Political equality,plural voting,and the leveling down objection

Challenging the reverse Pareto condition: OPOV necessarily makes some people better off

Consider first the possibility that the value of improving people's welfare is what grounds the egalitarian distribution of votes recommended by the principle of OPOV. The claim here is that the example of plural voting I sketched in the introduction of the paper is conceptually impossible: there are no circumstances under which an unequal distribution of votes will be better for some, all things considered, and at least as good for everyone else, and so no conditions under which OPOV will satisfy the reverse Pareto requirement—OPOV necessarily makes some people better off. ¹⁸ Consider the following argument, perhaps voiced by someone who is on the losing side of an outcome due to plural voting:

Government will only be responsive to my interests if votes are distributed according to OPOV.

On the one hand, voters might have an interest in correspondence (Kolodny, 2014a). This interest is satisfied whenever political decisions match my own judgement—what I myself would have chosen had I the opportunity and ability to choose (Kolodny, 2014a). On the other hand, voters might have substantive interests (Kolodny, 2014a). Our substantive interests are satisfied whenever political decisions advance something we have reason or would have reason to care about—an interest in having good schools for our kids, or in being happy or healthy, for example. One's interest in correspondence might differ from one's substantive interests ¹⁹ —and, at any rate, it is neither the case that OPOV will ensure that at least one of these interests of ours will be satisfied, nor is it the case that plural voting guarantees that either interest of ours must go unfulfilled.

Part of the reason why my having an equal vote fails to guarantee that democratic governments will satisfy my substantive interests or my interest in correspondence is because individual votes have very little instrumental value. ²⁰ Perhaps the rationale behind the first premise relies on the thought that politicians must heed my interests if they wish to get my vote, which they presumably want given that they need a sufficient amount of votes to remain in power or get elected into positions of authority. But insofar as this is the right sort of account of (i), then it seems like politicians only have a reason to heed my interests when my vote is relevantly needed: and votes under or above the sufficiency threshold are in an important sense not needed. ²¹ Conversely, a system of plural voting—even one that operates under non-ideal conditions—might just turn out to satisfy my interests as a matter of brute luck. This is perhaps unlikely, but the formulation of (i) claims that my interests cannot possibly be served—as a matter of conceptual necessity—unless votes are distributed equally, which is surely false.

Perhaps the right thing to say, then, is that governments will probably not be as responsive to the interests of those who hold fewer votes under a scheme of plural voting. If I have an interest in lowering restrictions on who can legally purchase guns and ammunition and the Millian electorate has the opposite preference, then it is less likely that my preference will be translated into law. But two things are worth noticing here. On the one hand, if we take the relevant interests of the population to be their substantive interests, then it is not obvious that premise (ii) is true: presumably, the fact that government is more inclined to respond to the enlightened preferences of the Millian electorate is what guarantees the best results—in this case, the interest we all have in being safe from harm. On the other hand, if we take the relevant interest expressed in (i) and (ii) to be an interest in correspondence, then it seems like the objection is no longer an objection about being straightforwardly well or badly off—and so not really an objection about the way in which OPOV is related to the reverse Pareto condition. Suppose that, upon reflection, I am now inclined to believe that my all-things-considered well-being is better served by having stronger regulations on gun ownership, but nonetheless conclude that those restrictions make me in some important sense worse off. My judgement here might mirror a smoker's judgement who opposes banning cigarette sales even if doing so would leave her all-things-considered better off. Presumably, the objection is about the way in which the law fails to take into account my will—or perhaps about the way in which the law prevents me from carrying out my will. In some cases, of course, being restrained from doing what I want to do will leave me worse off. But not in others. ²² So if there is a sense in which I can properly be said to be worse off by not having my way when it comes to the content and character of political decisions that leave me better off, all things considered, then it seems that the interest at work here is explained by other considerations: perhaps by our interest in not having to obey certain laws, or an interest in freedom or non-paternalism. Whether these sorts of considerations succeed in grounding the value of OPOV is, however, a separate matter—and I will discuss these different possibilities shortly. The point I want to make for now is simply that these types of objections fail to establish that an unequal distribution of votes must necessarily make some people all-things-considered worse off (better off).

I should also mention that this last claim seems to be true even if we think that votes are positional goods or even if political power turns out to be a central component of some people's well-being. Think about education, for example, another positional good. If college were canceled overnight, this might make people who were on the lower end on the distribution of educational opportunities better off in at least one respect: they can now compete for a job that they would have been otherwise disqualified for. But notice that in other respects, these same people might be made worse off still: perhaps with respect to technological innovations they care about, for example, or with respect to the production of vaccines they need, or access to adequate medical care, and so on. With respect to voting power, the claim is that even if OPOV turns out to be really good for some people in at least this one respect—and even if people value voting power above everything else—the election of certain bad candidates under an egalitarian distribution of votes might leave them and the rest of us worse off—worse off once all relevant considerations are factored into an all-things-considered judgement. Compatibilists who think that the value of OPOV is grounded on welfare must show this is somehow conceptually impossible.

Still, it might be argued that the problem is not so much that there are no circumstances under which OPOV fails to satisfy the reverse Pareto requirement (and no circumstances under which plural voting manages to be a strongly Pareto superior alternative), but rather that, over time, a system of plural voting tends to do worse than OPOV. One way of putting this point is to say that once we take into account some historical or even just some likely facts about people and elections, it becomes clear that the conditions under which Plural voting produces better results are very rare: candidates are sometimes indistinguishable from one another, there are no meaningful epistemic differences between electors, and, more often than not, people's commitment to the common good is more apparent than real. So if OPOV does a better job of increasing people's welfare under non-ideal conditions, then perhaps distributing votes equally is the best thing we can do to ensure that political decisions are sensitive to our substantive interests in the long run, and therefore required on welfare-promoting grounds.

These revised statements seem partly right to me. But one question is whether this reply suffers from the same “inconsistency” problems that rule consequentialists face. ²³ Even if this response does give us a reason to generally adhere to OPOV, it provides little guidance with respect to intermittent deviations from the rule. It tells us very little, for example, about what exactly might be objectionable about the United States celebrating its centennial by, say, giving three more votes to residents of the thirteen original colonies. ²⁴ If that practice is objectionable, an answer that emphasizes the good outcomes that follow from general adherence over the long run has trouble explaining what might be objectionable about fairly contained exceptions to the rule.

Challenging the egalitarian condition: OPOV and our interest in correspondence

Consider now a challenge of a slightly different kind. Those who endorse the compatibility view might concede that OPOV will sometimes make us all worse off, all things considered, but argue nonetheless that OPOV is needed to protect some other important interest or value—an interest we all have that is not grounded on the value of equality. So let me first consider, and quickly discard, the possibility that the value of OPOV may be in fact grounded in our interest in correspondence. Some people might think there is something objectionable about trying to bring about the substantively best results if people either do not accept the laws required to bring about what is substantively best or, in some other cases, if people disagree—even mistakenly—about what is substantively best ²⁵ The claim here might be that people have an interest in avoiding political decisions they do not—or could not—accept. Perhaps people have an interest in seeing their political communities correspond to their own views about what is good and right, for instance ²⁶ —in living their public and private lives with a certain integrity—and either plural voting schemes are an obstacle in the way of that interest or, more to the point, OPOV is needed to promote that interest a fair share of the time.

However, there are two problems with this way of grounding the value of OPOV. One initial difficulty is that our interest in correspondence seems to be a second-order interest, an interest that can normally be traded-off against or simply sacrificed for the sake of other values—values like running an efficient government or fostering social cooperation. After all, we do not normally think that people should be exempt from laws with which they disagree, for example, nor do we think that our interest in correspondence justifies granting an individual power of veto to all eligible voters. So there are at least some reasons to doubt that our interest in correspondence is weighty enough to trump the welfare considerations that a scheme of plural voting could potentially bring about and tip the balance in favor of OPOV.

The second problem with this way of grounding the value of OPOV is even more pronounced. Consider two presidential candidates: Ms. Virtue and Ms. Vice. The thought here is that supporters of Ms. Vice disagree with Ms. Virtue's supporters about what is best (or might have reason to dislike the laws that would bring about what is substantively best), and OPOV is then presented as valuable given our interest in correspondence. But the worry is that this sort of grounding for OPOV will be hard to square with a common fact of democratic life: the existence of perennial losers, so to speak—minority voting blocks that are consistently outvoted in free and fair elections despite having equal individual voting power (Christiano, 1994; Kolodny, 2014a). If our interest in correspondence is said to give Vice supporters a claim against the state that votes be distributed according to OPOV, this same interest would seemingly give these persistent minorities the opposite claim. Perhaps it is true that people have a weighty interest in having a fair chance of shaping their political communities according to their own considered judgements about what is best, but then it is unclear in what sense might we properly say that OPOV manages to deliver on that aspiration. Given the presence of persistent minorities in democratic societies, it is doubtful that OPOV is ever sufficient or even necessary to promote our interest in correspondence: a prioritarian distribution of votes, for example, which accords more weight or votes to people or groups whose interest in correspondence has been traditionally neglected in the past might do a better job of getting these minorities what they want a fair share of the time. And, of course, if this is true, then an interest in correspondence could be better served by some alternative inegalitarian voting scheme.

Challenging the egalitarian condition: OPOV and legitimacy

Consider now a third grounding in support of the compatibility view: that OPOV is needed to ensure the legitimacy of the state or its governing body. The issue here turns not on the character and content of the law, but on its source—on the way in which the state, its laws, institutions, and political decisions are related to the people on whom they are enforced (Pettit, 2012). Take, as a kind of paradigmatic case of unauthorized, illegitimate government, a colonial power that has overthrown the local government and imposed its own rule. Most objections to actual, historical colonial powers have been predicated on the fact of their cruel and brutal treatment of the colonized subjects. But this is perhaps a contingent fact of colonialism. Colonial powers could be set up to track the interests of the colonized—whether substantive or in correspondence—and it seems like even in those cases there is still something objectionable about its claim to rule. Perhaps there is something objectionable about living under a regime that we have not authorized in a meaningful sense, and the question is whether OPOV is necessary to extend that kind of authorization and block whatever is ultimately objectionable about the paradigm case.

There are of course certain features of colonialism that seem irrelevant to the question at hand: most obviously, there are issues of occupancy rights and territorial rights violations that occur in one case but not the other. ²⁷ There are also important questions of self-determination—questions that I will be putting aside for the moment: these worries are closer in form and substance to autonomy-based worries, and I will be addressing those shortly. I take it the relevant analogy here turns rather on whether or not the authority of colonizers over the colonized is in some way like the authority of the Millian government over Ms. Vice supporters who find themselves on the wrong side of Plural voting. And so the question is in what way exactly might it be said that the moral standing to rule of public officials elected through a system of plural voting could potentially be like the moral standing of officials appointed by a colonial power.

Perhaps one thing to say here in support of that analogy is simply that neither was authorized in the relevant sense: people have an interest in consenting to government, and officials elected through a scheme of plural voting can, like colonial rulers, attain power without meaningful support from the people themselves. Suppose, for example, that we have a plural voting scheme in place that distributes ninety percent of the total share of votes to just ten percent of the population, leaving the remaining ninety percent of the voting population with just ten percent of the total share of votes. If one were to take a sociological conception of legitimacy (Weber, 1947)—which makes legitimacy assessments depending on whether or not the political order is widely accepted by the governed—then it might be said that the wrong of plural voting is that, unlike OPOV, the government it produces need not reflect the actual consent of the ruled (or at least an ample majority of them). Here is one attempt at making that case:

Legitimate governments are legitimate if and only if they are authorized to rule by the governed (or a majority of them).

Consider one final way of linking OPOV to legitimacy assessments. Perhaps what is distinctively objectionable about colonial rule—and, to the extent the analogy holds, objectionable about plural voting—is rather that it fails to respect our interest in influencing the political decisions that affect us. This need not be bad in terms of what the law does for us. But the objection here is directed at process by which the law is made—at how and why it comes about—and the thought is that laws that do not meaningfully spring from those who are ultimately bound by them are illegitimate. So maybe two questions given this possibility: what is it about a government assembled through OPOV that makes its laws meaningfully spring from the people themselves? And, additionally, in what way would the laws of a government elected through a system of plural voting run counter to our interest in influencing the political decisions that affect us? In most cases, colonial powers legislate without any sort of input from the colonized subjects, and so, straightforwardly, the laws it produces do not have the people as its source of influence. But if this is the challenge, then it is unclear in what sense a system of plural voting might be vulnerable to it. If we all have an interest in influencing the political decisions that affect us, the objection seems best suited against proposals that would disenfranchize some sectors of the population. Mill (1991, 329. Emphasis added) himself thought that his way of distributing votes avoided this type of objection: “… it is a personal injustice to withhold from any one, unless for the prevention of greater evils, the ordinary privilege of having his voice reckoned in the disposal of affairs in which he has the same interest as other people. If he is compelled to pay, if he may be compelled to fight, if he is required implicitly to obey, he should be legally entitled to be told what for; to have his consent asked, and his opinion counted at its worth, though not more than its worth.” If anything, it seems like the grounds for the objection are given by the inequality itself: it is an objection against having an unequal voice—an unequal influence. But then it seems that we are no longer in the same territory as before. To be sure, the inequality between voters generated and sustained by a scheme of plural voting is, or at least seems to be objectionable, but this objection needs to be expressed in some other way: by putting the wrongness of plural voting in terms of procedural legitimacy we risk begging the question against sociological and substantive accounts of legitimacy, both of which seem like plausible alternatives.

Challenging the egalitarian condition: OPOV and autonomy

Let me now turn to a fourth possible grounding for OPOV suggested by my discussion of colonial rule. I’m thinking here in particular about the claim that OPOV is justified by some autonomy-based consideration—either as a requirement of individual autonomy or as being necessary to promote our interest in collective self-determination.

Consider first the individual autonomy claim. According to a familiar Kantian formulation, an agent is fully autonomous only if she acts in accordance with a law she has prescribed for herself. So one possibility is that those on the wrong side of Plural voting—individual members of the political community with fewer votes—will not be able to prescribe laws for themselves: more often than not, they will be forced to obey whatever laws the Millian electorate wills. Proponents of the compatibility view might then claim that because OPOV is necessary to protect the autonomy of individual voters—the self-legislating interests of people like Smith, Williams, Johnson, and so on—no leveling down takes place by distributing votes according to OPOV (even if alternative inegalitarian voting schemes do sometimes make us all-things-considered better off).

The argument is an instrumental argument to autonomy, but attributing such instrumental effects to OPOV is problematic. To clarify why, we might distinguish here between two autonomy-based interests. First, people have an autonomy-based interest in not living under laws that would curtail certain essential features of their personal lives—this is a gay man's interest in same-sex marriage legislation, for example. Second, people have an autonomy-based interest in being the authors of the laws they prescribe for themselves. This latter interest is presumably shared by all members of the community, regardless of whether or not the law under consideration would prevent them from carrying out some of their other autonomous aims: when it comes to same-sex marriage legislation, this an interest shared by us all regardless of sexual preference. The question now is how might OPOV help promote either understanding of that autonomy-based interest—how might it help us craft laws that are friendly to our ends or allow us to become authors of the laws we prescribe for ourselves? Clearly, OPOV is not sufficient to protect our autonomous choices in the first, personal sense: there is no guarantee that Smith will be able to legally marry Jones under an egalitarian distribution of the franchise. And, clearly, OPOV is not necessary here either: certain plural voting schemes might give us candidates who are proponents of same-sex marriage legislation.

So perhaps it is the second, less personal understanding that the autonomy-based interest is tracking—the interest we might have in being the authors of the laws. But the problem with this way of grounding the value of OPOV is that it is difficult to see how my having an equal vote successfully protects my autonomy in this sense—how it might effectively give me some kind of authorship over the laws and political decisions of the state that I would otherwise lack under some versions of plural voting schemes—and therefore difficult to see any sort of meaningful difference between plural voting and OPOV. OPOV does not grant any individual voter control over the laws promulgated by the state: it is the majority who holds that power. And, of course, there is a considerable difference between prescribing laws for oneself, having the majority prescribe those laws, or even having a fair amount of contributory influence over the majority's decision regarding what laws we ought to prescribe for ourselves (Kolodny, 2014a; Viehoff, 2014). Autonomy or authorship is usually thought to require some form of control over decisions that concern us: I should be the one to decide whom to marry or what career to pursue. It would strike us as a loss of autonomy if control over any of those decisions were given to a collegiate body, even if we had equal contributory influence as members of that body: even if, suppose, I, like other members of my family, had an equal say over whom I should marry or what career I should pursue (Brennan 2016; Christiano, 2008; Kolodny, 2014a). The value of autonomy therefore cannot justify a move from plural voting to OPOV: if autonomy requires control, there is no plausible distribution of the franchise short of authorizing a strongman that would accomplish this for any one person.

I want to conclude this discussion of autonomy by considering the possibility that plural voting defeats the community's interest in self-determination—or, more precisely, that OPOV is in some way necessary to preserve that interest Self-determination refers to the political community's claim against both external and internal actors to organize its own affairs independently (Stilz, 2016). “Externally,” this is the interest the community has against a colonial invader, for example. “Internally,” it refers to a people's right to choose their own government (Stilz 2016). If I were to mount a coup on the democratic institutions of my political community, then I would have taken something away from the group as a whole, such that every member of the group would have a complaint against me (Altman and Wellman, 2009). The challenge for those who argue for the compatibility view on grounds of self-determination is to specify the way in which unequal voting schemes violate the people's right to self-rule. Here's one illustrative case. Suppose that the United States granted a prize every election to the three or four or five “most outstanding citizens” in the form of more votes. Presumably, if justice does indeed require OPOV, this would be a pro tanto violation of justice, and that violation needs to be explained in some way. The question is whether this case is analogous to other cases in which the community's claim to autonomy has been violated—cases like colonial invasion or coups against the democratic state. Would all remaining voters have a complaint roughly of the form: “this is a violation of the community's right to organize its own affairs independently?” It is indeed hard to see how that might be true. In purely instrumental terms, it is unlikely that the additional votes given to a handful of citizens would have any effect on the laws, institutions, or political decisions of the state, such that we might say that whatever “the people” choose under OPOV will be chosen under this more constrained version of plural voting. In a more principled vein, it is difficult to see how the right of all other voters to choose their own government has been violated by the existence of these more influential voters: the people themselves retain ultimate authority over whom will govern—they can veto the decision of the most outstanding citizens and can elect leaders in direct opposition to wishes of the three or four prize recipients. Collective autonomy therefore seems to do no better than individual autonomy does in providing firm grounding for OPOV.

Challenging the egalitarian condition: OPOV and freedom as non-domination

Consider now the possibility that OPOV's grounding is provided by the value of non-domination rather than autonomy. According to Philip Pettit's (1997) account of non-domination, one person or group of people is free only insofar as they are not exposed to or liable to suffer the arbitrary interference of another person or group. For Pettit, this means that domination obtains whenever one party has a relatively costless, ready ability to interfere according to his or her or its own wishes in the choices of another. ²⁹ The claim then is that OPOV is necessary to protect our interest in being non-dominated members of the state: by granting a higher share of votes to some but not to others, the state might expose minority shareholders to the arbitrary will of majority shareholders, regardless of whether the Millian electorate is inclined to use their increased political influence in benevolent or malevolent ways. Plural voting might make all of us better off, but that depends on the goodwill of the people in possession of more votes.

I take it that the link between non-domination and OPOV is most intuitively plausible when considering cases in which two groups of people with opposing interests have different voting power and is therefore less costly and easier for the larger group to arbitrarily interfere with the choices of the disadvantaged group or any of its individual members (Brennan, 2016). If Hispanics lack the voting power to enact laws meant to protect them from exploitative or abusive practices, for instance, then this makes it easier for members of other groups to exploit and abuse Hispanics. But there are two problems with this way of grounding the value of OPOV. First, as was the case with the failed attempt to ground the value of OPOV in our interest in correspondence, OPOV will distribute voting power equally to all social groups only in those rare instances in which all groups are composed of the same number of individuals. A non-domination equilibrium in all other cases seems to favor a prioritarian distribution of votes for minority groups who lack the necessary power to block the arbitrary interference of majorities in electoral democracies (barring other institutional arrangements that would protect minority interests: individual rights, judicial appeals, etc.). Second, if what matters is preventing domination through the ballots, why should we care, for example, if the recipients of the “most outstanding citizen award” have more votes than all other members of the community? Votes, as I have said before, have very little instrumental value, and for the non-domination view to fit our intuitions about OPOV we would have to insist that in any large-scale democracy, an additional three or four or five votes given to just one person would have the effect of putting all other voters in danger of being dominated by the preferences and interests of this single voter. So what we need is an explanation of why an unequal distribution of votes would be wrong or objectionable even if the instrumental value of individual votes is negligible. The relational egalitarian account I and others have defended is in a better position to do this, as I elaborate in the second part of this paper.

Challenging the egalitarian condition: OPOV, non-paternalism & self-respect

Before moving on to that defense, however, I want to discuss one final, influential alternative grounding of OPOV. So far, most of the different groundings I have discussed make an instrumental connection between the pattern of distribution of votes and the good equality is said to produce. In so many words: distribute votes equally and what you get is either more welfare, legitimacy, autonomy, some form of freedom as non-domination, or some other value we have an interest in bringing about. The grounding I want to consider in this final subsection makes a different case for OPOV. The issue here turns not so much on the good consequences an egalitarian distribution of the franchise might bring about, but on our obligation to treat people with respect. Perhaps OPOV is required given some criteria of what it means to treat people in ways that acknowledge our shared humanity, for instance. This second approach might be best understood as a status-based justification for equal voting rights: we begin here with an account of people as agents of a certain kind, and then ask what voting scheme is required to treat people thus understood with the respect they deserve: because we are all relevantly alike in this important (but here unspecified) way, it follows we should all have an equal say, an equal vote.

This grounding seems abstract. But think about two more concrete things that might be said against plural voting supported by this status-based grounding of OPOV. First is the idea that Mill's plural voting scheme is objectionable because it treats people in unacceptably paternalistic ways. Second is the closely related notion that plural voting insults us by making a judgement, evaluation, or statement regarding who is more or less fit to pass judgement on matters that affect us all. ³⁰ The issue raised by these two more concrete illustrations of the claim is whether OPOV is required to treat people with the respect that Mill's scheme is said to lack.

There are of course several questions that might need to be addressed first to fully make sense of this account, and perhaps once we do that we might be able to say some things on Mill's behalf. But I want to ignore this line of thought simply because I think that vindicating the compatibility view would still need a defense of a different kind: even if this status-based grounding does in fact undermine Mill's proposal, in order to establish this particular grounding of OPOV's value we need to do more than simply show that his plural voting scheme is paternalistic or that it fails to treat people in non-objectionable ways. In principle, if OPOV is indeed a requirement of justice, it would be objectionable to distribute votes not only according to Mill's plural voting scheme but, depending on what exactly we think OPOV is good for, also according to any number of alternative proposals: including disenfranchisement lotteries, distributions according to alternative epistemic or ethical credentials, and so on. So what needs to be shown by proponents of the compatibility view is that OPOV is actually required—given a number of alternative unequal distributions of votes—to treat people appropriately.

Consider Claudio Lopez-Guerra's (2014) “enfranchisement lottery” as one possible alternative proposal. ³¹ As I understand it, the enfranchisement lottery has two different stages. The first stage consists of a lottery that would disqualify most voters from casting a ballot in the upcoming election. From the sample of voters that remain—and this is the second stage—all “pre-voters” would then have to participate in a “competence-building-process” meant to provide them with relevant information and a moment of reflection about the best alternatives on the ballot. After clearing these two hurdles people would finally be able to vote in the upcoming election for a candidate of their choice.

Lopez-Guerra's proposal promises to deliver some of the same good results that Mill's plural voting scheme promises to deliver: pre-voters can approximate the epistemic qualities that Mill's enlightened electorate is presumed to have thanks to the capacitation process these pre-voters must submit to. Whether those who make it through to the next round will, by virtue of becoming better voters, be more inclined to vote for the common good is a further empirical question. But in any case—whatever we might ultimately say about the merits of the enfranchisement lottery on welfare-promoting grounds—the two relevant questions for my purposes are whether the inequalities produced by the enfranchisement lottery are objectionable on grounds of non-paternalism, and, two, whether they are objectionable on related grounds of dignity, self-respect, or something along those lines.

To begin by addressing the second question first, I take it we have a strong interest in being recognized as agents of a certain kind—as, for example, being capable of forming judgments and acting on them. This seems to be an absolute rather than a relative interest—or, perhaps more accurately, a “non-comparative” rather than a “comparative” interest: I want to be recognized in such a way regardless of whether other people are so recognized (Lopez-Guerra, 2020). Once the point is put like that, it does not seem to me that the enfranchisement lottery undermines our interest in self-respect. If you were to say to me, “listen, explaining the details of this healthcare bill to you is just not worth it”—meaning, you lack the necessary cognitive capacities to understand the proposal—then your attitude would certainly be “demeaning towards me” (Lopez-Guerra, 2020: 778). But the enfranchisement lottery proposal does not seem to disrespect voters in that way: rather, it recognizes that we all share a similar capacity for competency if we think about the issues in a certain way. It says to us something more like “gather information, weigh the relative merits of relevant alternatives, and then act on your own considered judgement” (Lopez-Guerra, 2020).

Is this paternalistic? Well, at least with respect to pre-voters selected by the lottery, I think not. In Seana Shiffrin's (2000) account, paternalism obtains whenever P substitutes his or her judgment for C's on matters that affect C's interests or concern her “legitimate sphere of authority” on the basis of P's presumed superior judgment. On this latter point, the enfranchisement lottery does not take any individual voter's judgment as better than another's. Rather, the starting assumption is that everybody's judgement is equally inadequate. Of course, it is certainly true that after the competence-building process the judgement of fully enfranchised voters is taken as superior, but this is only the consequence—not the basis—of the lottery (López-Guerra, 2014). I tend to think of the proposal as sharing some important features with informed consent in medical ethics—the requirement to have patients know what their options are before willingly submitting to or refusing treatment—what inadvertent consequences might they face after surgery, what related side-effects might they be risking, what benefits can they expect, and so on. The presence of informed consent is often taken to be an indication that the patient has made an autonomous choice—a protection against paternalistic practices such as deceiving or manipulating the patient for his or her own good. And my suggestions is that the competence-building process can be thought of as analogous: if informed consent is necessary for “discovering,” in some sense, the patient's own considered judgement, the competence-building process—by giving voters time, resources, and opportunity to gather all relevant information to make up their own minds—might be plausibly said to be necessary, too.

The harder question is whether the enfranchisement lottery is paternalistic with respect to those who are ultimately left out by the lottery. On the one hand, the question is difficult because, to reiterate, the basis for becoming an eligible voter does not depend initially on one's epistemic credentials. But, on the other, it seems to be the case that what fundamentally distinguishes voters from non-voters in the last stage of the enfranchisement lottery is that the former group is now a better judge than the latter group—and in that way the proposal is similar to Mill's scheme (which I am inclined to think is paternalistic). But there is nonetheless a broader point to be made here that might have been lost by my discussion of the enfranchisement lottery. Regardless of its merits with respect to these two points, we can imagine an alternative enfranchisement lottery in which people's judgment is not taken into account at all—not in the initial criteria specifying who participates in the lottery nor in the final criteria for deciding who gets to vote. We might simply enfranchise at random because—suppose—doing so would save the community a modest amount of money otherwise spent on materials for voting booths and man-hours tallying up the votes (Kolodny, 2014a). The proposal might not be as substantively reliable as the two-stage lottery or even as reliable as Mill's plural voting. But the question rather is whether the inequality that would result from this stripped-down version of the enfranchisement lottery is objectionable because it disrespects people or treats them in unacceptably paternalistic ways. So while it is indeed plausible to imagine that certain distributions of votes are ruled out by the sorts of agents we are and have an interest in being, it is unclear that OPOV—as a characteristically egalitarian distribution of power—is the sole arrangement that might adequately promote that interest The claim defended by proponents of the compatibility view is that OPOV is strictly necessary to protect, promote, or simply treat people with respect—and enfranchisement at random, despite its inegalitarian character, seems to be as sensitive to people's self-respect as OPOV. ³²

OPOV and relational equality

As a starting point, consider once again Mill's proposal—giving more votes to the educated. Would the uneducated have a complaint on the grounds that the scheme undermines relational equality by giving them a subordinate social standing?

Perhaps it would be helpful to distinguish here between two classes of complaints that might be voiced by the uneducated, each class of complaint corresponding to one of two aspects of relational equality that need to be disentangled. So far I have been focusing on power and influence as a way of describing what relational equality entails, but according to the relational egalitarian ideal, relations among equals are relations in which not only is power equally distributed, but also relations in which parties to the relationship show equal concern for the claims, interests, and preferences of one another. One way in which Mill's scheme might be said to violate our equal standing is by failing to show equal concern for the legitimate claims of the uneducated. I want to briefly consider this point and raise some doubts about this way of specifying what is wrong with Mill's proposal.

The particular type of complaint I have in mind is directed at the chosen criteria for distributing votes unequally. On the one hand, the distribution of educational opportunities in society are highly correlated with wealth, and this unfairness will likely carry over to the distribution of votes in Mill's scheme. The uneducated might reasonably complain about not having an equal or even a sufficient opportunity to attain more votes. On the other hand, the educated tend to come from similar racial and social backgrounds, biasing their perspective in ways detrimental to minorities and the overall quality of political decisions (see Estlund, 2003). The specific complaint here might simply be that education is not a good proxy for “good judgment:” the uneducated might reasonably point out that not all of the educated have it, and not all of the uneducated lack it.

With respect to both complaints, the general objection is that Mill's plural voting scheme does not give the uneducated a shot at the relevant good (being in a position to get more votes), and that inequality of concern for their legitimate claims is what seems to be objectionable. In the first case, the complaint is specifically against the fact that the uneducated could come to possess the requisite good judgment if only educational opportunities were fairly distributed—a claim that the state is reasonably expected to fulfill. In the second case, the complaint is that the state fails to treat the equal claims of people with good judgement and bad judgment relevantly alike: some people who possess good judgment are being overlooked simply by virtue of the fact that they did not go to school. Because our standing as equals depends in part on how the state and citizens respond to our interests and claims—as being important and worthy of respect—failing to show equal concern in these ways for the uneducated is what might be said to violate relational equality: the claims and interests of the uneducated are placed in some ways below the claims and interests of the educated.

As I said before, however, I want to record some doubts about this way of arguing against Mill and, more importantly, about this way of grounding the value of OPOV. Both of these complaints are persuasive, of course. But, for one, it is unlikely that all inegalitarian distributions of votes fail to show equal concern for the interests or legitimate claims of those who are left with fewer votes. Part of the appeal of a scheme of plural voting, after all, is that it might sometimes make us all better off. Besides, if the complaint is specifically about some people's claims to an adequate education going unfulfilled, better candidates in office, better laws, and better policy proposals are exactly the sorts of improvements that might fix the unfair distribution of educational opportunities that gave rise to their complaint in the first place. What Mill denies is that opportunities for influencing the political decisions of the state should be distributed equally—without regard to merit, wisdom, capacity, intelligence, or virtue—if there are others who are more competent judges of what is best and more disposed of making good use of their superior judgement for the benefit of the many. And both of these complaints, neither of which requires an equal distribution of votes, seem compatible with Mill's principled point—a feature this approach has in common with alternative non-egalitarian groundings of OPOV.

I think the best thing to say in response to Mill's proposal is that distributing political power unequally is itself objectionable, not that it fails to show equal concern for our legitimate claims. Consider a slightly different case: an example of limited or constrained plural voting. In tweaking Mill's proposal in this way I am simply trying to clean up some of the confounding objections that might otherwise be raised on behalf of minority voters:

Citizen of the Year Award: Suppose that government decides to reward one or two or three “outstanding citizens” with additional votes in the context of an otherwise egalitarian distribution of the franchise regulated by the principle of OPOV. Recipients of the award have performed some extraordinary feats of civic engagement throughout the electoral term (they have picked up trash in their community, set up animal shelters, or rescued people from a burning building) and every four years—to coincide with every presidential election—new awards are distributed.

Suppose everyone had an equal opportunity to perform these extraordinary acts of civic virtue—everyone had more or less the same amount of free time and resources at their disposal. Would citizens on the losing end of the award nonetheless have a complaint roughly of the kind: “it's fine that such a recognition exists, but the award shouldn’t take the form of additional votes to some because strict equality of (at least) formal political power—understood as equal opportunity for contributory influence—is necessary for relational equality?”

As the example makes more or less clear, what needs to be explained here is why the pressure towards strict equality in voting given that the inequalities introduced are unlikely to affect who gets elected and what laws get passed: why should we care about relative power when our absolute power—measured in terms of having our way regarding who gets elected or what laws get passed—is virtually unaffected by the distribution of the award?

Some of the previous accounts I have discussed in support of the compatibility view will have a hard time answering this question. But those who appeal to more explicitly egalitarian considerations can specify what is objectionable about the inequalities generated in cases like Citizen of the Year in a rather straightforward way—without referencing the outcome of the election. Consider three different replies a hypothetical objector might offer against distributing these awards:

Causal explanation. On a first reading, the objector might say that equal power is causally necessary for relational equality: people need equal votes in order to relate to one another as social equals. When some of us have greater political power, we relate to those with less power as our social inferiors.

Causal explanations of the kind offered in (a) are essentially empirical questions, and a lot depends here on how differences in political power—from small to large—are likely to affect our interactions with one another. At the far end of the distribution, where vast amounts of power are concentrated in a few hands, one can imagine those who wield power might begin to see their position as entitled, deserved, or appropriate: they might think their higher influence comes as a matter of right. Under those circumstances, causal explanations will have more force (think about Mill's proposal in the context of Victorian England, for example). But this type of complaint seems unlikely to provide firm grounding in most circumstances. For one, the empirical explanation underpinning this objection is unlikely to obtain where differences in power are small—as they are in Citizen of the Year. Additionally, suppose that extra votes to some serve to counteract overall power differentials. I suspect that there is an objection on relational egalitarian grounds that would remain to the distribution of the award even if its all-things-considered effect was neutral—even if the two or three additional votes made no difference (or even had a positive effect) in our attitudes towards one another. ³⁷

What might this objection that remains consist of? One possibility is simply that there is an expressive wrong associated with the distribution of the awards that is independent of its causal effects. Voters who fail to win the award might worry that their having less power would express their inferiority (even if government did not intend to communicate this), or they might find it objectionable that the state thinks that power is something that can be awarded to some as a way of expressing or communicating (unequal) appreciation.

For present purposes, perhaps this second complaint is enough to carry the day. Even if one were to think that complaints on expressive grounds are usually rather weak, it bears mentioning that the reasons for distributing the awards are themselves rather weak. ³⁸ Perhaps there are other ways in which “outstanding citizens” could be recognized without inducing anxieties about social superiority and inferiority: winners could be given a diploma or a modest financial reward, for example. In these cases, the complaint—if one remains—would be, it seems, different.

In a way, the force of considerations (a) and (b) will depend on the particular form of plural voting we adopt. And, in my view, (c)-type considerations get us closer to what we think is wrong with any inegalitarian voting scheme. Suppose that the distribution of the award does not have the empirical effects associated with (a), and suppose the awards are distributed at random, such that (b)-type worries regarding the expression of social inferiority are at least attenuated. As the objector points out, the awards would be objectionable nonetheless because the ideal of social equality requires equal opportunity for influencing, alongside others, the political decisions that structure the terms of our relationship. Because I have an interest in being your equal, and a corresponding claim to equality with you, I could say that the awards treat me unfairly by giving me a lower standing.

To motivate this idea consider how political power, in contrast with some other forms of recognition, might be special for social equality (Kolodny, 2014b). Suppose, for example, that a lord and his servants, or a master and his slaves, agree to set the terms of their mutual relationship, both parties exercising equal opportunity to influence the content and scope of those terms (the example is Kolodny's, 2014b: 305). In such a case, the fact that each servant or slave has as much influence as the lord or master over the terms of their relationship plays an essential role in moderating the social inferiority that unequal power would otherwise entail. As Niko Kolodny (2014b: 305) suggests, when both lord and servant set the terms of their relationship together, with equal influence, “this [makes] the labels ‘lord’ and ‘servant’ less applicable, [and] that is the point.” Their equal power at least evokes the thought that we are now dealing with something else: a relationship that more closely resembles what we would otherwise normally associate with certain types of relationships between employers and employees, each negotiating on equal footing the terms of their contract. Things would already be different if the lord retained slightly more power than the rest of the servants. This would give rise to recently familiar concerns about bosses being like petty dictators and offices like small fiefdoms (see Anderson, 2017). To the extent that we have equal opportunity to influence how these relationships are structured, worries about social inequality are diminished. To be sure, this might not be enough: some deep inequalities in status, concern, or regard might remain regardless of the distribution of power. And, in such cases, we would hesitate to call the servant his lord's (full) equal. So the suggestion is simply that equal opportunity for influence is one constitutively necessary component of social equality (see Kolodny, 2014b; Viehoff, 2014). When power is equally distributed among members of a group, “there is no one whom we can point to and say: ‘because you had a greater opportunity than me to influence the shape and form of a decision that binds us both, I am, in virtue of being subjected and bound to obey that decision, a subordinate of yours’” (Koldony, 2017: 106).

Now, as I hope is clear, the objection is not merely that the servant can secure for himself a better deal given an equal opportunity to influence the terms of his employment. I don’t doubt this is largely true. But even if we could imagine the opposite being true—a situation in which, suppose, the kindly master is already inclined to give her servant more free time, better meals, and better lodgings but, due to incompetence on the servant's part or miscommunication, his participation on equal footing would end up affecting his substantive interests in important ways—the servant would nonetheless have a reasonable complaint if he lacked that equal opportunity. The servant would have a reason to complain about a world in which the lord held more power over him, even if an unequal distribution of power—and thus his unequal standing—would be better for him. The servant might of course go along with the arrangement: he might prefer his lower standing, all things considered, given the alternative. But there is a complaint here should he change his mind—a complaint that is in some sense similar in strength to the complaint voters would have living under an electoral system that distributes votes unequally.

Let me try to strengthen the case for this constitutive explanation by considering the question of whether or not this way of grounding the value of OPOV in relational equality is not also available to people who endorse the compatibility view: couldn’t one also say, for example, that OPOV is constitutively necessary for autonomy, even if not causally necessary? Or couldn’t one similarly say that OPOV is constitutively necessary for non-domination?

To develop this thought further, one might argue, for instance, that it is because autonomy just is (in part) a matter of deciding things for ourselves that having a procedure by which we come to our own decisions constitutes a form of autonomy. So OPOV, the thought might run, is constitutively necessary for autonomy. But this type of reply is unpersuasive for two reasons. There is first the semantic issue that even if one could properly say that OPOV is indeed partly constituted by the value we attach to deciding things for ourselves, autonomy is essentially a matter of deciding things for myself. Second, notice that this sort of explanation makes no reference to what is distinctive about OPOV (distinctive at least relative alternative inegalitarian distributions of votes): an equal distribution of power. Even if one were to understand the claim being advanced here as a kind of collective self-determination to which OPOV is said to make a constitutive contribution, it is choosing for ourselves that matters here, and, as I tried to argue in section III.D, an inegalitarian system of plural voting is one such mechanism by which we are given the opportunity to make our voices heard and collectively choose for ourselves.

Now, as to the question of whether or not OPOV makes a constitutive contribution to non-domination, I think the answer is most probably “yes.” Non-domination is not only essentially about power over others rather than power over ourselves, but it is also fundamentally about how power is distributed: part of what we mean when we say that A dominates over B is that A has the power to dominate B. But it is important to distinguish here between two possible meanings of non-domination. On the one hand, power differentials might be objectionable on non-domination grounds because they could potentially give some people an improper control over the lives of others. Among other things, this is objectionable because we have an interest in leading our own lives according to our own choices. I take it this is what is usually meant by freedom as non-domination. On the other hand, power differential might be objectionable because we have an interest in standing in relations of non-domination with one another. As Philip Pettit has argued, this objection is not an objection to how people come to exercise power: it is objectionable that we stand in relations of domination even if people abstain from exercising that power to foreclose some of our choices. Unlike this second understanding, the constitutive claim in support of the first understanding of freedom as non-domination runs into the same problems as the autonomy claim (how does OPOV make a constitutive contribution to my interest in being self-legislating?). But notice that this second way of grounding the value of OPOV in non-domination does not compete with the relational egalitarian grounding I and others have proposed: it is merely a different way of stating the same worry. What we have here is an objection to standing in an inegalitarian relationship, and OPOV does make a constitutive contribution to our equal standing.

Consider, then, the following argument in indirect support of premise (iii) above: ³⁹

Equal opportunity for formal contributory influence (equal power) is a constitutive component of social equality.

Objection: Some unequal distributions of votes are consistent with social equality

Let me now consider some objections to this way of grounding the value of OPOV. One possibility here is that certain inegalitarian distributions of votes might be compatible with relational equality, making premise (b) in the previous argument false. This is a claim that Andreas Bengtson (2020) has recently argued for. Think once more about the case of persistent minorities, voting blocks that are consistently outvoted in free and fair elections. Bengtson argues it will be difficult for these people to relate on a footing of equality with persistent majorities, the empirical premise being here perhaps that those who lose elections on a regular basis will come to feel lesser than those who win most of the time (and vice versa). Bengtson thinks this psychological fact about winning and losing is a problem for relational egalitarian accounts of OPOV's value, a problem that perhaps shows that OPOV is in fact not necessary for social equality: after all, if the causal mechanism leading people to feel as superiors and inferiors is given by the outcomes of elections, it might be the case that by adopting a scheme of plural or prioritarian voting (which accords more weight to the votes of the worse off in terms of electoral results), we could balance electoral victories and defeats in such a way so as to get these two groups to treat each other as one another's equals, not as winners and losers.

Bengtson's discussion illustrates the fact that the demands of relational equality can sometimes pull in opposite directions. Think back, for example, to the three objections I mentioned against Citizen of the Year. It might be the case that insulting Smug is an effective way of boosting Meek's confidence (or reducing Smug's), causally improving the way Smug relates to Meek. With respect to voting, the case of persistent minorities points to a conflict between the causal and constitutive explanations I offered against an unequal distribution of votes. But if the conclusion is that, given that tension, an equal distribution of power is therefore not necessary for social or relational equality, then it seems to me like this is the wrong conclusion to reach. To focus now on the state's obligation to promote egalitarian relations among citizens, it seems like the objection here is analogous to an objection one might have against the state making derogatory comments about Smug as a social group in order to encourage better relations with Diffident or Meek. Or think about the relational egalitarian ideal more generally, and what appears to be a similar tension between regarding someone as an equal and treating him as an equal. It might be the case that for me to eventually regard as an equal someone whom I now regard as my superior I ought to force myself to treat her as an inferior, but I doubt the right conclusion that follows from this is that treating people as equals is not necessary for relational equality. Perhaps what we ought to conclude with respect to the case of persistent minorities is just that an equal distribution of power is insufficient for relational equality—a conclusion that I think is right. Of course, there are important questions here about what would be better to do, all things considered, under these non-ideal circumstances by relational egalitarian lights. But if the constitutive grounding of OPOV does indeed provide a good explanation for the objection we have against cases like Citizen of the Year, an equal distribution of power seems to be necessary for relational equality in the same way that treating someone as an equal in my previous example does: these are, or so it appears to be, necessary conditions for full relational equality—which is the sort of explanation I am after. ⁴¹

I think this point also travels across different unequal distributions of votes in which the inequality itself would have no causal effect on the way in which voters treat one another (or even if it had a positive causal effect). But perhaps there is an alternative defense of certain unequal distributions of power as being compatible with relational equality that does not depend on the causal effects such a distribution has on our attitudes towards one another. Think about López-Guerra's enfranchisement lottery proposal, for example. As currently designed, the lottery distributes power unequally at time t, and as such it might be suggested that there is an objection on expressive and constitutive ground the remains here independently on the effects it might have on how we treat or regard one another: the arrangement falls short of what is required for us to live in a community of equals, mutually expressing our equal standing to each other. This thought becomes perhaps clearer if we think about a modified version of the enfranchisement lottery, one in which only one person (perhaps he or she with the best understanding of the issues at stake) makes it out of the pre-voter stage of the lottery. In that case the objection is easier to see. Even if this one person could be any one of us, deciding for all of us only once in a lifetime, the fact that the enfranchisement lottery grants this kind of unequal power to one voter surely elicits doubts about our status as equal citizens. But there might be a revised version of the lottery nonetheless that I want to consider here as an objection to my view that OPOV is necessary for relational equality.

To sharpen the objection I have in mind, suppose a group of friends take turns deciding what to do together. On Mondays, John decides. On Tuesdays, Paul does. On Wednesdays, George does. On Thursdays, Ringo does. This situation does not appear to be straightforwardly objectionable: even if the distribution of power between them is unequal at time t, it is not obviously the case that such an arrangement threatens their standing as equals (perhaps because we are imagining that power is distributed equally between the four of them over the course of their entire lives, at time t + n). Of course, thinking now about the enfranchisement lottery proposal, this type of alternating influence is unlikely to obtain in large-scale democracies, each person getting to be a voter and non-voter as many times as everybody else, but we can approximate this by design: we can, for example, forbid voters from reentering the lottery until everyone else has been selected as a voter (or, in a similar spirit, we could we could opt for an age-differentiated voting weight system, which accords more votes to people at certain points in their life, and so distributes power unequally at time t but equally at time t + n). What might we conclude from this?

Well, one possible conclusion suggested by this example is that not all inegalitarian distributions of votes are incompatible with social or relational equality: perhaps all that relational equality requires is that power be distributed equally over the course of people's lives, not that power be distributed equally at all times. And, of course, if this is true, some unequal distributions of votes at time t might be compatible with relational equality. Compatibilists might therefore suggest that because distributing votes according to the principle of OPOV is not grounded on the value of relational equality, it is not obvious that OPOV is a relevant case of leveling down. Is this challenge plausible? I do not think it is, so let me try to register some doubts about drawing this conclusion from the previous example.

Notice, first, that time-slice inequalities do seem to matter on relational egalitarian grounds in a wide variety of cases—raising the question of how well this conclusion about relational equality and the unequal distribution of power at different points in time generalizes beyond what seems to be true of these four friends. Consider two examples from McKerlie. The first concerns a feudal society in which nobles and peasants exchange places every ten years. The second concerns a marriage in which both partners trade positions of authority. He says:

Imagine a new kind of egalitarian society. It contains great inequality, with happier lives attached to certain social positions. But at a fixed time people change places and switch from a superior position to an inferior one or vice versa. One example would be a feudal society in which peasants and nobles exchange roles every ten years. The result is that people's lives as wholes are equally happy. Nevertheless during a given time period the society contains great inequality, and in one sense this always remains true. I will call this system “changing places egalitarianism.” If equality between complete lives were all that mattered, an egalitarian could not object to it. But I think that many egalitarians would find it objectionable (McKerlie, 1989: 479).

Similarly:

Most of us believe that there should be equality in a marriage. A marriage in which the husband makes all of the important decisions, and achieves his own goals with his wife's help, can be criticized for not being an equal partnership. Rather than having the inequality continue throughout the marriage, it would be better if the wife had her turn at being the leading partner—if, later in the marriage, she were able to achieve her own goals with the assistance of her husband. Then, if we think about the marriage as a whole, the husband and wife have equal shares in decision-making and responsibility. Even if the marriage does change in this way, we would not regard it as the best kind of marriage. It is not an ideal marriage because the husband and wife are never equal partners during the marriage. There is no time at which they share decision-making and responsibility (McKerlie, 1992: 289).

McKerlie raises some issues about the implications of equality and time in the distribution of income, wealth, or welfare that I will be putting aside. As many people have noted, what strikes us as immediately objectionable about these two examples is how the relevant groups stand in relations of inequality—sometimes acting as superiors and other times as inferiors.

There seem to be two further issues at work here. First, power inequalities in both examples seem to be working to the disadvantage of the powerless. Part of what is troubling in both cases is that the interests of one person or group are in effect subordinated to the interests of the other person or group. In this sense, McKerlie's examples are unlike the revised version of López-Guerra's enfranchisement lottery and the comparable arrangement between friends (though perhaps relevantly analogous to an age-differentiated voting weight scheme). Second, the way in which McKerlie describes his two examples raises the question of how these inequalities—between nobles and peasants and husband and wife—might causally contribute to them treating each other in objectionably hierarchical ways, given the long stretches of time in which one group or person has power over the other. One question is whether these two points are enough to explain what we might find troubling about the situation in which the disadvantaged find themselves in. Would the objection simply disappear if, for example, people were to trade places after every couple of months (or weeks) subject to the constraint that those in positions of authority must rule with the best interests of the powerless at heart? If not, we need an explanation of why time-slice inequalities in the distribution of power between peasants and nobles—and maybe husbands and wives—seem more objectionable than the same kinds of power inequalities between friends.

The disanalogy, I think, seems to revolve around two issues. First, there is a power-over-what question that might be important for social equality and needs to be teased apart here. One way in which the four-friends example seems different from McKerlie's two examples has to do with the sort of authority we might expect the powerful to exercise over the powerless. Perhaps we are less troubled by the way these friends have decided to distribute power between them because more often than not we would be right in assuming that the decisions over which John has authority on Mondays, say, are both relatively trivial and fairly constrained: what movie to see, where to eat, etc. Returning to the paradigmatic case of the kindly master, it seems that unequal power over a limited range of minor decisions does not—or, rather, need not threaten our equal standing, at least not in the same way. Suppose lord and servants (however many there may be) have equal opportunity to influence the terms of the contract under which the servants are to be employed by the lord, but the lord nonetheless retains sole authority over certain aspects of uniform choices—things the lord expects the servants to wear while the duration of their employ. It seems that so long as opportunities for influencing the general terms that structure their mutual relationship are equally distributed, then anxieties about social inferiority are diminished (though, to be sure, power over trivial choices can accumulate—and, to the extent that they do, worries about social inferiority and superiority will kick back in). Things would look different for me if the arrangement between friends would allow whoever is in charge that day to unilaterally change the terms of their original agreement and distribute power differently without consulting each other, for example.

Second, notice that being in a position to exit a relationship is one way of “moderating” or even perhaps entirely avoiding “what would otherwise be a relation” of subjection and subordination (Kolodny, 2014b: 304). If one can exit a “slave contract” at will, as Kolodny (2014b) points out, it is not clear in what sense one is a slave to begin with (provided that one's exit option is not to land in a different slave contract). One way in which the four-friends example is unlike the feudal society example—and, to a degree, unlike the marriage example—is that any of these friends can easily refuse to obey the dictator-of-the-day's directives by simply exiting the relationship. Perhaps this is an additional reason why we don’t find the lord's greater power with respect to uniform choices as troubling as we otherwise might: under normal circumstances, the servants can just refuse to comply and look for work elsewhere.

It is important to take these two points together. To the extent the power I have over you grants me an increased opportunity to modify the terms of our relationship in ways that we both have reason to care about and there are limited exit options available to you, then the stronger the case for thinking you have a subordinate status to mine. This is true regardless of whether or not I exercise that power and whether or not my exercising that power would have a positive effect on your life. Imagine these four friends are business partners with equal shares in a joint venture. And suppose the authority granted by these friends’ agreement allows whoever is in charge that day to make whatever decision he sees fit without consulting the other three. Finally, suppose that neither of them has a decent fallback: they can try to make the business work or they can each return to their previous low-paying jobs with long hours. Given those starting assumptions, I see no fault in someone refusing to agree to an entire-lives distribution of equal power on relational egalitarian grounds: the objector might point out that all four of them are in it together as equal partners, and the agreement would make each of them subordinate to the other depending on what day it is. When it comes to living together as citizens (and as spouses or business partners) with corresponding claims to equality with one another, there is something to be regretted if parties to the relationship are never actually equal partners (McKerlie, 1992), and a time-slice distribution of equal power seems to be necessary for that.

Returning now to the revised enfranchisement lottery proposal and the age-differentiated voting scheme, it should be clear that the sort of moderating influences that made the friends’ synchronically unequal distribution of power more tolerable will be difficult if not impossible to reproduce. First, it is unclear how the state can allow citizens to simply refuse to comply with the decisions of the state or help dissenters exit the political community at will. Second, the political decisions of the state are neither trivial nor fairly constrained: they are wide-reaching, final, and have profound consequences on how we are to live together. It is doubtful that a political association can exist for long by refraining to engage in lawmaking of the kind that would not be relevant for social equality. Finally, it is hard to see how both of these moderating conditions would work together in anything that would resemble a political community that is recognizable to us. So while it is possible for an entire-lives distribution of equal power to be compatible with relational equality, this is true exceptionally. Most of the times, relational equality puts a strong pressure on distributing power equally at any given point in time—and this is especially true for relationships among citizens.

Objection: Some lotteries are compatible with social equality

I want to address now one final objection. Consider the case of traditional lotteries—lotteries in which political offices (or laws) are distributed (or chosen) by lot: if relational equality requires that power be distributed equally at time t, couldn’t an equal absence of power at any given point in time satisfy the relational egalitarian ideal? Let me call these types of traditional lotteries “egalitarian lotteries” in order to distinguish them from lotteries that distribute power unequally at time t. An egalitarian lottery deprives us all of the opportunity to positively influence the political decisions of the state and, as with OPOV, it seems like there is no one whom I can point to and say “because you had a greater opportunity than me to influence the shape and form of that decision, I am, in virtue of being subjected and bound to obey that decision, a subordinate of yours.” Therefore, the challenge goes, OPOV is not strictly necessary for relational equality.

There are perhaps some relational egalitarians who might want to resist the objection by showing conceptually that relational equality in fact requires a positive opportunity to influence the political decisions of the state, an opportunity that lotteries might deprive us of. In an interesting paper, Sam Arnold (2017) has argued that in order for there to be a society of equals, members of that society have a duty to participate in certain kinds of political activities. Among the reasons he offers, Arnold says that showing equal concern for the equally important claims, interests, and preferences of all requires (i) having some idea of what the other person's interests are, and (ii) using that knowledge to reach outcomes that are mutually acceptable to all—or at least outcomes that no one could reasonably reject. ⁴² Arnold (2017: 37), thinks that both conditions are likely to go unfulfilled when refraining to participate in certain political processes: if you fail to participate, then I have no idea what your interests are: “you force the decider to guess, rather than know, which policies or outcomes you could live with. This runs contrary to your interests (you might be saddled with a choice you cannot accept) but also to those of your partner, who—being an egalitarian—wants (among other things) to author decisions that are acceptable to all involved.”

Arnold's emphasis is on the outcomes themselves and the reasons we as relational egalitarians might have for wanting them to be of a certain kind—outcomes that are consistent with equal concern. But there is a different way of reading this point—a reading that emphasizes the fact that, as a committed egalitarian, I not only want your interests satisfied but I may also want to be in some sense responsible for your interests being satisfied. For helpful contrast, think about a relationship in which we showed equal concern for each other only as a matter of luck, or only because someone or something external to our relationship was forcing us to treat each other in that way (like a boss forcing her racist employee to treat his black co-workers with respect, for instance). ⁴³ The thought then might be that OPOV is strictly necessary for relational equality because it is the only arrangement by which we are able to both distribute power equally and collectively author decisions that are acceptable to all involved. Lotteries might sometimes produce mutually acceptable outcomes, but they do so in a way that runs counter to our interest in being the authors of (and bearing responsibility for) those decisions: they deprive us of the opportunity to choose for ourselves the laws and candidates we think are consistent with equal concern. So while there is a sense in which even though lotteries manage to preserve the equal power desiderata of relational equality, they might nonetheless fail to promote the interest we have in authoring the political decisions of the state: put differently, as a matter of relational equality, we want to show equal concern, not merely have laws that are consistent with equal concern.

I think this extended and revised version of Arnold's claim is suggestive, but I also think it is ultimately ill-suited as a response to the objection I want to consider here. Part of the problem with this account seems to be that even if it were true that OPOV is the only sort of arrangement that potentially satisfies both relational egalitarian requirements simultaneously, we can nonetheless imagine different types of arrangement in which these two requirements are satisfied successively. This could be the case of a society with high levels of contestation, one in which even though candidates are selected on the basis of an egalitarian lottery, citizens actively contest policy decisions and hold their representative accountable for failing to show equal concern. It is unclear to me that a society like this one would have gone awry by relational egalitarian lights.

Though there is much more to be said here, I want to put this particular discussion aside: as it happens, I do not think that my argument depends on successfully proving that these kinds of egalitarian lotteries fail to live up to the demands of relational equality. In fact, what I think is required by relational equality—particularly in the political domain—is an equal distribution of political power at time t, and both egalitarian lotteries and “one person, one vote” seem compatible with that point. Still, it seems to me sufficient for my purposes to say the following two things in response to the challenge presented by lotteries:

At the first level, the reason why we think it is inappropriate for some to have more opportunities for influence than others is because an equal distribution of power is constitutively necessary for social equality.

Now, if these thoughts about lotteries and relational equality are right, the argument I presented before stands in need of revision. Here's perhaps one way of modifying the argument:

(a₁) Equal opportunity for formal contributory influence at time t (equal power) is a constitutive component of social equality.

The question is whether this revised conclusion is still compatible with the premise I set out to defend at the beginning of Section IV, namely:

Equality is the central or fundamental value promoted by OPOV (the egalitarian condition).

The issue here seems to turn on how best to interpret the fact that OPOV is all-things-considered justified (if it is) not only by virtue of the contribution it makes to social equality but also partly because of the contribution it makes to other values. As (b₁) makes clear, lotteries and OPOV are contingently necessary for equal power. If lotteries are ruled out in case X for whatever reasons then OPOV is necessary for social equality in case X. If voting is ruled out in case Y for some other reasons, then lotteries are necessary for social equality in case Y. Would it be right to say then that “equality is the central or fundamental value promoted by OPOV,” given that a full explanation of why we ought to distribute votes according to OPOV in case X is given by social egalitarian reasons as well as by some other considerations?

Consider the following illustration:

Brown Sweater: It is cold outside. I decide to wear my brown sweater because it matches the pants I’m currently wearing.

My view is that OPOV stands in relation to social equality in more or less the same way my brown sweater stands in relation to the weather outside. What explains most fundamentally why I am wearing the brown sweater I am currently wearing is the fact that it is cold outside, details about the sweater being secondary to that first-order reason. In the same way, what explains most fundamentally our interest in OPOV is given by the first-order reason requiring that opportunities for influencing the political decisions of the state be distributed equally given our interest in social equality, details about efficiency relative lotteries being secondary to that first-order reason. Without it being cold outside I would normally lack a good justification to wear my brown sweater. And without an interest in social equality, we would perhaps similarly lack a good justification for OPOV. ⁴⁶

In sum, the fact that there are a plurality of reasons that will ultimately point us in the direction of one policy over another does not, I think, undermine the relational grounding of OPOV: if OPOV is required or permitted by justice, the requirement is most plausibly explained by the contribution OPOV makes to our standing as equals. To make one final remark, consider that other, more traditional cases of leveling down will also be justified (assuming that they are for the sake of argument) by a plurality of reasons. Let me return to Incentives. Suppose we have implemented a policy of equal post-tax pay for everybody, and we want to know what justifies that policy. Just as with OPOV, there seems to be a two-tiered justification here. On a second or lower level of justification, the reason for implementing this policy over, say, a policy of equalizing all salaries pre-tax is because this way people who want to work jobs that benefit everybody can have a clear market signal of what those jobs are: knowing that being a software engineer makes more, pre-tax, than a professional bowler will tell me that society values the work done by software engineers more highly than the work done by professional bowlers. However, higher up in the justification order, as it were, the reason that ultimately justifies that policy (and again: assuming that the policy is justified) is simply because justice requires equality. If OPOV does not level down the relevant population because there are further non-egalitarian reasons for distributing votes equally, then neither does equalizing salaries post-tax level down. But this conclusion seems clearly wrong.