1. Introduction
Jonathan Quong’s Liberalism Without Perfection (Reference Quong2011) revamped interest in political liberalism and the theory of liberal neutrality, even though some philosophers have considered it long gone (see Arneson Reference Arneson, Wall and Klosko2003, 191–219). Quong forcefully showed that the debate between comprehensive and political conceptions of liberal theory, as well as between perfectionist and anti-perfectionist propositions for duties of liberal institutions, is far from over.
The book makes many important arguments about contemporary liberalism. It is particularly critical of the comprehensive variant of the doctrine. Two arguments against it, from the first part of the book, strike me as particularly interesting and worthy of debate.
In the first argument, Quong suggests that the position called the ‘comprehensive anti-perfectionism’ is not tenable. That is, a liberal theory anchored in a conception of the good, such as personal autonomy, cannot coherently propose anti-perfectionist political actions. In the second argument, he claims that liberal perfectionism cannot avoid being paternalistic because it necessarily implies negative judgments about individual capabilities to decide what is best for them.
In this article, I challenge these two arguments and move the debate forward. I think that Quong’s objections to comprehensive liberalism are built on a narrow understanding of personal autonomy. I examine this in detail and show that a different (and more plausible) way to conceptualize personal autonomy undermines his objections in both arguments. However, I do not follow some of the existing critiques of Quong on similar grounds (see Colburn Reference Colburn2012, 17–29). My argument attacks his claims from an alternative direction, and ultimately yields a different normative proposition.
The main aim of the article, however, is not (only) to defend comprehensive liberalism from Quong’s charges. I aim to show that this discussion reveals something rather unassuming, yet plausible nonetheless: that liberal perfectionism and neutrality are not mutually exclusive, but convergent liberal positions. Moreover, I suggest that conceptualizing liberal perfectionism and neutrality as complementary rather than exclusive contributes to better understanding, and hopefully resolving, some of the issues discussed by Quong. This is a novel position that I try to develop and defend in the third section of the paper, which differs substantially from some of the similar arguments in the recent literature (see Merrill and Weinstock Reference Merrill and Weinstock2014). Its main thrust is borne by the idea that a valuation of personal autonomy within liberalism can coherently accommodate both perfectionist and neutralist normative proposals.
2. Is comprehensive anti-perfectionism coherent?
2.1. Quong’s claim
In order to distinguish between different varieties of liberalism, Quong asks two questions. First, must liberal philosophy be based on a particular idea or a conception of the good life? Second, should a liberal state promote certain ideas or conceptions of the good life on the basis of belief in their inherent value or other metaphysical claims (Quong Reference Quong2011, 15)?
Different conceptions of liberalism part ways in relation to how they respond to these two questions. For example, if we answer affirmatively to the first, then ours is a comprehensive conception of liberalism. On this view, as Quong says, ‘there is a particular liberal conception of what constitutes a flourishing human life’ (Reference Quong2011, 16). If we answer negatively, then it is a political conception. Within this conception, liberalism represents a ‘justified political response’ to the fact that people disagree about what constitutes a good life.
Liberals who respond affirmatively to the second question endorse a perfectionist conception of liberalism, while those who respond negatively sustain an anti-perfectionist or neutralist conception.Footnote 1 So, perfectionists believe states should, while neutralists hold they should not, promote certain conceptions of the good or particular ways of life.
However, answering one way to the first question does not commit us to a particular answer to the second one. One could plausibly believe that liberal philosophy is based on a certain idea of the good life, yet still argue that institutions should remain neutral between competing conceptions of the good life. Similarly, one could believe that there is no single conception of the good that underpins liberalism, yet still hold that institutions should promote certain ways of life and not others. Quong dubs the first of these combinations ‘comprehensive anti-perfectionism, ’ and the second ‘political perfectionism.’ Following the logic of this argument, one can also find ‘comprehensive perfectionist’ and ‘political anti-perfectionist’ conceptions of liberalism.
I don’t wish to question Quong’s distinction between different varieties of liberalism, though some doubts could be raised.Footnote 2 What I do wish is to assess one of his claims based on this distinction, which is that the ‘comprehensive anti-perfectionist’ conception is untenable. I will use this assessment to develop a further argument about comprehensive conceptions of liberalism in the third section of the paper.
Quong offers two arguments for this claim. First, he invites us to imagine two friends, Mike and Sara, having an argument about the value of recreational drug use. Mike is a comprehensive perfectionist, believing that doing drugs is an unworthy activity, and that this activity should be criminalized. Sara, on the other hand, believes that there is nothing unworthy about using drugs, provided that it is a result of an autonomous choice. She, unlike Mike, does not support drug criminalization because she thinks institutions should not prevent people from making their own decisions about valuable activities.
Quong thinks that comprehensive anti-perfectionism, as expressed by Sara, runs into great difficulty because if Mike were to push Sara to answer why leading autonomous life trumps other considerations, she would have to answer that being autonomous is simply more important than other things. Mike could point out that this view depends on a problematic view of human flourishing and makes Sara no less perfectionist than Mike. Quong suggests that, in this case, Sara is ‘trying to use coercive power of the state to impose a particular perfectionist judgment, ’ just as Mike does (Reference Quong2011, 24). Her appeal to autonomy, Quong believes, does not yield an anti-perfectionist liberal state, but just the opposite.
Second, Quong goes further and examines what would happen if Mike adopted Sara’s point of view. Mike could plausibly come to believe that being autonomous is important, so it should not be jeopardized for other considerations about the good life. But, as Quong suggests, Mike would not necessarily need to reject the view that there are other equally important considerations about the good life. His perfectionist state could easily promote these without infringing on individual autonomy because there are other (soft) forms of government intervention into individual choices of the good that do not involve coercion, such as taxes, subsidies, advertising, and other forms of encouraging citizens to live in accordance to a comprehensive view of the good.
But even this argument would run into problems because Mike would have to claim that either the soft forms of government intervention do not infringe individual ability to lead autonomous lives, or that personal autonomy implies that it could be sometimes curtailed by the government for other perfectionist reasons. Obviously, holding either of the two views is not plausible for an autonomy-minded anti-perfectionist. If autonomy implies choosing one’s own conception of the good, then any form of external interference (no matter how soft it is) by which the state pushes individuals into selecting certain options will violate autonomy. The implication of this, says Quong, is that there seems to be no way for a comprehensively conceptualized liberalism to be anti-perfectionist. ‘Once liberalism is tied to some specific views of the good life, the liberal state will unavoidably be acting for perfectionist reasons’ (25).
The problem with comprehensive anti-perfectionist conception of liberalism, according to Quong, is the fact that it is anchored in the value of autonomy. This is because if anti-perfectionism is construed exclusively in terms of autonomy, it will have to place a non-conditional value on the autonomous status of persons. Thus,
[s]o long as autonomy is understood in this way, it will be difficult to justify one of the central features of liberal perfectionism: subsidies, incentives, and other means of manipulating citizens into making good choices. On the other hand, if the manipulation of others is consistent with the value of autonomy, it is not clear that a theory based on this conception of autonomy will be recognizably liberal. (Reference Quong2011, 71)
Clearly, Quong wants us to believe that one cannot consistently believe that autonomy is more important than other value considerations, yet refrain from imposing it through perfectionist measures. We are thus forced to choose between:
(1) Being perfectionist about autonomy and promote it through ‘soft’ institutional measures, and
(2) Being anti-perfectionist about autonomy and refrain from giving it a special status through institutional action.
This problem poses a dilemma for comprehensive anti-perfectionists, as Quong believes, because if they choose (1), they run into danger of betraying the meaning of personal autonomy. If autonomy implies the principle of being the author of one’s life, then even soft measures that manipulate individuals into choosing options they would not usually choose violate the principle. If they choose (2), they cannot consistently claim that their view is anchored in autonomy as the comprehensive conception of the good.
2.2. Critique
But, is personal autonomy a concept that is vulnerable to this kind of charge? Of course, it depends on how one understands it, but one particular understanding does not necessarily exhaust the interpretative possibilities the concept itself can offer. Quong follows the interpretation from Joseph Raz, according to which the value of autonomy consists of individual ability to choose worthy ways of life among various alternatives. For Raz, autonomy is valuable only if a person chooses a valuable option, and not if she chooses an unworthy one. So, a person’s autonomy is valuable if she chooses to go to the opera, and not to a mud wrestling competition. Thus, for Raz, autonomy is a substantive principle: only certain choices qualify as valuably autonomous (see Raz Reference Raz1986, 372).
Quong doesn’t leave much space for alternative ways to conceptualize personal autonomy. In his view, the only possible alternative to Raz’s substantivism is the formal view, in which autonomy gains independent value. But, as Quong believes, opting for this understanding changes the way autonomy is valued and used by the liberal and undermines the comprehensive anti-perfectionism nonetheless. Here’s the full range of autonomy-based options for liberalism, in Quong’s view:
If the value of personal autonomy is made to depend on using our autonomy to choose valuable options, the appeal to personal autonomy justifies a perfectionist state, but not a recognizably liberal state. If on the other hand, personal autonomy has value independent of ends that are autonomously chosen, this may justify a liberal state committed to something approximating Mill’s harm principle, but it will preclude most perfectionist state action. The value of personal autonomy, I conclude, is not a stable base on which to construct a liberal perfectionist theory. (135–6)
The coherency of the comprehensive anti-perfectionist position does not stand or fall with Raz, of course (for a comparable critique see Kulenovic Reference Kulenovic2014, 35–46). But it is not difficult to see why the substantivist view of autonomy falls prey to the coherency objection. One cannot be an anti-perfectionist if one believes that only a certain set of autonomous choices qualify as valuable, while others do not. The inherent logic of the view will demand perfectionist, rather than anti-perfectionist measures.
Quong’s mistake in following Raz’s substantive conception of autonomy (in order to show that the comprehensive anti-perfectionism is incoherent) is that he uses the view of autonomy that is strongly perfectionist by default. If autonomy is valued only when particular options are chosen, then the state has no other way to promote autonomy but through promotion of these appropriate choices. This makes the claim about the incoherency of comprehensive anti-perfectionism question begging because it tacitly assumes the claim it wishes to defend. The conception of autonomy employed to prove that comprehensive liberalism cannot be anti-perfectionist is strongly perfectionist itself.
In the alternative perspective, the coherency objection loses its grip. Almost any other non-substantive conception of autonomy – such as, for example, Harry Frankfurt’s or Gerald Dworkin’s procedural theories – would undermine Quong’s objection because none of them imply substantivism of Raz’s type. For Frankfurt and Dworkin, the substance of autonomous choices is normatively irrelevant as long as the person satisfies a certain set of procedural requirements, such as a second-order identification with her first-order desires (Dworkin Reference Dworkin1988, 3–21; Frankfurt Reference Frankfurt1988, 11–26). From this perspective, perfectionism is not necessarily entailed by the comprehensive commitment to autonomy. Liberal institutions only need to make sure that individuals identify with their choices, and do not endorse or impose any particular option.
However, the real challenge in addressing Quong’s objections here is not to take the easiest way out by adopting a conception of autonomy that will cause problems for his argument. Rather, the challenge is to adopt the most plausible conception of autonomy and then reflect back on the prospects of Quong’s theory through this lens.
Both procedural and strong substantive conceptions of autonomy have been vigorously criticized in the literature, and none seems fully acceptable. Critics objected to procedural approaches for being internally incoherent (Noggle Reference Noggle and Taylor2005, 93; Watson Reference Watson1982, 96–111), failing to account for autonomy-diminishing influences, and failing to distinguish authentic from non-authentic choices (Abrams Reference Abrams1999, 805–846; Beauchamp Reference Beauchamp and Taylor2005, 318). Substantive theories, however, face another challenge. By requiring a specific substance of individual commitments, substantive theories place overly demanding conditions on what counts as an autonomous choice, rendering them paternalistic (Kristinsson Reference Kristinsson2000, 257–286; Westlund Reference Westlund2009, 26–49).
A better way to conceptualize autonomy is to go beyond procedural and substantive limitations while being sensitive to both: the formal sovereignty of the individual’s choice, and the substantive character of the good he selects. This means that we have to view it as a complex principle, combining both formal and substantive properties. It doesn’t mean, however, that autonomy is not a conception of the good, or a way of life. On the contrary, we can still consider it a particular way to conceptualize the good life, but one that is structurally different from other conceptions of the good.
2.1.1. Autonomy, authenticity, and social position
Unlike other conceptions of the good, personal autonomy is an internally complex and somewhat equivocal conception. Although it does tell us what kind of a life for a person is valuable, it doesn’t tell us yet the content of life a person will choose to lead; we do not know the substance of her choices in advance. It can equally be a life of an opera fan, a chocolate connoisseur, and a life of a professional mud wrestler or a vagabond. We should be careful not to conflate the possible objective worth of a conception of the good life with its being the object of autonomous choice. Mud wrestling may be less valuable than opera, but the judgment of its inherent worth should not simultaneously be a judgment on the autonomous character of the personal choices for either of them.
What kind of conception of the good is autonomy? We have two possible answers to this question. First, we could follow Ben Colburn in adopting a distinction between two different types of values: first- and second-order ones. According to Colburn, the first-order values are content specific because they fully specify the states of affairs that are valuable. For example, a statement that what is valuable in life is being able to play Bach’s Cello Suites flawlessly is content specific. Contrary to that, a statement that what is valuable in life is satisfaction of desire is content neutral because it does not specify the kind of desire one considers valuable. For Colburn, the distinction between first- and second-order values enables us to see what is different about autonomy as a value: if autonomy is defined in terms of persons deciding for themselves what is valuable in life, then such definition refers to secondary judgments of what is valuable. This is what makes autonomy a second-order statement about value, ‘since its specification must be able to contain a second order variable’ (Colburn Reference Colburn2010, 253). Therefore, Colburn suggests, comprehensive anti-perfectionism is coherent: the promotion of autonomy means promotion of its second-order, not the first-order properties.
However, Colburn’s distinction faces a number of problems. Thomas Porter, for example, argued that the distinction doesn’t hold because a plausible specification of autonomy will also include:
not only the second-order variable that Colburn highlights, but also straightforward specifications of states of affairs that are valuable in the manner of a first-order value. (Porter Reference Porter2011, 7)
In other words, defining autonomy will include a range of content-specific values and facts, from social conditions, individual positions, to choices at individual’s disposal. Porter’s objection to Colburn implies that autonomy is a ‘second-order value with first-order characteristics, ’ which thus runs counter to Colburn’s claims that anti-perfectionism only prohibits promotion of its first- but not second-order properties. It shows that anti-perfectionism cannot be based in the value of autonomy because the normative assumptions about it are inadequate. This argument about the internal structure of autonomy does not vindicate the anti-perfectionist claim, and thus, has no case to make against Quong (see more criticism of Colburn in Nye Reference Nye2012, 81–99).
The second option is to take a positional understanding of autonomy that emphasizes relational authenticity of individual commitments as the constitutive element of autonomy. The positional approach suggests that personal autonomy is a conception of the good that refers to individual identity constituted by the person’s regulative self-governance, her social relations, and the position in society, rather than to a particular choice substance.
According to this view, an autonomous person exercises de facto control over her commitments; they represent her own choices. This means two things. First, it implies that the person is competent (rational, mentally capable, reflective) to make her own decisions. This view has been a prevalent in procedural understandings of autonomy. However, it does not necessarily give full picture of the phenomenon. A person may be competent, and make her decisions in a procedurally satisfactory way, yet her decisions may still derive from social structures that reflect commitments other than her own. Secondly (and more importantly), it implies that the person’s identity and her commitments are authentic. They are not only endorsed by the person, but also satisfy other normative requirements of authenticity. Such requirements could include coherence, whereby person’s commitments must fit well with the rest of the facts about her; depth, whereby the person must have ‘access to his or her own psychic dynamisms and to reflect such awareness in the construction of his or her identity’ (Ferrara Reference Ferrara1998, 96); equality, whereby persons’ identity must not assume its unequal worth against other persons; or open future, whereby individuals must have freedom and access to means of change of their commitments and identities. We could say that if these conditions are satisfied, persons are autonomous regardless of what particular shape their commitments take (for a comparable view see Stoljar Reference Stoljar, Mackenzie and Stoljar2000, 94–112).
The example of ‘Mrs H’ that Catriona Mackenzie lays out in her paper describes this conception of authenticity well. Namely, Mrs H is a patient who suffers from aggressive bone cancer and has just had her leg amputated. Although the doctors are fairly positive about her survival prospects, she informs the medical team that she desires no further treatment if the cancer spreads; that is, she wants to die. In addition to the illness, her husband has recently left her because he considered her illness too burdensome and embarrassing (Mackenzie Reference Mackenzie2008, 518). Mrs H’s conception of herself is formed by traditional views of femininity that are authoritative in her culture, where marriage confers meaning to a person’s life, which makes Mrs H’s situation particularly hard on her. Now that her husband has left her, she feels her life has no meaning anymore.
According to Mackenzie, Mrs H cannot be considered autonomous because she internalized oppressive and non-egalitarian social norms that determine her notion of the self-worth. Although she is a competent and rational person,
[t]hese seem to be precisely the kind of factors that compromise autonomous agency and hence an agent’s normative authority over her decisions, her actions, and her will. The problem then is that just because an agent acts in accordance with her settled character or reflectively endorses the value commitments that define her practical identity, this is no guarantee that she does so autonomously. (513)
These factors constitute objective and content-specific facts about personal autonomy that do not depend on the person’s subjective views and dispositions (see Baumann Reference Baumann2008, 422). The content of Mrs H’s conception of herself matters for her autonomy, but not in a Razian fully substantive view. The positional conception of autonomy doesn’t require that Mrs H commit to a certain range of choices, but only that the way she does so reflects certain normative requirements, such as equality, depth, and freedom to change. The content of her commitments matters relationally, as a result of a particular interaction she has (or had) with her environment, in terms of particular doctrines and traditions in her society that may conform or reject these normative requirements.
However, this view is not purely formal either, in terms of being content neutral, because it has to enumerate the content-specific facts about the person, such as the specific forms of her identity (whether it assumes basic equality, how malleable, coherent, and deep is it), social position, and her relation to others in a given social context.
We could think of this as a middle-ground view of autonomy; it is neither strongly substantive in terms of requiring the particular content of commitments, nor fully content neutral in terms of refraining to specify certain facts about the person. It still, though, represents a comprehensive valuation of autonomy,
in the sense that it claims that a just society has an obligation to promote autonomy by ensuring that its basic social, legal, political, and economic institutions provide the recognitive basis for its citizens to realize their autonomy. (Mackenzie Reference Mackenzie2008, 524)
If Quong adopted this conception, his dilemma about the coherency of comprehensive anti-perfectionism would be seriously undermined. Namely, Razian perfectionism simply doesn’t follow from a comprehensive commitment to autonomy conceptualized like this. To see this more clearly, go back to Quong’s imagined debate between Mike and Sara. Recall Quongs’ suggestion that Sara’s position is problematic because it falls prey to Mike’s objection that her commitment to autonomy, based on her belief in its asymmetric importance in comparison to other goods, is as equally perfectionist as Mike’s. If Sara’s conception of autonomy were more akin to Mackenzie’s rather than Raz’s understanding, then this would not necessarily commit her to institutional perfectionism similar to that of Mike. His proposition was, if you recall, that drugs should be legally prohibited because consuming them represents an unworthy activity. His institutional perfectionism is substantial: it identifies a particular type of activity that should be legally prohibited. If Sara continued to uphold a comprehensive valuation of autonomy, but in the alternative way, her institutional measures would not propose prohibiting or promoting any kind of substantive activity. All individual activities, as long as they are autonomous in accordance to the outlined autonomy standard, and are not in violation of some other normative principle, would be permissible. As long as the person’s commitments are fully reflective of his personality, derive from non-oppressive or non-egalitarian norms, and are open to change or abandonment, he should be permitted to hold any kind of substantive conception of the good and engage in any kind of (otherwise legal and morally permissible) activity. Clearly, this is not perfectionism of Mike’s type, and Quong is wrong to equate the outcome of Sara’s with the outcome of Mike’s position.
It is a different matter to discuss what kind of institutional promotion or personal autonomy Sara’s position would entail, and how is this different from the content-neutral conception of autonomy. I will have to say more about it later in this article. Here, I only wanted to show that Quong’s argument that a comprehensive commitment to personal autonomy cannot be anti-perfectionist is limited only to those views that define autonomy substantively, and does not have a larger purchase. While he is probably right that Raz’s conception of autonomy cannot yield institutional anti-perfectionism, other conceptions can. Next, I will show that his argument about the necessarily paternalist outcomes of liberal perfectionism shares a similar fate.
3. Is liberal perfectionism necessarily paternalistic?
3.1. Quong’s claim
The second claim I will examine in this article is Quong’s argument that liberal perfectionism cannot avoid the charge of paternalism. Quong argues that liberal perfectionism, in order to distinguish itself from Rawlsian type of liberalism, must claim that perfectionist policies are necessary even if, in accordance to Rawls’ conception, everybody has been given their fair share of rights, liberties, opportunities, income, and wealth (Quong Reference Quong2011, 85). If this is so, then Quong has a question for liberal perfectionists:
Although this claim is an essential part of all contemporary theories of perfectionism, it is seldom explained. Why should state action be required even if resources have been fairly distributed to individuals? What reasons might a liberal perfectionist offer for this somewhat puzzling view? (85)
He exercises several possible answers and claims that all are guilty of paternalism, or that all require the state to make negative judgments about individual abilities to advance their interests.
First, the rationality argument, which supposes that there will always be a sufficient number of people failing to advance their interest, implies that perfectionist measures need to be undertaken to help them do so. This implies a negative judgment of institutions towards certain groups of people, and therefore is paternalistic. Second, weakness of the will argument, according to which perfectionist measures are needed to help people overcome their weak will, also implies negative judgment and is guilty of paternalism. Third, the free-riding argument implies that since not enough people want to consume certain goods (such as public parks, libraries, and similar), perfectionist measures are needed to cover their maintenance costs. Quong thinks this is also paternalistic because it assumes that left on their own, people will not make the best decisions on how to spend their resources, which implies a negative judgment about their abilities. Similar judgment underpins the prohibitive costs argument, which suggests that some things will be overly expensive for individuals without subsidies, making perfectionist measures necessary. This is also paternalistic because it implies that the high cost of certain goods is caused by the lack of individual preferences for them, which drives the price up but also reveals a poor judgment of those who are not ready to pay for these worthy goods. Finally, the experience argument suggests that perfectionist measures are needed so individuals could experience certain goods for themselves, and then willingly decide to pursue them, provided that they accept their inherent worth. For Quong, even this is paternalistic, since it involves a judgmental re-ordering of incentives for individuals to pursue certain goods (86–96).
Ultimately, Quong believes to have shown that perfectionism is incompatible with the liberal idea of persons as free and equal beings. He suggests liberals should be anti-perfectionist instead because
[t]he aim of anti-perfectionist liberal theories is to fairly allocate resources or advantages amongst citizens, and then allow citizens to use their fair share of resources or advantages as they see fit. (106)
Provided that everyone gets their fair share or resources, persons should be free to decide what to do with them, that is how to spend them and what options to choose. There seems to be no need for any non-justice-based perfectionist activity of the state. Making individuals choose particular options against the background of equal distribution is paternalistic and inherently wrong because doing so denies them equal status they deserve as human beings. Even soft measures, such as manipulation of personal choices or re-ordering of the ‘choice architecture’ (Sunstein and Thaler Reference Sunstein and Thaler2008), violate this status because they imply negative judgments about individual abilities to make best decisions for themselves (for a similar argument see Coons and Weber Reference Coons and Weber2013, 23).
3.2. Critique
In order to see what is wrong with this reasoning, we need first to understand the internal structure of Quong’s claim. First, he distinguishes between perfectionist justice and non-justice-based perfectionism. Perfectionist justice is when distributive principles of justice derive from a substantive definition of the good life, for example, a religious doctrine. Non-justice-based perfectionism is when a state promotes ends independent of the established principles of justice, say when a state promotes religious principles without having them as part of the consensually accepted principles of justice (Quong Reference Quong2011, 85).
This means that liberal perfectionism can be either about the basic principles of justice or about substantive goods independent from it. That is, perfectionism either applies when we define the principles of justice, or when we promote certain goods long after the principles of justice have been established. The second case implies that the ends promoted have not been the object of justice, or that justice itself does not promote these ends so the state institutions must do it independently of its principles.
But this is a rather narrow view of the conceptual possibilities for liberal perfectionism. If we follow the initial distinction between perfectionist justice and non-justice-based perfectionism, we can also establish the corresponding elements. If there is a perfectionist, then there is obviously a non-perfectionist justice. Justice is non-perfectionist when its distributive principles derive from public reason, or from the mutual agreement of the participants in the original position. Furthermore, if we allow for non-justice based-perfectionism, we have to allow for justice-based perfectionism, based either on a perfectionist or non-perfectionist conception of justice. This entails state perfectionism aimed to support application of the principles of justice that would follow through after the adoption of the principles of justice in the original position and ensure that the principles are sustained by the ongoing distributions.
Quong thinks that there cannot be such a thing as a justice-based perfectionism derived from a non-perfectionist conception of justice. For him, institutional perfectionism is either based on perfectionist conception of justice, or on conceptions of the good independent of the principles of justice.
I think this belief is unwarranted because two problematic assumptions are built into it, both of which exclude the distinct possibility of liberal (unlike other forms of) perfectionism. The first assumption is that autonomy is just another conception of the good, no different than others, so if participants in the original position haven’t built autonomy into the principles of justice, any institutional promotion of it will be a case of non-justice-based perfectionism. But, as we already established, autonomy is a different kind of value. It is assumed that participants of the original position are autonomous (able to sustain conceptions of the good and principles of justice). They do not have to agree that autonomy is a worthy conception of the good and a way of life, yet they do have to be autonomous in order for their principles of justice to be valid. They have to stand in autonomous positions and thus formally and indirectly endorse autonomy as a comprehensive conception of the good. So, even a non-perfectionist conception of justice will have to be comprehensively committed to personal autonomy. Sara’s position, from Quong’s example, represents this view. Hers may be a non-perfectionist conception of justice, shareable by other liberals, but it does imply a comprehensive valuation of autonomy, either directly (as she exemplifies in Quong’s case), or indirectly (as it would be the case for somebody who does not substantively value autonomy but thinks institutions should not impose ways of life on individuals out of respect for autonomy, say an Amish person).
So, unless we assume that something like the Rawlsian initial distribution has already taken place sometime in the past, we have reasons to promote autonomy continuously so particular conceptions of the good and the principles of justice, defined by non-perfectionist means, could be appropriately sustained by social agents. If the agents are not autonomous when choosing their conceptions of the good and the principles of justice, then the entire system lacks legitimacy. Thus, we have reasons to want that individuals are autonomous when they do so.
Since we cannot assume that a just distribution has already taken place, it makes sense to conceptualize autonomy-based liberal perfectionism as a redistributive principle, guiding governmental provision of appropriate amounts of opportunities and resources for individuals to acquire autonomy so they can choose their own substantive commitments by themselves and sustain even non-perfectionist principles of justice.Footnote 3 This kind of liberal perfectionism would be diachronic, and it would work to ensure that an equal (or whatever normatively appropriate) distribution of resources, goods, and opportunities through time obtains. This qualifies as perfectionism, since state institutions are involved in promoting autonomy as a conception of the good, not as the content of personal commitments, but as a form within which other valuable ways of life are conceptualized.
The second assumption is that social positions persons take on after the imagined initial distribution has taken place do not matter for their autonomy, and hence give no perfectionist reason for state intervention. However, consistent with the positional understanding of autonomy elaborated earlier, we could say that liberal perfectionist measures are needed to ensure that individuals are in appropriate social positions to exert regulative control over the authenticity of their commitments even after they have, presumably, been given equal share in the initial non-perfectionist distribution. This is because there are no guarantees that the social roles persons would choose after receiving their fair share of resources will be amenable to their exercise of regulative control over their commitments. This does not have to imply paternalism, yet it does imply perfectionist government action, given the character of autonomous social position as a distinct conceptualization of the good. In addition, it implies a more realistic understanding of the imperfect human rationality and the possibility of error. By suggesting that non-justice-based perfectionism is superfluous, Quong implies that individuals are perfectly rational, as well as that the validity of options chosen at one particular time must stretch indefinitely. Take the case of voluntary slavery, for example. Quong’s claims would commit him to saying that an individual’s choice to submit to the absolute will of another, made after the initial distribution of resources has been completed and the individual given appropriate amounts of whatever currency we choose, does not warrant any kind of state perfectionism to reposition the individual. That is obviously wrong.
Consider some other familiar examples. Say, Joe may value mud wrestling only because there are no other alternative conceptions of the good he can choose, or he is unaware of them, or he committed all of his resources to a life in which mud wrestling is the only cultural activity available. If he has never seen or experienced opera, or the library, because he has no means to experience them, then he is not appropriately social positioned to choose the option of mud wrestling in full authentic fashion. Subsidizing opera is thus not necessarily a paternalist judgment on Joe’s views; it is simply an intervention that repositions Joe so he could exercise more control over his comprehensive commitments. He may still continue valuing mud wrestling more than opera after the repositioning, but that would not be a matter of perfectionist concern any more. Similar logic can be applied to Mackenzie’s example of Mrs H. If she has no access to other, non-traditional conceptions of women, marriage, and her own position in society, her commitment cannot be considered authentic.
Therefore, subsidies to public parks, libraries, opera concerts, or concepts of femininity do not necessarily need to be justified exclusively by a reference to their inherent worth, as some perfectionists would have it, and as Quong assumes it to be. They could be justified by a reference to the contribution of the activities of going to the park, reading, and attending an opera concert to the social standing of individuals that enables them to regulate their lives and commitments. This would be a reference to the comprehensive good of standing in an autonomous social position and having an authentic identity.
So, negative judgments liberal perfectionists might have towards certain individuals in a liberal society are not necessarily judgments about their abilities to make autonomous decisions, even if it may appear to be so. They could be judgments towards persons’ particular social positions that prevent them from assuming control of their lives to qualify as autonomous. Quong’s charge that liberal perfectionism is paternalistic because it violates the ideal of equality of human beings misses the target. Liberal perfectionism does not necessarily degrade the equal status of autonomous beings. It does, however, hold a more realistic understanding of the ways individuals are socially positioned to own their actions and exert control over their lives.
4. Neutrality with perfection: the convergence thesis
In previous sections, I tried to show that Quong’s attack on comprehensive anti-perfectionism and liberal perfectionism does not significantly undermine the plausibility and coherency of these theoretical positions. By doing so, I argued for two distinct claims. First, that the comprehensive commitment to autonomy can yield anti-perfectionist government actions. Second, that the comprehensive commitment to autonomy can be perfectionist without being paternalist.
But, do these two claims cancel each other? How can a comprehensive valuation of autonomy coherently be perfectionist and anti-perfectionist?
To address what seems like another dilemma, I want to build on the previous discussion to outline and defend the main proposition of this paper: that liberal perfectionism and anti-perfectionism (or neutrality) are not mutually exclusive, but convergent philosophical doctrines, and that both have a role to play in contemporary comprehensive liberalism.Footnote 4 I believe the discussion of Quong has given us at least some reasons to consider the validity of this proposition. I also wish to suggest that understanding perfectionism and neutrality as non-exclusive helps us address problems with Quong in a useful way.
Recall again the discussion between Mike and Sara. As I suggested, Sara’s comprehensive valuation of personal autonomy does not need to commit her to institutional perfectionism similar to Mike’s. While Mike holds that the unworthy status of doing drugs warrants its institutional prohibition, Sara disagrees because she believes that institutions should not violate personal autonomy. But, given the conception of personal autonomy adopted earlier, Sara’s position does not imply institutional promotion of any particular substance of individual commitments, only of autonomy of persons. So, it does not imply the type of perfectionism Quong assumes. But what does it imply?
The coherency of the position Quong dubs comprehensive anti-perfectionism hinges on our ability to show that Sara’s commitment to autonomy holds liberal institutions responsible for some kind of intervention in individual lives. While this doesn’t imply a substantive perfectionism similar to Raz, it assumes that institutions must ensure that individuals are socially positioned in such a way to sustain authentic commitments. But is that perfectionism?
The problem of understanding the institutional outcome of the comprehensive commitment to autonomy may not necessarily be due to the intricacies of this argument. It could also be due to the problematic nature of the philosophical categories of perfectionism and anti-perfectionism. Namely, these categories, understood in terms of institutional support of, or the exercise of neutrality towards particular conceptions of the good, are taken as mutually exclusive. Either a state supports a particular conception of the good, or it is neutral towards it. It cannot be simultaneously both because it would endanger its philosophical coherency. This is why we have a trouble of conceptualizing Sara’s position. If she values autonomy and suggests that institutions should do something about it, yet thinks they should refrain from imposing anything because doing so violates individual autonomy, it is not exactly clear what her normative proposition about the role of liberal institutions is. No wonder Quong questions the coherency of her view.
However, consistent with the positional understanding of autonomy discussed earlier, I believe the conceptual relation between perfectionism and neutrality should be reconsidered. If a liberal committed to the protection of personal autonomy in society, like Sara, understands autonomy positionally, she will also understand that the institutional duty to protect personal autonomy is twofold. First, institutions have a duty to respect individual autonomy by not imposing coercively (or non-coercively, through manipulation) particular conceptions of the good upon individuals. Doing so not only violates their autonomy; it also sends them a message that they are not considered normatively authoritative and respect-worthy beings. Quong is right that this kind of behavior constitutes paternalism and that it is wrong.
But, she will also understand that institutions must support and promote the autonomy of persons by engaging in a certain kind of promotional activities. These activities, in order to remain liberal, would not be directed at the content of individual commitments, but at the way the persons’ commitments are socially constituted. If individuals are in positions that prevent them from adopting and sustaining commitments that will be truly their own, according to the normative standards of authenticity, the institutions must step in and reposition them (not necessarily coercively) so they can do that. This is a particular kind of intervention that does not qualify as perfectionist in the sense Quong uses it because it doesn’t target any substantive conception of the good, but it does represent a form of autonomy-based government interventionism and targets authenticity of individual commitments as a formal conception of the good life.
It may be questionable if one should call this a type of perfectionism at all. There are some reasons for that, since it involves government intervening into individual lives. But, that would be somewhat incorrect because what makes a government intervention perfectionist is not the fact that it intervenes, but that intervenes from a vantage point of considerations that certain ways of life are more superior to others. Still the use of the term would not be too far off, since this proposal has, admittedly, some perfectionist qualities. The reason for government intervention may not be fully substantive, in terms of a particular content of the conception of the good life, but it may be weakly substantive because it will not be content neutral: it will assume that certain social positions and identities qualify for autonomous status of individuals, while others do not. Basically, this claim qualifies as perfectionism of the authentic identity; it implies that some identities and social positions are superior to others in terms of their relation to personal autonomy, and should therefore be promoted by government institutions. For example, Mrs H’s identity that implies traditionalist values of marriage (which imply authoritative and meaning-conferring role of husbands) and renders her life meaningless after her husband has left her cannot count as an authentic identity. The government has a duty to help Mrs H adopt and sustain an identity that will not be reflective of non-egalitarian or oppressive social norms.
The problem one may have with this proposal is that it seems to imply that the government, committed to the comprehensive value of personal autonomy, must be both perfectionist and anti-perfectionist in relation to individual conceptions of the good. If it wants to preserve coherency towards autonomy as its justifying value, it needs to act in ways that will favor it. However, if it also wants to stay true to the meaning of autonomy – individuals having regulatory control over their commitment choices – it should not impose a particular conception on individuals because that will decrease their control over their lives.
Although this sounds like another dilemma, it is actually not. What determines the character of this proposition is the complex nature of personal autonomy. As already indicated, autonomy has a dual character. It is simultaneously formal and substantive. It is formal because it doesn’t specify the content of individual commitments. It implies that a person is capable of making certain choices about the ways of life he or she wants to lead. She may even change her conception of the good throughout her life, so no specific content of her commitments will remain intact. Persons can start their lives as devout believers and then change into militant atheists. They can be committed to mud wrestling and then change to opera. Most individual lives are fluid and rich in experience, so it makes no sense to assume that one particular conception of the good will determine the character of a person throughout her life.
Autonomy is also (weakly) substantive because every autonomous person will sustain a particular content of her conception of the good. Everybody is committed to some substantive views about the good life. This means that even those who reject autonomy as a value can qualify, ceteris paribus, as formally autonomous. They could be formally autonomous, while the substance of their commitments would reject the way of life associated with autonomy.
This means that personal autonomy will require a complex set of liberal institutional actions towards individuals and their conceptions of the good. Being exclusively perfectionist or anti-perfectionist will not be sufficient because a liberal state will have to do both simultaneously: it will have to be perfectionist and promote the formal value of autonomy by enabling individuals to hold social positions that will make them able to exert control over their lives, but it will also have to stay neutral to the content of the commitments autonomous individuals end up choosing. The promotion of social position formally resembles promotion of certain ways of life – after all, how is a person socially positioned is constitutive of her way of life in a certain way – and that’s what makes it perfectionist. However, this being associated with social position and her identity rather than the substance of a person’s commitments, it is not an objectionable paternalist version of perfectionism that Quong assumes it to be, and it also entails a neutralist stance of institutions towards commitments individuals make from their autonomous social positions.
The implication of this for liberal theory, such as the one Quong employs in his book, is that the use of philosophical categories of perfectionism and neutrality in an exclusive manner does not fully reflect the complex nature of individual autonomy as a value. Moreover, this is one of the main reasons why Quong has problems with comprehensive anti-perfectionist conception of autonomy, as well as with alleged paternalism of the liberal perfectionist position. If we understand that perfectionism and neutrality should not necessarily exclude one another, we will be better poised to understand that Sara can coherently believe in the value of autonomy and still abstain from proposing that any individual is made to choose any particular substantive commitment. Also, we will better understand that the perfectionist interventions of the liberal state do not necessarily have to be paternalist because institutions concerned over personal autonomy will make sure that individuals hold appropriate social positions that will enable them to sustain authentic commitments, and will not pass any judgment on the choices individuals adopt in such positions.
The claim that perfectionism and neutrality are not necessarily exclusive of one another is not too far-fetched, and has been somewhat entertained in the newer philosophical literature. For example, Steven Wall recently argued that neutrality and perfection are consistent as long as neutrality is restricted to ideals that are of equal or incommensurable value (Wall Reference Wall2010, 233; also see Merrill and Weinstock Reference Merrill and Weinstock2014). If we presume that both perfectionists and neutralists accept the thesis of value pluralism, there will be a number of equally or incommensurably worthwhile definitions of the good life. According to Wall, liberal institutions should favor these definitions over others deemed unworthy, but be neutral among them. The result is something he calls the restricted neutrality principle, implying that neutrality is owed towards equally worthwhile goods that have adherents in society.
My claim is somewhat different, however. Unlike Wall, I do not think that one needs to be committed to perfectionism first and then apply neutrality against that background. I argue that the non-exclusivity of the two perspectives is much more fundamental because it derives from a proper interpretation of personal autonomy. Namely, since both liberal anti-perfectionism and perfectionism are based on a prior appreciation of the value of personal autonomy, they share a significant common ground by mere default. But upon a closer examination, one could perhaps argue that the plausibility of neutralist and perfectionist arguments, when taken separately, depends on the fact that both simultaneously obtain. In order for the anti-perfectionist requirement (that institutions stay neutral to individual conceptions of the good) to obtain, the perfectionist requirement (to promote the autonomy of individuals) must obtain as well.
Personal autonomy does not occur in individuals without any involvement of their environment. Humans are not born autonomous: they become autonomous through a whole series of events that substantially depend on social structures. Without basic health, education, and a network of social relations that help development of capabilities for autonomy, there can be no autonomous individuals. It is plausible to say that social determinants of autonomy are sufficiently provided only within a perfectionist liberal context, where institutions have a duty to sustain provision of conditions for autonomy through provision of social positions that enable individual regulation of their commitments (see discussion in Anderson and Honneth Reference Anderson, Honneth, Christman and Anderson2005, 127–150).
Therefore, in order to be able to exercise neutrality toward autonomous individuals, institutions need to promote the development of their autonomy. The requirement of neutrality is institutionally satisfied in a meaningful way only against the background of autonomous individuals. Otherwise, neutrality is a shallow ideal, aimed to preserve the status quo under the guise of sustaining a liberal institutional structure.
Similarly, if perfectionist promotion of autonomy (through repositioning of individuals so they can be autonomous) is to be purposeful, institutions must exercise neutrality against choices autonomous individuals make from those positions. Autonomy is a valuable ideal for liberals, but its value is manifested only in social contexts that are bound by it. Otherwise, autonomy has no meaning. Therefore, the most important political implication of the perfectionist promotion of autonomy is the subsequent exercise of neutrality. From the institutional perspective, having autonomous individuals is not the final end of autonomy-promotion efforts. Liberal institutions promote personal autonomy so they can exercise impartiality towards individual choices. Paradoxical as it may sound, the institutional exercise of neutrality is the ultimate aim of liberal perfectionism.Footnote 5
Understanding the consistency between neutrality and liberal perfectionism is important because it bears on the legitimacy of institutional decisions that touch upon their requirements taken separately. Supporters of exclusive understanding will diverge in pointing to the sources of institutional legitimacy. They will point either to the institutional duty of restraint from imposing conceptions of the good or to the need to contribute to human well-being and development of human nature. However, a better way of conceptualizing legitimacy of institutional decisions in such contexts suggests that it derives from simultaneous satisfaction of both neutralist and perfectionist requirements. Government’s regulation is legitimate and properly justified only if institutions support individual social positions from which they can assume control over their lives, and at the same time exercise neutrality towards choices individuals make from those positions. Because liberal perfectionism is based primarily on the value of autonomy, institutions owe neutrality toward individual choices only under the condition of their autonomous acquisition and pursuit.
The legitimacy of institutional decisions in such cases will derive from constitutive co-dependence of neutralist and perfectionist requirements. The institutional exercise of neutrality toward particular individual choices is fully legitimate only if it occurs against the background of institutional promotion of personal autonomy. Similarly, institutional promotion of personal autonomy is fully legitimate only if it happens against the backdrop of institutional neutrality toward autonomous choices. Liberal neutrality and perfection provide background conditions for each other’s legitimacy.
5. Conclusion
In this article, I defended the view that liberal perfectionism and neutrality should not be understood as exclusive of one another. I did so by criticizing two Jonathan Quong’s objections to comprehensive understanding of liberalism. First, I targeted his claim that comprehensive anti-perfectionism is incoherent. Second, I criticized his claim that liberal perfectionism will necessarily assume a paternalist stance.
In relation to both claims, I argued that Quong’s underlying assumptions about personal autonomy undermine the wider validity of his claims. Namely, he adopts Joseph Raz’s understanding of autonomy, which assumes that autonomy is a strongly substantive principle valuable only if it leads to specific individual commitments. By employing a different and more plausible conception of autonomy, I showed that Quong’s objections against comprehensive liberalism lose the bite. The view that comprehensive anti-perfectionism is untenable does not hold because personal autonomy is more determined by the individual’s social position and authenticity rather than the substantive content of her commitments. This makes state interventions targeted at her social standing and the way she relates to her commitments, instead of her substantial conception of the good. This is not perfectionist in the way Quong argues it to be.
Second, I criticized Quong’s claim that liberal perfectionism is paternalist and unjustified when applied against the background of equal distribution of resources and opportunities. I suggested this claim fails because it assumes that liberal perfectionism cannot have a comprehensive justice-based redistributive role, aimed at ensuring that individuals have access to appropriate amounts of autonomy-relevant resources through time. I argued that it also wrongly assumes that the social positions individuals choose after the initial distribution will be amenable to full control of their comprehensive commitments. If a person chooses an option that puts her in a position in which she is unable to exercise control over that and other possible options, then the state has a perfectionist role to reposition her accordingly. Being focused on persons’ social positions, and not their substantive commitments, this interpretation, as I claimed, avoids Quong’s paternalist charge.
Finally, I took a step further from discussions about Quong to suggest that the conceptual assumption we make in comprehensive liberal philosophy about mutual exclusivity of liberal perfectionism and neutrality needs to be reconsidered. I argued that we have reasons to think these positions are mutually supportive and convergent rather than exclusive to each other. I showed that such a proposition is entailed by the arguments made against Quong.
This is, admittedly, a novel thesis about comprehensive liberal theory, which goes against the grain of much of contemporary writing in political philosophy. Let it be noted that I have not argued for its unconditional validity, but only for conceptual plausibility for the comprehensive conception of liberalism, based on considerations of problems initiated by the discussion in and around Quong’s book.
However, if I’m right about it, this could have significant consequences on the ways liberal political theory frames its future discussions. It would mean that we should give up on one of the staple distinctions in contemporary liberal theory and frame some of the old problems anew. I have no particular views on such prospects at this point, nor do I have expectations that a single paper can achieve just that. But if this reconsideration helps us pose new questions about some of the old ideas and moves us forward in finding new ways to solve policy problems by following liberal principles and values, it might be useful after all.
Acknowledgments
This article is based on my doctoral dissertation, defended in June 2014 at the Central European University (CEU) in Budapest. I wish to thank Janos Kis, Nenad Dimitrijevic and Ben Colburn for their kind help and support during this process. I am also grateful to the reviewers and editors of the Canadian Journal of Philosophy for comments and suggestions that have helped me make this article better than it was in the beginning.
