Respecting Children’s Choices

Abstract The traditional liberal view on conflicts between care for wellbeing and respect for choice and desire is that we should look to degrees of competence and voluntariness to determine which moral imperative should take priority. This view has likely influenced the common view that children’s choices should be considered only to the extent that this promotes their future autonomy and helps us determine their best interests. I reject both the general traditional liberal view and its application to children. Competence and voluntariness, as well as maturity, are at best proxies for what really matters, which is wellbeing, choice and desire. We typically have reason to respect children’s choices, irrespective of any further positive consequences. If we should more often make children do what they do not want to do, this is mainly because, though we should care about respecting their choices, we should care even more about their wellbeing and future autonomy.


Introduction
What reason do we have to consider what children want for themselves? The philosophical literature is dominated by two main answers. First, there are considerations of short-term wellbeing. Though generally not as competent as adults, children might have some idea about what is good for themselves, in their present circumstances, and so hearing them out can provide information relevant for judging what is in their best interests. In addition, allowing children to do what they want avoids the need to force them to do otherwise. Such forcing may in itself impact negatively on their wellbeing. It may also impact negatively on those doing the forcing, but my focus here is on children.
The second main consideration in the literature is future autonomy. Allowing children some measure of control over their own lives will teach them responsible decision-making, a skill that one needs to learn, preferably before one's 18th birthday. We might have a special responsibility to ensure that each person reaches some threshold of autonomy, so that she can take her place in adult society as a free and equal person. In addition, being a more capable decisionmaker may have a positive impact on the child's (and future adult's) long-term wellbeing.
These two reasons for considering what children want are in stark contrast to the reasons typically given for why we should consider what adults want for themselves. For an adult, in matters that regard herself, sheer respect for her judgment and agency is typically considered sufficient reason to not only take her view seriously, but to let it be decisive. This is so even if the same reasons that are invoked for considering what children want are also available for adults. Adults too typically have information to share about what is best for them, and they will generally fare better if they are not forced. Adults too can practice their decision-making ability, and this may make them more autonomous in the future and may improve their long-term wellbeing in other ways.
In this article, I argue that the reasons we have to consider what other people want for themselves, and to allow them to direct their own lives, are the same whether those others are adults or children. If we should sometimes let an adult do what she wants, but not a child in similar circumstances, this must be because the reasons weigh differently, not because different sorts of reasons apply, or apply differently. Call this the Same Reasons View. My main aim is to support this view. However, I will in large part do so by way of arguing for the more specific view that we typically have reason, for all people, to both promote their wellbeing (including long-term autonomy) and respect their choice.
The article is divided into three main parts. In the first, I describe the traditional liberal view of the two moral imperatives of respecting choice and promoting wellbeing. I discuss some problems with this view, focusing on what it says about the role of these imperatives in general. In the second part of the article, I describe how the traditional liberal view has come to influence the debate on children's rights and how some authors have endorsed it and others amended it. I argue that similar problems arise as for the general version of the view. In the third part of the article, I give a brief positive argument for the Same Reasons View both in general and as it relates to adults and children.

The Traditional Liberal View on Wellbeing and Respect
I take it to be essential to a liberal moral outlook that we have (objective) reason to care about other people's wellbeing as well as to respect their choices and desires as far as they concern their own lives. 1 Our reasons of respect may be further explained in terms of, for example, respect for autonomy (Feinberg 1986), respect for persons (Darwall 2006) or respect for self-expression (Cullity 2018). 2 Irrespective of any such further grounds, however, the common idea is that a person wanting something for herself provides an immediate reason to not prevent or interfere with her seeking it and indeed may provide a reason to support her in getting it, whether or not it promotes her wellbeing. While, on a traditional liberal view, this moral requirement only applies to the choices of some people, on the Same Reasons View it applies to all people. For consequentialist liberals, our reasons of respect are ultimately based on their contribution to long-term wellbeing. John Stuart Mill is the most well-known liberal of this sort, though it is controversial whether his consequentialism is consistent with his liberalism (see, for example, Gray 1996Gray /1983. For a consequentialist liberal, it is always an open question whether or not a person's choice should be decisive or whether we should instead interfere in some manner. In settling on action-guiding rules of thumb or institutional rules, it may be optimal to treat as decisive only the choices of adults, or only those of some other category of persons, depending on complex psychological, social and legal circumstances. Since I have no room to discuss such empirical circumstances, I will for simplicity assume that we have reason to respect people's choices and desires irrespective of any further consequences, though these reasons need not be decisive but may be outweighed by other reasons, including outcome-based reasons. Depending on the details of any consequentialist liberalism, my arguments may or may not be also relevant for such a view. Reasons of respect and reasons of wellbeing are in conflict whenever we can make a person better off by treating her in some manner that she does not want to 1 We may not have reason to care about everyone's wellbeing or everyone's choices and desires, and we may have reason to be partial. What is important is that we have reason to care about some people's wellbeing as well as these same people's choices and desires. 2 By citing these authors, I do not mean to endorse their complete views. In particular, I do not share Feinberg's and Cullity's view that respect for choice can 'trump' (Feinberg) or 'undermine' (Cullity) care for wellbeing. be treated. 3 The traditional and still mainstream liberal position is that such conflicts should be resolved differently depending on how competent the person is or how voluntary her choice or action is. 4 If the person is sufficiently competent and her choice or action is sufficiently voluntary, then her will trumps her wellbeing interest and so we have no right to go against her will in order to promote her wellbeing. To do so, it is sometimes said, would be 'hard paternalism'. If, on the other hand, the person is insufficiently competent, or if she is sufficiently competent in general but her particular choice or action is insufficiently voluntary, then her wellbeing interest trumps her will and we should disregard her will in order to promote her wellbeing. To do so is sometimes called 'soft paternalism' (on the concepts of hard and soft paternalism, see Feinberg 1986: chapter 17). 5 I will call this view the Different Reasons View to emphasize that it gives priority to different kinds of reasons for different people and in different circumstances.
I have now shifted from speaking of respect for choices and desires to talking of what people want and them having a will. These similar concepts can be used in different ways. I am mainly interested in and will mainly discuss the standard case where we choose to do approximately what we desire to do. This does not mean, however, that I presume choices to be optimific in relation to our complete set of desires. On the contrary, I take it that we want things, make choices, and have a will also when we clearly fail to maximize our own desire-satisfaction. In fact, I assume that also relatively incompetent people make choices and that competent people make choices also when they are acting relatively non-voluntarily, such as when we are very upset or under a lot of stress. In particular, I am assuming that young children have a will and make choices. I believe that this is in line with common usage of these terms. By using these terms in this way, I am of course not 3 Many liberals hold that we have no reason to respect a choice or desire not to listen to rational arguments. I see the connection between this restriction on our reasons of respect and the liberal tradition of seeking the rational. However, I think that such a restriction is ultimately unacceptable. If a person does not want to hear an argument, we have reasons of respect not to make her hear it (see Dworkin 1983;Tsai 2014). 4 Voluntariness can be analyzed in different ways, but the two main ingredients are typically informedness and instrumental rationality. For a rich, plausible and influential account, see Feinberg 1986, chapter 20. 5 This dichotomous position goes back at least to John Stuart Mill. After formulating his famous liberty principle in On Liberty (1859), Mill immediately clarifies that the principle is of course not applicable to children: 'It is, perhaps, hardly necessary to say that this doctrine is meant to apply only to human beings in the maturity of their faculties. We are not speaking of children, or of young persons below the age which the law may fix as that of manhood or womanhood. Those who are still in a state to require being taken care of by others, must be protected against their own actions as well as against external injury' (Chap 1, para 10). assuming anything normative, such as that choices should always be respected. On the contrary, this is the thesis that I will argue for.
The Different Reasons View combines rejection of hard paternalism with acceptance of soft paternalism, or in other words rejection of unwanted interference for some people with acceptance of such interference for others. Both components of the view have received friendly (liberal) criticism in the scholarly debate on paternalism. The rejection of unwanted interference with the sufficiently competent and voluntary has been criticized for being too rigid, seeing as rights against interference need not be absolute; there is a distinction between rights violations and mere infringements (for example, Arneson 2005). It has also been criticized by pointing to concrete cases in which even rather well-informed and instrumentally rational adults should be made, against their will, to avert suboptimal, possibly disastrous consequences (for example, Dworkin 1972).
The acceptance of unwanted interference with the insufficiently competent or insufficiently voluntary has also received friendly criticism, though not as much. 6 Richard Arneson (for example, 1980) has warned against setting too high a threshold for sufficiently voluntary choice and against allowing character traits such as impulsiveness and foolhardiness to imply lack of competence or voluntariness. Danny Scoccia (for example, 2013) has argued that people's aversion to interference even with their own quite involuntary and harmful behavior sometimes makes such interference unjustified. Outright rejections of soft paternalism are uncommon. The best candidate is arguably the child liberationist case for equal legal rights for children and adults. Samantha Godwin (2011) provides an overview of this position and arguments supporting it, as well as an extensive criticism of many arguments and assumptionspractical, legal and philosophicalmade for disregarding the will of children.
I see two main problems with the Different Reasons View, in addition to its implausible recommendations in some cases. The first is that it introduces normative jumps, and the second is that the factors that are supposed to underpin these normative jumps are not morally relevant. To see the first problem, note first that the Different Reasons View entails a dichotomy between on the one hand competent and voluntary, and on the other incompetent or involuntary. Clearly, there are no natural thresholds of either competence or voluntariness when these 6 Libertarianism has the resources to reject soft paternalism, since ownership rights are not affected by ignorance or irrationality. However, I am not aware of a libertarian who has argued against all interference with reasons that actions are ignorant, irrational or self-harming, other than to say that if the government gets involved it will probably only make things worse. This may be because of the difficulty with finding a plausible position on paternalism from a libertarian perspective (Wall 2009). are understood as empirical concepts (and not, for example, legal). Any lines drawn are arbitrary, however well-measured. This is problematic because the moral difference between the two categories is supposed to be stark. Under the threshold (of either competence or voluntariness), respect for will plays no role, but over the threshold (of both combined) respect for will is trump. At no point does respect for will warrant respect without being decisive, though this seems like a plausible intermediate possibility. It is often unavoidable to draw lines, but it is rarely, if ever, unavoidable to introduce these kinds of jumps into the normative terrain. 7 Normative jumps can be softened by introducing indeterminacy or vagueness around the thresholds, 8 but this comes at the familiar cost of restricting the domain of application. The idea would be that, at some point of competence and voluntariness, will trumps wellbeing, while at some lower point, wellbeing trumps will, but in between there is a sort of twilight zone. To my mind, this is no improvement on the view. Indeterminacy and vagueness are plausible in some contexts. They are plausible when the applicability of some normative principle depends on whether or not some empirical concept is instantiated. For example, the principle that we should respect choices only applies when some behavior is a choice, and there may be difficult borderline cases. They are also plausible when we need to make trade-offs between two values that are not precisely comparable. For example, it may be indeterminate whether we should go against our reason to respect choice in order to act on our reason to promote wellbeing, or the other way around, when both reasons are about as strong.
However, it its not plausible to introduce a cut-off point or a threshold and shield it with indeterminacy or vagueness for normative purposes. In our present case, the claim is that some degree of competence and some degree of voluntariness is required for choice to warrant respect. These thresholds do not depend on uncertainty regarding the applicability of the concepts of competence or voluntariness, or any other empirical concepts (though it is of course possible to introduce a normative conceptsuch as substantially voluntaryin order to identify the threshold). In this case, therefore, indeterminacy would only serve to obscure the lack of a plausible explanation for why wellbeing and respect for choice cease to be relevant considerations at some more or less arbitrary threshold.
The second and more fundamental problem with the Different Reasons View is that it is unclear why either competence or voluntariness are relevant to whether or not a person's choice should be decisive. They may of course be useful indications. People who are more competent and act more voluntarily typically make choices 7 For a longer argument along these lines, see Grill 2010. 8 As suggested by a reviewer for this journal. that are more in line with their desires and more conducive to good outcomes. High degrees of competence and voluntariness arguably do not, however, guarantee desire-satisfaction or (otherwise) good outcomes. At least, they do not do so on the most plausible understanding of these concepts.
Degrees of voluntariness can be understood in two main ways. The most natural, arguably, is to look to the different factors that reduce voluntariness. These may include, as on Joel Feinberg's account, coercion and duress, manipulation, ignorance, and temporarily distorting circumstance such as fatigue, intoxication and time pressure. The more there are of any of these factors, the less voluntary a choice is. On this understanding, the fact that a person draws the wrong conclusion about some fact, however important, does not automatically reduce voluntariness to zero, or to below sufficient. We can, in other words, voluntarily perform an action even though, if we had full information, we would not perform it. On this understanding of voluntariness, it is a mistake to assume, as some liberals are prone to do, that just because some person is factually mistaken, her choice is not deserving of respect. Mill makes this mistake in On Liberty when he says that we can stop a man crossing a dangerous bridge unawares coercively 'without any real infringement of his liberty; for liberty consists in doing what one desires, and he does not desire to fall into the river' (1859: chap 5, para 5). 9 We may overall have stronger reasons, even stronger liberty-based reasons, to interfere, but this does not imply that we have no reason not to interfere.
An alternative way to understand voluntariness is in relation not to optimal capacities and circumstances, but in relation to optimal outcomes. Feinberg sometimes indicates this is his understanding, and Danny Scoccia at one time (2008, p. 358) took this to be Feinberg's considered view. 10 On this understanding, a choice is more voluntary the closer its outcome is to the outcome of a certain hypothetical choicethe choice the person would have made if she were perfectly informed and rational. This understanding allows, in contrast to the first, that the choice to cross a dangerous bridge is non-voluntary simply because the person does not desire the outcome of the choice, that is, falling into the river. I take it that this alternative understanding of voluntariness is a minority position and not what liberals typically have in mind when they defend or assume the Different Reasons View. One reason is that actual people are not perfectly informed and rational and so it may seem a 'monstrous impersonation' to respect this hypothetical extrapolation, rather than whom we in fact are (Berlin 1969(Berlin /1958. The outcome-oriented understanding implies that we have reason to interfere with people who are acting quite sensibly and with very little of the typical voluntariness-reducing factors in 9 Mill's mistake is thoroughly analyzed by JP Day (1970). 10 Scoccia has since retracted this interpretation (personal communication). order to make them do what they would have done had they been perfectly informed and rational. It also implies that a choice that is made under whatever extremes of coercion, intoxication or ignorance can be perfectly voluntary, as long as it happens to coincide with what the person would have chosen if perfectly informed and rational. Consider a person who is deterred from harming others by fear of its consequences. Suppose this person harms another person by mistake, for example by opening a door in her face. It would seem that on the outcome-oriented understanding, this action is perfectly voluntary if it is just what the person would have done if she were perfectly informed and rational, even though there are no realistic circumstances under which she would have caused such harm intentionally.
Given that we reject the outcome-oriented understanding of voluntariness, under what circumstances should we interfere with a man crossing a dangerous bridge unawares? It seems to me that this must depend on other aspects of the situation than how voluntarily the man is acting. Suppose he is intelligent, calm and collected, emotionally stable, and under no pressure from others. Suppose he can see the bridge and its cracks very well and knows a lot about the durability of bridgeshe just fails to conclude from all this information that this particular bridge is dangerous. Is any of this relevant? What is the threshold of voluntariness such that I should let the man fall to his death? It seems to me that there is no such threshold. What is important is rather whether there is any upside to crossing the bridge and what is the nature of this person's inclination to cross it. If there is no upside and the person has no particular reason to cross this bridge rather than another one nearby, then I fail to see what level of voluntariness would imply that I should remain passive.
Suppose instead that the person knows the bridge is dangerous and still wants to cross it. This makes all the difference, to my mind. This is not, however, because the person scores higher on voluntariness, but because crossing will produce desired outcomes. I would have more reason to allow a person to cross if this was very important to him. Perhaps he is a daredevil who has prepared for this challenge for many years. Crossing then seems reasonable enough, if a bit eccentric. That he is very excited about fulfilling his long-time ambition, to the point of being emotionally unstable, does not seem to matter, as long as this does not affect the outcome of crossing.
The philosophical paternalism debate is substantial, and I cannot do justice to it here. Based on my brief treatment, I venture to propose that many participants in this debate have come to treat voluntariness as a powerful proxy for what really matters, that is, the probable value of outcomes and the probable direction and intensity of a person's will or desire. 11 Against this background, I propose that we reject the Different Reasons View on wellbeing and respect. We should reject, that is, the idea that either a person's general competence or the voluntariness of her choice or action should determine whether wellbeing or respect should have priority. Instead, we should recognize that both of these moral imperatives are relevant and that conflicts between them must be resolved by considering how much of each is at stake. This is the most plausible view both for the most competent people making the most voluntary choices and for the least competent people making the least voluntary choices.

The Traditional View Applied to Children
In an influential article, Harry Brighouse (2003) defends the Different Reasons View of wellbeing and respect as applied to children in particular. Brighouse holds that children have moral rights that protect their wellbeing as well as their happiness. However, he believes that what children want for themselves should be taken only as 'consultative', not as 'authoritative', while the reverse is true for adults. In other words, children's views are only relevant as potentially useful information about how best to promote their wellbeing. It is us adults, and in particular 'those adults who are morally charged with protecting the interests of those children,' who should make decisions for them in order that their interests are better promoted. I believe these views are rather common among liberals, which is unsurprising given their congruence with the more general traditional liberal view as described in the previous section.
According to Brighouse, if a view is authoritative 'it is regarded as the view that must be taken as defining the person's interests for the purpose of decisionmaking' (p. 692). This seems implausibly strong, as it implies that we should not only let adults do what they want, we should also assume that they are infallible in identifying their own best interests as well as the options that best promote them (though Brighouse indicates that there are exceptions to this general rule). Be that as it may, what is important is that respect for choice clearly trumps promotion of wellbeing for adults, at least under normal circumstances.
To treat a view as consultative, as opposed to authoritative, 'is to treat the person who expresses it as having a right to express her own view of her own 11 For an excellent overview over the philosophical debate on the nature and moral status of soft and hard paternalism in connection to voluntariness, see Hanna (2018).
interests, but not to treat that expression as sufficient grounds for action.' For the rest of us, this person's view is something 'I might find it useful to know … or I might not' (p. 693). In other words, there is nothing morally wrong with disregarding children's views on what they want for themselves, once we have heard them out.
Note the sharp dichotomy between authoritative and consultative views. When a view is authoritative, respect for it is the only relevant consideration. In other words, the view provides decisive reason. When a view is consultative, in contrast, it provides no independent reason at all, though it may indicate the presence of other reasons. On neither perspective does a person's view provide a pro tanto or defeasible reason.
As with the Different Reasons View in general, its application to children faces objections both against the arbitrariness of threshold-setting and against the relevance of the threshold-defining properties. The most obvious objection to Brighouse's view is its presupposed strict threshold between children and adults. 12 It seems implausible that when a person goes to bed on the eve of her 18th birthday, we should take her intention to move to another city in order to pursue an education merely as information on what might be good for her, but when she wakes up the next morning we should consider her same intention, for all practical purposes, as decisive to the point of defining her best interest. 13 A view with this implication may seem too extreme to be held by Brighouse or any other sensible liberal. I suppose the idea behind it is that once a person approaches the threshold of adulthood, her views on her own wellbeing become so accurate that the step to treating them as authoritative is in practice not so great. However, on a theoretical level the step is unavoidably large, as different reasons apply depending on whether a person is a child or an adult. Brighouse does not defend this implication of his view; perhaps he has not considered it. However, the implication should, I believe, be either acknowledged and accepted or else avoided by some development or modification.
One modification is to propose, as some authors have, that children gain authoritative say in matters that regard them gradually and partially, rather than all at once. As children mature, their sphere of authority expands, in terms of what sorts of things they have the right to control. For example, older children 12 For more on this objection and its relation to different views on autonomy, see Archard 2015. 13 Towards the middle of the article, in considering the objection that some children are more competent than some adults, Brighouse suddenly invokes the state's inability to consider all cases individually as a justification for setting a threshold between children and adults and treating them differently. However, otherwise the article clearly concerns not only the law, but how we should treat children quite generally.
can gain the right to decide who should care for them, if there are available alternatives, though not the right to completely abstain from being cared for. In fact, Feinberg indicates a gradualist account of autonomy rights for children in his seminal paper 'The Child's Right to an Open Future', though this is rarely noted (1992/1980, p. 95).
The gradualist approach has been further developed by Hugh LaFollette and by Samantha Brennan. LaFollette (1998) takes a gradualist position for instrumental reasons, mainly to promote children's future autonomy as adults. Brennan (2003), in contrast, declares a 'worry about grounding rights on consequentialist considerations' (p. 65). Brennan is not clear what should be the grounds of children's rights other than consequences, but clearly believes that there is some sort of independent reason to respect children's choices. Another kind of gradualism is endorsed by David Archard and Marit Skivenes (2009), who propose that, at some ages at least, denying children an authoritative say on their own interest should be presumptive only, such that individual children can prove themselves more competent than others and thereby be granted authority over their own interests (p. 19). Just as for Brennan, the grounds for this view seem at least in part to be nonconsequentialist.
Introducing degrees, nuances and exceptions improves on the traditional dichotomous view that Brighouse defends. However, there is a risk that the dichotomy remains at each degree and in each context such that, in any particular case, what the child wants for herself is either decisive or can be disregarded. Consider some child and some situation such that there is something substantial at stake, but it is difficult to say whether or not the child's view is authoritative. Perhaps the child wants to move to another parent or quit school. If we have decisive reason to grant her request, conditional on her being competent enough, then the moral cost of making even a very small mistake in judging her competence can be very great. If we grant authority when we should not have, we may allow a person to cause serious harm to herself when we had no reason to do so. If we do not grant authority when we should have, we may violate a person's autonomy when we have no reason to do so. This dichotomy is inherent in any view on which we have decisive reason to respect other person's choices if and only if certain thresholds are reached, even if the thresholds are sensitive to personal characteristics.
It is important in this context to distinguish two questions: (1) When (if ever) do we have independent reason to consider what a child wants for herself? and (2) When should we, all things considered, act in accordance with what a child wants for herself? The answer to question 1 is 'always' according to the Same Reason View and 'when she is sufficiently competent and her choice or action sufficiently voluntary' according to the Different Reasons View. Answering question 2, in contrast, does not adjudicate between these views. While question 1 and 2 are obviously different, they are not always kept apart in the debate on children's rights. 14 We have, of course, stronger reason to act in accordance with a child's view the wiser it is in terms of securing (or probabilizing) good outcomes, and weaker reason the more foolish. However, our reason to respect a child's view or choice need not vary with its wisdom. Indeed, non-consequentialist liberals (at least) would typically consider it unacceptably elitist to propose that we have independent reason to consider what adults want for themselves only if they are wise. It is not clear, on the face of it, why this would be any less problematic in the case of children.
Inspired by Norwegian law on child custody and protection, Archard and Skivenes (2009) propose that we have independent reason to listen to all children who have views on what they want for themselves, not only those that qualify for having an authoritative say. However, the authors are careful to point out that we need not accord the content of all children's views any moral weight (p. 18). Because, on their view, our reasons to listen are independent of whether or not doing so provides any useful information, Archard and Skivenes claim that they are improving on Brighouse's view to the benefit of children's voices. However, though he does not emphasize this point, Brighouse says, as we saw above, that a person whose views are merely consultative has 'a right to express her own view,' so I am not sure the difference in views is substantial.
In other publications, Archard (2018) seems more inclined to award greater importance to children's views of what they want for themselves, saying, for example, that 'children have a voice over their own affairs that must be appropriately weighted' (p. 143). Archard presupposes that the weight of a child's choice should be proportional to her competence (or 'maturity'). This could be interpreted as a non-dichotomous proposal such that we always have reason to respect children's choices, where the strength of this reason varies with the child's degree of competence. If competence is what distinguishes children from adults, morally speaking, it makes sense that our reasons of respect should vary in strength with degree of competence, incrementally, up to the point of maximal competence, if 14 Consider for example Brennan's (2003) statement of her view: 'The picture I prefer is one in which children move gradually from having their rights primarily protect their interests to having their rights primarily protect their choices. This reflects the transition of the child from being a creature whose interests are of moral concern, and hence deserve the protection of rights, to being a creature who can choose for herself ' (p. 63). It is clear that Brennan holds that all children have some moral rights, and it is clear that she holds that people who can choose (well) for themselves should be allowed to do so, even if they are children, but it is not clear why she endorses this second claim, and in particular whether or not she holds that we have independent reason to respect children's choices. there is such a point. Hence, though it is only gestured at in the literature, this is the gradualist account that I find most plausible.
So far in this section, I have discussed the problem with setting thresholds for when we should respect children's choices or treat them as authoritative, and how gradualist accounts may or may not ease these problems. I turn now to the more fundamental objection to the Different Reasons View as applied to children, which questions the very relevance of the property of being a child, as well as underlying properties such as competence. Note first that the fact that children tend to want things that are not good for their wellbeing or for their future autonomy is not a reason to disregard what they want. This is at most a reason against counting their view as decisive. What then, we might ask, could justify that either a threshold or a gradual scale of competence determines whether, or to what extent, a person's choice has moral weight? A number of authors have pointed to ways in which children lack rationality over time, such as their lacking the ability to make systematic plans for their own future (Scarre 1980), their lack of 'a unified regulative perspective' (Schapiro 1999), or their lack of the self-awareness necessary for identifying as a coherent and self-creating person over time (Richards 2010). It is not clear, however, at least not to me, why having long-term rationality should be a prerequisite for deserving respect for one's choices. Clearly, if I have no long-term desires and make no long-term choices, then there is nothing of that sort for you to respect. But that does not imply that you may disregard whatever choices I do make and desires I do have.
Imagine a person that magically comes into being, as an adult, to live for one day only. Imagine also that this person is very spontaneous, living fully in the moment (perhaps as a psychological reaction to her very limited lifespan). When this person wants to spend her one day in meditative contemplation, or rushing between art museums, or getting to know one other person as deeply as possible, do we have less reason to respect her choice than we would another person's because she lacks systematic long-term plans or because she does not (for obvious reasons) identify as a self-creating person over time? These may be good abilities to have, without doubt, making lives that last longer than one day richer in different ways. It may also be that having these abilities makes it more harmful to a person to have her choices and desires disrespected. An interference with my choice may be more harmful if it not only thwarts my immediate plans, but also undermines my long-term plans and my project of self-creation. None of this, however, indicates that we lack reason to respect the choices of people who live only for the short-term.
In sum, the case against intrinsic respect for children's choices seems to rest either on an invalid inference from our overall reason to interfere with their choices, because they are unwise, to our lacking any independent reason to respect their choices, or else on the dubious claim that long-term rationality is a precondition for independent respect. Though the debates are in large part independent, the matters at stake are very similar in the general debate on paternalism and in the debate on the moral status of children's views. 15

Respectful Paternalism for Adults and Children
The Different Reasons View on conflicts between care for wellbeing and respect for choice is that we should look to degrees of competence (or maturity) and voluntariness in order to determine which moral imperative should take priority in which cases. Applied to children, this view has been taken to imply one of two things. Either that care for wellbeing always has priority for children, along with the possibly independent imperative to promote development of future autonomy. Or that children sometimes have sufficient competence and voluntariness, in which case respect for choice should take priority, and sometimes do not, in which case care for their wellbeing (and future autonomy) should take priority.
I should admit that the thesis that respect for choice is conditional on competence and voluntariness might make sense on some views of the ultimate moral basis of respect. I have indicated why I am not impressed with some common arguments against respecting children's choices, but I cannot of course survey all views of autonomy to prove that they are either mistaken or that they do not imply that respect for choice requires substantial competence and voluntariness. I will be content, therefore, to propose the Same Reasons View as a possibility to be considered.
The mainstream view on wellbeing nowadays is that beings with much more limited cognitive abilities than human adults have wellbeing and thus, if we care about wellbeing, we should care about the wellbeing of these beings. There may be higher pleasures and achievements that only adult human beings can experience, but many beings can certainly experience pain. When they do, this is bad, regardless of what kind of being has the experience. What is controversial is how much and what sort of wellbeing different beings have and how their wellbeing is affected by different circumstances. My proposal is simply that we adopt the same approach to respect for choice. Though some beings, such as young children, have much more limited cognitive abilities than adults, they clearly make choices and have desires. They may not have desires as advanced as adults, but their choices can certainly be disrespected. When that happens, this is bad, regardless of the age or maturity of the child. What may be debated is how bad it is and how exactly we should determine what people want.
Liberal debate on children's rights, and to some extent liberal debate on paternalism towards adults, is often conducted from the perspective of allocating authority among individuals. The question is presumed to be 'Who should decide?' This is an important question in many contexts, mainly when laws, rules and regulations are designed and developed. However, in most everyday moral contexts the main question is rather 'What reasons apply?' 16 This moral question is also relevant when laws, rules and regulations are in place, since it is not obvious that we should always abide by such norms, even when they are enacted in legitimate ways.
When my child wants to do one thing and I believe he will be better off if I make him do something else, I should not ask myself 'Who decides?' but 'What should I do?' I may have the legal right to decide, but this does not settle the moral matter. The relevant question is what reasons apply to my choice of either letting him do what he wants or making him do otherwise. Central among these reasons, I propose, is care for his wellbeing, which in my view includes his present as well as long-term autonomy, along with many other dimensions, but the exact nature of wellbeing is not my topic. Central among my reasons is also, I propose, respect for his choice and desire. If my or anyone else's desires and wellbeing are affected, these are of course relevant considerations as well.
One explanation for why reasons of respect for choice have been taken to be either decisive or non-existent may be the difficulty involved in measuring the strength of such reasons if they do not belong to one of these clear-cut categories. For example, Archard and Skivenes (2009) say: 'It is not easy to understand how one might "weight" a child's views ' (p. 18). To some extent, this concern is relative to the framework of authoritative and consultative views inherited from Brighouse. As Archard and Skivenes go on to ask: 'But how can a set of views be authoritative but not completely? ' (p. 18). This is indeed difficult, given how the notion of an authoritative view is defined. However, we do not need this notion and its associated dichotomy between children and adults. Reasons can certainly have strength without being decisive. And when reasons conflict, we must assess their relative strength.
My own view is that the strength of our reasons to respect choice should be assessed in terms of how important the choice is to the person who makes it, and that we should consider not only the person's particular choice, but also her desires more generally. Sometimes people choose some alternative without realizing that this will frustrate their ambitions in other matters. In such circumstances, I believe that it can be most respectful to honor these other ambitions by preventing the choice, depending on how important the choices and desires affected are relative to each other. We need to establish a sort of vector of all relevant choices and desires in order to establish our overall reasons of respect, which can then in turn be weighed against other reasons. However, I mention this view primarily to provide an example of one available view. 17 The point is that we need some theory of how to determine the relative force and the direction of a person's will in order to establish the strength of the reasons of respect it provides.
This may all seem very complicated, but I believe it is not more complicated than aggregating the moral import of wellbeing. In some theories of wellbeing, wellbeing has only one component, such as pleasure. On such views, it is in theory (though not in practice) easy to establish which alternative best promotes wellbeing. However, many people endorse pluralist theories of wellbeing. Indeed, the commitment in the liberal debate on children's rights to promoting children's future autonomy indicates that many authors endorse a plurality of values of some sort. The value of future autonomy must somehow be weighed against the value of present wellbeing, since they often conflict. It does not then introduce a radically new sort of difficulty to also compare these two values to the value of respect for choice and desire.
As I have indicated, my view of wellbeing and respect for choice is not particular to children. When an adult person wants to do one thing and I believe she will be better off if I make her do something else, I should consider how reasons of wellbeing weigh against reasons of respect. This may seem very paternalistic. Indeed, on some views, it is paternalistic merely to consider a person's wellbeing as a possible reason for interfering with her (for example, Feinberg 1986). However, I do not agree with such views and I am not even sure what they say exactly. On almost any liberal view, the extent of harm makes some kind of difference. For Feinberg, for example, the more seriously a person aims to harm herself, the more voluntary she must be in order for us to allow her choice to be decisive. This makes sense. It is difficult to see how a view could be plausible that sets the threshold of voluntariness at the same level for failing to floss as for climbing Mount Everest without supplemental oxygen. It seems, therefore, that wellbeing considerations can justifiably enter one's deliberation in some ways.
I submit that the main concern of most liberals is not with whether or not facts about wellbeing are at all considered, but rather with the relative weight they are granted in relation to facts about adult's choices. If we reject the Different Reasons View, we are free to endorse the Same Reasons View. Taken literally, this view is consistent with rejecting the promotion of wellbeing completely and endorsing respect for choice as our only moral consideration, or the other way around. However, I have only discussed and I endorse the alternative view that we have reason both to promote wellbeing and to respect choice, for all people. I have not taken a stand on the relative strength of these two kinds of reasons, and I do not mean to. 18 For all I have argued in this article, it is fine to ascribe much greater importance to wellbeing than to respect, and it is fine to ascribe much greater importance to respect than to wellbeing. It is also fine to endorse more moderate views. 19 My point is not that children should get what they want, but that what they want matters, independently.

Conclusion with Illustrative Examples
I have argued that common views on the reasons we have to consider what children want for themselves are intimately related to a more general liberal view on the relationship between reasons of wellbeing and reasons of respectthe Different Reasons View. On this view, when these values conflict with regard to the same person, we should either promote wellbeing or respect choice, and which we should do depends on the person's competence and voluntariness. However, this view is mistaken, since competence and voluntariness are at best proxies for what really matters, which is wellbeing, choice and desire. This charge is not meant to undermine the liberal view, but is rather a friendly amendment, or perhaps a friendly reconstruction.
Given that respect for choice is but one moral imperative to be balanced against others, notably the imperative to promote and protect wellbeing, respect can be granted to children without radical or implausible implications. Consider some examples and how they are assessed on the Same Reasons View, filled in with what I take to be some relevant reasons.
For starters, should a seventeen-year-old be allowed to join her country's armed forces if this is what she wants to do? On the Same Reasons View, the reasons relevant for answering this question are the same reasons that would be relevant if an adult was in this same situation. Arguably, there are several such reasons. Respect for choice is a strong reason in favor of allowing this young woman to serve. Protection of her wellbeing is a strong reason against, given the probably great risks involved. Depending on the details, the balance of reasons may well be against. If the nation is fighting a just war and more soldiers are needed, this may be an additional reason for allowing her to serve, possibly tilting the scales. In any case, the fact that allowing her to make her own choices will foster her autonomy is arguably not a very strong consideration and does not alone explain why her choice should be respected. If the balance of reasons is against allowing her to serve, this in no way implies that there is no reason to respect her choice. It only implies that reasons on the other side are stronger.
For another example, consider a custody case related by Archard and Skivenes (2009). In Supreme Court ruling 2007-376 the Norwegian court declared that it was in the best interest of two siblings, aged 12 and 9, to live with their father, in part because this would allow them to stay in their local community. However, the children preferred moving to live with their mother. The court decided in their favor. It seems that the court deliberated in a manner consistent with the Same Reasons View. In weighing the conflicting considerations, the court took into account the children's consistency of opinion, the reasons they gave for this opinion, and their appreciation of the consequences of getting what they wanted. All these three factors make it more likely that the children's view was coherent with their larger set of desires. If the court had found that moving would have been positively harmful, however, it would likely have ruled, as it did in other cases, against the children's preference.
For one final example, consider a five-year-old who does not want to eat supper before she goes to bed. A child of this age typically has an extremely limited understanding of her own nutritional needs and the short-and long-term consequences of skipping supper. Her motives for refusing may be very banal, or 'childish'; suppose she is upset with her parents for some unrelated and superficial reason and aims to punish them by making them concerned over her refusal to eat. Still, I propose, we have some reason to respect her choice, just as we would have reason to respect a similar, and similarly childish, choice made by an adult. Not only may we have reason not to force-feed her, which may be very unpleasant and may violate her bodily integrity, but we may even have reason not to excessively pressure her or manipulate her into eating. These reasons, furthermore, need not have anything to do with promoting her future autonomy. Instead, sheer respect for her choice may be sufficient, given that very little is at stake in terms of wellbeing or other values. However, if the child is not eating properly over a longer period, so that her health is potentially affected, there are strong reasons of wellbeing to try to get her to eat, one way or other, and in extreme cases she may be given nutrition intravenously, against her will.
These examples hopefully help to indicate that respect for children's choices is consistent with comprehensive protection of their interests, including their interest in healthy development and future autonomy. When we interact with children, as well as when we design and implement laws and rules to govern their lives, we can and should recognize that we have reason to respect their choices, independent of any further positive consequences. In this regard, children have equal moral status with us adults. If we should more often make children do what they do not want to do, this may be in part because they are less invested in how their lives go in the long term, but mostly because, though we should care about respecting their choices, we should care even more about protecting and promoting their wellbeing.