Deciding for Oneself: Considering Wolff’s and Raz's Conceptions of Autonomy

(1)

Considering Wolff’s and Raz's Conceptions of Autonomy

Master’s Thesis

Philosophical Perspectives on Philosophy and the Economy Leiden University

Xanthe Sluijs s2328240 17 February 2020

Supervisor: Bruno Verbeek Word count: 20.534

(2)

Abstract

In In Defense of Anarchism, Robert Paul Wolff claims that a legitimate authority cannot exist, as any authority will necessarily prevent its subject’s autonomy. Jospeh Raz, in contrast, argues that a legitimate authority can exist and should support its subjects’ autonomy. In the argumentation to defend their conclusions, both focus on formulating their conceptions of (legitimate) authority. In this paper, I argue that the difference in their conceptions of autonomy is the basis of the

disagreement in Wolff’s and Raz’s conclusions. After a comparison of the two conceptions and the role autonomy plays in the authors’ argumentations, I conclude that Raz’s conclusion is the more plausible of the two.

(3)

§1 Introduction

In In Defense of Anarchism, Robert Paul Wolff argues that legitimate state authority does not appear to be possible due to the incompatibility of individual moral autonomy and authority. Wolff bases this argument on the Kantian conviction that individuals can only be responsible for their actions if they are free to determine their own life policies. In other words, moral autonomy follows from being able to determine your own policies and actions (Wolff, 1970: 8). He argues that if individuals have the possibility to be morally autonomous, they have the obligation to be so. In contrast to autonomy, Wolff defines legitimate authority as the right to command and right to be obeyed (Wolff, 1970: 6). Such a right, he argues, implies a subject’s surrender of judgement to others. However, this is incompatible with those subjects being autonomous. Wolff concludes that a

αὐτός [au̯.tós] self νόμος [nó.mos] usage, custom, law

statute, ordinance made by authority αὐτονομία

[u̯.to.no.mí.aː]

freedom to use its own laws, independence

Henry George Liddell, Robert Scott, A Greek-English Lexicon

autonomy (noun)

[ɔːˈtɒn.ə.mi]

the right of an organization, country, or region to be independent and govern itself,

the ability to make your own decisions without being controlled by anyone else

(5)

legitimate authority, a moral right to rule, does therefore not appear to be possible (Wolff, 1970: 10).

This challenge to the possibility of the existence of a legitimate authority has garnered a variety of responses. Among them, Joseph Raz has tried to formulate political authority in such a 1 way that it does not necessarily negate the personal autonomy of its subjects. To do so, he has proposed what he calls ‘the service conception of authority’. According to this conception, authority is legitimate if it serves its subjects to better conform their actions to the reasons that apply to them (Raz, 1986: 67; 2006: 1014). Raz considers personal autonomy as people having, to some degree, control over their own life; fashioning it through successive decisions throughout their lives (Raz, 1986: 369). He argues that a legitimate authority facilitates, rather than negates, the conditions for personal autonomy of its subjects (Raz, 2006: 1015; 1986: 407-10, 415-8). Raz's conclusion is radically different from Wolff’s, in both envisioning the co-existence of personal autonomy and political authority, and in allowing the possibility of a legitimate authority.

Despite the differences in their conceptions of authority, both Wolff and Raz commit to a conception of authority in which subjects submit to authority independent of the content of its directives and in a pre-emptive fashion. However, they differ in their conclusions as to whether political (legitimate) authority is compatible with personal autonomy. Whilst Wolff rejects the possibility of an autonomous individual being commanded, Raz does not consider being autonomous and at the same time following commands as mutually exclusive. In this thesis, I argue that the difference in their conceptions of autonomy is the basis of Wolff’s and Raz's disagreement on the possibility of a legitimate authority. The central question of this thesis is therefore: How do the conceptions of autonomy, that Wolff and Raz have, affect their disagreement on the possibility of its existence under political authority?

See for example Ani, 2014; Perkins, 1972; Raz, 1989; 2009.

(6)

To answer this question, I summarize Wolff’s challenge to the possibility of legitimate authority and Raz's answer to this conception of authority in the second section. This leads me to reject Wolff’s and Raz’s conceptions of authority as the (sole) basis for their disagreeing conclusions. It follows then, that autonomy needs to be the main focus when comparing the two theories. Therefore, I consider the two authors’ conceptions of autonomy in section three (Wolff) and four (Raz). This is followed by a comparison of the two conceptions of autonomy in section five. In section six, I trace how the authors’ conception of autonomy informs their different conclusions.

Finally, I conclude that Wolff’s and Raz's difference in the way they conceive autonomy indeed affects their disagreement on the possibility of its existence under political authority. 2 Wolff’s conception does not allow for any outside coercion of an individual, dismissing the possibility of a legitimate authority. For Raz, an autonomous person requires the conditions of autonomy, which a legitimate authority has a duty to facilitate for its subjects. The disagreement between Wolff and Raz is based mainly on the different requirements of —, and purposes ascribed to, autonomy. The more nuanced nature of Raz’s conception of autonomy enables him to put forth a more coherent account of the legitimacy of authority. 

From this point onwards, I shall use authority when I consider political authority.

(7)

§ 2 Wolff’s Challenge and Raz's Reply

§ 2.1 Wolff’s Challenge to Justified Political Authority

As stated in the introduction, Robert Paul Wolff challenges the existence of legitimate political authority in his In Defense of Anarchism. Political authority is incompatible with “the moral autonomy of the individual” (Wolff, 1970: 3). Wolff starts his challenge by first describing the properties of a hypothetical legitimate authority of a state. He continues by arguing that any individual has the obligation to be morally autonomous. Finally, he asserts the incompatibility of both having autonomy and giving up final judgement of one’s actions, which would be required under a legitimate authority (Wolff, 1970: 10).

Wolff describes a state as “a group of persons who have supreme authority within a given territory or over a certain population” (Wolff, 1970: 4). A state can have de facto authority, but only a legitimate state has de jure authority. A legitimate authority’s right to be obeyed must be understood as a normative claim, i.e. de jure in the sense of “the moral right to rule” (Wolff, 1970: 6), rather than a statement of a de facto situation. De facto authority exists when authority is acknowledged, whilst de jure authority is to have the right to authority (Wolff, 1970: 5). That people accept a de facto authority and obey its commands due to their acceptance, is not enough to prove that a de jure authority can exist. Wolff claims that, as it is a non-empirical concept, the existence of a legitimate authority must be proven a priori, rather than shown through pointing out its instances. Wolff questions if there is any possibility of (an a priori) existence of a “moral right to rule” (Wolff, 1970: 6).

Wolff, considering a state with supreme authority, claims that since de jure authorities would have the moral right to rule, a legitimate state would have a supreme moral right to rule. Wolff

(8)

argues that a de jure, moral right to rule should be perceived as an authority having the right to issue commands that correlate with its subjects’ duty to obey (Wolff, 1970: 4). Subjects’ obedience to 3 commands is based on the authority possessed by a person or persons, rather than on their agreement with the content of a demand (Wolff, 1970: 5). In other words, authority is content-independent. Wolff does not list possible limitations to a legitimate state’s power, he merely notes that the limitations of affairs over which a state may claim authority, is one of the questions political philosophy must answer (Wolff, 1970: 5). Instead, he seems to imply that a legitimate state would have a nearly unlimited authority to make decisions for its subjects in a wide range of topics. On the basis of this assumption, he argues that an authority negates its subjects’ autonomy.

Wolff argues for this by offering an interpretation of the Kantian argument for the necessity of the moral autonomy of individuals. Wolff claims that people can only be held accountable for their actions when they are responsible for them. This requires that they are autonomous or, in other words, self-legislating (Wolff, 1970: 8). This is the only way in which they can be held morally responsible for their actions. Wolff extends this argument by claiming that whilst individuals can temporarily surrender their autonomy, they cannot give up the responsibility for their actions or the consequences that follow from those actions. For him, “responsibility for one’s actions means making the final decisions about what one should do. For the autonomous man, there is no such thing, strictly speaking, as a command” (Wolff, 1970: 9). Wolff asserts that whilst autonomous individuals can take others’ arguments into account in their deliberation, they cannot obey another’s command (Wolff, 1970: 9). He concludes that there is no possibility of a legitimate authority, as no moral right to rule over autonomous people can exist.

However, it can be questioned whether Wolff’s conclusion can be sustained. He argues that the negation of autonomy automatically follows from a state’s supreme authority. It can be Not everybody assumes the automatic connection between a right to command and a subject’s obligation to obey.

3

(9)

questioned if this also necessarily follows for any form of authority whatsoever. Wolff’s legitimate state seems to be authoritative in nature, as it has supreme power to dictate its subjects’ behaviour, which results in an absence of autonomy (Wolff, 1970: 5, 10). The authority Wolff grants to a state, covers such a wide range of matters it may determine for its subjects within the borders of its territory, that if it were real, very few people would consider it legitimate. It results in a state with a potential for such invasive power, that its subjects would experience no sense of freedom. Wolff’s conflation of any form of political authority with supreme state authority leads to questions about the usefulness of Wolff’s conception of authority. I will further discuss this problem in section four of this essay.

For now—returning to Wolff’s argument—the autonomous agent’s rejection of the duty to obey authoritative commands, is the crux of the dilemma he presents. At the end of In Defense of

Anarchism, he concludes that “the concept of a de jure legitimate state would appear to be vacuous”

(Wolff, 1970: 34). The impossibility of a legitimate state is neither due to an imperfect rationality, nor due to a subject’s limited knowledge. Even in a democratic Utopia of “a society of rational men of good will,” Wolff cannot imagine a possibility for a de jure authority to exist (Wolff, 1970: 34). He therefore sees no other alternative than a doctrine of philosophical anarchism for any reasonable person. Philosophical anarchism, Wolff explains, is the categorical denial of “any claim to legitimate authority by one man over another” (Wolff, 1970: 35). Consequently, Wolff cannot 4 discern an a priori possibility for the existence of a legitimate authority, as his view of the incompatibility of political authority with individual moral autonomy precludes this.

The challenge Wolff sketches is a serious one. If there is no such thing as an a priori justified political authority, it seems only right we should question our commitments to the de facto

Philosophical anarchism can be contrasted with political anarchism. Advocates of the former argue that there is no

4

moral reason to act in accordance to a law, advocates of the latter often make the stronger claim that there is a moral obligation to oppose states and ultimately replace them with a non-statist organisation of society (Stilz, 2016: 2).

(10)

authorities that do exist. However, whilst Wolff’s entire argument is based on authority’s incompatibility with individual moral autonomy, he does not provide a specific definition of that autonomy. In In Defense of Anarchism, Wolff refers to Immanuel Kant’s self-legislating man and asserts that a moral autonomous agent is both free and responsible. However, these requirements remain vague, as he does not explicate their scope or precise relationships.

§ 2.2 Raz's Defence of Legitimate Authority

Joseph Raz’ work could be seen as answering Wolff’s challenge. He argues in favour of a certain conception of authority, the service conception of authority, and asserts that a legitimate authority does not negate, but is instead compatible with autonomy. Raz makes an even stronger claim: legitimate authority is required to provide an maintain conditions for personal autonomy for its subjects.

In agreement with Wolff, Raz endorses that “authority is a right to rule” (Raz, 2006: 1012). In opposition to Wolff, however, Raz argues that political authority can be legitimate. For it to be so, it must adhere to what he calls the ‘service conception of authority’ (Raz, 2006: 1014). According to this conception of authority, a political authority is legitimate when it facilitates its subjects’ well-being better than they could do themselves. A legitimate authority helps its subjects align their actions better to the reasons that apply to them than if they were to do this on their own (Moore, 1989: 842; Raz, 1986: 425). Raz's conception of a legitimate authority is instrumental, as an authority’s legitimacy is dependent on the outcome of its actions (Green, 2012: 15-6). Raz claims that legitimate authority is legitimate due to it serving its subjects.

Raz's theory of authority is a part of his general moral theory which concerns “the advance-ment of well-being of individuals” (Raz, 1986: 267). Rather than an assumption of pure self-inter-est, an individual’s well-being depends on her success in accomplishing her comprehensive goals.

(11)

Comprehensive goals are those goals that are imbedded deeply in one’s life and that affect a longer period and multiple aspects of it (Raz, 1986: 293). These goals are based on social conventions (Raz, 1986: 289, 309). An example of such a goal is marriage. A person might want to get married, or not, or non-traditionally, but all these goals exist in reference to the convention of marriage. Raz’s service conception of authority claims that a legitimate authority helps align its subjects’ ac-tions to reasons applicable to them, which are mainly determined by the subjects’ comprehensive goals.

Raz's argumentation for the service conception of legitimate authority is built on three theses. He claims that combined, these theses explain that individuals have “a duty to subject one’s will and judgment to those of another” (Raz, 2006: 2012). These theses are the normal justification thesis, the (in)dependence thesis and the pre-emption thesis. The relationship of the theses could be 5 envisioned as a triangle: they all suppose and depend on each other. Raz assumes that if an authority is proven to adhere to these theses, its authority is legitimate.

The normal justification thesis states that subjects “would better conform to reasons” that already apply to them when guided by an authority’s directives than when they would not (Raz, 2006: 1014). Raz assumes that people normally justify the existence of authority in this way, hence the name ‘normal justification’ thesis (Raz, 1986: 53). When proven, Raz argues that it demonstrates that there are instances in which there are compelling reasons to accept authority. However, the normal justification thesis does not shed light on the relation between an authority’s directive as a reason for an individual to act and other possible reasons for (different) actions (Raz,

Throughout his work, Raz has employed different structures of his argument. In The Problem of Authority the first two

5

theses are introduced as ‘conditions’ to legitimate authority, whereas the pre-emption thesis is introduced afterwards as something inherent to authority’s nature (Raz, 2006: 1014, 1019). In Morality of Freedom the pre-emption thesis is in-troduced first and as a condition to legitimate authority. The conditions of the dependence and normal justification the-sis are presented subsequently (Raz, 1986: 46-7, 53). I have chosen to adhere to neither, instead I adopt the ordering from The Problem of Authority and the structure of Morality of Freedom, as I think it allows for more clarity and sup-ports the claim Raz makes that authority’s pre-emptive force should only be granted to legitimate authority (Raz, 1986: 46).

(12)

1986: 56-7). This is what the (in)dependence thesis contributes to Raz’s service conception of authority.

Raz refers to the second thesis as the dependence thesis in earlier work and as the independence thesis in more recent articles. They differ slightly. The focus of the independence thesis is on the matters for which an authority may issue directives. It claims that a legitimate authority should be concerned with only those matters in which its subjects require help in aligning their actions to reason (Raz, 2006: 1014). The dependence thesis on the other hand claims that:

all authoritative directives should be based on reasons which already independently apply to the subjects of the directives and are relevant to their action in the circumstances covered by the directive (Raz, 1986: 47; emphasis added).

The focus of the dependence thesis is on the reasons an authority should take into consideration for its directives. It claims that an authority’s reasoning must be dependent on the reasons for action which subjects would consider when deciding how to act. 6

Despite the difference in focus, the dependence and independence theses share the same function. They both direct the scope of matters an authority can decide upon and the suitability of its directives. For Raz such directives can only concern those matters that subjects are aided by in aligning their actions to reasons that already apply to them. Both theses therefore claim the need for some alignment in an authority’s considerations to the reasons applicable to its subjects. Due to their shared function and slight difference in focus, in this article, I refer to the thesis as the ‘(in)dependence thesis’. 7

Later, Raz adds a caveat to this thesis. Rather than only based on reasons already applicable to its subjects, an authori

6

-ty’s directives should at least have taken some of these reasons into account. In this formulation, it is not necessary that only or all the reasons applicable to its subjects are considered. Rather, an authority’s reason to act might and may differ from its subjects’. But directives can still only be given for matters that its subjects need assistance with.

It is open for question how feasible Raz’ (in)dependence thesis is. Could an authority have the possibility to both de

7

-termine in which matters its subjects require assistance and de-termine the reasons applicable to its subjects? If this tion is answered with a no, one could doubt the usefulness of Raz’s service conception of authority. I set aside this ques-tion for now, but it might be good to consider it in further research.

(13)

Where the normal justification thesis considers the reasons applicable to subjects for conformity to authority, the (in)dependence thesis concerns the matters about which directives can be given by a legitimate authority. The normal justification thesis and (in)dependence thesis ground Raz's justification of an authority’s claim to be obeyed, but they do not explain how subjects should regard its directives.

Raz's explication of subjects’ attitudes to an authority’s directives is based on the pre-emptive thesis. This thesis states that a directive of a legitimate authority must replace other reasons subjects might have for acting in a certain way (Raz, 1986: 46). More precisely, reasons that apply to subjects, which could normally challenge the reason for acting in accordance with the authority’s directive, must no longer be considered as possibilities to pursue, on the basis of the directive coming from the legitimate authority (Raz, 2006: 1019).

According to Raz, only when its subjects consider an authority’s directives as pre-emptive, can one speak of the existence of authority. First of all, any non-pre-emptive command or directive would only be considered advice or as one of many reasons considered for acting. On top of this, Raz claims that regarding directives as non-pre-emptive, would lead to double-counting a reason to act. Raz argues that an authority’s directive provides a new reason for a subject to act in a certain way, based on other reasons. However, if both the directive and the reasons that are the basis for the directive are a part of a subject’s reasoning for a certain course of action, the latter is double-counted (Raz, 1986: 58). For Raz, directives of a legitimate authority pre-empt other reasons for action.

Having elaborated his three foundational theses, Raz continues his argument by stating that a legitimate authority will make mistakes. Still, he claims that even when it does, its directives should be followed and it should still be considered legitimate (Raz, 2006: 1022-3). He does so from the claim that the probability that a legitimate authority’s overall contribution to subjects’ lives

(14)

is higher than if subjects should everything decide for themselves. Thus, whilst Raz's conception of authority is instrumental, as asserted above, it is also content-independent. It is content-independent in the sense that subjects should not consider the reasoning or merit of every directive of a legitimate authority. Instead, they should accept its directives, as they originate from the authority. Raz argues that if subjects would reflect on each directive they have doubts about, they would not approach the authority’s directives as pre-emptive (Raz, 1986: 61). There would no longer be an actual authority to speak of.

However, this claim results in an inconsistency in Raz’s theory. Raz’s instrumental service conception in effect does not allow subjects to question directives and act against their authority as soon as they perceive it as legitimate. This exposes a problematic aspect of Raz’ conception of authority. It leaves little space for subjects to contest directives without immediately contesting an authority’s legitimacy. Even further, if the subjects of an authority should not reflect on that authority’s legitimacy (by questioning and not acting on its directives), then how can they ever determine if a de facto authority is also a de jure authority?

Asserting the need of unreflective acceptance for the functioning of a legitimate authority, Raz effectively claims that a subject must follow directives of an authority as soon as she considers it legitimate. After all, he claims that when a subject has at one point considered as authority as legitimate, this prevents her from acting against that authority’s directives. Consequently, if subjects themselves should not determine legitimacy of an authority based on the service it gives provides them, how can legitimacy be questioned once it has been given to an authority. Should such a task be given to outsiders to determine whether an authority serves its subjects’ adherence to reasons applicable to them? This seems disputable. Why would one assume that outsiders, rather than subjects themselves should not question an authority’s directives, which are the basis for that authority’s legitimacy?

(15)

Leaving little space for subjects to question and act against the directives of the authority they consider legitimate, Raz’s service conception seemingly grants legitimacy to any de facto authority due to the strength of his pre-emptive thesis. As soon as a subject assumes that that a authority has the right to rule, she therefore assumes its legitimacy since she follows the normal justification thesis. Due to the preemptive thesis, however, that subject is unable to question or act on any further doubts about the authority actually meeting both the (in)dependence and the normal justification thesis, going forward. Raz proposes a service conception of a legitimate authority which is instrumental due to its focus on outcome (service to subjects), founded upon the normal justification thesis, the (in)dependence thesis and the pre-emptive thesis, and is content-independent. Though Raz prescribes a weak or limited conception of authority, he does not offer subjects much room to decide when an authority is not legitimate, leaving his theory of legitimate authority with a serious (authority) problem. I return to this problem in paragraph 6.2.

Having described his conception of authority, Raz continues his answer to Wolff by claim-ing that personal autonomy is not excluded by a legitimate authority. This step is based on both his service conception as well as his assumption that, contra Wolff, autonomy should not be seen as just the absence of outside coercion (Raz, 1986: 207). Rather, autonomy is the actualization of three conditions: the mental ability to choose and plan, having an adequate range of options, and inde-pendence as freedom from other’s coercion and manipulation. Only if an agent actualizes the condi-tions for autonomy by choosing the kind of life she wants, does personal autonomy come about (Raz, 1986: 373). As we shall see, for Raz, a legitimate political authority must promote personal freedom, which presupposes both value-pluralism and autonomy (Raz, 1986: 367). He claims that a legitimate authority does not exclude personal autonomy, rather, it must actively try to bring about the conditions of autonomy for its subjects.

(16)

§ 2.3 Two Different Conceptions of Autonomies

Wolff and Raz thus have different conclusions to the question if autonomous subjects can exist under authority. However, there is considerable overlap in their conceptions of authority; both assume authority as right to rule and as tied to subjects’ duty to obey. Both also assume that directives of authority are accepted on the basis of them coming from the authority, rather than from the reasoning on which the directives are based. Of course, Wolff’s description of authority is more expansive, whilst Raz's authority conception seems quite limited. At the same time, both Wolff and Raz propose, effectively, a content-independent conception of authority. The disagreement in their judgement of the possibility of autonomy under authority must therefore be based on more than just the differences in their authority conceptions. Their disagreement might be based on their conceptions of autonomy.

Before continuing, however, it must be noted that Wolff and Raz use different adjectives to describe their autonomy conceptions. Wolff uses ‘moral’ and Raz ‘personal’. Raz's acknowledges that there is a difference in concepts of autonomy between Wolff (moral autonomy) and himself (personal autonomy). Raz argues that instead of a “doctrine about the nature of morality,” like moral autonomy, personal autonomy is concerned with individuals’ capacity to be (partially) in control of their own life (Raz, 1986: 370n2). Regardless, he does seem to claim that his theory can 8 be seen as overcoming Wolff’s challenge to the existence of a legitimate authority (Raz, 1986: 97; 2009: 27). Raz does not focus on the different concepts of autonomy and instead continues with the formulation of authority. However, this difference in the conception of autonomy explains for a large part the difference in their assessment of the possibility of a legitimate (state) authority.

Due to the partial overlap in Wolff’s and Raz's conceptions of authority and due to Raz's acknowledgement of difference in their conceptions of autonomy, a closer look at the two

At the beginning of section four, I explain the difference between moral and personal autonomy.

(17)

conceptions of autonomy is required. Therefore, the main focus of the rest of this paper will be on Wolff’s and Raz's conceptions of autonomy and the implications it has for (legitimate) authority.

(18)

§3 Wolff’s Moral Autonomy

In this section, I start with a consideration of Wolff’s conception of moral autonomy in In

Defense of Anarchism. Wolff’s conception of autonomy is moral, as he introduces it as the

requirement for individuals to be held accountable for their actions. I continue by exploring

Autonomy of Reason, a work in which Wolff explains Kant’s Groundwork. From this book, I

characterize Wolff’s conception of autonomy more extensively, through Wolff’s opposition to Kant. Moral autonomy for Wolff is having the possibility to rationally decide on one’s own policies. Furthermore, I review some criticisms on Wolff’s conception of autonomy, to both consider others’ portrayal of Wolff’s conception of autonomy and to see if there are possible problematic aspects in his conception.

§3.1 Moral Autonomy in In Defense of Anarchism

Wolff argues that an agent must act autonomously in order to be morally accountable, he claims that therefore, autonomy is a duty. Being morally accountable implies responsibility for one’s actions. An agent’s actions must be her own to be accountable for them. Wolff continues his argumentation by stating that such responsibility can only exist if agents are free to choose their actions and able to reason about their choices (Wolff, 1970: 8). For Kant, since rational beings are morally accountable, they have the duty to be autonomous. Wolff claims to follow Kant in assuming that a person who is free to choose and able to reflect on her choices has the obligation to do so. A moral person must be the sole judge on any constraints on her actions; put differently, she must be morally autonomous.

(19)

Wolff characterizes moral autonomy as follows:

Since the responsible man arrives at moral decisions which he expresses to himself in the form of imperatives, we may say that he gives laws to himself, or is self-legislating. In short, he is

autonomous. As Kant argued, moral autonomy is a combination of freedom and responsibility;

it is a submission to laws which one has made for oneself. The autonomous man, insofar as he is autonomous, is not subject to the will of another. He may do what another tells him, but not

because he has been told to do it. He is therefore, in the political sense of the word, free

(Wolff, 1970: 8; emphasis in original).

In other words, in order to be autonomous an agent has to be the final judge and author of her actions. To be able to do so, she must be free to decide. This freedom comes from the absence of outside coercion. For Wolff, autonomy is not a matter of degrees; either an agent makes the final decisions in her life, or she does not. An agent is either autonomous or not. Only when an agent has the freedom to decide for herself, without another infringing upon that freedom, can she reflect on 9 her options and is she responsible for her actions (Wolff, 1970: 8). Since we assume that people are responsible for their actions, Wolff concludes that agents have the duty to be autonomous.

Wolff argues that this duty remains throughout an agent’s life, although one can suspend one’s autonomy in a certain context. For example, if you are in a burning house and a you obey a firefighter who tells you to jump out of a specific window without further reflection, you temporary give up your autonomy. Since there is no reflection on which course of action to take, you do not act autonomously. Your final decision is based on being told to do something by another individual with authority, not through your own reasoning. You could have jumped out of another window, but to act on the basis of the firefighter’s authority without reflection, you follow her command.

However, for Wolff, such a temporary surrender of decision-making is both bound by time and place, and does not absolve an agent from the responsibilities for her actions. So, even when you act because another says so, it is still your responsibility. Autonomy as a duty is an absolute factor in one’s life, according to Wolff (Wolff, 1970: 9). Only an agent taking final responsibility for Wolff’s account of freedom as the absence of outside interference in decisions, can be seen as an example of negative

9

freedom. Put shortly, negative freedom is freedom from obstacles and restrictions (Carter, 2019). I return to this obser-vation in paragraph 5.1.

(20)

her actions through formulating her own decisions, adheres to the duty of autonomy and is truly autonomous.

Wolff expands his argument by claiming that an agent remains autonomous when she is following another’s command in the case of a life-threatening situation, as long as she reflects on her decision to obey that other. The example he gives here is of an agent following the orders of the captain of a sinking ship. All the other passengers follow the captain’s orders, it is therefore prudent to follow them also. According to Wolff, if the agent decides for herself to follow the captain’s orders, even without further reflection on those orders, the agent remains autonomous. He defends this position by pointing out that the agent decides for herself to follow the commands (she is not forced) and that her adherence to the commands is not based on the captain giving them. It is simply the most reasonable course of action; she would have followed another’s command, whether or not that person is the captain (Wolff, 1970: 9). In short, for Wolff an agent is autonomous if she, based 10 on reflection, decides her own policies in life. At the same time, autonomy can remain even in situations where another makes the final decisions for her actions.

§3.2 Moral Autonomy in The Autonomy of Reason

In The Autonomy of Reason: A Commentary on Kant’s Groundwork of the Metaphysics of

Morals, Wolff attempts to provide further insights into Kant’s Groundwork, whilst also rectifying it

where he thinks it is flawed (Wolff, 1973: 4-5, 84-5). As the title of his work suggests, Wolff considers autonomy the pivotal element of Kant’s moral theory (Wolff, 1973: 178-9).

Exploring this book is of interest due to Wolff’s juxtaposition between Kant and himself. It uncovers an analogy between Wolff’s conception of an autonomous state and of an autonomous in-dividual. The juxtaposition also facilitates further description of Wolff's conception of autonomy as

Both Harry Frankfurt and I notice that Wolff’s expansion seems to lead to an inconsistency in his conception of auto

10

(21)

a certain discretion to rationally determine one’s own policies and actions. To be able to extract Wolff’s conception of autonomy, I consider his explanation of Kant’s conception of autonomy and, based on Wolff’s commentaries, I lay bare some of the features of his own conception of autonomy.

Wolff identifies Kant’s principle of autonomy as: “Never choose except in such a way that in the same volition the maxims of your choice are also present as universal law” (Wolff, 1973: 177; citing Kant, Ak. 440). Wolff criticizes this formulation as too vague. However, the main source of his critique stems from his dissatisfaction with Kant's explanation of the formulation: “Kant turns out not to be saying what I want him to say” (Wolff, 1973: 178). Kant begins his explanation of au-tonomy by saying that an agent must always be considered as the maker of the law for herself. This is in line with Wolff’s conception of autonomy (Wolff, 1973: 178; based on Kant, Ak. 431). How-ever, Wolff disagrees with Kant’s emphasis on the need for disinterest in making laws for oneself as an essential part of autonomy. For Kant, an agent should not legislate from her particular interests, but independently from them (Wolff, 1973: 179).

Wolff points out that that Kant’s explanation of autonomy as making disinterested laws for oneself is based on Kant’s presentation of autonomy as a formulation of the categorical imperative. An imperative is a law for oneself, a policy for one’s life. An imperative is hypothetical if it is based on a rational agent’s interest; adherence to such an imperative is based on reasons that are not inher-ent in that law. This means that abiding by the law is conditional on the right circumstances, mind-set, etc. of the agent (Wolff, 1973: 179). A categorical imperative is the opposite of that. Adherence to a categorical imperative is based on reasons inherent in the law. No personal interest is the reason to adhere to that imperative. A categorical imperative must therefore be universal; every rational being has the same reason for adherence, which is found in the law (Wolff, 1973: 126-33). Kant in-sists that moral autonomy revolves around choosing an imperative that is not dependent on contin-gencies, and, as a result, can be universally applied. Wolff asserts that for Kant, if an agent can

(22)

submit to an imperative without adhering to her interests, this is only possible if the imperative would be self-legislated. Following another’s laws will necessarily remain conditional, as the fol-lowing would be based on the agent’s interests (for example, an agent wanting to prevent that she comes to harm by not following the law or wanting to please the one who established the law). Therefore, a categorical imperative must be self-legislated. (Wolff, 1973: 180). According to Kant, moral autonomy, presented as a formulation of the categorical imperative, must be both self-legis-lated and universally applicable.

Wolff further explains why Kant depicts moral autonomy as universally applicable, through Kant’s distinction between heteronomy and autonomy. Wolff explains that for Kant, a het-eronomous agent’s will is bound to objects of volition rather than wills external to the self (Wolff, 1973: 187). An agent bound by objects of volition acts from a hypothetical imperative. Bondage to external objects of volition arises due to an agent’s own interests in bringing about certain objects, which are outcomes that occur outside herself. For Kant, heteronomy occurs due to desire, for ex-ample a desire for happiness. It can also occur due to an agent’s rational conception, for exex-ample the principle of perfection (Wolff, 1973: 187). If an agent makes policies for herself from the desire to be happy or to ensure something is perfect as measured by predetermined, external factors, her will is heteronomous.

By formulating Kant’s explanation of heteronomy as such (as bound by objects of volition instead of external wills), Wolff implies an alternative interpretation of heteronomy. Instead of ing bound by objects of volition, he seems to claim, a heteronomous agent could also be seen as be-ing bound by external wills. He does so in reference to Kant’s assumption that a person must legis-late an imperative herself, if it is to be categorical. Kant’s conception of will comes from ‘willed,’ which is similar to choosing from a certain principle (Korsgaard, 1998: xi). According to Kant, a will is a rational being’s “capacity to act according to the principles provided by reason”

(23)

(Mc-Cormick, n.d.: §8b). Wolff concludes that for Kant, autonomy being the opposite of heteronomy, an agent’s will is autonomous only when it is not influenced by outside objects of volition, it must therefore be universal (Wolff, 1973: 187).

That Wolff’s interpretation of autonomy is distinct from Kant’s, becomes more apparent by his dissatisfaction with Kant’s argumentation for autonomy’s universal applicability. He claims that Kant’s explanation does not explain why self-legislation by an agent creates an unconditional necessity to follow one, universal law for that agent (Wolff, 1973: 181). Wolff argues that Kant engages in two incompatible doctrines. The first is Kant’s assumption that all rational agents recognize and adhere to “objective, substantive, categorical moral principles” (Wolff, 1973: 181). For Wolff, this leaves the notion of agents’ self-legislation vacuous, since they cannot decide for themselves what law to legislate. There is, in effect, nothing to choose from, as there is only one universal law according to Kant. Wolff argues that such a formulation predetermines an agent’s policy choice. The other doctrine Wolff discerns, is Kant’s belief that only to the extent that a rational agent’s free choice is the reason for her adoption of a policy, should she considered herself bound by it. Wolff argues that a belief in the latter necessarily dismantles a belief in the former doctrine. Either there is free choice in policy, or there is the same moral law for all rational agents. Wolff argues that the content of moral principles must be determined on the basis of commitment to freely self-formulated goals (Wolff, 1973: 181). In contrast to Kant, Wolff champions the need for free choice of policy for rational agents.

Wolff concludes his explanation of Kant’s formula of autonomy by stating that Kant assumes objective laws that bind the content of autonomy for every rational agent. In opposition to Kant, Wolff claims that such objective laws are merely formal criteria that ensure that policies are not inconsistent. They are negative laws, excluding the incorporation of non-rational policies, never contributing to selection of specific rational policies. Recapping, Wolff agrees with Kant that agents

(24)

are bound by life-policies only when these are legislated by agents themselves. However, Wolff argues that agents should not do away with the influence of their interests. Rather, according to him, agents must have an actual choice when formulating their own rational policies (Wolff, 1973: 181). They should reject the influence of others’ wills upon one’s own. Wolff, contra Kant, considers an autonomous agent as free from bondage to external wills (Wolff, 1973: 187). Rather than freedom from objects of volition, Wolff assumes that autonomy requires freedom from “wills external to the self” (Wolff, 1973: 187). Freedom from others’ wills, is freedom from influence of others’ choices on one’s own. Wolff argues that autonomous agents will be influenced by their own interests, 11 since to be truly free to decide, agents must be able to choose different (rational) policies.

It is noteworthy that Wolff’s conception of autonomy shows an analogy with his conception of (state) authority. For Wolff, an autonomous agent has a nearly limitless power to determine policies for herself. The only restriction the autonomous agent encounters is rationality. Wolff conceives authority, or rather a state authority, as having a nearly limitless power to determine policies for its subjects (Wolff, 1970: 4). However, this is not a perfect analogy, as Wolff does not seem to place restrictions on the power of authority, as he does for agents. Rather, Wolff seems to be convinced that the only restrictions such an authority has, are its territorial borders. Within those borders, it seems, anything goes for an authority’s directives. All the same, it can be claimed that for Wolff an autonomous agent has nearly limitless authority to determine her own rational policies and actions.

Combining insights from Autonomy of Reason and In Defense of Anarchism, Wolff considers an autonomous agent as deciding for herself what life-policies to act from. According to him, there are objective principles that exclude an autonomous agent from selecting inconsistent,

As in In Defense of Anarchism, Wolff’s understanding of freedom here can be labeled as negative (negative freedom

11

is freedom from obstacles and restrictions (Carter, 2019)). Here, Wolff seem to argue that autonomy is freedom from the restrictions of others interfering in an agent’s choice.

(25)

non-rational policies. Wolff argues that this consequently results in a freedom to choose her own actions, as long as she reflects on their rationality in relation to her goals. In opposition to Kant, for Wolff there is a variety of valid reasons for an autonomous agent to reason from when determining one’s own policies and actions. Wolff does not adhere to Kant’s view of autonomy as freedom from external objects. Instead, he considers autonomy as freedom from other’s wills. The freedom to choose from a variety of policies and actions, is presented by Wolff as a duty of all those who can be autonomous. Due to that duty to freely choose, Wolff excludes the possibility of legitimate authority, because obeying an authority is obeying another’s will. Any (state) authority, due to having the power to make policies for its subjects, necessarily infringes upon the authority of an autonomous agent over herself. Under an authority, an agent’s policies cannot be determined by herself. Therefore, Wolff concludes, authority negates any agent’s moral autonomy.

Wolff lays bare some of the features of his conception of autonomy by taking distance from Kant and by contrasting autonomy to authority. However, the above interpretation of Wolff’s conception of autonomy is not based on Wolff clearly explaining what it is or on his explicit argumentation, but on my deduction from his opposition to Kant and from his minimal description in In Defense of Anarchism. This leaves his conception of autonomy unclear still.

§3.3 Critiques of Wolff’s Moral Autonomy

Since Wolff does not give a complete conception of autonomy, though one can deduce its central aspects by combining his arguments of In Defense of Anarchism and The Autonomy of

Reason. Therefore, it seems worthwhile to consider another avenue to shed light on it. I do so by

considering some critiques of Wolff’s works on autonomy and authority. These critiques show how others interpret Wolff’s conception of autonomy. On top of this, if their arguments hold, this might influence the validity of Wolff’s conclusion that autonomy precludes the possibility of legitimate

(26)

authority. The critiques presented here fall into two categories. The first and most important shows an inconsistency in Wolff's autonomy conception, already shortly discussed above. The second category questions if Wolff's conception of autonomy is ‘Kantian’.

§3.3.1 Critique of Wolff’s Moral Autonomy: Inconsistency

As mentioned in a footnote above, both Harry Frankfurt and I notice an inconsistency in Wolff’s explanation of autonomy. In In Defense of Anarchism, Wolff argues that autonomy is something that exists in absolute only: “moral autonomy is simply the condition of taking full responsibility for one’s actions. […] Taking responsibility for one’s actions means making the final decisions about what one should do.” (Wolff, 1970: 9). This requires that an agent reflects on the consequences of her possible actions and makes her decisions based on such reflections. Wolff’s conception points to him perceiving autonomy as an absolute: an agent is either autonomous and makes the final decisions for her actions, or she is not.

The inconsistency that both Frankfurt and I observe, presents itself in Wolff’s example of an agent retaining autonomy when following a captain’s commands on a sinking ship. This example expands the scope of his conception of autonomy considerably. Wolff claims that an agent can remain autonomous if she follows the captain’s commands out of prudence, as her life is in danger and others already act accordingly. Wolff claims that the agent does not have to reflect on every forthcoming command. Instead, autonomy’s required reflection is present in her earlier consideration whether to follow the commands or not. On top of this, Wolff asserts, her following the commands is based on prudence, not the perceived authority of the commanding person (Wolff, 1970: 9). In such a life-threatening situation, according to Wolff, an agent can still be autonomous whilst following commands.

(27)

Frankfurt notes, however, that whilst it might be prudent to follow the orders, the actions in the given example cannot be said to be in line with an agent always taking the final decision for her actions (Frankfurt, 1973: 409-10). Wolff claims that having autonomy requires making all decisions about ones actions oneself on the basis of one’s reflection (Wolff, 1970: 9). However, if it is sufficient to make a single autonomous decision to hand over one’s final decisions to another, it is not clear how Wolff can maintain that authority negates autonomy.

Wolff might argue that only if an agent consents to obedience, she may act unreflectively on command. An agent would still be autonomous and therefore accountable for her actions. However, repeating Frankfurt’s argument, if this would be the case, autonomy would no longer require an agent to make her final decisions (Frankfurt, 1973: 409-10). Furthermore, if consent to obedience would not negate autonomy, it is unclear why authority negates autonomy. Taking the above argument to its logical conclusion, one could argue that as long as agents, after reflection, decide for themselves to follow an authority’s commands, they preserve their autonomy. Therefore, Wolff’s conclusion —authority necessarily negates autonomy— would not follow from autonomy conception, as long as he maintains that an agent can remain autonomous when following an authority’s command in a life-threatening situation.

Either autonomy means making the final decisions about one’s actions and is therefore incompatible with authority, or autonomy allows surrendering some of the final decisions on one’s actions, as was shown in the example. The latter conception might allow the possibility of autonomy’s compatibility with authority. However, this conception flies in the face of Wolff’s earlier statement that an agent can never be absolved from the responsibilities of her actions.

(28)

§3.3.2 Critique of Wolff’s Moral Autonomy: Kantian

On top of the possibility of inconsistency, doubts can also be cast on Wolff’s claim to have a

Kantian conception of autonomy. Wolff uses his explanation of autonomy to reject the possibility of

autonomous agents existing within a political authority. Since such an authority will claim the right to rule, not all the actions an agent would perform under such an authority are expressions of her own policies. He argues that, on the basis of Kant, autonomy must be seen as a duty to all rational beings (Wolff, 1970: 10). Therefore, Wolff concludes, a legitimate political authority is an impossibility (Wolff, 1970: 11).

Those somewhat familiar with the Kantian notion of autonomy might be surprised by Wolff’s explanation and the consequences he supposes the duty of autonomy has on the possibility of legitimate authority. This surprise is shared by some of Wolff’s commentators (for example Oberdiek, 1977; Riley, 1978). Hans Oberdiek, for example, states in a review on Wolff’s The

Autonomy of Reason: A Commentary on Kant’s Groundwork of the Metaphysics of Morals that

Wolff’s work, though inspired by Kant, “does not adequately reflect Kantianism” (Oberdiek, 1977: 482). Rather than a mere commentary, what Wolff presents is a reconstruction of Kant’s moral theory (Oberdiek, 1977: 483). Oberdiek notices the difference between Kant and Wolff in their consideration of autonomy.

Oberdiek contrasts Kant and Wolff, stating that the former assumes autonomy to be freedom due to reason, whilst the latter assumes autonomy as be “the freedom from reason” (Oberdiek, 1977: 484; emphasis in original). Whilst in Kant’s autonomy reason is the way to rise above influence from one’s outside world interests, Oberdiek claims that Wolff’s autonomy is the freedom to follow one’s self-formulated principles, be they reasonable or unreasonable. However, I do not agree to this statement, as Wolff emphasizes reason as a requirement for autonomy, claiming that

(29)

there are objective laws that autonomous agents must follow in determining their policies (Wolff, 1973: 181).

Oberdiek is correct, however, in contrasting the two conceptions of autonomy. Wolff introduces his conception of autonomy in In Defense of Anarchism as Kantian, whilst at that time already having worked for almost two decades on The Autonomy of Reason in which he distances himself from Kant’s notion of it. Wolff seems to be cherry picking in Kant’s theory, whilst simultaneously wanting to enjoy the authority that the Kantian title gives his argument.

Taking the critiques together, Wolff’s conception of autonomy is rendered inconsistent and not Kantian. As it stands, his conception does not preclude the existence of an autonomous agent that obeys the commands of an authority. Wolff claims that an agent remains autonomous whilst not making the final decisions in her life, even when, in such a situation, the agent hands those decisions over to an authority. Finally, without the backing of the Kantian label Wolff ascribes to himself, his argumentation loses some of initial apparent strength.

(30)

§4 Raz's Personal Autonomy

I start this section with a consideration of Raz's conception of autonomy in more recent work (The Problem of Authority: Revisiting the Service Conception), after which I move to older work (Morality of Freedom) in which autonomy is discussed more extensively. Next, I consider some forms of criticisms which focus on Raz's conception of autonomy. All these steps lead to a extensive formulation of Raz's conception of autonomy.

Before I do so, however, it is important to note that in opposition to Wolff’s moral autono-my, Raz focusses on personal autonomy. These different adjectives point to a difference in the func-tion of autonomy in the authors' theories. Wolff’s moral autonomy revolves around objectively ra-tional, self-developed choices. For him, moral autonomy is necessary for an agent to be held re-sponsible for her actions. From the obligation to be rere-sponsible, he infers the obligation to be au-tonomous (Wolff, 1970: 8-9). Raz's personal autonomy is not as focussed on responsibility. Rather, Raz describes personal autonomy as an ideal for guiding one’s own life. To him, it is a requirement for individuals living a good life (Raz, 1986: 391).

§4.1 Personal Autonomy in The Problem of Authority

In The Problem of Authority: Revisiting the Service Conception, Raz's conception of auton-omy seems to resemble Wolff’s at first. Equalizing autonauton-omy and independence, he describes au-tonomy as the “general capacity to guide our conduct (and our life more broadly) by our own judg-ment” (Raz, 2006: 1017). However, Raz emphasizes autonomy’s connection to an agent’s reasons for action more explicitly than Wolff. For him, the capacity of autonomy allows agents to act in ac-cordance with reasons that are applicable to them. This is due to the agents’ exercise of judgement. Being able to judge and act upon their own judgement is valuable, but the actual value of the

(31)

capac-ity of autonomy is due to the purpose it has. For Raz, that purpose is alignment with reason. Ac-cording to him, autonomy means acting in accordance with the reasons as they apply to you. This acting in accordance with the reasons has to be intentional. The capacity needed for this is the com-bination of internal and external factors that make this possible for you (Raz, 2006: 1017).

Raz claims that an agent’s rational capacity to autonomy consists of the alignment of actions with reasons applicable to them by reflecting on situations in the world and deciding how to re-spond. He argues that autonomy has value due to the effect it has when successful: conformity to applicable reasons. Forming and executing their own judgements, however, is not the only option for agents to conform to reason. Instinctive or emotional, non-reflexive reactions can sometimes be preferable to a lengthier process of reflection (Raz, 2006: 1017). For example, if there is a lion in your hallway, quickly running away might be a more preferable reaction than standing around for a while to consider possible reactions. Alternatively, for Raz, agents can sometimes conform better to reason by obeying authority and basing their actions on the commands of the authority (Raz, 2006: 1017). The Problem of Authority gives some insights in Raz’s conception of autonomy, it is not pos-sible to distinguish a careful explanation of it from this article.

§4.2 Personal Autonomy in Morality of Freedom

In Morality of Freedom, Raz discusses, amongst other things, the interplay between authori-ty and autonomy. Raz argues that significantly autonomous agents make their own choices. Howev-er, the act of agents choosing is not sufficient for autonomy. As described above, agents must act in accordance to the reasons as they apply to the agent. Autonomous agents "adopt personal projects, develop relationships, and accept commitments to causes, through which their personal integrity and sense of dignity and self-respect are made concrete” (Raz, 1986: 154). Raz describes such choices as “comprehensive choices” (Raz, 1989: 309). In this process of forming and pursuing

(32)

these, agents determine their own conception of a successful life, i.e., they determine what sort of life they consider a life well lived (Raz, 1986: 154, 387). Being able to answer this question on one’s own terms is a criterium of a good life according to Raz. Autonomy is not achieved through a single choice, but can only be achieved by making successive choices throughout one’s life in order to successfully pursue goals, projects and relationships that are chosen by the agent herself (Raz, 1986: 369-71).

To be sure, Raz does not take such self-determination as existing in a void. He argues in-stead that people’s comprehensive goals are at least partially based on what he calls ‘social forms’. For Raz, an agent can form “a comprehensive goal only if it is based on existing social forms, i.e. on forms of behaviour which are in fact widely practised in his society” (Raz, 1986: 308). This is not to say that agents can only copy the behaviour seen in their society. Rather, their goals will be based on behaviour in a society, whether they are copied straightforwardly from, outright rejections of or variations on perceived behaviour in their society (Raz, 1986: 309). Raz argues that only through the connection of comprehensive goals to social forms, an agent can value such goals (Raz, 1986: 310).

Social forms, along with natural forces (like mountains, tsunami’s and rain), will at the same time confront agents in such a way that they can never be wholly autonomous. Raz approaches au-tonomy as an ideal that cannot be fully reached, as both natural and social forces will require some compromises in agents’ choices and fulfillment of their goals. According to him, autonomy is there-fore a matter of degree. An agent can be more or less autonomous, depending on her (history of) decisions in regard to goals, pursuits and relations that she has chosen for herself as valuable (Raz, 1986: 154, 319, 371).

According to Raz, on top of an ideal one tries to achieve, autonomy is also a value that en-ables one to become the author of one’s life (Raz, 1986: 204, 370). Raz hardly explains his

(33)

concep-tualization of either ideal or value. However, it can be inferred that the former conveys that auton-omy is something to strive for, put differently, autonauton-omy is good for all. The latter conveys that au-tonomy is the object of admiration and thus an ambition for individual (Raz, 1986: 369, 377).

For an agent to be autonomous, certain conditions must be met. Raz distinguishes three 12 such conditions: an agent must have “appropriate mental abilities, an adequate range of options, and independence” to be able to be the author of her own life (Raz, 1986: 372). Concerning the first condition, an agent’s ability to rationally formulate complex intentions and also plan according to them, is essential here (Raz, 1986: 372-3). To achieve the former, an agent needs to have integrity: an agent must be aware of both her available options and understand the influence these choices might have on her life (Raz, 1989: 381-2). An agent must also be loyal to her goals: she has to both identify with the choices she makes and commit to those choices. This does not mean that one can never change one’s mind. However, the reason for changing one’s mind must not come from a (mi-nor) setback in reaching one’s goal (Raz, 1986: 383). The condition of having mental abilities en-sures that an agent is truly the agent of her own life, determining what motivates herself.

Raz's extensively explains the second condition of autonomy: the importance of an adequate range of options to choose from. To illustrate this necessity, he gives two examples: the man in the pit and the hounded woman. In the first example, a man has fallen in a pit and will never be able to leave it. There is enough sustenance to survive without suffering, but he is unable to move around. His choices are severely limited; he can decide when to sleep or eat, and “to scratch his left ear or not” (Raz, 1986: 374). In the second example, the hounded woman can roam around freely over an island. However, she is constantly pursued by a animal set to devour her. She must continuously use

Raz uses capacities for and conditions of autonomy interchangeably. Unless I am quoting him, I employ conditions

12

rather than capacities. I do so, since describing the second condition of autonomy (having an adequate range of options) as a capacity does not cover the external factor at work in this condition. It seems strange to claim that an agent has the capacity of an adequate range of topics if the agent is the one choosing from the options.

(34)

all her energy, wit and resources to be able to escape it (Raz, 1986: 374). Both have some available options to choose from but neither is autonomous.

In the case of the man in the pit, the effects of his choices on his life are inconsequential. In the case of the hounded woman, it is the opposite; every choice she makes has far-reaching conse-quences for her life. Neither have the possibility to use their mental ability to choose outside of the confined conditions of their life. Raz, therefore, argues that an adequate range of options to choose from must be a condition of autonomy, without specifying when a range is adequate, as this de-pends on an agent’s context (Raz, 1986: 373-4).

For Raz, only when an agent has multiple valuable options to choose from, concerning choices which can be consequential for the course of her life, but are not necessarily so, can she be said to determine her life. Whilst both the man in the pit and the hounded woman act from reason that are applicable to them, they have no control over their lives. Instead, an agent must have a range of options to choose from and the choices must vary in the importance of their consequences. Not every choice should affect her life path drastically, but not all choices must be trivial either (Raz, 1986: 374). Rather, Raz argues that an agent has an adequate range of options that she can choose from a range of options both for consequential long-term and non-consequential short-term commitments or projects.

Raz claims that such a range requires multiple good options to choose from. One’s choices must be free from worries about personal needs required for having a worthwhile life. On top of this one’s choices must “not be dominated by the need to protect the life one has” (Raz, 1986: 376). Pro-tecting the life one has implies a lack of freedom to choose from multiple good options, due to ei-ther social or natural forces. An example of protecting the life one has, is an agent who has worked all her life to become a house owner who, due to some bad luck, is now threatened by the mob. They will demolish her house, unless she gives it away for free. Although she is sure that she would

(35)

be able to live somewhere else and all her other needs are met, the life she has is under threat, as she would not be able to buy another house. She has no good options to choose from.

Raz also emphasizes that options available are adequately varied. The range of options must contain options that are truly different from each other (Raz, 1986: 153, 204). Since people’s goals, pursuits and projects vary immensely and are not always compatible, the range of options available should include options that are incommensurable to each other (Raz, 1986: 375, 396). Raz also claims that amongst these varied options, some collective goods must be offered, as they are a re-quirement for a good life. He describes collective goods as those goods provided for all in a society whose benefits cannot be controlled by anyone other than a potential benefactor (Raz, 1986: 198, 206). An example he gives is the existence of homosexual marriage. He argues that this is a collec-tive good, since such a marriage can only exist through the existence of a society and the benefits are only distributed by those who benefit from its existence; nobody is forced to marry somebody of the same sex (Raz, 1986: 203). An adequate range of options must therefore contain both truly dif-ferent options and collective goods.

Another requirement for an adequate range of options is that the available options should stimulate the pursuit of good. For Raz, a range of options is only valuable when it allows an agent to choose between multiple options that facilitate an agent to maintain or bring about something of value. Options that are morally bad cannot be erased from agents’ lives, he states, but they do not have to be present in the range of options to ensure this condition of autonomy (Raz, 1986: 379-81). It should be noticed as well that the condition of having an adequate range of options to choose from should not be conceived as a singular choice, but rather as a succession of choices throughout one’s life (Raz, 1986: 374n1). So, according to Raz, one of the conditions for an agent to be au-tonomous is that the agent can choose form a range of varied and good options throughout her life.

(36)

Raz's third condition of autonomy considers an agent’s relation to others: “an autonomous person is not subject to the will of another” (Raz, 1986: 155). An autonomous agent must be inde-pendent. This implies both absence of coercion, as well as another form of outside control; manipu-lation (Raz, 1986: 377-8). According to Raz, the necessity of independence from others shows that the ideal of autonomy is partially a social ideal. An agent should be able to determine what to do by herself, whilst at the same time being formed by her society. Raz assumes that facilitating and main-taining another’s independence is an aspect of the way people should relate to each other (Raz, 1986: 378). For example, the social relation of independence is necessary for the existence of mod-ern democracy. If a society perceives manipulation and influence on people’s votes as normal and acceptable, no democratic claim can be upheld. Therefore, independence is one of Raz’s conditions for autonomy.

At the same time, Raz argues that the degree and area of influences due to coercion or ma-nipulation should be taken into account in determining the rightness or wrongness of such actions (Raz, 1986: 378). He claims that not all instances of coercion and manipulation are equally bad. For example, if an agent is manipulated into buying a certain brand of socks, this should not be consid-ered as bad as her being manipulated to vote on a certain candidate in national elections. Therefore, according to Raz coercion and manipulation both threaten autonomy and an autonomous agent therefore requires independence, but its importance exists in degrees.

Raz continues by stating that the three conditions of autonomy do not suffice for an agent to be autonomous. For him, an autonomous life is created only if an agents makes use of those condi-tions; in the process of making decisions and pursuing her comprehensive goals, an agent achieves autonomy (Raz, 1986: 372). Raz argues that if you have the required mental ability, have an ade-quate range of options, and are independent from coercion and manipulation, you can still fail to be