In this chapter, David wants to distinguish his epistemic argument for democracy from what he calls the democracy/contractualism analogy (I’m going to refer to this simply as the analogy or the analogy argument). The analogy rests on two key claims. The first is that justice or moral rightness is best understood via some version of moral or political contractualism. The second claim is that democratic outcomes have the capacity to track the requirements of justice or morality because democratic institutions can be arranged in a manner that is sufficiently similar the structure of the hypothetical choice situation of whatever theory of contractualism is favoured. David rejects the analogy because he believes this latter claim is false. If democracy does track justice (something David obviously doesn’t want to deny) it is not because democratic institutions mimic the features of a hypothetical contract scenario.
One of the attractive features of the analogy argument is that it can be used without appealing to substantive claims about justice over which there is likely to be qualified disagreement. Provided some version of a hypothetical choice scenario would yield true or valid principles of justice or morality, we don’t need to know what those principles are – we only need to design our actual democratic institutions so that they are sufficiently similar to the contractual scenario and will thus deliver roughly the same results. The analogy thus relies on the idea that similar procedures will produce similar outcomes. This analogy argument is attributed by David to John Rawls, Brian Barry, William Nelson, and Joshua Cohen (I think the attribution to Rawls is questionable, and I’ll say more about this later). Because Scanlon’s contractualism has been most influential for advocates of the analogy argument, David focuses on Scanlonian contractualism, though I think the objection he makes will be robust across the major versions of contractualism.
In order to know whether our actual democratic procedures can be sufficiently similar to those of an ideal contract scenario, we need to know what the key features of contractualism are. David provides the following list (drawn mostly from Scanlon) on pp. 241-242:
(a) The task is to choose general rules that shall apply to all members of society.
(b) Agreement is not forced or coerced.
(c) Everyone affected by the chosen arrangements is a party to the choice.
(d) Participants interact on terms of equal opportunity for input and no unequal bargaining power.
(e) Agreement is defeated if it is rejected by any of the individual participants.
(f) Participants will not reject any proposal unreasonably.
(g) Participants are all motivated to come to some agreement rather than live without any rules.
(h) The proceedings are not bound by, and the decisions are not subsidiary to, prior tenets of justice or right.
David then notes a number of ways in which actual democratic procedures quite obviously deviate from this contractual ideal:
(not –c) Not everyone affected does in fact participate (e.g. children and future generations).
(not –D) There will always be some unequal political power in real world democracies.
(not –h) Democracies are almost always constrained by a higher law, for example, a constitution.
(not –e) Nothing like an individual veto is appropriate in large-scale democracies.
It’s not necessarily a problem if real democracies do deviate from the contractual ideal in some ways since the analogy argument doesn’t claim the procedures must be perfectly analogous, only sufficiently so. However, David argues that the last of the differences (not –e), which he also refers to as the veto gap, is a crucial disanalogy, one which fatally undermines the analogy argument. The basic problem is that a veto for individuals in real democracies is not only impractical, but undesirable. We wouldn’t want individuals to have such a power, since even if all individuals were well-intentioned, they are bound to misuse the veto due to the problems of limited information and rational errors. The veto, likely to be misused, would thus create a status quo bias if deployed in real democratic systems.
The difficulty, of course, is that the veto is a crucial feature of the hypothetical choice scenario – it is what prevents some people from being used as mere means, or in Rawls’s terms, it is how contractualism models the separateness of persons. The veto is essential to contractual theories in this way because those theories must also posit that parties are reasonably self-serving. That is, the parties in a contractual theory don’t engage in justificatory arguments with one another about what morality or justice requires because that would be to presuppose conclusions about justice or rightness that are prior to the contractual device. This would either make the contractual device circular or superfluous, or it would create an infinite regress. In Estlund’s terms, the primary question such theories are meant to answer is: what does justice (or morality) require or forbid? But in order for the theories to avoid circularity or regress, the parties in the choice situation cannot address this primary question. Instead they must address some subsidiary question, for example, which rules will maximize my share of primary goods? Because of the way the choice situation is designed, if each person addresses themselves to this subsidiary question, the agreement reached will provide us with the answer to the primary question while avoiding the charge of circularity or redundancy.
Essential to Scanlonian contractualism is the idea that parties can only reject proposals for ‘personal reasons’ (e.g. appeals to the harm, cost, or undue burden any given individual would have to bear under a proposed rule). Impersonal reasons (e.g. the intrinsic beauty of the Grand Canyon) cannot be appealed to except insofar as they contribute to someone’s personal reasons. David also claims it is an important part of Scanlon’s theory that parties cannot reject proposals on the grounds that someone else has a reasonable objection – you can only reject proposals if you yourself have reasonable grounds for doing so (it’s not clear to me that this feature is a necessary part of Scanlon’s theory, and I’ll say something about this later). In this way each party in Scanlon’s theory is reasonably self-serving.
The difficulty for the analogy argument should be clear. If these features were mirrored in real democracies, but there was no individual veto, then many proposals that would be reasonably rejected by Scanlonian contractualism could easily be passed by a democratic majority. Without the power to veto, a person, or a minority group which had reasonable grounds to reject a given proposal could be defeated by a majority since no one is supposed to reject a proposal merely because others have reasonable grounds for rejecting it. Provided the majority of people don’t have reasonable grounds to reject a proposal, they can pass it even though it may be reasonably rejectable by a minority. The problem can be presented as a dilemma:
(1) either we mimic Scanlon’s contractualism (absent the veto), and citizens in a democracy can be moved to reject only for personal reasons, in which case many proposals that would be rejected in Scanlon’s contractualism might be approved by majority-rule in a democracy, or else
(2) we declare that citizens in a democracy should reject any proposal when they think that anyone, not just themselves, has a reasonable reason to reject it. This would no longer look anything like contractualism since citizens would be addressing the primary question, that is, they would be asking themselves whether this proposal can be reasonably rejected by anyone. So even if this model of democracy does track what justice or morality requires, it would not do so because it mimics the choice situation in Scanlon’s contractualism. It would do so because citizens were addressing themselves to the primary question of what justice or morality requires.
The chapter concludes with David’s take on how parties in Scanlon’s contractual scenario can take reasonable account of others’ claims without falling into the trap of addressing the primary question. David believes this can be done since a Scanlonian contractor only asks whether they themselves have reasonable grounds for rejection, and this is determined by seeing if the strength of their complaint is weightier than the complaint anyone else could press against an alternative proposal. If their complaint meets this test then it can ground a reasonable rejection (this is explicated using two figures on pp. 252-253). By asking this question, contractors don’t ask the primary question since they are only asking whether they have a reasonable complaint against a proposal – no individual contractor asks whether anyone has a reasonable complaint, which of course is the primary question.
In sum, the problem for the analogy is that unless citizens are given a veto (which David argues is undesirable) or unless citizens are allowed to reject proposals purely because they believe others have reasonable objections, majority rule will simply permit many proposals that would be reasonably rejected in a contractual scenario. The key difficulty is that there is no plausible way for real democracies to mirror or mimic the central feature of contractual scenarios: reasonable rejection for personal reasons only. It seems the idea of reasonable rejection can’t be replicated in real democracies, partly because it’s not clear how ‘the idea of [reasonably] rejecting a proposal [can] be translated into some approach to [reasonably] voting for proposals’ (p. 254) in a manner that doesn’t require citizens to address the primary question. If democracy does do a relatively good job tracking the requirements of justice or morality, it must be because citizens do directly consider the primary question, and not, as the analogy argument suggests, because democratic institutions somehow mirror the procedures of moral or political contractualism.
Let me start by saying that I find David’s argument against the democracy/contractualism analogy persuasive – I think he must be right that the analogy argument as he sets it out is bound to fail. I’ll make one small (and picky) exegetical point, before trying to raise one or two possible difficulties for David’s position, though I’m not convinced they really are serious difficulties.
First, the picky exegetical comment is that I’m not sure David is right to suggest, at least on the basis of the quote he provides, that Rawls is a proponent of the analogy. The quote from Rawls is on pp. 239-240, and it reads as follows:
‘The guarantee of fair value for political liberties is included in the first principle of justice because it is essential in order to establish just legislation and also to ensure that the fair political process specified by the constitution is open to everyone on the basis of rough equality. The idea is to incorporate into the basic structure of society an effective political procedure which mirrors in that structure the fair representation of persons achieved by the original position.’ [emphasis added by David]
The first sentence might seem to imply the analogy, but it needn’t do so. Rawls says the fair value of political liberties is a necessary condition for just legislation, but believing this doesn’t commit one to the analogy argument. I might believe, for example, that just legislation cannot emerge from unfair procedures, and procedures that don’t guarantee the fair value of political liberties are obviously unfair. Holding that fair procedures are a necessary, though not sufficient, condition for the creation of just legislation does not imply a commitment to the analogy. The second sentence, which David emphasizes, clearly does draw an analogy between democratic procedures and the original position, but again, it’s not clear that Rawls is thereby proposing the analogy argument as it is set out by David. Rawls might just as easily be making a different point, which is that certain values (like freedom, equality, and fairness) are modelled in the original position. We might want our democratic procedures to mirror the original position not because we think this means they will track the truth about justice in the same way, but rather because we think this is a way of instantiating those same values in our real political practices. By mimicking the original position, we instantiate the value of fairness in our political procedures, and that’s a good thing even if we don’t think the procedures will track justice in the same way. I’m not sure which way to read Rawls, but I thought it was worth pointing out that the passage quoted by David is open to a different interpretation, and also draws our attention to a very different sort of democracy/contractualism analogy.
Second, consider something that David says towards the end of the chapter: ‘to avoid the primary question, voters should somehow vote against proposals on the basis of burdens to themselves, though qualified by some reasonable accommodation of others, and in a way that supports a strong tendency for reasonably rejectable alternatives to be electorally rejected. I can see no way of making this work’ (p. 254). As David says, one of the difficulties is that in democracies citizens and legislators vote for proposals, not against them, and it’s hard to see how the reasonable rejection idea could be modified to fit this feature of actual democratic practice.
I want to propose a way it might be made to work, that is, a way real democracies could try and ensure that reasonably rejectable alternatives would tend to be electorally rejected, without having citizens address the primary question. Suppose the primary question for democracy is: what would be the optimal policy on this issue for our society, where optimal is defined as the most preferred option of a majority or plurality of voters (citizens or legislators) out of those options that cannot be reasonably rejected by anyone? Now suppose that voters do not address themselves to this primary question but rather to the following subsidiary question: which proposal do I most favour out of the options that cannot be reasonably rejected by anyone? Voters haven’t asked themselves the primary question since they don’t ask what would be most preferred by the largest number of voters – they only ask what they themselves prefer. But voters still need to make this choice only from those policies that cannot be reasonably rejected by anyone. Voters determine whether a proposal could be reasonably rejected by anyone by asking themselves whether, for each option, there is a person whose objection is weightier than any objection to any alternative proposal.
This model provides a way of for democracy to mimic the reasonable veto without actually giving a veto to anyone, and also offers a way to mimic the idea that contractors/voters should be reasonably self-serving: they should pursue the proposals they most favour while taking some account of the reasonable interests of others, and without asking anyone to address the primary question. So in key respects this model of democracy does look like a pretty good analogue to the contractual choice scenario. There is, however, an obvious difficulty, which is that my model requires voters to consider whether anyone has a reasonable objection against a given proposal, and not merely whether they themselves have such an objection. This might seem problematic for two reasons.
First, it deviates from the idea that parties in the contract device cannot reject proposals because other people have reasonable grounds for doing so – contractors are not supposed to join in other contractors’ reasonable rejections (David calls this rejection joining). I’m not sure, however, that the ‘no rejection joining’ rule is a necessary feature of contractualism. David claims that allowing rejection joining would be ‘artificial’ and that there is ‘no independent rationale’ for it (p. 250). Even if these claims are true, rejection joining does not seem inconsistent with the basic spirit of contractualism since the grounds for reasonable rejection are still based on some individual’s personal reasons. Because democracy does not have to be perfectly analogous to contractualism in order for the analogy argument to work, I don’t think this is a fatal objection to my proposal.
The more serious objection is that, by asking citizens to consider whether anyone has a reasonable objection to a proposed rule, we seem to be asking citizens to consider the primary question. I have tried to avoid this by claiming that the primary question in democracy is not, what rules could no one reasonable reject?, but rather, what rule would the majority or plurality prefer out of the options that are not reasonably rejectable? While this does, perhaps, ensure that citizens in my proposal are not asking the primary question, it still marks a major disanalogy between contractualism and democracy, since I am conceding that the primary question for each of these two projects is different. I think this draws attention to another problem for the analogy argument, one not explicitly mentioned by David, which is that it seems implausible that democratic decision-making should be addressed to the same question as Scanlon’s contractualism. I don’t think democracies should be deciding which policies are morally wrong or unjust (i.e. reasonably rejectable by someone). Those policies (at least the clearly unjust ones) should be beyond the pale in a well-functioning liberal democracy. Democracies don’t decide, for instance, whether or not persecuting a minority religion, or denying women the right to vote are reasonably rejectable rules. Those questions are already taken care of at a constitutional stage. Democracies, I think, are more plausibly seen as addressing the question: of the options that are not obviously unjust (i.e. those that don’t violate anyone’s basic rights or entitlements) which do we think will be best in terms of efficiency, or equality, or priority to the worst-off etc…? So, to me, a fundamental problem with the analogy argument is that I don’t think democracies are supposed to be addressing the same questions as contractualist theories, and thus any attempted analogy model (including the one I’ve proposed above) is bound to run into difficulties.
My third and final comment has to do with David’s proposed interpretation of what it means for someone to have grounds for reasonable rejection. David, recall, claims that a given contractor must decide ‘whether the alternatives to the option they would veto are subject to objections as weighty as theirs. If they all are, then a veto would be unreasonable and so they would refrain from exercising it’ (p. 251). Only if you have an objection weightier than anyone else’s objection to any alternative can this ground a reasonable veto. The difficulty with this proposed interpretation of reasonable rejection is that it appears to deliver the wrong answer in various cases involving the aggregation of burdens and benefits. Here I draw on an argument made by Derek Parfit in a Ratio article commenting on Scanlon’s book, ‘Justifiability to Each Person’. Parfit argues that Scanlon should drop what Parfit calls Scanlon’s
Individualist Restriction: In rejecting some moral principle, we must appeal to this principle’s implications only for ourselves, or for any other single person.
As Parfit quotes Scanlon: ‘the justifiability of a moral principle depends only on individuals’ reasons for objecting to that principle and alternatives to it’.
Parfit makes a pretty plausible argument that this individualist restriction prevents Scanlon’s contractualism from getting the right answer in some cases where we must decide how to distribute benefits and burdens. Parfit’s argument is, of course, complicated, and I won’t try and summarize the whole thing here – I’ll just use one of his examples to illustrate the problem. Suppose we are faced with a choice between saving five people or saving one person (called White). Most people believe that in such cases, other things being equal, we ought to save the five. But contractualism has a notoriously hard time explaining how this can be true. This is because White looks to have valid grounds for reasonably rejecting a rule directing us to save the five. If we follow a rule that says, ‘save the greater number’ this imposes a burden of certain death on White. If, however, we follow a rule that says ‘give each person the greatest equal chance of being saved’ (e.g. by tossing a coin to decide which way to go) we do not impose a burden of certain death on anyone – we only impose a 50% chance of death on everyone. White thus has a weightier complaint against saving the greater number than any other single person could have against a rule proposing greatest equal chances (note this follows from David’s account of what it would mean to have reasonable grounds for rejection). But Parfit claims, plausibly, this conclusion cannot be right – we ought to save the greater number.
The individualist restriction, Parfit claims, prevents us from taking into account the kind of aggregative reasons that ought to sometimes figure in any plausible theory of moral wrongness. Scanlon worries that if we drop the individualist restriction, then contractualism would no longer represent a distinct alternative to utilitarianism, but Parfit disagrees. He argues that utilitarianism fails to take account of distributive reasons (like equality, or priority to the worst off) and thus a version of Scanlon’s contractualism that dropped the individualist restriction, but retained the idea of justifiability to each person, would represent a distinct and compelling alternative to utilitarianism.
As I said, I’m not certain what I think about this, but I’m curious to know how (or if) David thinks his account of reasonable rejection can be squared with our intuition that we ought to save the greater number.