Why rights are not optimisation requirements

ABSTRACT In this article I pursue the implications of the statement that constitutional rights are – as Alexy’s principles theory argues – optimisation requirements, and show that they are not. I argue that, applied to moral rights, optimisation obfuscates their nature, their relationship to human well-being, and the work they do in practical thought. As to constitutional rights, the fact that they belong in an institutional framework suggests some reasons for treating them like optimisation requirements in circumscribed cases. But these reasons are far from conclusive; and treating rights like optimisation requirements in other scenarios (such as cases of structural discretion) indicates that optimisation, as defined in the principles theory, does not assist us in thinking well about the structure of constitutional rights. Constitutional rights demand compliance with whatever the interests on which they are based demand – whether or not what they demand is antecedently clear – not with some purported optimisation requirement.


Introduction
In this article I pursue the implications of the statement that constitutional rights areas Alexy's influential theory arguesoptimisation requirements, and show that they are not. Alexy's theory originated as a reconstruction of German constitutional rights practice. Nonetheless, Alexy has always conceded that this reconstruction incorporated a general theory of constitutional rights to the extent that the latter concerns itself with 'the necessary structure of constitutional rights'. 1 Furthermore, it was argued early on that 'his theory is applicable more widely', 2 and Alexy's view on rights now informs a prominent strand of contemporary constitutional theory across different jurisdictions. 3 Thus, it is both important and timely to consider to what extent Alexy's conception of rights as optimisation requirements provides a sound account of the 'necessary structure of constitutional rights'.
I start by clarifying that in targeting the statement that rights are optimisation requirements, my critique in this article takes issue with Alexy's theory only to the extent that it purports to inform us about the nature or structure of rights. I do not question the limb of the principles theory that concerns itself with the method to resolve conflicts involving rights, namely, proportionality. While Alexy argues that conceiving of rights as optimisation requirements implies a commitment to proportionality, I deny this is the case: we can, and should, prise optimisation off proportionality in Alexy's theory, for a commitment to the idea of optimisation distorts proportionality analysisthat is, it distorts the process through which conflicts between rights, or between them and other interests, are resolved. The rest of the paper is devoted primarily to showing why this is so. It begins by testing the veracity of the statement that rights are optimisation requirements when it is applied, mutatis mutandis, to moral rights. My argument is that it obfuscates, rather than clarifies, their nature, their relationship to human well-being, and the work they do in practical thought. I go on to test the veracity of the statement when applied to constitutional rights. I argue that at first blush there may be reasons for treating constitutional rights like optimisation requirements in the case of coercive paternalistic interference, or cases that can be assimilated to it. But these reasons are far from conclusive; and treating rights like optimisation requirements in other scenariosinvolving conflicts of rights and structural discretionindicates that there are better ways of making sense of the nature or structure of constitutional rights. Constitutional rights are best understood as demanding compliancewithin the factual and legal possibilitieswith whatever the interests in which they are grounded require, whether or not the content of those requirements is antecedently clear; not with some purported imperative to realise their object to the greatest extent possible. It will be clear from this argument that I take Raz's conceptualisation of rights (as grounding duties in the interests of other people) as the most successful account of the nature of all rights. 4

Rights as optimisation requirements and proportionality
In denying that rights are optimisation requirements, I am not interested in questioning proportionality as a method of resolving conflicts between rights or between rights and other interests. This may seem odd, as Alexy claims that the nature of rights as optimisation requirements implies the necessity of proportionality analysis and vice-versa, and that he considers the claim that rights are optimisation requirements equivalent to the claim that rights call for proportionality analysis. 5 If Alexy were right on the co-implication of optimisation and proportionality, in discussing optimisation one must discuss proportionality, and vice-versa. I argue, however, that Alexy is not right on this point.
Möller has argued that optimisation does not imply proportionality on the basis that conflicts involving optimisation principles can be resolved in ways that do not invoke proportionality in the way Alexy articulates it. For example, Möller argues, some such conflicts should be resolved by appealing to deontological considerations. 6 Möller's argument is contingent on ascribing to proportionality, as articulated by Alexy, a narrow understanding that excludes deontological reasoning from its purview.
In a 2007 article, Kumm agrees with Möller on the point that the resolution of conflicts involving rights can, but need not always, take the form articulated by Alexy in his account of proportionality centred on balancing. 7 Contrary to Möller, however, Kumm takes Alexy's idea of the co-implication of optimisation and proportionality-as-balancing at face value. Accordingly, Kumm conveys the idea that the resolution of conflicts involving rights can and should take different forms by stating that rights are not optimisation requirements. 8 By this he means that when rights prevail over conflicting interests it can, but need not, be as a result of Alexy's optimisation model of proportionality-as-balancing. Similarly, Kumm maintains that rights do not necessarily trump conflicting interests, though they do sometimes work like 'trumps'; 9 and that they are not 'shields', 10 because they do not always work in accordance with the shields model (whereby rights take priority by default, subject to giving way to particularly powerful conflicting interests).
The key point is that both Kumm and Möller ascribe to proportionality analysis, as articulated by Alexy, a narrow meaning and conclude that, thus understood, proportionality is only one way of resolving conflicts involving rights. The difference is that while Möller reaches this conclusion from what he believes is a proper understanding of the idea that rights are optimisation requirements, 11 Kumm finds that the conclusion is precluded by the idea of optimisation, and so rejects that idea in the first place. 12 My own strategy in this article is different. Specifically, I take a broader understanding of Alexy's articulation of proportionality analysis than Kumm and Möller do. I then sever my discussion of optimisation from a discussion of proportionality, largely de-problematising the latter (in this section), and foregrounding the former (in the rest of the article).
I think that proportionality analysis, as rendered by Alexy specifically in terms of balancing, is not the problem that Kumm and Möller argue it is. Alexy argues that to determine whether a limitation on a right 13 is justifiable in accordance with proportionality analysis, the Law of Balancing should be applied, whereby the greater the limitation on the right, the greater must be the importance of realising the competing interest (or 'principle', in Alexy's terminology). 14 He further specifies this law by articulating the so-called Weight Formula, which takes into account the weights of the competing interests (or, as he calls them, principles) in the abstract, the intensity of the interference with each, and the reliability of our predictions about the detriment to each interest if it gives way to the other in the particular case. 15 This quantitative emphasis on weights might lead one to believe that, on Alexy's rendition, proportionality analysis rules out ever giving qualitative precedence to rights over competing considerations, either in accordance with a deontological model (as when rights are treated like trumps), or an 'excluded reasons' model (as when rights are treated like shields), or some other model. But just as Raz's interest theory of rights (despite its terminological emphasis on weighing and outweighing) accommodates cases of rights taking qualitative priority over other considerations, 16 so does Alexy's account of proportionality. As he explains, cases of excluded reasons (where a competing interest simply does not qualify as a justification for limiting the right) are 'cases in which the abstract weight of a principle is zero', and cases in which rights are treated as absolute deontological constraints on the pursuit of other goods are 'cases in which the abstract weight of a principle is infinite'. 17 I do not want to go so far as committing in a definitive way to the view that the Law of Balancing and Weight Formula lend themselves to accommodating all conceivable forms of sound moral reasoning relevant to resolving conflicts involving rights. But I believe Alexy makes a good prima facie case for that proposition. For the purpose of my present analysis, this suffices: thus, I will assume that proportionality analysis lends itself to helping us resolve conflicts involving rights in morally sound ways.
If the ground on which Alexy's theory has been found wanting by his critics (namely, the inability of proportionality/balancing to account for deontological and similar considerations) is less solid than Alexy's critics think, 18 it does not follow that the principles theory is without problems. The problem, as I argue in this article, lies with the optimisation limb of the theory, not with the proportionality limb. The problem, as I see it, is this: conceiving of rights as optimisation requirements is inconsistent with, and detracts from, Alexy's (at least prima facie defensible) account of proportionality. Far from implying proportionality analysis (understood as the correct method of working out relationships between competing interests in accordance with sound morality, given applicable legal constraints), the optimisation view is at odds with it. The idea that rights demand realisation of their objects to the greatest extent possible given relevant possibilities (=optimisation) disables sound moral reasoning about rights, for all the reasons the rest of this article sets out. Thus, realising the potential that proportionality analysis, as rendered by Alexy, has for assisting us in resolving conflicts involving rights in accordance with sound morality (given applicable legal constraints) is contingent on giving up the view that rights are optimisation requirements.

Alexy's claim
'Key to the entire theory' 19 developed by Alexy is the idea that rights, as protected in Constitutions and Bills of Rights, are principles, in the sense of 'norms which require that something be realised to the greatest extent possible, given the legal and factual possibilities'. 20 Once a righthitherto an abstract principle enshrined in constitutional documents, and thereby unrelated to concrete legal and factual possibilitiesis related to such possibilities in a concrete case, and is found to carry the day, then it turns from a principle into a rule. On this view, the idea of 'optimisation' is built into the very concept of 'principle', 21 given Alexy's own stipulation of what a principle is, which contains the formula 'realised to the greatest extent possible'. The key point is that, on this view, constitutional rights require the optimisation of the objects they protect. 22 By requiring the optimisation of the objects they protect, rights require their realisation to the greatest extent possible given the legal and factual possibilities. In what follows, for reasons of linguistic economy, I will sometimes replace the expression 'realisation to the greatest extent possible' with 'maximisation'. Note that maximisation, thus understood, is not the same as optimisation; rather, optimisation is maximisation (=realisation to the greatest extent possible) as conditioned by the legal and factual possibilities.
An example can be used to clarify what is meant by legal and factual possibilities in Alexy's account. Take the right to freedom of expression. Certain factual possibilities (the state of technological advancement) necessarily constrained free speech rights to non-virtual spaces before the invention of the worldwide web and allied technologies. Now that virtual communication is possible, if governments have a justifiable reason for suppressing certain kinds of speech that circulate through the web, for Alexy they must find the least restrictive means of doing soavoiding, for example, shutting down the web in its entirety, or, say, targeting non-problematic kinds of online speech if these can be disentangled from the problematic ones. All of this is implicit in the idea of maximisation (of speech) given the factual possibilities.
Legal possibilities, on the other hand, affect the extent to which a right is realised in a number of different ways. Thus, a legal system may limit the maximisation of a right's object by giving constitutional protection to competing values, or by failing to provide for constitutional review, or by limiting the remedies at Courts' disposal when they ascertain a breach of rights, etc. The first of these examples is the most important: 'Rules aside, the legal possibilities are determined essentially by opposing principles'. 23 The fundamental point is that constitutional rights, for Alexy, demand optimisation, that is, given these external factual and legal constraints, they call for their objects to be realised to the greatest extent possible.
The teleology of moral rights: optimisation or equilibrium?
Virtually all mainstream understandings of constitutional and human rights (from which I exclude critical accounts that view them as tools, say, of disciplinary or class power) share the view that the point of giving them constitutional protection is to manage certain relationships between individuals, or between individuals and institutions, or between individuals and organisations, in accordance with certain requirements of justice. 24 21 See also Klement (n 3) 173, 180. 22 This is the classical formulation and understanding of Alexy's theory. Alexy later claimed that what is to be optimised is the right itself, as a principle: Robert Alexy, 'On the Structure of Legal Principles' (2000) 13 Ratio Juris 294. As Poscher incisively argues, however, this move is ill-conceived, misconstruing the object and logic of optimisation: Ralph Poscher, 'Theory of a Phantom: The Principles Theory's Futile Quest for its Object' in Stanley L Paulson, Alexandre TG Trivisonno, and Julio Aguar de Oliveira (eds), Alexy's Theory of Law (Franz Steiner Verlag 2015) 129. Since the only reason Alexy revised the point was to maintain a distinction between rules and principles that was purely a matter of theoretical stipulation on his part, we can and should ignore the revision. Though not all constitutional rights need to be very important, the central case of a constitutional right is that in which this right institutionalises an important moral right. One may hypothesise that, consequently, the structure of constitutional rights should mirror, at least broadly, that of the moral rights they institutionalise, and that any divergence should be accounted for. Alexy himself agrees with this, where he speaks of the 'existence of human rights' (by which he obviously means our important moral rights) 25 and of the 'ideal dimension' of constitutional rights, which, he argues, 'lives on' despite their nature as positivised human rights. 26 As it happens, however, if constitutional rights are, as Alexy claims, optimisation requirements, then there is a significant divergence between their teleology and that of the moral rights they institutionalise.
The teleology of constitutional rights, on Alexy's view, is one of optimisation: the right requires its object to be realised to the greatest extent possible given the factual and legal constraints it necessarily has to reckon with, which limit the extent to which this outward, expansive movement can proceed. Moral rights, however, do not work like that. To appreciate this, let us adapt Alexy's optimisation view to moral rights. Specifically, let us test the hypothesis that moral rights require the maximisation of their objects given relevant possibilities. Since we are speaking of moral rights, the relevant possibilities here would be factual possibilities (as with constitutional rights, these are inescapable) as well as, intuitively, the moral possibilities. 27 In what follows, for the sake of simplicity, I will not, in the main, keep mentioning factual possibilities.
Onenot necessarily the most straightforwardway of interpreting the statement that 'Moral rights require their objects to be realised to the greatest extent possible given the moral possibilities' is: 'Moral rights require their objects to be realised to the greatest extent possible given the moral possibilities in respect of the requirement to realise it'. Thus interpreted, the optimisation definition of moral rights seems true, but is also unnecessarily redundant. To say 'Moral rights require their objects to be realised to the greatest extent possible given the moral possibilities in respect of the requirement to realise it' is the same as saying: 'Moral rights require their objects to be realised to the extent that it is morally justified'. Or: 'The appropriate realisation of moral rights is determined by relevant moral considerations'. But this is true of any moral norms, so it does not tell us anything distinctive about the nature of rights. Nor does it really tell us anything insightful about the nature of moral normsit is just a tautological restatement of the moral nature of moral ought-statements.
A more straightforward, non-redundant, interpretation of the statement 'Moral rights require their objects to be realised to the greatest extent possible given the moral possibilities' is: 'Moral rights require their objects to be realised to the greatest extent possible given the moral possibilities in respect of their realisation'. But there are several problems with such an understanding of moral rights. Moral rights justify duties that help towards (without, generally, guaranteeing) 28 the realisation of their objects not to the greatest extent possible, but only insofar as they dono less, but also no more. It is entirely In some cases, these considerations may include the need to comply with the law, so that the moral constraints on the realisation of moral rights may include some of the 'legal possibilities' (to use Alexy's terminology) that constrain the corresponding constitutional right. 28 Raz (n 4) 194, 199. possible that the relevant moral possibilities permit the realisation of the object of the right quite beyond the point at which the right requires it to be realised. Indeed, the moral possibilities may recommend that the object of the right be realised beyond that point. But the right itself has nothing to say about that, for rights are one of the grounds for duties (which are peremptory), not for supererogatory activities (which are voluntary). At most, we could say that the interest that grounds the duties whose performance the right requires may justify supererogatory activities not required by the right itself.
To clarify through an example: I am about to drown in strong currents and you happen to pass by. My moral right to life is a ground for your moral duty to throw me one end of the rope lying at your feet, in an attempt to help preserve the right's object (my life). But I am weak and cannot reach the rope. Factual possibilities (you being there, able-bodied, and a decent swimmer) permit you to do more, and risk your own life by jumping into the water to rescue me, thus 'realising' the object of my right to a greater extent. The moral possibilities (resulting from considerations about courage being a virtue, selfsacrifice being morally praise-worthy, and my life being valuable) may even recommend that you do so. My right to life, however, has nothing to say about that. If my right to life required the realisation of its object to the greatest extent possible, given only how far it is factually and morally possible to realise it, people would have unreasonably far-reaching duties towards me.
Thus, to say that moral rights require the optimisation of their object seems fatuous at best (first interpretation) and absurd at worst (second interpretation). Moral rights require their objects to be realised no less, but also no more, than relevant moral considerations require. Such considerations will take into account not only aspects of the well-being of concrete and abstract others (including those who have duties correlative to my right, people whose interests happen to conflict with mine, as well as future generations), but also aspects of the well being of the right-holders themselves.
The latter circumstance gives us another reason to doubt the optimisation view (in the secondthe non-redundantsense I articulated) when applied to moral rights. Rights are grounded in interestsnamely aspects of the right-holders' well-being, 29 and their point is precisely to serve those interests. It is therefore unlikely that it is in their nature to extend in directions that fail to promote relevant interests of the right-holder, even if this is wholly within the bounds of the factually and morally possible. A right, after all, is a ground for other people's duties to protect or promote in particular ways that aspect of the rightholder's well-being in which the right is grounded; it would, therefore, be odd to hold other people duty-bound to do (or refrain from doing) something in the name of another's well-being, if the required action or omission did nothing to advance it.
Let me illustrate. You saved me from drowning, but I am now lying in bed at the hospital, brain-dead, with no possibility of recovery. I may be made to persist several years in a painless vegetative state by being drip-fed nutrients via a needle, being kept clean, and being turned around every couple of hours to prevent pressure sores, until I die a natural death. The factual and moral possibilities allow for the object of my right to life to be optimised in this way: there are doctors and nurses and resources enough. Being kept in a vegetative state will not harm me, but neither will it benefit me. Assume, for the sake of simplicity, that it is not possible to establish what my preferences might have been if I anticipated that this might happen to me. It seems counterintuitive here to say that my interests require that I be kept alive. It seems equally implausible to claim that they require the oppositethat is, that I be let die. 30 Yet, if my right to life requires my life to be realised to the greatest extent possible given the factual and moral possibilities, then under the optimisation view we would be forced to conclude that the doctors and nurses have a moral duty to keep me alive.

Moral rights, well-being and self-interest
Another reason for denying that the teleology of moral rights is one of optimisation hinges on the distinction between self-interest and well-being. To say that people are rationally self-interested is to say that they want to maximise their utility, that is, to employ efficient means to pursue their freely chosen ends. 31 If rightslike (many) material goodsare seen merely as resources at the disposal of rationally self-interested individuals, it may be plausible to speak of moral rights as requiring the maximisation of their object given relevant possibilities. The problem with this is that such a view of rights would only make sense in a morally impoverished universe.
Suppose we have a moral right to hold wealth we acquire through work, or that is bequeathed to us. This moral right will not be absolutefew if any rights are. Assume that the wealth people acquire is subject to redistributive policies that aim to ensure everyone a tolerable standard of living, and that these policies are not only morally justified, but morally obligatory, in the sense that no or less redistribution would fail to meet the requirements of justice. In Alexy's terms, we would say that there are competing norms (relating to people being ensured a tolerable standard of living) that limit the realisation of the object of the right to hold wealth. Now suppose that, were welfare and redistributive policies more robust, there would be somewhat less wealth overall (because of the disincentivising effect of higher tax rates), but society would turn out to be more cultured and virtuous in various ways. Suppose, however, that we have no moral right to live in such a society: though each of us, objectively speaking, has an interest in living in such a society, with all its advantages, this interest may not be enough to require the government to take steps to realise it.
In such a scenario, if the government has a duty to maximise the object of our right to hold wealth given the factual and moral possibilities, it is precluded from increasing tax in the name of living in a more cultured and virtuous society. Since the government has no duty to bring about such a society (for by hypothesis we have no right to live in it), it is morally possible for it to maximise our wealth by not increasing the tax rate with a view to bringing about such a society. And since it is morally (and, we shall assume, factually) possible, then the optimisation logic requires that it be donethat the tax status quo be maintained.
This is a startling conclusion. It seems to only make sense under the stipulation that moral rights are there to serve people's self-interest rather than their well-being. Rationally self-interested actors, as we have seen, seek to maximise their utilitythat is, to pursue 30 These interests may, of course, require other thingsthat I be cared for in a way that does not facilitate the onset of disease, for example. their goals in efficient ways. On this view, liberty rights are resources: it is in people's selfinterest to maximise the objects of these rights, as they enable each individual's pursuit of what she cares about (be they the pursuit of virtue and culture, or something less lofty). The right to hold wealth, from this perspective, is a resource whose object must be maximised given whatever is factually and morally possible in a universe that values self-interest over well-being. In such a universe, the right to hold wealth would preclude the government from raising tax in the name of a more cultured and virtuous society.
But we might well reach a different conclusion if, rather than in self-interest, rights are grounded in aspects of our well-being, 32 where our well-being captures the idea of leading a successful life. 33 If the point of rights is to help us lead successful lives, rather than maximise our self-interest, our moral right to hold wealth would not be a reason to hold the government duty-bound not to increase taxes. For our quality of life would be likely to be higher if we lived in a somewhat less prosperous, but more cultured and virtuous society. The right may still ground duties not to redistribute wealth beyond a certain point (we need not pinpoint where), but it would otherwise leave room for the government's pursuit of goals that increase people's well-being, even in circumstances where the government has no duty to pursue them.
In a moral universe in which well-being matters (as it should) more than self-interest, moral rights do not require the maximisation of their objects given the factual and moral possibilities. They require their objects to be realised no less and no more than the interests in which the rights are grounded demand, even when it is morally and factually possible to realise them to a greater extent. When it is morally and factually possible to realise the object of a right to a greater extent, the right itself does not require such realisation, but relevant considerations related to human well-being (including the value of the interest in which any given right is grounded) may or may not recommend that it be realised to a greater extent, including to the greatest extent possible. On this view, a requirement to realise the right's object to the greatest extent possible (given factual and moral possibilities) does not adequately capture the teleology of rights.
My argument here is not that if you think that the optimisation view accurately describes the nature (not just of constitutional rights, but also) of moral rights, you necessarily espouse a view of rights centred on self-interest rather than well-being. It is, rather, that if you define rights that way, your definitional moves arewhether or not you intend them to becongenial to the self-interest view, and hence they authorise exercises of practical reason that are consistent with that view, rather than ones informed by the richer concept of well-being. Understanding moral rights as optimisation requirements seems to gesture towards a moral universe where self-interest matters more than well-being.
In a related way, it is also suggestive of a moral universe in which the individual and society are pitched against each other, echoing the hackneyed picture of individual rights that communitarian critiques are fond of attacking. Rhetorically, the optimisation view on moral rights depicts a two-stage process: first, there are rights in the abstract, demanding the fullest realisation of their objects; secondly, there are the factual and moral possibilities supervening on rights in the real world, curbing their potential for blossoming. This makes it look like a loss of value is always necessarily involvedthe added value the right would have had for the right-holder, if only those external constraints were not the necessary evil that they are. All this ratifies an unnecessarily conflictual view of morality, in which individual interests and the rights they ground are inevitably or routinely opposed to each other and to those of society. 34 But individual interests and the rights they ground do not emerge in a normative vacuum, and they are not outliers to the rest of normativity. They belong in the same moral universe as other moral goods, and, like them, they come into being with inbuilt limitations that are determined by the fact that they derive their value from their contribution to human life and its quality. 35 Those who take an optimisation view on moral rights may well agree with this; presumably, they are convinced that conceiving of rights as requiring the maximisation of their objects given relevant possibilities assists us precisely in the goal of determining the extent of each right relative to competing values within this unified moral universe. But conceptualising rights as optimisation requirements is just not serviceable to that goal. This is because for pro-optimisation theorists, as Poscher puts it, rights, 'in an ideal world, … would satisfy their unconditional formulations, because they would not have to be relativized with regard to the legal and factual possibilities'. 36 To this extent, the idea of realisation to the greatest extent possible disables a sound appreciation of the interrelationships between rights and other interests. It does so by obscuring the point that, at least in many cases, the limitations which rights are subject to are part and parcel of the conditions that make rights valuable, rather than a necessary evil we must put up with in a non-ideal world.

From moral rights back to constitutional rights: moral constraints
My argument so far is that Alexy's definition of rights as optimisation requirements, when applied, mutatis mutandis, to moral rights, is either redundant, or invites an impoverished understanding of rights and their place in morality. Specifically, it obscures both rights' relationship to well-being, and the distinction between peremptory and non-peremptory reasons for action. There are, after all, few if any goods whose maximisation given relevant possibilities is a moral requirement, even when we do have moral rights that have those goods as their objects, and even when their realisation to the greatest extent possible may be a very good thing.
An objector might grant all this about moral rights, but argue that this says nothing about the adequacy of Alexy's theory. That theory is, the objector will point out, about constitutional rights. These are positivised rights that are part of an institutional framework, with its own distinctive demands, so that optimisation may be meaningful and useful in the context of constitutional rights, even if it is not when applied to moral rights. If so, then a possible disjunction between moral and constitutional rights may always have been on the cards in Alexy's theory. In practice, this disjunction would mean that the following two scenarios are always possible. First, the 'legal possibilities' that constrain the realisation of a constitutional right's object may prevent its realisation to the extent that the corresponding moral right would require (one need only imagine a competing 34 I follow Raz in arguing that they are not: Raz (n 24) 193-216. 35 See Raz (n 24) 317-19 (arguing against the view that morality and prudence are normally at odds). constitutional value that happens to be legally overprotected). Secondly, and conversely, the constitutional right may require its object to be realised to a greater extent than the corresponding moral right requires; and even to a greater extent than relevant moral considerations make permissible.
This may sound odd, our objector will concede. Ideally, she will say, the requirements of constitutional rights would map onto those of corresponding moral rights; but we live in a non-ideal world, populated by non-ideal decision-makers, and the optimisation framework is supposed to cater for that. On this view, a requirement to optimise the objects of constitutional rights: (a) relieves judges from some of the hard moral issues that deciding rights cases involves, economising their epistemic resources; and (b) it bolsters their legitimacy by making their job more transparent and less moralised. 37 I will address (a) in the next section. In this section I want to dispute (b), by arguing that considerations about legitimacy militate in favour, rather than against, aligning as far as possible our understanding of constitutional and moral rights. One need not be a natural lawyer to accept this. It is enough to accept that all authorities, even illegitimate ones, claim legitimacy. 38 When it comes, specifically, to those authoritative directives consisting in guarantees of constitutional rights, to claim legitimacy is equivalent to claiming that these rights (or at least many of them) mirror the especially important moral rights we have, as a matter of justice and sound moral reasoning, vis-à-vis the State (as well as, at least indirectly, other individuals and organisations). If we take the possibility of legitimate authority seriously, and if constitutional rights are (as they must be) central to any political authority's claim to legitimacy, our understanding of constitutional rights must therefore make it possible that constitutional rights be capable of correctly channelling moral requirements into constitutional law. But this is just what the optimisation theory fails to do in demanding that the object of rights be maximised given legal and moral possibilities: this requirement disables a sensible use of practical reason on the part of decision-makers adjudicating conflicts involving rights.
Let me illustrate. Suppose a country's Constitution gives unqualified protection to the right to bear arms, and that the wording of its guarantee of the right to life is limited to a prohibition against the State depriving people of life without due process. Suppose, too, that this country has a high rate of murders, and that empirical research demonstrates a strong correlation between it and the general availability of weapons in that country. Finally, suppose that the interpretive conventions of constitutional judges in this country are relatively contested. All judges agree that they primarily need to engage with the constitutional text (rather than, say, unwritten principles), but some judges favour originalist interpretations, others literal ones, and others still a moralised constitutional hermeneutics. If, as an optimisation requirement, the constitutional right to bear arms demands that people's ability to bear arms be realised to the greatest extent possible given the legal and factual possibilities, then the optimisation view seems biased towards the status quo and against the introduction of stringent gun laws. For it tells us that by their very nature (as optimisation requirements) constitutional rights call for their object's maximisation given the (factual and) legal possibilities, and the legal 37 Thank you to Dimitrios Kyritsis for putting the point so lucidly to me. possibilities include literal approaches to the constitutional text, which give the right to bear arms unqualified protection (in contrast to the very qualified protection afforded to the right to life). This outcome, apparently recommended by the optimisation view, seems to go against what a sensible use of practical reason would recommend as a matter of people's moral rights.
Alexy is, however, a non-positivistspecifically, an non-positivist of Radbruch's persuasion: he thinks, in other words, that patently unconscionable laws are not genuine (valid) laws. 39 In light of this, Alexy's account of the nature of constitutional rights makes room for moral unconscionability as one of the background 'legal possibilities' that always condition the maximisation of the objects of constitutional rights. Thus, on the assumption that it is morally unconscionable to facilitate much unnecessary loss of life by giving unqualified protection to the right to bear arms, an understanding of constitutional rights as optimisation requirements may be able to account for the imposition of certain limits on this right without forcing on us the conclusion that such limits necessarily violate the right qua constitutional right that demands the optimisation of its object.
Let us assume, then, that this qualification is built into the optimisation view: namely, that the reference to 'legal possibilities' always rules out the expansion of a constitutionally protected right beyond the point when it would be patently unconscionable to protect it. But since unconscionability sets the bar so high, this still leaves the relevant constitutional right, potentially, on a course to expand well beyond the point at which the corresponding moral right would. Ultimately, as the example used illustrates, the extent to which the legal possibilities that condition the realisation of the objects of constitutional rights (understood as optimisation requirements) incorporate moral limits more stringent than unconscionability would depend on constitutional sources and the interpretive practices and conventions of relevant decision-making communities. In jurisdictions with constitutional review, these communities will be primarily constitutional courts. Because canons of constitutional interpretation are often contested, it may not be easy to come to a straightforward conclusion even within a single jurisdiction. The problem with Alexy's theory, as I argue below, is that it biases constitutional interpretive practice towards overprotection of constitutionally enshrined rights.
Granted that the realisation of the object of a constitutional right is constrained by the legal possibilities, the maximum level of realisation these possibilities allow for may fall short of the level of realisation morality (or justice) requires. This does not follow from Alexy's definition of constitutional rights; it is an inevitable consequence of positivising moral rights into constitutional rights (or it must seem so to all but classical natural lawyers). Thus, the extent to which such divergence between moral and legal requirements occurs will be a contingent matter of fact in any given jurisdiction. What is, however, purely an artefact of the optimisation limb in Alexy's theory is the reverse scenario: if (apart from the factual possibilities) all that stands in the way of a constitutional right's object being realised as much as possible are the legal possibilities, these (depending on their content, as established following decision-makers' interpretive conventions) may allow the object to be realised: (1) Over and beyond what the corresponding moral right requires (thus creating excessive burdens on duty-holders); (2) Over and beyond what supererogatory considerations recommend (thus creating excessive burdens on duty-holders and failing to serve any useful purpose); (3) Over and beyond what justice permits, so long as the threshold of Radbruch-type unconscionability is not reached (thus creating excessive burdens on duty-holders, failing to serve any useful purpose, and violating moral requirements).
One may legitimately doubt a definition of constitutional rights that has the potential for automatically ratifying all these scenarios. Let me be clear here. Mine is not a natural law argument to the effect that we need to define constitutional rights in such a way that allows us to say that morally incorrect instances of balancing (which overprotect rights, as in (a), (b), and (c) above) are legally invalid. Rather, starting from the point that morally sound legal balancing of constitutionally protected rights must be central to claims to legitimate authority, I argue that our definition of rights must be such that it is amenable to that claim to legitimacy. It does not follow that when judges get the balance wrong, their decisions are not valid law. But a definition of rights is inept if it demands that we (and institutional decision-makers) favour balancing outcomes that overprotect rights when the legal possibilities would be equally compatible with balancing exercises that do not overprotect them. Yet, this is precisely what the optimisation view demands of us in its insistence that rights require their objects to be realised to the greatest extent possible given the factual and legal possibilities. Key to my argument is that the legal possibilities (depending as they do on contingent aspects of individual legal systems) may well permit more protection of constitutional rights than sound morality demands; and if they permit it, then the optimisation view requires such overprotection by mandating the realisation of the right's object to the greatest extent possible.
Here, the objection might be raised that Alexy claims that optimisation implies proportionality. If this is true, then the legal possibilities that condition the right's object's maximisation are to be determined, in every constitutional system, in light of a prescription that a right's object be realised, other things being equal (that is, to the extent that other legal possibilities permit), no more than the duties grounded in the right morally requirean extent determined by proportionality analysis. The problem is that it is not true that optimisation implies proportionality, much less the reverse (which Alexy also maintains). To understand 'legal possibilities' in light of a prescription that a right's object be realised, other things being equal, no more than the duties grounded in the right morally require, is to replace the concept and logic of optimisation with something else.
It follows that the potential for overprotecting rights that a definition of constitutional rights based on optimisation licenses is not forestalled by what Alexy calls the Law of Competing Principles. This is the idea that, in the light of the concrete circumstances of a case, we decide, via balancing, on the morally correct scope of any given right visa-vis any opposing norms that condition its realisation to the greatest extent possible, thus turning the right from an abstract principle offering prima facie protection into a dispositive rule offering or denying definite protection in a particular case. 40 The Law of Competing Principles belongs to the proportionality limb of Alexy's theory, which, I 40 Alexy (n 1) 50-54, 406-08. have argued, is independent of, rather than implied by, the optimisation limb. If the Law of Competing Principles helps avert the scenarios listed above at (1-3), it is only on condition that it be taken to invalidate the optimisation precept according to which the right's object needs to be realised to the greatest extent possible, given only the legal and factual possibilities.
I suggest, therefore, that we should reject the idea of optimisation as central to our understanding of constitutional rights, and that we should bring closer our understanding of constitutional rights with a normatively sound understanding of moral rights. Constitutional rights require their objects to be realisedfactual and legal possibilities permitting no more and no less than morally demanded by the interests in which the rights are grounded.

Overprotecting constitutional rights
My argument so far is to the effect that the optimisation view is irrational. Its baseline (maximisation given relevant possibilities) is arbitrary. Ironically, Alexy's theory is intended as an alternative to theories that 'overprotect [rights] relative to what political justice requires'; 41 yet, given how it is formulated, the optimisation limb of that theory (if not the proportionality limb) naturally tends towards overprotecting rights in just that way.
There is no good reason why the realisation of the objects of constitutional rights should default to the upper limit of what facts and the law permit. Or isn't there? Perhaps the fact that constitutional rights are not just moral rights, but institutionalised moral rights, makes a difference. Perhaps the whole point of institutionalising our especially important moral rights is not just to codify them, but to give them enhanced recognition and protectionmore protection than they actually, morally speaking, call for. For we need to guard against official decision-makers' overreach and abuse, and the errors that they night make in the process of considering and balancing competing values. On this view, we have good reasons to relieve judges from some of the hard moral issues that deciding rights cases involves, for too much is at stake in the event that they erroneously under-protect rights. If it is true that others (or society) are poor judges of the value that particular options, activities, occupations, or courses of action have for a right-holder, 42 then the right-holder's constitutional rights, as the institutionalised counterparts of her important moral rights, perhaps need to be more generous than the latter. Perhaps they need to be biased towards the right-holder's self-interestthe maximisation of her utility, as she herself sees it and defines itand against her wellbeing, for we cannot trust decision-makers to correctly work out what the latter means.
On this view, even if an aspect of my well-being is not important enough to morally require the State to abstain from interfering with it in a particular way, or to take a certain action to promote it, it may be constitutionally appropriate, in light of the risks of under-protecting it, to artificially inflate its importance, and treat it as if it does ground those duties. 43 But if constitutional rights, as institutionalised moral rights, protect more than their un-institutionalised counterpart (so that they really do require the realisation of their objects to the greatest extent possible given legal and factual possibilities), it seems likely that this is so only in a circumscribed set of cases: those involving coercive interference in the name of the right-holder's own interest, or in the name of the general interest 44 (particularly, perhaps, when appeals to the general interest are relatively amorphous, and likely to mask paternalistic interference). For reasons that I will explore in the next section, the idea that constitutional rights require their objects' optimisation is often unhelpful in cases where the reason for restricting a constitutional right is someone else's constitutional right. As I argue in this section, it is also implausible to apply this idea to cases of interference in the name of either the right-holder's own interest, or the general interest: (a) Where the purported State duties that the right grounds are positive, rather than being negative duties of non-interference; and (b) Where the State's interference is not coercive in nature.
As to (a), since positive duties are generally quite burdensome, it is irrational to expect that a decision-maker should hold the State bound by them just because the factual and legal possibilities do not rule out doing so, when the decision-maker is not, in conscience, convinced that the relevant right actually requires the imposition of such duties. As to (b), non-coercive interference is much less concerning than its coercive counterpart. If the idea is that we should err on the side of not interfering with an option that the right-holder may (for reasons we fail to appreciate) care very much about, the methods and outcomes of the interference seem to matter a great deal. 45 The disfavour for coercion (say, being seized by the police) has to do (at least in part) with the immediacy and success (at least temporary) with which it removes the option whose utility to the individual the State fails to appreciate. Non-coercive interferencesay educational measures, or failure to financially sponsor my option on a par with other optionsis unlike coercive interference in these respects.
In sum: if it is at all plausible to overprotect constitutional rights by treating them like optimisation requirements, it seems more so in cases of coercive interference in the name of paternalism or the general interest. I want to conclude this section by casting doubt, however, on the strength of these reasons based on risk-aversion even in these circumscribed cases (to reiterate, the risk to be averted here is that of decision-makers' discounting the value of the right's object to the right-holder). I will argue that adopting the optimisation view in such cases may well obscure the role different constitutional rights should have in managing the relationship between individual freedom and coercive State action.
Suppose we must decide on the constitutionality of laws criminalising animal sacrifice in religious rituals. Suppose we are concerned with their constitutionality insofar as they apply even to instances of quick killing of an animal, and suppose we agree that animals do not have rights. From an optimisation perspective, the logical way to proceed is to start by 44 See ibid. 45 Think of anti-smoking educational campaigns. These, unlike coercion, have no guaranteed outcomes. Those to whom smoking continues to bring sufficient utility (as they themselves see it) will continue to smoke; if the others give up the habit, then it is because their own appreciation of the utility they derive from smoking will have altered as a result of the campaign. treating these instances as falling under the ambit of the relevant right 46 (freedom of religion), and then consider whether the legal (and factual) possibilities enable us to say that they are constitutionally protected, and that their criminalisation is unacceptable. This approach, however, carries with it a rather crude framework of analysis. It is not conducive, specifically, to making the argument that people may not have a religious freedombased right to do these things even if their criminalisation may well be unconstitutional.
Abandoning the view that constitutionally protected goods such as religious freedom need to be realised to the greatest extent possible given relevant possibilities enables decision-makers (and the rest of us) to make more nuanced judgements about what genuinely contributes to human well-being and what does not. It may lead us to realise that what is really objectionable in the criminalisation of these instances of ritual killing of animals is the criminal nature of the interference, rather than the fact of interference itself. We might conclude that the reason why the relevant laws are unconstitutional is not that they interfere with a purported religious right to sacrifice animals, but that they violate the right-holders' right not to be subjected to the inhuman and degrading treatment inflicted by coercive interference in the absence of harm to other rightsbearing subjects. As Raz explains, in the absence of such harm, coercion used as a means of removing morally unacceptable options is unjustifiable because it symbolically demeans the person coerced upon, as well as indiscriminately interfering with her other options tooincluding the acceptable ones. And even in the presence of harm, coercion may be unjustifiable if the harm it inflicts exceeds that which it averts. 47 There is a difference between concluding that concerns about the general interest are not sufficient to justify coercive removal of the right to perform quick ritual killing (a conclusion that might well obtain under the optimisation view, insofar as it requires freedom of religion to be realised to the greatest extent possible), and concluding that my right not to be subjected to degrading treatment makes it unconstitutional for the State coercively to remove that option from me, even if the right to religious freedom does not cover it. If the optimisation view on rights is not conducive to drawing such distinctions, it seems that it does not properly assist our practical reasoning in thinking well about rights. But remember that our reason for endorsing the optimisation view in cases of coercive interference was precisely scepticism about decision-makers' ability to think well about rights. Even if that scepticism is justified, it looks like the touted remedythe optimisation viewmay compound the problem of poor decision-making, rather than resolve it.

Conflicts of rights
Whether or not the idea of treating constitutional rights like optimisation requirements is plausible in the circumscribed cases described above, it may obfuscate hopelessly the task of decision-making in conflicts between opposing constitutional rights. Suppose X and Y are conflicting moral rights, and constitutional rights [X] and [Y] their institutionalised counterpart. X is the right to free speech, including the right to make allegations that damage others' reputation; and Y is the right to one's reputation, grounded, say, in a 46 A necessary consequence of Alexy's principles theory of rights is 'a wide conception of the scope of constitutional rights'. Rivers (n 2) xvii, xxix. 47 Raz (n 24) 412-20. broader right to private life. Now suppose that the truth of one's defamatory allegations correctly sets, morally speaking, the boundary between X and Y for people in the public eye. Y, that is, grounds the moral duty of X-holders to refrain from making defamatory allegations unless they are true. X grounds the moral duty of Courts and other State agencies not to suppress or punish defamatory allegations when they are true. X and Y, here, are in equilibrium. Now suppose that, in a relatively new constitutional system (with limited precedents and contested interpretive conventions), the legal possibilities are consistent with any of these outcomes: (a) People in the public eye forego [Y]; (b) [X] includes no right to make defamatory allegations about anyone at all, regardless of their truth; and (c) The truth of one's defamatory allegations sets the boundary between [X] and [Y] for people in the public eye. If, as the optimisation view tells us, both [X] and [Y] require their realisation to the greatest extent possible, beyond the duties morally justified, respectively, by X and Y, these duties will no longer offer guidance on how to reach an equilibrium between the rights. Decision-makers will be facing two conflicting (optimisation) requirements, and find it logically impossible to resolve the conflict.
The Law of Balancing and the Weight Formula can be used to resolve the conflict in favour of (c), but remember that they belong to proportionality analysis, which, pace Alexy's co-implication claim, is logically inconsistent with conceptualising rights as optimisation requirements. Proportionality postulates that a correct balance between rights can be found precisely because (on the understanding ascribed to it in this paper) it assumes that neither right extends (even potentially) any further than it morally should. Conversely, optimisation, requiring maximisation of a right's object subject only to factual and legal possibilities (which, as a matter of contingent social fact, may permit more than the morally required), licenses outcomes in which constitutional rights are overprotected relative to the moral rights they institutionalise (as long as overprotection does not fall afoul of Radbruch-style unconscionability). 48

Structural discretion
I have argued above that there is a mismatch between the logic of optimisation and the logic of proportionality. One implication of this is that even assuming (as I do in this paper) that the proportionality limb in Alexy's theory facilitates rights determinations in accordance with sound morality, proportionality cannot function simply as a qualification or corrective to the optimisation limb. Optimisation has its own distinctive logic (following from how it is articulated in the definition of rights as optimisation requirements); because of this, even when wedded to proportionality analysis (the Law of Balancing, the Weight Formula, the Law of Conflicting Principles, etc), it continues to interfere with such analysis, distorting a sound assessment of the interests at stake. This is particularly apparent in cases of so-called structural discretion.
When Courts defer to Parliament's discretion, they often mention Parliament's superior expertise on particular matters, or its democratic legitimacy to resolve socially controversial problems. In these cases, the assumption is that there is one resolution to 48 In the free speech v reputation case discussed in this section, while it is clear to me that (a) and (b) are inferior to (c), it is unclear that settling on either of them would be patently unconscionable. In other words, the unconscionability constraint built into the concept of 'legal possibilities' does not seem to rule out (a) and (b), which means that we can realistically envisage the scenario in this example. the problem that is the correct one -Parliament being better placed to find it. These are cases of epistemic discretionwhether factual or normativewhose existence Alexy's theory readily acknowledges. 49 But deference can also be used as a method of dispute resolution in situations in which practical reason indicates that the values at stake do not require one uniquely correct outcome.
Suppose I have a right to same-sex relationship recognition (grounded in my more general right to family life), and that I live in a society in which relationship rights follow the performance of either religious or civil ceremonies of the appropriate kind. If I have such a right under these circumstances, I must have access to at least some of these ceremonies, which means that someone is under a duty to perform such a ceremony for me if I meet the eligibility requirements. Let us assume that the right to freedom of religion rules out forcing religious organisations to perform the ceremony. Now suppose that a public official who happens to be religious refuses to officiate at my civil ceremony, in a context in which another official is available to officiate. Morally speaking, under these circumstances there may be no uniquely correct option between the relevant public authority requiring its religiously-minded official to officiate at my ceremony in their civil capacity (under pain of being dismissed), and permitting them to decline on conscientious objection grounds. Assume that either option is equally consistent with both factual and legal possibilities. A judge asked to resolve this conflict of rights could legitimately defer to the course of action chosen by the public authority.
Alexy recognises these scenarios. He calls them instances of 'structural', or 'substantial', or 'substantive' discretion, arguing that they are not problematic: 'That the legislature [or, by extension, a public body] is free where no obligations exist is not in need of justification'. 50 The distinction drawn by Alexy between epistemic and structural discretion has become commonplace. 51 Yet, ultimately that distinction collapses, for there are parts of Alexy's argument that tend to reduce structural discretion to epistemic discretion. Specifically, Alexy argues that (the most important kind of) structural discretion arises because, in the process of balancing, our judgments cannot be sensitive enough to accurately gauge the relative importance of competing values and the seriousness of interferences with them. 52 Only a relatively crude assessment is possible, because of 'the limits of ordinary language', 53 which we use to express (and which, hence, shapes) these judgements. This seems to assume that, if only our categories of language/thought were refined enough, we would be able to grasp the real moral significance of the alternative outcomes and be able properly to adjudicate between them. It is almost as if Alexy is saying that structural discretion is something of a second best scenario, forced on us by the limits of our cognition of the worldwhile in an ideal world we would always be able to come up with one uniquely correct answer. Klatt reinforces this view, speaking of discretion as 'a universal and unavoidable problem' (emphasis added). 54 This is puzzling. After all, morality is best understood as immanent in our human practices, 55 and if we conscientiously come to the conclusion that the values at stake justify a range of equally acceptable options, why mistrust our practical reason? Our exercise of practical reason is of course always fallible; but why assume that there must be shortcomings (however unavoidable) only when we reach the conclusion that more than one option is acceptable under the circumstances, and not when we reach a clear-cut conclusion? 56 Why view situations of multiple morally permissible alternative options as somehow out of the ordinary in law, when they frequently occur in everyday life?
I think these difficulties in thinking well about structural discretion stem from the very starting point of the theory that rights are optimisation requirements, bearing out my point that a commitment to optimisation distorts a proper understanding of proportionality and balancing. According to Klatt, Alexy recognises that in cases of structural discretion 'the law neither commands nor prohibits following either of the two conflicting principles'. 57 This means that the lawthat is, the balancing exercise that determines the legal possibilitiespermits that either be followed. Yet, logically speaking, permission cannot be the result of two colliding optimisation requirements.
Alexy intuits this where, in discussing structural discretion, he speaks of a 'stalemate' in balancing competing principles, 58 which, according to him, means that the principles have the same 'concrete weight '. 59 No wonder that, under this view, structural discretion becomes a problem and a failure of practical reason: for giving effect to one principle (maximising its object given relevant possibilities) means frustrating the other without proper justification (that is, without a weightier counter-consideration than the defeated principle). It is like randomly awarding a prize to a contestant when the match was a draw. This is of course very unsatisfactory. Unsurprisingly, having conceptualised structural discretion as a stalemate, Alexy and his followers feel that something has gone wrong, that if only it were possible to refine the scales, we could truly decide which principle should prevail.
But the real problem is with the theory that tries to account for our use of practical reason in these instances, rather than with the use of practical reason itself. For the impression that there is stalemate is just that, an impression, and an artefact of the way in which the optimisation theory wrongly conceives of rights. In reality, in cases of structural discretion there is no stalemate; nothing regrettable has happened at all. It is simply a case where the purported duties grounded in the supposedly conflicting rights do not exist (that is, are not justified by the relevant rights), so that there are no clashing requirements in the first place. In the example I used above, although I have a right to family life, and the official a right to religious freedom, neither requires the realisation of their object to the greatest extent possible given the legal and factual possibilities. In particular, my right to family life does not ground a duty in the public authority to guarantee that a specific celebrant officiate at my ceremony; and the official's right to religious freedom does not ground the authority's duty to refrain from disciplining them through dismissal if they refuse officiating. Simply, neither right is important enough to ground those duties.
Nor do the conflicting rights have to have the same 'concrete weight' in order for us to reach these conclusions. It is enough that they do not have sufficient weight to ground the relevant duties. It may be, for example, that the interest in religious freedom is never important enough to justify civil servants' declining to perform their official duties (whenever these are morally permissible), because, say, of considerations about the authority's efficiency in delivering services, combined with considerations about not fostering divisiveness in the workplace. Employers are, thus, under no duty not to dismiss those who do decline; though neither are they obliged to dismiss them. On the other hand, the right to family life, because of organisational reasons, may not be important enough to ground a right to access relationship recognition via the offices of a particular public official (as long as another is at the ready), whatever the reasons for their failure to officiate (be it religiously-inspired homophobia, or the need for a toilet break). In such a situation, public authorities are under no duty to dismiss the nonperforming employee, though they may choose to do so, in order to pursue legitimate interests (avoiding divisiveness; fostering efficiency; supporting a non-homophobic culture). But here the interests in freedom of religion and in family life need not be of equal concrete weight, and there need not be a stalemate between them. They just need to be, in the circumstances, not important enough to ground the relevant duties.
We can only properly understand cases of structural discretion if we take constitutional rights to extend just as far as they need to and no further, given relevant moral considerations, as well as legal and factual possibilities. Depending on the outcome (and we are assuming either outcome is equally defensible) either family life or religious freedom will have been realised more than was required by the right that protected them (though no more than was morally acceptable). Perhaps, in light of the legal and factual possibilities, we will even conclude that it was realised to the greatest extent possible. But if that is so, it was not the corresponding right that required that optimising outcome; it was the relevant interest that justified the outcome without requiring it.

Conclusions
Constitutional rights are not optimisation requirements because the moral rights that constitutional rights institutionalise are not optimisation requirements.
The reason why moral rights are not optimisation requirements is their connection to human well-being. Understanding human well-being means, among other things, appreciating the difference between duties grounded in justice and supererogatory actions (a distinction flattened by the optimisation view), and discriminating between that which contributes to well-being and that which does not (a task the optimisation view is not conducive to). Moral rights do not require their objects to be realised to the greatest extent possible given relevant possibilities, but for their objects to be realised no less and no more than relevant moral considerations require. This is tautological, but tautologies are true. An optimisation perspective on moral rights obscures this tautology by, among other things, conflating the morally possible with the morally required.
If we take the possibility of legitimate authority seriously, our understanding of constitutional rights must allow that they be capable of correctly channelling moral requirements into constitutional law. Unfortunately, depending on what the legal possibilities contingently happen to be in any one legal system, the optimisation view may instruct that system's decision-makers to realise the object of a constitutional right over and beyond what the corresponding moral right requires, over and beyond what supererogatory considerations recommend, and over and beyond what justice permits (as long as Radbruch-style unconscionability is not reached).
The fact of institutionalisation perhaps justifies some divergence between the structure of constitutional rights and the moral rights they institutionalise. But any such divergence should be capable of being properly accounted for, and it should not result in making nonsense of constitutional rights. I have argued that only when paternalistic coercive interference (or interference that can be assimilated to it) is at issue, may there be an advantage in treating constitutional rights like optimisation requirements. In these cases, overprotecting the objects of rights may be a way of warding off the risk of under-protecting them. But I am not sure the case is clear-cut even here. It seems that in these cases, as in all others, the optimisation view ultimately does not assist us in thinking well about rights. It is particularly inept at making sense, and guide decision-making in the context, of conflicts of rights, including conflicts which legally and morally admit of more than one correct solution.
A better view than the optimisation view holds that constitutional rights demand that their objects be realisedwithin the factual and legal possibilitiesno more and no less than morally required by the interests in which rights are grounded. This is the understanding on which any sensible method of resolution of conflicts involving rights must be premised. Indeed, I think this (and not a conceptualisation of rights as optimisation principles) is the understanding of rights tacitly informing the proportionality limb in Alexy's own theory.
Alexy says that there is a qualitative difference between principles and rules because, unlike the former, which demand optimisation, 'if a rule is valid, it requires that one do exactly what it demands, nothing more and nothing less'. 60 But this is as true of constitutional rights (understood as principles) as it is of rules. 61 Rights that take the form of statements of principle may require more work than rules generally do to work out what they demand; but, like rules, they demand compliance with whatever they demandwhether or not what they demand is antecedently clearnot with some purported imperative to realise their object to the greatest extent possible.