ETHICS AND THE DYNAMIC VULNERABILITY OF CHILDREN

In this paper,we want to examine the particular vulnerability of children from an ethical perspective.We want to defend three claims: Firstly, we will argue that children’s vulnerability is best understood as a dynamic quality, meaning that as children progress through childhood, their vulnerability also undergoes particular changes.To capture this, wewant to discriminate among physical,mental, social, and symbolic vulnerability,which vary according to certain features, such as age,maturity, gender, and race.These different traits are furthermore important in order to understand what we owe children from an ethical perspective. In a nutshell, children have moral claims not to be harmed and to be protected against threats to their well-being and well-becoming, and these claims have to be explicated via the dynamic vulnerability of children. Finally, we will argue that one of the main issues is to balance the protection of children and their autonomy claims, which both enhance and diminish their vulnerability.


ABSTRACT:
In this paper, we want to examine the particular vulnerability of children from an ethical perspective. We want to defend three claims: Firstly, we will argue that children's vulnerability is best understood as a dynamic quality, meaning that as children progress through childhood, their vulnerability also undergoes particular changes. To capture this, we want to discriminate among physical, mental, social, and symbolic vulnerability, which vary according to certain features, such as age, maturity, gender, and race. These different traits are furthermore important in order to understand what we owe children from an ethical perspective. In a nutshell, children have moral claims not to be harmed and to be protected against threats to their well-being and well-becoming, and these claims have to be explicated via the dynamic vulnerability of children. Finally, we will argue that one of the main issues is to balance the protection of children and their autonomy claims, which both enhance and diminish their vulnerability.

INTRODUCTION
Vulnerability is a key concept in the ethics of childhood. Many authors use it as an important characteristic necessary for grasping the special moral status of children, for clarifying what their well-being consists in, and for identifying and combating certain threats towards children (Mullin 2014a;Benporath 2003;Macleod 2015). While we fully agree with the importance that is attributed to these and with other works related to the concept of vulnerability, we believe that some features pertaining to the special vulnerability of children have not received sufficient attention-in particular, a systematic differentiation of different sources of their vulnerability is lacking so far, as is a recognition of the importance of their developmental capacity, for that matter. In this paper, we would like to address some of these shortcomings. We do not want to provide an analysis of what vulnerability is-in general or in childhood in particular-but we want to explore important characteristics and sources of vulnerability for children, and we will focus on those that are particular to children. In the first part, we will draw attention to the fact that children's vulnerability is best understood as a dynamic quality, meaning that as children progress through childhood, their vulnerability also undergoes particular changes. Children, as developing beings, do not possess one generic or homogeneous vulnerability; rather, there are many different forms of vulnerabilities, which depend both on the way children are and on the context in which they live. To capture this, we want to discriminate among physical, mental, social, and symbolic vulnerability, which vary not only according to age but also according to child-unspecific features, such as gender and race. Moreover, those latter features can show childspecific effects. In the second part, we will explore the ethical significance of children's particular vulnerability. Assuming (a) that children have moral claims, (b) that their vulnerabilities must not be exploited, and (c) that they should be protected against such potential harms, we will argue that their vulnerable condition is essential in order to explicate the content of these children's moral claims in detail and to clarify that what counts as a morally wrong action towards children is interwoven with their dynamic vulnerability. In the third part, we will turn to the relationship between children's autonomy and their vulnerability. Children's autonomy can be considered both a vulnerability enhancer and a vulnerability diminisher. In line with the consideration of autonomy as a vulnerability enhancer, we will argue that children are entitled to practice and develop their autonomy in line with their level of development and maturity. Following the understanding of autonomy as a vulnerability diminisher, we also have to consider those claims that may jeopardize children's well-being and autonomy in adulthood. Therefore, carefully balancing this double dimension of autonomy will be crucial to determining what liberties should be granted to children.

CHILDREN'S DYNAMIC VULNERABILITY
Vulnerability is a widely discussed concept and there are different approaches and definitions available in the literature. In this paper, we want to approach the special vulnerability of children via a very broad definition provided by Macken-zie and colleagues, who define vulnerability "as being at increased risk of suffering harm and/or having diminished capacity to meet one's needs, safeguard one's interests or protect oneself from harm" (Mackenzie 2017, p. 88;Mackenzie, Rogers and Dodds 2014). While this concept is clearly relevant to all human beings given their corporal neediness, dependence on others, finitude, and fragility, children have characteristics that make them particularly vulnerable. But what are the factors contributing to their increased risk of suffering harm or of diminishing their capacity to meet their needs? Before we start formulating an answer to this question by distinguishing different dimensions of children's special vulnerability, we have to make four general remarks that are key to our account. First, we choose a pragmatic approach, applying the definition from the Convention on the Rights of the Child by the United Nations, which defines a child as any human being from birth to the age of 18 (UN 1989), in order to determine the group of persons we refer to as "children." It is therefore a very heterogeneous group we have in mind, and there are different and varying vulnerabilities present within it. Newborns are very different from teenagers, and the risks they are facing vary enormously, which brings us to the second remark: A crucial part of children's vulnerability is closely tied to how children are-their "nature" so to say-and one striking feature of this is their developmental capacity. Childhood is the period in human life where the most radical and dynamic changes are taking place. This becomes particularly clear in the first years of life, where a fast development occurs from a state of complete dependency to one where many skills, such as talking and walking, have been learned and where a child's agency has been formed to a considerable degree. An important aspect of children's vulnerability, therefore, has to be seen in this dynamic developmental process and in relation to their growing agency and autonomy, which introduce new risks. With this we do not want to say that every feature of children's vulnerability relates to their developmental capacity; there are certainly aspects that they share with other human beings that makes them vulnerable-e.g., the fact that they have bodies. However, here, we would like to focus on what differentiates children from other persons in terms of their vulnerability. The third remark concerns the relationship between a child's vulnerability and his or her social, cultural, and economic context. We have argued that a considerable part of this vulnerability is related to how children are. This, however, should not be understood as a purely naturalistic or biologic claim. Although we believe that such inherent features play an important role if we want to understand the vulnerability of children, contextual factors are equally important. There is some particular inherent vulnerability of children that relates to features of their bodies, minds, and emotional capacities in combination with their developmental potential. But how this inherent vulnerability manifests itself in the lives of children depends on how the institution of childhood is socially and culturally shaped-it is situational, since the concrete circumstances decide the risks a child is facing (Mackenzie 2017, p. 89). And, in this context, factors such as gender, race, and economic status are crucial, as we will subsequently demonstrate in more detail. Fourth, we are aware of the fact that some aspects of the vulnerabilities of children we are going to discuss can be found in other groups as well-e.g., (adult) persons with disabilities or health problems or the elderly. Some might therefore argue that we do not achieve our goal of characterizing a "particular" vulnerability of children but rather address aspects that are present in other groups. Our answer to this challenge is that, from the outset, we have in mind children and adults who do not suffer from disabilities, health issues, or other factors constraining them in their well-being and well-becoming. Our claim then is that such children differ from equally fit adults in relation to their vulnerability, and one reason for this is their potential to develop. On the one hand child development is sensitive and can be distorted with potentially severe effects, while on the other hand child development comes with certain needs, which need to be addressed by others. At any rate, cases of severe disabilities or ill-health-in both childhood and adulthood-would need a different analysis, which we can only allude to in this paper.
Having these basic considerations in mind, let us dig deeper and distinguish among different-albeit in many ways related-dimensions of the dynamic vulnerability of children: physical, mental, social, and symbolic vulnerability. The characterizations we give are not exhaustive, but should be sufficient to underline our claim that there are special vulnerabilities of children in these dimensions.
Physical vulnerability relates to children's bodies and the potential harm they can suffer in this regard. Children's bodies are-especially in the first years of lifesmaller and weaker than the ones of adults, and this already implies that they have an increased risk of being harmed or hurt. A bodily interference that is harmless to an adult might be a serious risk to a small child. Two examples can illustrate this. The first one is the "shaken baby syndrome," also referred to as "abusive head trauma" (Mian et al. 2015). Shaking infants and young children forcefully is likely to cause severe brain damage. Young children's heavy and large heads (in comparison to the rest of their bodies), their weak neck muscles, and the delicate blood vessels in their brains constitute a physical vulnerability particular to them. The developmental dimension is crucial here, since the damage caused in the brain might translate into lifelong disabilities and hinder normal developmental progress across different dimensions. The other example stems from the study of environmental pollution. Research shows that exposure to environmental pollutants in early life is likely to endanger the well-being and well-becoming of the affected children. There is a risk of disease and death in infancy but also a danger that chronic, noncommunicable diseases that manifest at a later point in the lifespan might be triggered (Suk et al. 2016). Some of the main reasons for this relate to physiological processes operating in childhood. There are developmental windows where external influences, such as pesticides or other chemicals, have the worst effects, since metabolic pathways are still immature (Landrigan and Goldman 2011). As children grow older, the obvious differences in the strengths of their bodies in comparison with those of adults diminish. However, physical vulnerability remains a key issue, albeit in different variations. Teenagers, for example, already possess strong agency and a wide range of options to choose from. But with these developments, new physical vulnerabilities arise. They participate in or try "new things," often without former experience or knowledge, and expose their bodies to risks-e.g., in the field of sexuality. Here we see that, in the process of growing up, new vulnerabilities arise, vulnerabilities that are connected with the increasing autonomy of children, which entitles them to new rights, as we will argue in section 4 in more detail.
The next dimension we would like to turn to is mental vulnerability. Although there are many different approaches and definitions of mental health, it is indisputably a very basic element of well-being. According to the WHO, it is related to subjective well-being, perceived self-efficacy, and self-actualization of one's intellectual and emotional potential, and thus interconnected with the ideas of a stable sense of identity and self-worth. Moreover, it connects to social and behavioral skills as well (WHO 2001, p. 5). For children and adolescents, characterizations of mental health typically refer to sound family and peer relationships, an ability to be productive and to learn, and a capacity to use developmental challenges and cultural resources to develop adequately (Centers for Disease Control and Prevention 2013). Childhood is crucial in the development of the self and one's personality, while negative experiences during this phase are likely to cause lasting damage. Especially within the first years, children are completely dependent on others whom they find within their environment, which is conducive to mental health. And if they don't find others on whom to depend, the negative effects are extremely difficult to compensate later on. For example, there is evidence that early attachment security is related to measures involving emotional health, self-esteem, agency and self-confidence, positive affect, ego resiliency, and social competence in interactions with peers, teachers, camp counsellors, romantic partners, and others (Thompson 2007). Children who are not able to develop secure attachments are therefore disadvantaged in key areas of mental and emotional well-being and, in turn, extremely vulnerable in this regard. One element of children's mental vulnerability is related to their developing brain, which continues to grow across the whole of childhood, including adolescence, which is a critical phase for the development of mental illnesses. In fact, it is the peak time for the clinical onset of most mental illnesses, and one in five adolescents is estimated to have a mental illness that will persist into adulthood (Lee et al. 2014;Kessler et al. 2005). Important aspects of physical and mental vulnerabilities stem from features of children that are shared universally among them: their bodies are small and weak at the beginning, while their mental and emotional competencies only develop gradually. However, one must also emphasize that childhood is a social institution where social vulnerabilities arise (James and Prout 2005; Andresen 2014). There are many different ways in which a society might structure this phase of life-for example, in regard to what rights children possess, how educational facilities are organized, and which phases of childhood are recognized, and also in relation to its beginning and end. How these and related aspects of childhood are shaped in a society generates or influences vulnerabilities. Let us briefly comment on two such vulnerabilities, which are very common in many countries. First, most societies grant the sphere of the family a special status of protec- tion, privacy, and autonomy. There are good reasons for some privileges of families and parents, as Harry Brighouse and Adam Swift (Brighouse and Swift 2014) or David Archard (Archard 2010) have argued. But one has to point out that privacy and strong family autonomy also contribute to children's vulnerability and that thus every defense of the family buys into certain compromises. On the one hand, many cases of physical and sexual abuse happen within the family, and it often takes a very long time for them to be detected, if at all. Many children are thus suffering "in private" and do not have opportunities to seek help. Or, especially with young children, they do not even realize what is happening to them and that they do have a right to be protected (Graf and Schweiger 2017). On the other hand, children's developing autonomy and agency are highly dependent on their caregivers, who are in an extremely powerful position to be able to indoctrinate their children and to limit their identity resources (Macleod 2017). In some sense, the structure of society contributes therefore to the fact that some parents dominate their children and exploit their vulnerability further. Second, children are vulnerable to economic exploitation in various forms. If social security measures or children's rights monitoring are not in place or applied adequately in a country, vast numbers of children have to work in often unbearable conditions, as it is well known. But economic exploitation also happens if children are targeted by the commercial interests of big corporations, as is typical for modern consumer culture (Wild 2014). Finally, the protection of the family also contributes to the intergenerational transmission of inequalities (Archard 2010;Schweiger and Graf 2015). Against these reservations, we still believe that a defense of parental privileges and family autonomy is possible and is the best available option to protect and care for children (Graf and Schweiger 2017).
The last dimension of the special vulnerability is symbolic in nature. By this we mean that the status of "being a child" is usually connected to less prestige and less respect than that of adults, leading to inadequate attention paid to children or even to forms of disrespect or contempt. Two related phenomena illustrate these symbolic vulnerabilities: discrimination and epistemic injustices. When it comes to discrimination, one can state that many, probably most, societies are adult centred, meaning that the needs and preferences of adults are given priority compared to the ones of children, and that children's specific interests are not taken sufficiently into account. To give some examples, children are frequently excluded from adult spaces, not for their safety or well-being but for the convenience of adults; they are underrepresented in funding for research and development; and they are typically more often struck by poverty than adults (Webb 2004;Schweiger and Graf 2015). Furthermore, the practice of corporal punishment, which is still legal in many countries that have banned it for adults-such as the US-seems to be a deeply discriminating practice against children, since it denies them a basic right as a group that is in turn granted to all other members of society (Lenta 2012). In the last years, the concept of "childism," along with a profound analysis, was even introduced into the literature in order to grasp the many prejudices and the discrimination that children face in modern societies (Young-Bruehl 2012). With the notion of epistemic injustice, we draw on the work of Miranda Fricker, who has argued that there are specific injustices to people in their capacities as knowers if there is a bias against the credibility of a speaker's report or testimony or if a gap in collective interpretative resources puts a speaker at a disadvantage-for example, if a member of a minority uses concepts to describe moral wrongs that are not known to the (powerful) majority of a society (Fricker 2007). As was pointed out by Havi Carel and Gita Györffy, particularly in the health-care sector, children are often unjustly doubted in their capacities as epistemic subjects and are not taken seriously, which is a form of contempt and thus, in our terminology, a form of symbolic vulnerability (Carel and Györffy 2014).
The differentiation among the physical, mental, social, and symbolic vulnerabilities of children makes it possible to conceptualize their particularities more clearly than is usually done in the debate, and it also emphasizes the need to incorporate a developmental perspective. Children change significantly over time and there is therefore a great variety of vulnerabilities present in this group. However, one must stress that the vulnerabilities of children vary immensely not only from a vertical perspective (among children of different age), but also horizontally (among children of the same age). Factors such as gender, race, and economic position exert a significant impact on the manner in which childhood is experienced, the options available to children, the risks to which children are exposed, and therefore the way in which their vulnerabilities actually get exploited or protected. Poor children, for example, often face problems that nonpoor children do not have or that they experience to a much lesser degree (Felner and DeVries 2013). Gender, to give another example, is an essential category for analyzing childhood as well (Bhana 2016). Boys and girls still face different expectations; they are socialized differently and therefore different vulnerabilities arise-depending significantly on the social and cultural norms in place. And, of course, transgender or intersex children still face other challenges and have other vulnerabilities-for example, in relation to stigma, social exclusion, and the development of a healthy and stable identity. In conclusion, the framework we provide in this article points to particular vulnerabilities of children, but they have to be combined with other variables of social analysis to do justice to the variety of contexts children grow up in.

MORAL CLAIMS OF CHILDREN AND THEIR VULNERABILITY
In the previous section, we examined certain characteristics of children's vulnerability, with a particular focus on the different sources of vulnerability. Now we want to explore its ethical significance. Most theorists suggest that the vulnerability of children and their particular moral claims are somehow connected, and that an ethics of childhood needs to consider vulnerability as one of the main points of reference (see for example Mullin 2014a). Yet, the ethical significance of children's vulnerability is somewhat unclear, because many different understandings of vulnerability and of ethics exist. Our starting point is that children have some kind of moral claim to having their vulnerabilities not exploited by those with the intention of harming them, and that they should be protected against such harm. We want to apply a broad understanding of harm, which ranges from physical injuries to emotional suffering and from social exclusion to lack of respect. In a nutshell, everything that diminishes children's objective well-being (their level of well-being) and well-becoming (their healthy development to a future level of well-being) can be understood as a form of harm, whether or not the children actually experience it as being harmful or not (Bagattini 2014). Furthermore, our differentiation of physical, mental, social, and symbolic forms of vulnerability provides one framework to understand thisthese forms of vulnerability explicate what types of ill-being and ill-becoming children can experience. Our aim is not to show that vulnerability grounds (any of, some of, or all of, for that matter) the moral claims of children, but instead we are agnostic towards such justifications. Rather we assume only that children have such a moral claim not to be harmed and that this claim can be further explicated by exploring their vulnerability. Or, to put it differently: the content of children's moral claim not to be harmed relates to their vulnerability and is thus different from that of adults and changes over time as they develop. The justification of this claim itself-and related questions such as whether children have a stronger claim than adults-are not our concern here. It is important now, in our understanding, that the moral claims of children refer to both actions of harm and actions of protection. To make sense of that we need to consider that children's vulnerabilities, as we showed, are both inherent and situational. They are in danger of being exploited by others because they have weaker capacities than adults (or older children) and because there is a certain way the institution of childhood is framed in society. One paradigmatic case for this is child abuse. Every human being is vulnerable to being abused but for children this is clearly a more present threat because of them being physically, mentally, socially, and symbolically more vulnerable to abuse. They are less likely to protect themselves against a perpetrator because they are physically and emotionally weaker, and most cases of abuse also happen within a specific social setting, the family, which makes children more vulnerable because the perpetrator has power over them and is entrusted with their care (Graf and Schweiger 2017). While the first aspect refers to inherent aspects of children's vulnerability, the second one has a specific situational dimension, and is enhanced through social norms and practices. Particular forms of child abuse are even legalized and socially accepted, such as the case of corporal punishment in some jurisdictions, which enhances the vulnerability of children even more and makes it easier for perpetrators to harm them because they do not have to fear punishment or shaming (for a powerful critique of corporal punishment, see Lenta (2012)).
Children's vulnerability thus guides their moral claims, because in order to understand what can harm them, we need to determine the types of threats to which they are particularly vulnerable. We already mentioned the "shaken baby syndrome" as a severe form of abuse. Clearly the shaking of a baby is a violation of its moral claim not to be harmed, and leads both to ill-being and illbecoming. The shaking exploits a particular vulnerability of the baby, which the baby shares with other babies but not with older children or adults. Thus, if we assume that both adults and children have moral claims not to be harmed, the shaking only seriously harms babies. That says nothing about why they have such a moral claim in the first place, but as soon as such a claim is assumed, it says that babies deserve special protection from being shaken and that adults have a moral duty not to shake babies. Similarly, an abusing parent or an abusing coach exploits the particular vulnerability of a child being entrusted to his or her care, being both physically and socially more powerful than the child. Both adults and children have moral claims not be abused, but the vulnerability of children is greater and depends on some of their certain features and on their social surroundings, and from that we can assume that they deserve particular protection or that we should adapt their surroundings accordingly to be safe. Another way to make this point is to say that, although the moral claims of both children and adults are not grounded in their vulnerability, appropriate protections are, and the concept of vulnerability is crucial in order to determine the content or types of protections demanded. As we have argued in the previous section, children's vulnerability is dynamic; they become less vulnerable to some threats, but others arise and become more present, and the moral claims of children thus develop and change along with the children themselves. Experiencing a head injury, with or without devastating future effects, is always a form of illbeing, but a baby's vulnerability to this form of injury is much higher. Not to be harmed by the exploitation of particular vulnerabilities has a different meaning for a baby than it does for a teenager, and the concept of children's vulnerability helps us make sense of that.
Another aspect that can be explicated via children's particular vulnerability is the moral claim of children to be protected. This goes further than the claim not to be harmed by the exploitation of their vulnerability. Children are unable to protect themselves, not only in the sense of being unable to defend themselves against perpetrators, but also in the fundamental sense of being dependent on others. 1 Small children are severely harmed not only by being subjected to abuse or other forms of active infringement on their bodily integrity, but also by simply being left alone and not being cared for. This feature is also shared by some adults, such as people who are chronically ill, disabled, or severely injured. And although we acknowledge that autonomy is relational and that we are always intertwined in social relations, we think that the dependency of children, especially younger ones, is categorically different. They need caregivers to survive and to develop properly, and they cannot substitute for their absence, while adults, who might be harmed from being alone and on their own (see for example Brownlee 2013), are able to survive and to take care of themselves-it can be said that this is one key dimension of perceiving the modern self and its selfunderstanding as autonomous. The paradigmatic case that illustrates this dimension of children's moral claims is child neglect. Once more, in this case, we find inherent and situational aspects. Children's fundamental dependency is based on the lack or immaturity of physical and mental abilities, and on their social and symbolic positioning. Child neglect takes place if the particular needs of children are not met or if they cannot meet them themselves, and the reasons for them being unable to meet those needs can be found in both their limited capabilities and their social positioning. The needs of a teenage girl differ from those of a boy, but they also differ in their capability to meet their needs based on social norms and practices (e.g., gendered norms about sexuality, access to health care and knowledge). The vulnerability of children has, among others, a gender dimension.
Furthermore, what is perceived as neglect, as well as abuse, is based on social norms and practices, which differ greatly, while scrutinizing the moral claims of children to be protected refers to empirical knowledge. For example, the moral claim of children to education is well based (McCowan 2010), but its concrete meaning is once more also dependent on a substantial understanding of their vulnerability. Children are neglected if they are deprived of education, but this varies from basic education for young children in order for them to learn to read and write, and knowledge about personal hygiene, to the claims of older children to receive such education that enables them to become fully economically and socially integrated citizens. The vulnerability to educational neglect is not the same for all children, and thus the content of their moral claims differs in relation to it.
Another important ethical conclusion can be drawn from our elaborations on certain characteristics of the dynamic vulnerability of children, and its differentiation in physical, mental, social, and symbolic dimensions. The vulnerabilities in these dimensions do not develop automatically but instead rely on certain provisions such as care, resources, protection, and innate capacities. Moreover, they also do not develop independently or at the same pace. Children are particularly vulnerable, and have moral claims not to be harmed and to be protected accordingly, but they also have a moral claim to become less vulnerable, and to reduce the protection they need and the harm they can suffer from certain actions. To put it differently, one crucial point of the moral claim of children to be protected and cared for is that they should develop in such a manner that they become less vulnerable, and they are not able to do so alone.

BALANCING CHILDREN'S MORAL CLAIMS AND THEIR AUTONOMY
We have argued that children's vulnerability is dynamic and encompasses the different dimensions of physical, mental, social, and symbolic vulnerability, and that this concept can help us to scrutinize the moral claims of children. Now, we want to explore the particular relations among vulnerability, moral claims to protection, and children's autonomy. Autonomy, understood as a capability, does not only have a cognitive dimension but incorporates the ability to make individual deliberate choices and to act on them, thus encompassing sufficient agency, health, and resources or knowledge, to name a few conditions of autonomy. 2 This brings us to an important question regarding children's moral claims. How should we perceive the relationship of vulnerability to autonomy in childhood?
The main point we would like to make in this regard is that children have a moral claim to practice and develop their autonomy in line with their level of devel- A U T O M N E / F A L L 2 0 1 7 opment and maturity, but that this claim must not substantially jeopardize their autonomy as adults and their well-being as children in any case. Reference is thus made to their special vulnerability in this context. On the one hand, considerations about vulnerability can clarify the risks that are connected to children's choices and thus guide how certain freedoms should be restricted. On the other hand, vulnerability points to areas of children's lives unlocked by their increasing autonomy, something which has important normative implications as well, since their need for guidance by others decreases and, with it, the justificatory basis for certain types of paternalism.
We said that children have, in our account, a (limited) right to practice their autonomy. There are different reasons that ground this right, although we can mention them only briefly. Firstly, autonomy is regarded as a central value for adults (Adams 2008). But autonomy develops during childhood and is not present from birth on, which has two implications-namely, that children are by and large less autonomous than adults for the most time of their childhood (this is a relational claim, which compares children to adults) and that children are not autonomous enough to claim that their autonomy is granted protection (this is a nonrelational claim, which means that some adults also fall short of this threshold). We follow here, in particular, an argument brought forward by Monika Betzler (2011) and others stipulating that children should be brought up in such manner that they become autonomous (Feinberg 1980). Secondly, autonomy as a developmental goal tells only half of the story. Autonomy is both intrinsically and instrumentally valuable for children, especially for older ones. The intrinsic value of autonomy for children stems from the respect we owe them as moral subjects, who have a will and the capacity to express their wishes, preferences, and points of view (Archard and Skivenes 2009). Since these abilities are only very marginally present at the beginning of life and increase during childhood, the reason for a right to autonomy becomes stronger over time. The right to autonomy for a toddler is minimal, while it has a strong impact for a teenager. Another reason for granting such a right is connected to a different claim, which we made earlier-namely, that children are entitled to the development of autonomy during childhood. Now, this development is necessarily interwoven with exercising autonomy (Betzler 2011); being able to live an independent life and to enjoy the liberties associated with adulthood presupposes that experiences with free choice and free actions are lived in childhood. In this sense, the high value that is attributed to autonomy in adulthood also grounds children's rights to practice their autonomy, since without these opportunities they will never become fully autonomous beings later on. And finally-another argument we have already provided above-acting autonomously makes children less vulnerable in certain respects. Since they have moral claims not to be harmed, a (limited) right to practice their autonomy follows from this consideration. Thus, while autonomy is valuable for children, good arguments for a large degree of paternalism towards them also exist (Adams 2008). This means that their autonomy should be restricted considering their well-being and future autonomytwo relevant conditions from a normative perspective-which can be forfeited by the wrong choices and actions of a child. On the one hand, autonomy in chil- dren is still developing and their capabilities to make choices are limited. They need guidance and care from others. This reason for paternalism becomes weaker as children grow older and become mature. On the other hand, scholars such as Joel Anderson and Rutger Claasen (2012), as well as Andrew Franklin-Hall (2013), have argued that paternalism towards children should not be viewed only as capacity based, meaning that there are reasons to justify paternalism towards children also in cases in which they are mature enough to make choices for themselves, if not in all aspects of their lives. Or to put it differently: sometimes paternalism is justified even though children are as competent as adults.
The reasons for this are to conceptualize and institutionalize childhood as a safe space with limited responsibilities and to secure equality of rights and duties among children, regardless of their competencies. We do not wish to explore these types of arguments here but rather simply assume that children's autonomy is important for them but should also be limited for paternalistic reasons. We assume that also to be a noncontroversial approach towards children's autonomy even though the limits of paternalism and autonomy are under debate. Now, children's special vulnerability can serve as a guide to see what limitations can be justified. If we know, for example, that education plays a fundamental role in preparing someone for an independent life, a child's wish not to go to school can be overruled in light of future autonomy considerations. Let us take another example form the health-care setting. If we know that children are vulnerable to certain diseases, which often affect both current well-being and development, the child's will should play only a minor role in the decision when, for example, certain preventive measures are taken to avoid disease-such as regular checkups and vaccines.
A child's claim to exercise autonomy and to choose and act independently of other people's guidance (especially that of their caregivers) increases with age and maturity. So far, we have emphasized that children's special vulnerability implies that a focus on paternalism is justified if it secures their well-being and future autonomy. However, with children's increasing capacities their responsibilities also grow. Particularly with adolescents, very difficult questions arise as to the level of autonomy to which they are entitled. Still, the concept of children's vulnerability is helpful for framing the challenges within the different dimensions, as well as for justifying that certain levels of well-being must be secured and that they enter adulthood with the necessary conditions for an autonomous life. While many capacities, which are interwoven with autonomy, are already very strong in adolescence (e.g., cognitive development), other aspects typical of this phase of life are known to make children extremely vulnerable. Emotional imbalances and shifts in identity formation, to give two examples, provide reasons to insist on certain protections and limitations on adolescents' responsibilities.
One main reason why the relationship between vulnerability and autonomy is tricky is that-as we have already indicated briefly-autonomy is both a vulner- ability diminisher and a vulnerability enhancer. That complicates the moral claims of children. On the one hand, autonomy diminishes vulnerability because it makes children less dependent, and it implies that the child has developed certain physical and mental abilities. In order to be referred to as autonomous, the child, for example, needs to be able to reflect on risks and also on how to achieve ends without harming himself or herself. Children are autonomous only if they can at least to some extent provide for themselves-for example, being able to eat, drink, and move on their own. That makes certain forms of harm less likely and certain forms of protection less needed. On the other hand, autonomy is also a vulnerability-enhancer because certain types of vulnerability come into play only if the child is autonomous to some extent. Two examples might illuminate this point. A baby cannot engage in dangerous behaviours such as smoking or pedalling a bike without a helmet-at least not of its own will-but older children can certainly do that. Their increased capabilities allow them to do it, and their developing autonomy makes it possible for them to choose such activities willingly, avoiding parental supervision if necessary. Children cannot be controlled twenty-four hours a day, and they should not be, because, as stated before, they need to exercise their autonomy, to some extent at least, in order to develop properly. Furthermore, the respect we owe them entails giving them certain liberties and a certain privacy that separates them from their parent's will. This can be risky, and makes them vulnerable. They become vulnerable to threats towards their well-being and well-becoming to which younger children are not vulnerable because they cannot do these dangerous activities. This first set of examples refers to decisions that can be said to fall under the category of local autonomy (Franklin-Hall 2013). But as children grow up they also become vulnerable in respect to what can be called their global autonomy. Global autonomy implies that children become authors of their own biographies, that they make decisions and choices that alter their lives significantly because they decide who they want to be and what kind of life they want to live. So, this kind of autonomy is necessarily present during childhood, although often in a limited and still developing form. Certain decisions and actions are explicitly legally prohibited for children in modern Western democracies. For example, they are not allowed to marry and generally they are not considered to have full legal capacity. Other choices are discouraged but possible to make, such as having a baby or dropping out of school. What options are available to children depends on many factors, as these two previous examples show-factors that relate to the different dimensions of vulnerability we identified in section 1, and which develop and change as children become mature and more autonomous. Babies or pre-adolescent children cannot become fathers or mothers themselves-this is physiologically impossible; the possibility of dropping out of school is typically severely restricted up to a certain age by legal systems. Here, it becomes clear that young children are less vulnerable in this respect than older teenagers.
Regarding the moral claim of children, this double role of autonomy is crucial. If children have moral claims not to be harmed and to be protected, as we explored in the previous section, and the content of the claims can be scrutinized through the dimensions of vulnerability, then we should also acknowledge that these claims are autonomy sensitive. Autonomy sensitivity means that the moral claims of children not to be harmed (the paradigmatic example being child abuse) and to be protected (the paradigmatic example being child neglect) are (partially) determined by the level of their autonomy capacities. This autonomy sensitivity is not a simple one in which greater autonomy only widens children's moral claims to self-determination; it also changes what children are entitled to be protected from and what kinds of harm they might endure from others. This leads to balancing autonomy-based moral claims to preserve children's autonomy with autonomy-based moral claims to protect children from those threats that enter their lives because they are more autonomous. This balanced solution will be an uneasy one in many real-life cases, because the risks-or, the enhancement of vulnerability-and benefits of acting autonomously are not easy to determine. Let us consider the example from above: an adolescent who becomes a sexually active agent. The vulnerability associated with this situation-vulnerability in each of its physical, mental, social, and symbolic dimensions-is not so easy to determine, and also partially based on social norms and practices. In patriarchic societies girls face different dangers compared to boys, and those dangers are framed differently. The risks of unplanned pregnancy, of abuse, or of coercion are real, but a normative evaluation of them and their likelihood are less easy to determine. Only in abstract theory is it straightforward to claim that childrenin this case, adolescents-are allowed to be sexual agents to the extent that their actions and interactions are harmless to themselves; it is not clear from the outset what counts as harm and how to balance the risk of getting harmed with the claim to autonomy in this crucial area of private life. Some risks seem inevitable, such as having a relationship end, being confronted with wishes that one does not want to fulfill, or being exposed to external norms of sexuality, which are not an individual's own and which can be disturbing and frightening. The awakening and exploration of sexual desires are in themselves potentially troubling and upsetting adventures, during which children are confronted with themselves and others, and with the task of disentangling their own desires from those that confront them from the outside (for example, those of media, peers, partners, parents, etc.). The concept of autonomy sensitivity applied to vulnerability is fruitful here. It shows not only that childhood can never be a phase of total protection, but also that childhood is inherently risky-not only because children grow up in a risky world, but also because growing up and becoming autonomous are themselves risky. Autonomy, which is valued so highly, and which children should explore, makes them more vulnerable. The full restriction of autonomy is also not the answer to this, because autonomy sensitivity works in both ways. It also protects children if they are more autonomous and competent, so as to give them more space and enhance their capabilities. Regarding the example of sexual agency, this conception of balancing between autonomy and vulnerability implies that children can become responsible, authentic, and undistorted sexual agents only if they are also given the space, time, and privacy to explore their sexuality (alone and with others), which means giving them exactly that which also makes them more vulnerable. This is not to say that no restrictions of children's autonomy are justified when it comes to risky activities such as drinking or engaging in sexual activities. However, especially as they grow older, their autonomy claims involve some dangers, and cannot and should not be avoided completely for the sake of protection.

CONCLUSIONS
Children are particularly vulnerable beings-physically, mentally, socially, and symbolically. In this article, we have explored these different dimensions and argued that they provide a helpful framework for conceptualizing children's vulnerabilities. But the dimensions are not only conceptually and descriptively important. On the contrary, they also provide an indispensable basis for explicating the moral claims of children. Children have moral entitlements not to be harmed and to be protected, and these can be understood in detail and connected to concrete entitlements only if we understand what types of beings children are and what types of harm are dangerous to them, to their well-being, and to their development. But children do not only have moral claims not to be harmed and to be protected. They are also entitled to develop into fully autonomous adults and also to practice autonomy during childhood in a way that is compatible with their vulnerabilities, given that autonomy is both a vulnerability diminisher and a vulnerability enhancer. In this article, the main aim was a conceptual and normative one. We aimed to provide a more comprehensive analysis of the particular vulnerabilities of children than has usually been the case in philosophy and to explain how these vulnerabilities relate to children's moral claims and their autonomy. However, we also suggested how important the relationship between ethical reasoning and empirical facts is. Different disciplines need to work together in order to provide a rich account of the inherent and situational vulnerabilities, which is crucial to understanding what is owed to them.  (2017) has recently argued that this radical dependence, which is related both to vulnerability and domination, is one reason to think that the phase of childhood is overall bad. 2 We do not aim to dig deeper into the discussion about (children's) autonomy and its definition. See, for example, (Mullin 2014b;Noggle 2002).