1. INTRODUCTION
As is well known, Isaiah Berlin (Reference Berlin1969) mounted an attack on the Stoic concept of freedom as the ability to do what one wishes. He objects to this preference-dependent concept of freedom because it allows for a change in freedom that is due only to a change in preferences. This objection comes in two guises. First, the Stoic conception has the counterintuitive consequence that an agent can increase her freedom merely by extinguishing her wishes or changing her preferences. I refer to this as ‘the Stoic Retreat Problem’. Second, it has the abject implication that a dictator can oppress his subjects in the name of freedom by conditioning their preferences in such a way that they only want to do those things that he allows them to do. For reasons that will become clear below, I call this ‘the Deliberate Dictator Problem’.
According to what I call ‘the received view’, the negative concept of freedom on which freedom consists in the absence of external constraints solves both problems, because it does not take freedom to be conceptually related to preferences. The choice that remains to be made is between kinds of restrictions that count as constraints on freedom. One option is that all external restrictions count as constraints on freedom, as the neutral conception (NC) has it. Other salient options are to count only intentionally imposed restrictions or restrictions for which someone is morally responsible, as the intention-dependent conception (ID) and the moral responsibility conception (MR) have it respectively.Footnote 1 Dowding and van Hees (Reference Dowding and van Hees2007, Reference Dowding and van Hees2008) maintain in opposition to the received view that negative conceptions also suffer from the Stoic Retreat Problem. They go on to argue, however, that ID suffers less from preference-dependence than NC. Berlin provides another argument in favour of ID when he argues that a dictator such as the one in the Deliberate Dictator Problem exhibits ill will, which in turn betrays a lack of respect towards his subjects. If Berlin is right and the notion of freedom should be sensitive to respect, this provides a reason to prefer ID to other conceptions of negative freedom. In this paper, I criticize both arguments in favour of ID, and I defend MR instead.
Dowding and Van Hees (Reference Dowding and van Hees2007, Reference Dowding and van Hees2008) maintain that negative conceptions of freedom are sensitive to changes in preferences, just as the Stoic concept is, and that ID exhibits fewer possibilities for preference-dependence than NC. They conclude that, as the Stoic Retreat Problem condemns preference-dependence, ID is to be preferred to NC. In response, I point out that the Stoic Retreat Problem condemns only conceptual preference-dependence, and not the contingent preference-dependence that Dowding and Van Hees address (Carter and Kramer Reference Carter and Kramer2008). My core criticism, however, is that what I call ‘the argument from preference-dependence’ fails to establish that NC and ID differ in the extent of their preference-dependence. What is more, in spite of initial appearances, MR does not fare worse than ID in this respect. Hence, contingent preference-dependence is irrelevant for making a choice between conceptions of negative freedom (section 2).
The Deliberate Dictator Problem vividly illustrates how disrespectful restrictions are that are due to ill will. Whereas this counts in favour of ID in comparison to NC, it leaves open whether ID is to be preferred to MR. As intentions are of special relevance to moral responsibility – intended wrongs are often more blameworthy than unintended wrongs – MR can also do justice to the disrespect of ill will. The next thing to note, however, is that unintended restrictions can be disrespectful too. As an illustration, I present the case of Indifferent Dictator, a dictator who does not care at all about what happens to her subjects. Because of her indifference, the unintended restrictions that Indifferent Dictator imposes on her subjects are also disrespectful. Hence, they should also count as constraints on freedom. This they do on MR, but not on ID. In this way, this ‘argument from indifference’ reveals that MR can do justice to the disrespect that both ill will and indifference convey. The upshot is that MR best captures the ways in which respect bears on freedom (section 3). Thus, MR surfaces as the most attractive of these three conceptions of negative freedom.
But why, one might ask, should respect play such a central role in an analysis of freedom? Berlin (Reference Berlin1969) approvingly refers to Rousseau's claim that it is not the nature of things that maddens us, but ill will. In order to do full justice to his arguments, it is important to take this seriously. Stanley Benn provides further support for this claim when he argues that people are committed to respecting ‘the standing of every other person as an originator of projects’ (1988: 98). To disregard those projects, he argues, is to disregard what is distinctive of a person. Such disregard gives rise to resentment, which in turn is central to the complaints and grievances people express when they are made unfree. Now I do not assume that any conception of freedom should accommodate considerations such as these. It may well be that there are other things that are bad about a lack of freedom, and perhaps they can be better captured by other conceptions of freedom. What I argue instead is that MR does best justice to Berlin's two arguments.
2. PREFERENCE-DEPENDENCE
2.1 Negative Freedom and the Stoic Retreat Problem
The two versions of Berlin's (Reference Berlin1969) objection against the Stoic or preference-dependent conception of freedom are nicely captured in this passage:
If degrees of freedom were a function of the satisfaction of desires, I could increase freedom as effectively by eliminating desires as by satisfying them; I could render men (including myself) free by conditioning them into losing the original desires which I have decided not to satisfy. (Berlin Reference Berlin1969: xxxviii)
Berlin claims here that the preference-dependent concept allows for changing someone's freedom merely by changing this person's preferences. The kind of problem this raises depends on whether the agent changes his preferences, or whether someone else changes them. According to the first-person version of the objection, the Stoic Retreat Problem, the Stoic concept of freedom has the consequence that an agent can change his own freedom by changing his preferences. This consequence is counterintuitive, because it would seem that liberating someone requires lifting a constraint that is external to the agent in that it has another agent as its source.
The third-person version of the objection, the Deliberate Dictator Problem, targets a tyrant or dictator who uses the same strategy to manipulate the freedom of her subjects. A dictator could invoke this strategy to liberate his subjects by conditioning them into ‘losing their original wishes and embracing (‘internalising’) the form of life he has invented for them’ (Berlin Reference Berlin1969: 140). In this case, the problem is not so much that the consequence is counterintuitive, but that the manipulation that is at issue here is abject. Berlin makes clear how serious this problem is when he claims that what upsets people is manipulation. A lack of freedom is due to coercion and ill will. This in turn is problematic, because ill will reveals disrespect. In light of both of these two problems, Berlin concludes that the preference-dependent concept of freedom should be rejected.
The received view is that, as Berlin argues, the problem can be solved or sidestepped by adopting a negative concept of freedom as the absence of external constraints, which are restrictions that have their source outside the agent. The reason for this is that negative conceptions of freedom do not feature the preferences of the agent whose freedom is at issue. Dowding and Van Hees (Reference Dowding and van Hees2007, Reference Dowding and van Hees2008) criticize the received view and argue that negative conceptions of freedom do suffer from the Stoic Retreat Problem. They point out that the fact that negative conceptions of freedom do not refer to the preferences of the agent whose freedom is at issue leaves open the possibility that some of the constraints that people face depend in fact on their own preferences. It is an empirical matter whether, and if so, to what extent they do. They go on to argue that it is rather likely that ID registers fewer preference-dependences than NC, and they conclude that, because of this, ID is to be preferred to NC. As MR countenances more constraints than ID, their argument is likely to generalize and imply that ID is also to be preferred to MR.
Dowding and Van Hees (Reference Dowding and van Hees2007) construct an example that shows that NC allows for preference-dependences. Consider Ami and Fred. The question at issue is whether Fred is free to read a particular book. Only one copy of the book is available. By hypothesis Ami will be the one who succeeds if both try to get it. In this first scenario, Fred is indifferent between getting the book or not (see game 3 in Dowding and Van Hees Reference Dowding and van Hees2007: 154). What he wants to avoid is doing the same thing as Ami. Ami, however, wants to do the same thing as Fred. They know of each other's preferences, and because of that they do not reveal their actions to one another. As it turns out, Fred gets the book, and Ami does not even try to do so. As Ami does not prevent him from doing so, Fred is free to read the book.
The second scenario differs from the first in that Fred wants to get the book, which is what he sets out to do (see game 4 in Dowding and Van Hees Reference Dowding and van Hees2007: 154). Given that Ami knows this, the best thing she can do is to try to get the book. By doing so, she ensures that she gets it and neither one of them reads it. In this way, Ami and Fred will end up doing the same thing – not reading the book – which is what Ami wants. As in this scenario Ami prevents Fred from buying the book, he is not free to read it.
The two scenarios differ only with respect to Fred's preferences. In both scenarios, he wants to do something else than Ami. In the first scenario, he is indifferent between getting the book and not getting it. In the second, in contrast, he prefers to get the book. As it turns out, this neutrality entails a difference in freedom: in the first scenario Fred is free to read the book, in the second he is not. Thus, the book example reveals that, given the neutral conception of negative freedom, a difference in freedom can be due to a mere difference in preferences. Dowding and Van Hees conclude that the neutral conception of negative freedom fails to solve the Stoic Retreat Problem
Dowding and Van Hees believe that all negative conceptions of freedom involve preference-dependences of this kind, which means that none of them solves the Stoic Retreat Problem completely. They argue, however, that ID is affected by the problem to a lesser extent than NC. Whereas the change in preferences decreases Fred's freedom when conceptualized in terms of NC, they maintain that, when it is conceived of in terms of the absence of intentional prevention, Fred's freedom remains the same.
The key premise of their argument is this: Ami can legitimately claim that she does not intend to prevent Fred from buying the book, because all she wants is to do the same thing as Fred. Hence, Fred is free to buy the book in both scenarios, which means that, relative to ID, the difference in preferences does not translate into a difference in freedom. Granting for the sake of argument that contingent preference-dependence counts against a conception of freedom, this example supports ID at the expense of NC. As Ami is surely morally responsible for restricting Fred's actions, it also entails that ID is to be preferred over MR. This is the argument from preference-dependence.
Dowding and Van Hees (Reference Dowding and van Hees2007) defend three claims. First, the book example reveals that, when conceptualized in terms of NC and ID, an agent's freedom can depend on his preferences. As I agree with their analysis of the book scenarios, I accept this claim. Second, ID entails fewer possibilities for preference-dependence than NC, and hence it is to be preferred to NC. I take issue with this argument, the argument from preference-dependence, in section 2.2. Third, the Stoic Retreat Problem reveals why such preference-dependences are wrong.
This third claim is mistaken, I believe, because the Stoic Retreat Problem is concerned with conceptual preference-dependence (Berlin Reference Berlin1969: 140), and not with the contingent preference-dependences that Dowding and Van Hees (Reference Dowding and van Hees2007) address (Carter and Kramer Reference Carter and Kramer2008).Footnote 2 This means that Dowding and Van Hees (Reference Dowding and van Hees2007) are left without an argument as to why contingent preference-dependences are bad. To be sure, the restrictions that are due to preference-dependences may be bad. If that is the case, however, the thing to do is to remove those restrictions, and not to adopt a conception of freedom that does not recognize the restriction as a constraint on freedom. Pace Dowding and Van Hees, contingent problems should not be solved by conceptual means. The Stoic Retreat Problem draws attention to the fact that it is counterintuitive if an agent can liberate himself merely by changing his preferences no matter what his preferences are. Negative conceptions do not have this consequence. Resolving preference-dependent restrictions by adopting another conception of freedom rather than by lifting an external constraint is a matter of ignoring the problem rather than solving it.Footnote 3
All in all, I do not see why the fact that a conception of negative freedom allows for contingent preference-dependences would count against that conception. For the sake of the argument, however, I will grant this point to Dowding and Van Hees in the main body of the paper. My main concern is with the argument from preference-dependence, according to which ID is to be preferred to NC because it entails fewer possibilities for preference-dependence. I will argue in section 2.2 that, irrespective of whether the distinction between conceptual and contingent preference-dependence should carry any weight, Dowding and Van Hees (Reference Dowding and van Hees2007, Reference Dowding and van Hees2008) fail to establish that ID suffers less from preference-dependences than NC.
2.2 A Critique of the Argument from Preference-dependence
As Dowding and Van Hees (Reference Dowding and van Hees2007) present it, the argument from preference-dependence employs a notion of intentionality that is implausibly narrow. People intend the means they choose for achieving their aims (Searle Reference Searle1983; Adams Reference Adams1986; Bratman Reference Bratman1987; Moya Reference Moya1991; Enç Reference Enç2003). This uncontroversial platitude about intentions entails that the key premise of the argument from preference-dependence is mistaken: Ami can in fact not plausibly deny that she intends to prevent Fred from buying the book.
Recall that Ami's aim is to do the same thing as Fred. The means by which she achieves this aim is buying the book. This is a means to her end because by buying the book she prevents Fred from buying it. As people intend the means they choose for achieving their aims, this entails that Ami intends to buy the book and thereby prevent Fred from buying it. The upshot is that Ami makes Fred unfree also relative to ID. This in turn means that Dowding and Van Hees have made a mistake in applying ID to the two book scenarios discussed above.Footnote 4
Another way of making basically the same criticism consists of pointing out that a dictator could use the argument from preference-dependence to her advantage. Suppose a dictator prevents her subjects from doing all kinds of things just to increase her own pleasure, wealth and power. Perhaps she closes the road that connects her two palaces for her subjects on days on which she travels from the one to the other. Her only aim is to make her trip as convenient as possible. Does the fact that she does not directly aim at constraining her subjects imply that on ID she does not decrease the freedom of her subjects? Presumably it does not.
The thing to see is that she makes her trip convenient by preventing her subjects from using the road at issue. In light of this, the conclusion is unavoidable that she intends to prevent his subjects. When accused of restraining the freedom of her subjects, she might try to get away with maintaining that she intends to do no such thing, just as Ami might claim not to intend preventing Fred from getting the book. She can sincerely claim that her only aim is to increase her wealth and power. Dowding and Van Hees would have to accept this plea, given that they use the same defence for Ami. This would entail that on ID the dictator does not make her subjects unfree. Surely, that is an unacceptable conclusion. It is unacceptable because, just as Ami, the dictator cannot plausibly claim not to intend the means to his ends.Footnote 5
The thing to see is that the kind of preference-dependence Dowding and Van Hees (Reference Dowding and van Hees2007, Reference Dowding and van Hees2008) are concerned with is rather specific. It concerns pairs of cases with the following two features: (1) the cases differ only with respect to a preference of one of two agents, say agent B; (2) in one of the two cases, the presence or absence of the preference motivates the other agent, agent A, to prevent B from doing so, as this enables her to achieve her goal. The criticism that I have put forward boils down to the fact that, given that agent A achieves her goal by preventing agent B, preventing agent B is her means for achieving the goal. And just as ends, means are intended. In light of this, the preference-dependences Dowding and Van Hees are concerned with are bound to bear on the intentions people form. Hence, relative to ID agent A makes agent B unfree. As all restrictions that ID counts as constraints on freedom are also classified as such by NC and MR, the relevant cases are treated in exactly the same way by each of these conceptions of negative freedom. This means that none of them has an advantage over the other insofar as contingent preference-dependence is concerned.Footnote 6
3. RESPECT
3.1 The Deliberate Dictator and the Indifferent Dictator
The Deliberate Dictator Problem features a dictator who manipulates his subjects in the name of freedom. Berlin (Reference Berlin1969) argues that manipulation is bad, because it is due to an ill will or a bad intention. And bad intentions are bad because they convey disrespect towards the person concerned. As what is bad about a lack of freedom is related to a lack of respect, it is natural for Berlin to embrace ID, as it recognizes only restrictions that are imposed intentionally as constraints on freedom. Given a concern for respect, the neutral conception (NC) is somewhat arbitrary, as it countenances any external restriction. Hence, a concern for respect favours ID over NC, as only ID does justice to our human nature as purposeful beings (Berlin Reference Berlin1969: 202–203; cf. Morriss Reference Morriss2012).
Miller (Reference Miller1983: 73) and Kristjánsson (Reference Kristjánsson1996: 22–24, 32) argue, in turn, that Berlin's restriction to intended obstacles is arbitrary. Constraints on freedom should, for instance, also encompass restrictions that are due to negligence. Disrespect consists of a disregard for someone's interests and projects (Benn Reference Benn1988: 103, 107). Negligence also constitutes such disregard.Footnote 7 I provide a new argument for broadening the set of restrictions that count as constraints on freedom from those imposed intentionally to those for which someone is responsible. At the heart of my argument lies the fact that bad intentions are not the only attitudes that are disrespectful. Attitudes of indifference are so as well. In order to capture the relation between freedom and respect, both of these faces of disrespect have to be accommodated. Whereas ID captures only one, MR captures both.
Consider Indifferent Dictator. Indifferent Dictator does all kinds of things just for the fun of it. She acts on the spur of the moment not caring about the obstacles that her subjects face as a consequence. It is not just that she has no preference either way, but that she is not concerned with her subjects at all. Imagine that Indifferent Dictator is a bit of a health freak and wants to make sure that she and her family as well as all of her employers take enough vitamins. She buys more than enough for everybody making sure that the supply will last at least for another three years. Given that she has a large family and an immense number of employees, the amount of vitamins she buys is staggering. As a consequence, there is not enough left for those subjects that she does not employ. It is brought to Indifferent Dictator's attention that, as a consequence, some people with a vitamin deficit suffer from severe health problems. As is to be expected, Indifferent Dictator does not care either way.
Does Indifferent Dictator makes her subjects unfree? Compare her to Evil Dictator who does exactly what Indifferent Dictator does, but with the intention to let his subjects suffer. Surely Evil Dictator makes his subjects unfree. Would it be plausible to say that Indifferent Dictator does not make her subjects unfree simply because she does not intend to do so? Not if the underlying concern is with respect. Indifference amounts to a lack of moral concern, and as such it betrays a lack of respect (Arpaly Reference Arpaly2002). This implies that, in order to capture the relation with respect, a conception of freedom should condemn both Evil Dictator and Indifferent Dictator as making their subjects unfree. As ID distinguishes between them, it cannot adequately capture the relation between freedom and respect. This is what I call ‘the Indifferent Dictator Problem’.
In contrast to ID, MR treats Evil Dictator and Indifferent Dictator the same in that it supports the verdict that both make their subjects unfree. One might object and maintain that Evil Dictator is worse than Indifferent Dictator. Although it is far from obvious that this is indeed the case, MR can do justice to differences in verdicts about moral responsibility. The idea that intended preventions are at least sometimes particularly bad fits well with practices of moral responsibility ascriptions, as the fact that a wrong has been committed accidentally is usually treated as an at least partially extenuating circumstance (Wallace Reference Wallace1994). MR only insists on treating the two dictators the same insofar as freedom is concerned. It thereby solves the Indifferent Dictator Problem.
3.2 A Broad Conception of Intentional Action
The proponent of ID has an avenue of escape. She can solve the Indifferent Dictator Problem by adopting a broad conception of intentional action. According to a narrow conception, only intended actions are preformed intentionally (Searle Reference Searle1983; Adams Reference Adams1986). As discussed in section 2.2, this should be taken to encompass the ends an agent seeks to bring about, as well as the means he employs to achieve those ends. After all, agents have pro-attitudes towards means as well as ends. Proponents of broad conceptions of intentional action do not insist on a pro-attitude towards the outcome that the agent brings about intentionally. They might, for instance, accept that an agent can bring about an effect intentionally that she does not desire, i.e. towards which she has a con attitude. An agent can foresee an undesired consequence of her intended action, and perform the action because the benefits of the intended outcome outweigh the costs (Harman 1976; Bratman Reference Bratman1987).
In order to accommodate Indifferent Dictator, however, a further step needs to be made. There needs to be a third kind of case that shows that intentionality is consistent which indifference. Joshua Knobe (Reference Knobe2003, Reference Knobe2010) has experimentally established that people do in fact attribute intentionality to an indifferent agent when the consequence with respect to which he is indifferent is bad. Knobe's experiments feature a vignette concerning a chairman who favours a business strategy that will harm the environment. When this is pointed out to the chairman, he professes not to care about the environment. As it turns out, many people say that the chairman harms the environment intentionally when presented with this case (but not in the parallel case in which he benefits the environment).
To be sure, it has been argued that the intentionality attributions involved in what has become known as the Knobe Effect are mistaken (Malle and Nelson Reference Malle and Nelson2003; Nadelhoffer Reference Nadelhoffer2004; Nado Reference Nado2008). However, an increasing number of philosophers offer theories on which this is exactly what is to be expected.Footnote 8 Holton (Reference Holton2010), for instance, observes that the chairman violates a norm, and argue that this is why he harms the environment intentionally. Elsewhere I have argued that the agent's indifference with respect to the harm explains this type of attribution (Hindriks Reference Hindriks2008).Footnote 9 In the remainder of this paper, I will be concerned with conceptions of intentional action that are so broad as to include a condition that accommodates the intentionality attributions involved in the Knobe Effect.
In line with this distinction between narrow and broad conceptions of intentional action, I distinguish two versions of ID. On ID-narrow only restrictions that are due to a pro-attitude count as constraints on freedom. On ID-broad also preventions that are due to a con-attitude or to indifference limit freedom.Footnote 10 The thing to note at this point is that the case of Indifferent Dictator is structurally identical to that of Knobe's chairman. Just as the chairman, Indifferent Dictator expresses indifference with respect to a harmful side effect of his intended action. As a consequence, Indifferent Dictator limits the behaviour of his subjects intentionally. This in turn implies that, according to ID-broad, Indifferent Dictator makes his subjects unfree, just as MR. As a consequence, the Indifferent Dictator Problem, as presented above, affects only ID-narrow and cannot be used to defend MR against ID-broad.
In section 2.3 I present a second version of the Indifferent Dictator Problem in order to argue against ID-broad. Before doing so, however, it will be useful to discuss the claim that ID is to be preferred to MR because only the former is morally neutral. Dowding and Van Hees admit that ID-narrow is ‘weakly value-laden’, because the preferences of the agent who imposes a constraint will be reflected in his intentions (2007: 143).Footnote 11 This also holds for ID-broad, as it is sensitive to beliefs people have about the consequences of their actions – whether they are harmful or not – or about the norms they are subject to. MR, however, is value-laden in a stronger sense, because it resorts to objective moral norms or values for the ascription of freedom.
It is far from obvious, however, that the fact that MR is strongly rather than weakly value-laden counts against this conception of freedom. The thing to note is that, as it invokes objective norms or values, its pronouncements on freedom will be objective as well.Footnote 12 If freedom has to be a value-laden concept, it may be more attractive to invoke objective rather than subjective norms or values. Imagine that a restriction does not count as a constraint on freedom, because the agent who imposed it does not appreciate the harmfulness of his actions, or because he is unaware of a particular norm. The agent whose behaviour is restricted may resent the disrespect of the perpetrator, as he should recognize the harm or should have been aware of the norm. The perpetrator, she feels, has something to answer for.Footnote 13 Whereas MR can do justice to this sentiment, ID cannot.Footnote 14
So-called moralized conceptions of freedom are value-laden in an even stronger sense: when a restriction is just, it cannot count as a constraint on freedom. This entails that someone who is justly imprisoned is not unfree (Nozick Reference Nozick1974). It appears that moralized conceptions of freedom take the value-laden character of freedom a step too far. A lack of freedom indicates that some agent has to answer for imposing a restriction. It does not, however, entail that she was wrong to do so, and cannot have a satisfactory answer. In light of this objection, it is reassuring to know that MR is not a moralized conception of freedom. Someone can, after all, be morally responsible for something without having done something that is wrong or unjust (Miller Reference Miller1983: 68–69; Kristjánsson Reference Kristjánsson1996: 32).Footnote 15
The upshot of sections 2.1 and 2.2 is that, in addition to ill will (bad intentions), indifference can also make people unfree. The Indifferent Dictator Problem establishes that ID-narrow is inadequate, because it fails to classify Indifferent Dictator as someone who makes his subjects unfree. ID-broad remains a serious contender up to this point.
3.3 The Argument from Indifference
As it happens, Indifferent Dictator has an illegitimate child. He has always made sure that both she and her family have sufficient food and medication. Throughout the years, however, he gets increasingly fed up with taking care of them. Then the mother of his daughter informs him in the monthly letter she is allowed to send to him that she needs a new kind of medicine in order to keep his daughter's half-brother alive. Indifferent Dictator forms a half-hearted intention to have the requisite drug sent to them. The next day he does not quite remember whether he took care of this. Due to a lack of concern, he does not bother to check. As it turns out, the medication was never sent, and the half-brother dies as a consequence.
Relative to MR Indifferent Dictator makes the mother unfree. He is, after all, morally responsible for his omission, which prevents the mother from giving the appropriate medication to her son. Indifferent Dictator makes her unfree in this respect in spite of the fact that he does not prevent her intentionally from providing her son with the requisite medication. Indifferent Dictator's negligence does not excuse him. He should have been more careful and have checked whether he had done what needed to be done. This is what makes him answerable for the death of his daughter's half-brother. Indifferent Dictator acts in a way that is disrespectful with respect to him and his mother when he acts negligently in this way. His omission indicates that he is treating them with insufficient concern, even though it is not part of his intentions to do so.Footnote 16
This second version of the Indifferent Dictator Problem establishes a difference between MR on the one hand, and both ID-narrow and ID-broad on the other. In this example, Indifferent Dictator imposes a constraint accidentally on both narrow and broad conceptions of intentional action. In the first example, he decides to do something while ignoring a side effect. In this second example, he forgets to do something. Although he is indifferent in both cases, this is consistent with intentionality at most in the first case.
I believe that the fact that only MR issues in the verdict that Indifferent Dictator makes the mother unfree counts in favour of it at the expense of both ID-narrow and ID-broad. It is important to note that my line of argument goes beyond an appeal to intuition. The underlying point is that moral indifference is disrespectful and that, if a conception of freedom is to capture the connection with respect, it should also accommodate the Indifferent Dictator cases. I refer to the challenge that the two Indifferent Dictator Problems pose to ID as ‘the argument from indifference’.
Morriss (Reference Morriss2002: 118) argues that freedom is important because it matters to us how we are treated by other people. And we care about this because being treated in an inappropriate way affects our self-respect or our status as a person (see also Benn Reference Benn1975, Reference Benn1988; Pettit Reference Pettit2001). Given the considerations about ill will mentioned above, it is rather plausible that intentionally imposed constraints are of particular concern to our self-respect. There is little reason, however, to exclude unintentionally imposed constraints for which other agents are morally responsible.Footnote 17 The Indifferent Dictator example just presented reveals that someone's self-respect can just as well be affected by negligently imposed constraints or by omissions. Thus, the range of sources of unfreedom that MR captures is more appropriate than that of ID.Footnote 18
The two Indifferent Dictator examples concern foreseen side effects on the one hand, and unforeseen on the other. The vitamin example discussed in the previous section concerns foreseen side effects, whereas the medicine example discussed here concerns unforeseen side effects. They complement cases in which the agent has bad intentions, which are in fact covered by both ID and MR. The two Indifferent Dictator examples reveal that MR is to be preferred to ID.Footnote 19 Both of them are part of the argument from indifference. Only MR solves it completely. Hence, only MR captures both faces of disrespect, the face of bad intentions as well as the face of indifference.Footnote 20
4. CONCLUSION
I have argued that the moral responsibility conception of negative freedom (MR) is to be preferred to the intention-dependent conception (ID). The positive part of my defence of MR consists in the argument from indifference that concerns the lack of respect that behavioural restrictions can convey. Inspired by Berlin's Deliberate Dictator, I introduced Indifferent Dictator, a dictator who is indifferent regarding the restrictions that her subjects face, and ends up interfering with their actions without intending to do so. As Indifferent Dictator is disrespectful with respect to her subjects, her restrictions should count as constraints on people's freedom. ID can capture the unintended but foreseen restrictions only on a broad interpretation of intentional prevention. Such an interpretation, however, does not suffice for capturing unforeseen but negligent restrictions, which also convey disrespect. As only MR captures all of these restrictions, it is to be preferred to ID. Whereas ID only does justice to the relation between respect and ill will, MR also accommodates the connection between respect and indifference.
The negative part of my argument in favour of MR concerns the Stoic Retreat Problem. Recall the three claims that Dowding and Van Hees (Reference Dowding and van Hees2007) make about preference-dependence. First, negative conceptions of freedom exhibit preference-dependence. The reason for this is that interdependences between preferences of different subjects can give rise to constraints. Second, ID features fewer potential preference-dependences than the neutral conception (NC). Third, the Stoic Retreat Problem reveals why such preference-dependences are wrong. This third claim is problematic, because Dowding and Van Hees are concerned with contingent preference-dependence, whereas Berlin was concerned with conceptual preference-dependence. Whereas the first claim is true, the argument from preference-dependence in favour of the second claim fails. Pace Dowding and Van Hees, the restrictions that are due to interdependences between preferences are imposed intentionally. This means that these interdependences restrict freedom not only on NC and MR, but also on ID. Thus, ID does not have an advantage to NC and MR in this respect. And, as the argument from indifference reveals, MR does have an advantage over the other two.
ACKNOWLEDGEMENTS
I thank Ian Carter, Martin van Hees, Peter Morriss and Jan-Willem van der Rijt for useful comments, as well as two anonymous referees of Economics and Philosophy.
