

But why does the freedom concept give priority to the t₁ preference and not to the conflicting t₂ preference? One possible explanation is that A is more himself at t₁ than at t₂, and the preference he expresses at t₁ is therefore more his own than is his preference at t₂. This explanation discriminates between a higher and a lower self. Pettit (2012: 57, fn. 33) also suggests that the distinction between controlled and uncontrolled is based on a priority to a higher self when he considers the will expressed at the time of the instruction to be ‘stable or authoritative in comparison with the will that the instructions require [A] to frustrate’.

But by thus associating freedom with the will of a higher self, Pettit cannot maintain that the distinction between controlled and uncontrolled interference is not moralised, since it is based on an evaluation of mental states. If this is the correct understanding of republican freedom, then Keith Dowding (2011: 308) appears to be right when he suggests republicans bring individual welfare into their measure of freedom. Dowding understands the distinction between controlled and uncontrolled interference to be based on a welfarist comparison of two opportunity sets, the one before and the one after the interference. If B does not make A unfree in the alcohol cupboard case because we believe B thus best promotes A's welfare, then Pettit offers a welfarist account of freedom. And building welfarism into the definition of freedom is to moralise it.

But I doubt this is how Pettit would justify the priority to t₁ preferences. We have seen how he emphasises that his freedom ideal is neutral between different conceptions of the good and is not paternalistic. Perhaps we are misled to this association with welfare by cases like the alcohol cupboard example, as we might think A's interest in not drinking is his true interests because we believe alcohol consumption to be bad for A. It might therefore help to reverse the example so that A, at t₁, instructs B to make him keep drinking if he later wants to stop. At t₂, when A wants to stop drinking, B acts on A's t₁ instructions by making him drink more. To avoid giving priority to t₁ preferences on moral grounds, Pettit must see this as a case of freedom-compatible interference.

But we still have no explanation for why freedom gives priority to t₁ preferences. And at the collective level, Pettit again suggests that the priority is based on moral evaluation. He justifies the priority to t₁ preferences by saying we need to impose constraints on each other to make sure people behave predictively in accordance with their shared, or common, interests. Pettit appears to base this view on the importance of treating people with respect rather than enhancing their welfare. But this explanation does nothing to deny that the priority to t₁ preferences is based on moral evaluation of how people ought to treat each other. It is, in fact, hard to see any plausible reason for this priority independent of moral evaluation.

Levels of theorising

A moralised conception of freedom is logically posterior to the normative theory it is based on. We need this theory to tell us why one kind of interference makes us unfree while another kind does not. Pettit (2012: 21) sees republican freedom as the basis of a theory of justice. But when we notice that this freedom concept is moralised, we see why this is to take things in the wrong order. Nozick's concept of freedom is not the basis for his libertarian theory, it is based on this theory. Pettit's freedom concept, likewise, is not the basis for a theory of justified government under popular control, it is based on this theory.

If freedom is based on a theory of justice, it can serve no constructive purpose in that theory; it ‘falls out of the picture’, as Cohen (1988: 296) says (see also Carter, 1999: 72; Steiner, 1994: 15). ¹³ To see how, recall that republicans reserve the term ‘freedom’ for the status of living under a just and legitimate government – that is, a government under popular control. This obviously does not alter the fact that such a government imposes restrictions on what citizens can do. These restrictions stand in need of justification, and such justification is developed in a normative political theory. Republican freedom can perform no role in this justification since it is itself based on the theory.

At the basis of a theory, we find non-moralised concepts. We can start with a non-moralised freedom concept, and then consider which freedoms people ought to have, and ought not to have. We can describe our theories in part in these terms, and we can meaningfully compare them by looking at how they promote freedom. Aware of how our theories differ in these respects, we can use moral arguments in favour of how one theory promotes freedom and against the way another theory does it. A prerequisite for meaningful discussions of these matters is that we agree on what the term ‘freedom’ refers to. To meaningfully agree or disagree at the higher evaluative level, we must agree at the empirical level (Oppenheim, 1981: 154).

Pure negative freedom, as an empirical concept, enters at a more basic level of theorising than does republican freedom. ¹⁴ Republican freedom is based on a theory that pure negative freedom might serve a function in formulating. It is therefore unfortunate that republicans have devoted so many pages to discussing how republican freedom conflicts with liberal freedom as non-interference (Ingham and Lovett, 2019; Pettit, 1997: part 1, 2011, 2012: ch. 1, 2014: ch. 1; Skinner, 1998, 2002), and with the pure negative conception in particular (Pettit, 2008a; Skinner, 2008).

A slave, on the pure negative conception, is free to perform whatever action his master, or anyone else, does not prevent him from performing. Republicans point out that this view ignores how the power relation between the slave and the master affects the slave's freedom. Because the master has the unchecked power to interfere with the slave in a way the slave has not instructed, republicans deny that the slave can ever be free. But while the unchecked power the master has is morally problematic, there is no conflict between republican freedom and pure negative freedom. The difference is only that republican freedom is based on a theory that condemns slavery, while pure negative freedom is not – it might instead play a role in the formulation such a theory. To criticise pure negative freedom for implying – counterintuitively, in republicans’ view – that the slave can be free to perform whatever action no one prevents him from performing is therefore effectively to criticise this conception for not being moralised.

Ronald Dworkin (1987: 5) aims exactly this criticism at empirical, non-moralised freedom when he calls it ‘flat’ and worthless in normative theorising. It only describes states of affairs, and it does not, Dworkin argues, suggest any reason for endorsing one or another arrangement. The value we call ‘liberty’, for Dworkin, should consist of the freedoms people should have – or, more precisely, the freedoms they have in an ideal egalitarian society. ‘Flat’ freedoms include freedoms to treat others in ways that conflict with this ideal, and therefore cannot be useful in normative theorising. Liberty, in Dworkin's view, is therefore not a fundamental value, but rather based on the requirement that individuals are to be treated with dignity and respect.

Pettit implicitly also makes this point by preferring his moralised conception of freedom because of its harmony with a theory of justice, and by dismissing pure negative freedom for its unappealing implications. Implicit in his rejection of pure negative freedom is the view that a freedom concept implying that the prevention of unjust or illegitimate actions is a source of unfreedom, just as much as the prevention of just or legitimate actions, is surely of little use in normative theorising. In the next section, however, I suggest ways in which this empirical freedom concept can perform a central role in the formulation of Pettit's republican theory. And I show how clarifying this role will make the theory more precise.

The role of non-moralised freedom

The issue republicans should consider is not whether pure negative freedom conflicts with republican freedom, but rather how it features in the republican theory. The role of a non-moralised freedom concept in a theory depends on the value we give to freedom thus understood. Here the value of being free to do x is independent of the value of actually doing x. Carter (1999) distinguishes between the specific and non-specific value of freedom. On the former view, only freedoms to do certain specific things are considered valuable. On the latter view, on the other hand, a freedom is valuable regardless of its particular content – that is, what it allows you to do. Freedom as such is therefore understood to have value.

Pettit seems to take the first view – what Carter calls the ‘specific-freedom thesis’ – when he identifies the basic liberties as the freedoms institutions under popular control will ensure that all individuals can effectively exercise. Dignity and respect are the more fundamental values he bases this concern for the basic liberties on. The ideal of a free person, for Pettit (2012: 3), is someone who can relate to any other as an equal and ‘live without shame or indignity’. Without institutions protecting these liberties, people will not enjoy this status. Pettit does not seem to think other, non-basic liberties contribute to ensuring this status.

Pettit therefore appears to join theorists like Dworkin and Rawls in adopting the specific-freedom thesis. Dworkin and Rawls also see the basic liberties as valuable, and they find no value in freedom as such. As we have seen, Dworkin is not interested in ‘flat’ freedoms, and he even denies that there is such a thing as liberty as such. Only certain liberties are valuable, and that is because they contribute to the more fundamental values of concern and respect (Dworkin, 1977: 266–274). Rawls (2001: 44–45, 2005: 291), likewise, denies the importance of ‘liberty as such’, and instead points to the value of certain liberties because of their significance for developing the two moral powers – that is, the abilities to develop and act on a sense of justice and a conception of the good. ¹⁵ These are the basic liberties, and Rawls (1999: 53) and Pettit (2008b: 220, 2012: 103) have virtually indistinguishable views of which liberties are basic.

What Carter calls the ‘overall-freedom thesis’, on the other hand, finds value in freedom as such, and is therefore concerned not just with particular freedoms people have, or ought to have, but also with how much freedom they have, or ought to have. This concern with freedom as such is based on the view that freedom has non-specific value, though certain freedoms can still be considered particularly valuable because of the value of what they allow us to do. Carter defends the non-specific value of freedom and the overall-freedom thesis. Freedom as such is valuable, he argues, as our ends are better served not just by having certain freedoms but also by having more rather than less freedom overall. Carter (1999: 45) points out an important reason for this being that we are often unsure what our preferences are or will be in the future. ¹⁶

Carter (1999: ch. 2) distinguishes between different ways in which freedom can be considered valuable. On both the specific and non-specific views of freedom's value, freedom can be considered valuable instrumentally as a means to some other valuable end, or constitutively as a constituent part of some intrinsically valuable end. But according to the specific-freedom thesis, only certain freedoms have such value, while on the overall-freedom thesis, any freedom is valuable in this way (although some freedoms might be more valuable than others). Freedom's non-specific value can also be understood as intrinsic if freedom as such is considered an end in itself rather than a means to, or a part of, some other value.

Carter (1999: 60) understands Rawls's fundamental concern with self-respect to imply that freedom has non-specific constitutive value. Self-respect is of crucial importance, Rawls (1999: 386) says, because a person lacking it will see nothing as valuable or worth doing. For Rawls (1999: 477), an equal distribution of fundamental rights and liberties is the basis for self-respect. But Carter argues that a certain degree of freedom as such is a constitutive part of self-respect. ¹⁷ Although the basic liberties might be especially important for self-respect, individuals will not develop a sense of self-respect unless they have a range of opportunities to choose between (Carter, 1999: 60). This is especially so given Carter's observation that we often do not know what our future preferences will be. Freedom as such is therefore a constitutive part of self-respect.

This connection between freedom and self-respect is important also for understanding Pettit's theory. Like Rawls, Pettit defends institutional provisions for everyone's ability to effectively exercise the basic liberties, which is constitutive of the status of a free person (Pettit, 1997: 107). People enjoying this status, as we have seen, can develop a sense of self-respect and relate to each other as equals. Institutional protection ensuring this social standing gives individuals the opportunity to develop and lead ‘a full and meaningful life’ (Pettit, 2012: 103). Rawls and Pettit thus share a fundamental concern for self-respect that freedom as such seems to contribute to, or partly constitute, though the basic liberties might be particularly important. Even freedoms to interfere in ways Pettit and other republicans condemn can be considered valuable, since choosing the right action when faced with a choice between interfering or not interfering in this impermissible way can be dignifying. ¹⁸ Recall that the value of freedom is independent of the value of doing what one is free to do.

Pettit (2012: 107–110) tries to distance himself from Rawls by claiming Rawls, like most other liberals, is insufficiently concerned with the institutional protection each individual needs in order to feel like a respected and equal member of society. Rawls, Pettit says, does not demand the institutional protection required for making sure everyone is an independent person subjected to no other person's will. Here Pettit does not consider that protection itself involves restricting individuals’ (non-moralised) liberty, and he gives no indication of defending such restrictions beyond what Rawls argues for. ¹⁹ It is also doubtful whether such restrictions are compatible with a fundamental concern with self-respect, precisely because of the importance of unconstrained choices for individuals’ development of self-respect.

Pettit also underestimates Rawls's concern with institutional protection (Moen, 2022b: 261–263). Rawls is aware of the importance of protecting the basic liberties, and even says he shares the republican concern that this protection depends on a politically engaged citizenry (Rawls, 2001: 144). Pettit (2012: 11) seems unaware of this when he says liberals like Rawls see little importance in citizens being ready to contest institutions perceived not to function in their common interest. Samuel Freeman (2007) indeed understands Rawls's concern with protecting everyone's ability to effectively exercise the basic liberties to be based on a fundamental concern with ensuring everyone that no one should have to ingratiate oneself to another. This is the republican ideal of a free person. As Freeman (2007: 187) says, Rawls wants to make sure everyone is ‘socially and economically independent, so that no one need be subservient to the will of another’. Freeman (2007: 45) understands this independence as the primary end of justice in the high liberal tradition of which Rawls is an important part.

Since freedom as such seems to be considered valuable in the republican theory, then that raises the question of how it ought to be distributed. Freedom can, for example, be equalised or maximised, and a sufficientarian principle would demand that everyone has freedom at least up to a certain level (Carter, 1999: ch. 3). ²⁰ There is no clear discussion of the distribution of freedom in Pettit's work, or in the recent republican literature more broadly. But we might think the concern with elevating all citizens up to a level at which they can relate to one another respectfully as equals requires a minimal level of freedom, and therefore a sufficientarian distribution. The dignity and self-respect of the free person requires a certain degree of freedom secured by the provision of the protection and resourcing to which Pettit gives great importance. But beyond this level, freedom as such is not valuable (at least not as far as dignity and self-respect are concerned). Pettit (2012: 88) appears to endorse such a view quite explicitly when he says each citizen should be provided with ‘a certain threshold benefit in the currency of free or undominated choice, as required under the sufficientarian strategy’.

But a sufficientarian distribution does not restrict the inequality above the threshold, and Pettit makes it seem unlikely that his republican theory allows for much inequality of freedom. The aim is that everyone enjoys the resources and protection they need to relate to one another as equals with no feeling of dependence and inferiority. And what it requires to enjoy this status depends on how much resources others have to protect themselves against interference (Pettit, 2012: 91). One might feel inferior if others have significantly more freedom than oneself, perhaps because they have much more money with which to pay others not to interfere with them (Carter, 1999: 235–236; Cohen, 2011). What it means in terms of freedom to have a dignified and respectful standing in one's society therefore depends on how free others in the society are.

The republican theory requires that everyone enjoys a certain level of freedom, but since this level depends on others’ level of freedom, the requirement does not seem best expressed as a sufficientarian principle. Equal standing and self-respect do not seem to allow that some are left behind once a certain minimal level of freedom is realised. The worst-off might not feel respected if institutions stop working in their interest beyond this level. It is therefore more plausible that the fundamental concern for self-respect calls for an equal distribution, at least as long as inequality will not be to everyone's benefit. An unequal distribution that makes everyone better off, albeit some better off than others, might give everyone the feeling that the institutional arrangement functions in their interest. And that might plausibly give everyone a sense of being an equal and respected member of society.

This concern for everyone benefitting is Rawls’s (1999) ideal of a ‘cooperative venture for mutual advantage’. This view of society is associated with contractualism, and Pettit (2001: 157, fn. 1) acknowledges that his view of common interest, which we have seen is central in his theory, is inspired by Rawls and is ‘contractualist in spirit’. The contractualist concern with what everyone can accept typically leads to a special concern for the worst off (Scanlon, 1982: 123). After all, we seem most likely to have everyone's acceptance of a certain arrangement if it can be accepted even by those it benefits the least. And this reasoning leads towards a maximin distribution, under which the worst off are as well off as possible. For this reason, Pettit's republican theory – like Rawls's theory (Carter, 1999: 82) – appears to be a plausible home to a maximin principle for the distribution of freedom.

We would get a better understanding of the republican theory if it focused less on the alleged conflict with pure negative freedom and more on how this concept operates in the theory. We would especially get a clearer view of the required protection against interference conflicting with common interests. Such protection, as we have seen, itself entails interference, and with an empirical concept, we can consider how it promotes freedom.

Conclusion

In this paper, we have seen how republican freedom is moralised, as it is based on moral evaluation of different types of interference. It is moralised at the collective level, since it is compatible with government interference with citizens to protect common interests in accordance with a theory of justice. It is also moralised at the individual level, since denying that interference following an individual's instructions makes the individual unfree entails a priority to a preference expressed at the time of the instruction over the conflicting preference expressed at the time of the interference. And this priority is based on moral evaluation of these preferences.

The realisation that republican freedom is moralised has significant consequences for the republican rejection of freedom as non-interference, and particularly its non-moralised, empirical pure negative conception. It is misleading to say, as Pettit and other republicans do, that promoting one rather than the other produces different outcomes. When we recognise that republican freedom is based on a normative theory and pure negative freedom is not, we see that the two do not conflict. Instead, pure negative freedom is a more basic value that can be used in the formulation of the republican theory on which freedom as non-domination is based.

Republicans should therefore work with the concept of pure negative freedom and explain its role in their theory. I have argued that this theory does not just call for the protection of certain freedoms but also the promotion of freedom as such, as that will contribute to individuals’ self-respect, which is a fundamental value in Pettit's republicanism. And that means republicans should be concerned with how freedom is distributed. As I have shown, the republican concern with ensuring everyone's status as equal and respected members of society seems in line with the contractualist defence of a social arrangement that benefits everyone. A maximin distribution of freedom, beneficial even to the worst off, therefore appears to fit well into the republican theory. These considerations for making the republican theory more precise are more fruitful than insisting on a conflict between an empirical and a moralised conception of freedom.