Should We be Utopophobes about Democracy in Particular?

The Justice Particle

Imagine that we knew, with scientific certainty, what justice requires. Imagine, for example, that the scientific boffins at CERN, while looking for their ‘God particle’, actually found a ‘justice particle’ which contained everything we needed to know about the justice of institutions and interpersonal relationships, and made such information so readily available and easily verifiable that reasonable disagreement about justice ended. Now, you might think this little story about CERN and a justice particle is a bit silly. But the point of it is this – imagine a world in which we knew what justice requires, a world without reasonable disagreement about justice. What could we say about the value of democracy in such a world? Focusing on this question brings out different ways of understanding the relationship between justice ³ and democracy and, consequently, what we think the value of democracy is in a world like ours and what sort of value it is thought to be or instantiate. For example, democracy might be thought to be part of justice – one of the things we are justly owed is an equal say in the laws that govern us (call this Relationship 1). In which case, that is just the sort of information we should expect a justice particle to contain. Alternatively, we might think of justice as the output of democracy – justice is what democracy chooses (Relationship 2). In which case, we would expect the justice particle to tell us how to choose laws, but not which laws to choose. A third way of thinking about the relationship between justice and democracy is to think of democracy as being or promoting or instantiating a rival value (such as legitimacy) to justice (Relationship 3). On this view, it is good when laws are just, and it is good when they are democratically chosen. On such a view the justice particle is a boon to working out what justice requires (and we may hope that with the new scientific findings fewer trade-offs will be required, since, armed with such information, people will democratically choose what is just). But democracy's value is not associated with its justness, or its ability to produce justice.

On any of these three understandings of the relationship between democracy and justice, little would change in terms of how we see democracy and its value in the world of the justice particle – knowing what is just would either confirm democracy's essential relationship to justice (either as required by justice or as the producer of justice) or it would continue to be, promote or instantiate some rival value to justice, which must still be taken into account. This is not the case for some other views, however, where justice and democracy are seen as different kinds of value, and where democracy is seen as a response to a problem – the problem being that we do not know what justice requires, or reasonably disagree about what it requires. Here are two variants of that kind of view.

On one such view, we should prefer just laws to democratically chosen ones. However, when we disagree about, or are unsure about, what justice requires, we should instead heed an alternative value – and that is either democracy, or some value (such as procedural fairness or self-determination) that democracy is thought to promote or instantiate (Relationship 4). One variant of this view (in which fair procedure is the value that we should seek when disagreement about justice arises) is what Estlund calls fairness as retreat (pp. 71–2). On this view, democracy retains (or is thought to promote) inherent value, but that value is never sufficient to outweigh justice. In a world where we knew what justice required, democracy would not be needed. Democracy only becomes a live issue when there is (perhaps only reasonable) disagreement about what justice requires.

On a different view, where democracy is still, nevertheless, seen as a response to a problem, democracy has purely instrumental value. The only thing that has inherent value, so far as laws are concerned, is justice. We should have just laws. We only have democratic procedures, on this view, because we believe that, in the absence of knowledge of, or in the presence of disagreement about, what justice requires, democracy will get us closest to the truth of the matter. Democracy therefore has an epistemic function and an instrumentalist rationale – it is a solution to an epistemic problem (Relationship 5).

For the adherents of Relationships 4 and 5, we will see democracy very differently in the brave new world of the justice particle. Adherents of both views see the case for democracy as contingent on the fact that we do not know what justice requires. If we know what justice requires, we no longer need democracy, ⁴ or at least, if we need it at all, we will need it in a very different, and limited, way. ⁵

Both of these views see democracy as a response to a problem – that there is reasonable disagreement about what justice requires. When that problem disappears, they see democracy's point (and possibly value) disappear with it. They are, however, different types of response to a problem: Relationship 4 sees democracy as an independent second best response to the problem – a second best to be sought when the first best is not available, but whose value is not to be judged by the standards of the first best; while Relationship 5 sees democracy as an approximation response to the problem – a second best to be judged against the standards of the unavailable first best. Justice is what we want, and democracy should be judged against its ability to provide it. Democracy therefore not only responds to the problem of the unavailable first best; it aims to provide an (at least partial) solution to that problem – it is a problem-solving concept.

Estlund on the Value of Democracy

David Estlund opens his book Democratic Authority by stating that ‘My goal is to show how a concern for the quality of political decisions, properly constrained by other principles, supports democratic arrangements’ (p. 1). Estlund's case for democracy is unabashedly epistemic. This epistemic justification for democracy means that Estlund should be seen as having at least one foot in the ‘Relationship 5’ camp. Justice and democracy are, for Estlund, importantly different kinds of concept. When someone asks ‘why be just?’, there is not really much to say – laws should ideally be just. ⁶ But when someone asks ‘why democracy?’, we need a 309-page book to tell us why, and the answer will focus on democracy's ability to produce justice.

Although this epistemic element is central to Estlund's case for democracy, alongside this element we must take seriously the ‘other principles’ that ‘properly constrain’ the selection of an epistemic procedure. First and foremost among these is Estlund's ‘qualified acceptability requirement’. This requirement, which draws heavily from John Rawls’ Liberal Principle of Legitimacy (Rawls, 1993, p. 137), states that legitimacy requires justification in terms that are acceptable to all qualified points of view (ch. 3). As this constraining principle shows, Estlund does not see democracy as justified purely by its epistemic credentials, and therefore does not see the relationship between justice and democracy purely through the lens of Relationship 5 outlined above.

Therefore, Estlund holds a hybrid view of the value of democracy – democracy is best at finding justice in a certain kind of way, one that respects the qualified acceptability requirement (QAR). Estlund believes that the QAR must itself be a moral truth (pp. 53–8), but does not specify what type of moral truth it is. If it is a truth about justice, then Estlund's view will be a hybrid between Relationships 1 and 5 – democracy is the best way of finding what justice demands in such a way that respects the one truth of justice that we are prepared to assert: the QAR. In some sense, then, democracy responds to a demand of justice, as well as being the finder of justice. If, on the other hand, Estlund holds the qualified acceptability requirement to be a non-justice moral truth, and one that is worth respecting even when we know what justice is, then his view will be a hybrid between Relationships 5 and 3 – democracy is the best way of finding out what justice demands in such a way that respects a non-justice value of justification. Finally, if Estlund believes that the QAR is a non-justice value that is lexically inferior to justice, then his view is a hybrid between Relationships 5 and 4 – democracy is the best way of finding what justice demands in such a way that respects the value of justification which only applies when we do not know what justice demands.

Whichever of these combinations best captures Estlund's understanding of the value of democracy, democracy would, on such an understanding, lose much of its appeal should the justice particle world come about. Since there will be no reasonable disagreement about justice then we can just shoot straight for justice, and that will be justifiable to all qualified points of view (I assume here that the findings of the justice particle meet the QAR). ⁷

We can see that, for Estlund, the case for democracy seems to arise out of a problem, and we can see this by the way that the case for democracy goes away when we imagine a world in which the problem did not exist. The problem seems to be this: we require common rules to live by; common rules should be just; we do not know what justice requires, and reasonably disagree about it; and exercises of power must be justifiable to all qualified points of view (either always, or when we do not know what justice requires, or reasonably disagree about what it requires). Therefore, for Estlund, thinking about democracy seems to start when we acknowledge two important and non-ideal facts about the reality of common life: we are going to need coercive structures because some people are going to need coercing, and in choosing rules to govern these structures, we are not ideal moral reasoners (and, if some of us are, we do not have publicly acceptable ways of identifying them). There are passages in the book where Estlund explicitly acknowledges that he understands democracy to be a response to a problem, and the nature of that problem. ⁸ Estlund's answer to the problem is an approximation response (subject to certain limitations) – it is a response to be judged by its ability to approximate our unavailable first best (justice), on the condition that it does so in a way that is justifiable to all.

Horses for Courses: Political Theory and Utopian Thinking

Thus far I have tried to show that, on Estlund's view, democracy should be seen as a response to a problem. Moreover, since it appears to be a response to a problem that is judged against the standards of an (unavailable) first best, it should be seen as a problem-solving concept (i.e., an approximation response). Estlund's vision of democracy is within the deliberative tradition, making use of a highly idealised deliberative situation, which should be approximated in formal political deliberation. In this section, I will briefly outline Estlund's attack on ‘utopophobia’ before explaining some reasons to be sceptical of utopian understandings of problem-solving concepts.

Estlund's discussion of realism and utopianism in political theory is different to most discussions of these issues in that it is ostensibly a discussion of utopianism and democratic theory, rather than arising out of worries about justice or related concepts. However, in common with those discussions, it seems to begin as a discussion about the appropriate level of abstraction at which to theorise about a particular concept (in this case democracy), and yet quickly becomes one about the appropriate level of abstraction (or how abstract we can go) in doing normative political theory in general. Thus, Estlund quickly progresses from a question about how certain charges ‘bear on normative political theorizing about democracy’ to a defence of ‘an approach to normative political theory’ (p. 259, emphases added). This approach is clearly supposed to be one to normative political theory in general and Estlund describes it as follows: ‘Rawls speaks of a “realistic utopianism”. I prefer to speak, less eloquently, of the noncomplacent nonutopian range of normative political theories, the range in which most theorists would agree normative political theory should toil’ (p. 264).

Estlund defends an approach he calls aspirational theory (p. 259). Aspirational theory is to be found between the extremes of objectionable utopianism and complacent realism, and is an approach that ‘posits standards that are not generally met (as any noncomplacent normative theory seems bound to do), though they are possible to meet’ (p. 259). Estlund basically identifies two potential objections to a posited theory, standard or understanding of a concept – ‘people can't do that’, and ‘people won't do that’ – and argues that the first is a legitimate complaint, while the second is not. ⁹ It does not matter for Estlund whether the standards posited will never be met, or if the reasons they will never be met stem from ‘human nature’ – there are certain things we know human beings will not do, but that does not matter for ideal political theory. ¹⁰

This all sounds very sensible – ‘ought implies can’ is widely accepted to be true, under some understanding of ought and some understanding of can, and nobody thinks ‘ought implies will’. However, despite this prima facie plausibility, I am not one of the ‘most theorists’ who agree with Estlund that the proper site of labour for the political theorist lies solely between the inclusion of can'ts and exclusion of won'ts (p. 264). Sometimes can'ts should not count, and sometimes won'ts should.

I hold this muddle of views because I do not think that there is one single answer to how utopian or how realist we should be. I do not think that the utopian/realism debate should be conducted in terms of how utopian or realist ‘normative political theory’ should be. This is just too broad a question to ask. In many ways, this is obvious – if we are answering a question about how I should vote in a referendum, we would need to draw on lots of different information: about what the ideal would be, how the available options stack up against that and how other people are likely to vote. Estlund recognises this, calling this concessive political theory (p. 268). But I am not seeking here to make this obvious point – that applied normative theory need be more, well, applied than ideal theory. There is obviously no one level of abstraction at which political theorists should operate. But I actually think that how ideal ideal theory can get cannot be answered with reference to political theory in general.

I think how ideal or realist – that is, what facts we should allow to constrain our theorising – an ideal theory should be is entirely dependent on what it is an ideal theory of. Different concepts will have different restrictions on idealisations because they are different concepts. In order to work out the appropriate level of abstraction to theorise at, we must first decide what kind of claims we want to make, what kind of concept it is we are investigating, and what the job of that concept is taken to be (including what relationships it has with other concepts). To illustrate, allow me to introduce three different types of concept, and explain how they will differ on restrictions of the ‘people can't’ and ‘people won't’ type.

First, consider what I will call (following Parfit, 1997) ‘telic’ concepts. A telic concept is one whose job is to allow us to evaluate states of affairs. Telic concepts need not be all-things-considered ones – we can simultaneously affirm a plurality of telic values. The claim made when affirming a telic concept is that the world (or, indeed, the universe) is going in one way better when that concept is realised or promoted. Consider fairness, which I think of as a telic concept. I think things are, all other things equal, better when they are fair or, if not completely fair, fairer than the alternatives. In addition, I think that brute luck equality (with certain modifications) is a good understanding of the value of fairness. My claim here is that neither ‘people can't’ nor ‘people won't’ should count as an objection to a theory of fairness. For example, brute luck equality is impossible to achieve precisely, and that seems no block to me on thinking that it is a good understanding of fairness. We need concepts that allow and enable us to evaluate the world – and that includes ourselves and our capabilities. We need concepts that allow us to regret what we cannot do – in other words, to regret the facts of the world (see Cohen, 2008, chs 6, 7; Tomlin, 2010, pp. 240–6). Therefore, our telic concepts should not be constrained by the facts of the world – either facts about human capability (whether or not people can) or about human motivation (whether or not people will). Indeed, such theory should not be constrained by facts of nature either. If a hurricane comes along and kills 100 people, and there was nothing people could have done, we should nevertheless have concepts that allow us to regret the hurricane – a fact of nature – and what it did, even though the facts of human capability mean that nobody should be judged negatively for failing to stop it.

It might be objected (and, indeed, it has been) (for example, Miller, 2011) that this reduces political philosophy to lament, rather than action guidance. There are three things to say to this. The first is that recognising a role for lament in political philosophy (and thus for consideration of telic values) is not to reduce it to that. ¹¹ The second is that while telic values are primarily evaluative, they are not solely so. If the world is in one way better when a value is realised or promoted, that means there is a reason (albeit a pro tanto one) to promote that value (possibly constrained by the extent to which one can). ¹² So considering telic concepts does not pull us away from action guidance. Third, lamentation can itself be action guiding. It is a false dichotomy to set up a contrast between lamentation and action guidingness. ¹³ For example, if there is some unfairness between me and my fellow citizens that I recognise but either cannot (or should not, due to other values) do anything about, recognition of that unfairness does not direct me to reduce that unfairness, but it can nevertheless direct me in the way that I present myself and interact with the other citizens in the public sphere.

The second type of concept (again, following Parfit, 1997) is deontic concepts. Deontic concepts are more directly normative. The idea of a deontic concept is to say ‘you ought to ϕ’. Some people think of justice, or rightness, as this type of concept. Now, if ought implies can, and deontic concepts consist of ought claims, then ‘people can't’ objections will rule out a posited understanding of such a concept. Because people can and do do wrong actions, the fact that they won't do something does not automatically disbar a posited right action from being the right action.

Finally, we have problem-solving concepts. ¹⁴ Problem-solving concepts are tools to be used to achieve some independently valuable state of affairs (that state of affairs may be defined in deontic terms – i.e., everyone does as he or she should). Such concepts have only instrumental value – their value derives from their ability to secure other values (although there may be important moral limits on how they solve problems and secure other values). I think these are the concepts that should be defined in the most ‘realist’ terms. In political life, thinking about a solution to a problem involves thinking about the nature of the problem (which must be defined in telic or deontic terms) and the beings and societies that have them. As such, we are already deep into the realities of life, and far from idealism, before we even begin to theorise about the concept. In addition, solutions to problems that posit unrealistic solutions do not seem to be very good solutions – they do not seem to be taking the role of problem solving very seriously. When we want to give people a tool to help them solve a problem, we need to take seriously their abilities and inclinations – not only whether they can use the tool, but how they will actually use it. It is no use giving them a tool that we know will be misused, and will therefore fail to solve the problem.

Consider a group of soldiers who need to cross a river. One of them happens to know a political philosopher, and calls her for advice. If the political philosopher says ‘the best solution is for you all to fly over the river’ we might surmise that she was not much good at thinking through solutions to problems. This example seems to serve as a good reason to allow ‘people can't’ objections to be decisive against posited solutions to problems. Indeed, the range of theorising should be even narrower – it is not just that ‘people can't’ should rule out a proposed solution, but that the specific people who are having the problem can't should also rule out a proposed solution, even if, in general, people can. Thus, imagine that the people who have some problem are all wheelchair users – no proposed solution should contain an injunction to walk, even though people in general can walk.

A particularly bad form of proposed solution to a problem is one that contains what I will call a circular recommendation. A circular recommendation is a recommendation made as part of a solution to a problem that assumes away the nature of the problem we are trying to solve. For example, imagine our problem is that our car (which is the only one we have access to) has run out of fuel. Any posited solution that involves driving the car anywhere (‘;there's a petrol station on the A44, you could drive there’) would be a circular recommendation. If we could do that, we would not have this problem. And if we could do that, we would not pursue this solution, we would pursue the (currently unavailable) first best. Approximation responses have no value if the first best is on the table.

So far I have spoken of ‘people can't’ (or ‘these people can't’) objections. But Estlund is happy to say that these objections should be decisive in political theory (I actually disagree with this as a blanket statement, as my discussion of telic concepts makes clear). But what of objections of the ‘people won't’ variety? It is these that Estlund wants to defend political theory against having to take into account. I am not sure if some ‘people won't’ objections should not count for problem-solving concepts. For example, some ‘people won't’ objections may show up circular recommendations. Consider our band of soldiers who need to cross a river. Imagine that the problem arises because they won't jump across because they're too frightened (even though they could). Any solution that involves jumping across the river, or some other, similarly sized, gap will not be a good solution – because we know that they won't. If they would do that, we would not have the problem in the first place. Thus, some ‘people won't’ objections look like they might be relevant. Even if they do not disbar jumping from being considered a ‘solution’ they seem to count against it as a good solution, or at least as a useful one – and solutions, qua solutions, surely need to be judged according to their usefulness.

Whether or not a recommendation is circular or not may not be a binary classification. It may be that some recommendations are not strictly circular, but given the nature of the problem it comes close to circular to suggest certain solutions. Imagine that our problem stems from the fact that the men will not swim through crocodile-infested waters. A solution that recommends swimming through shark-infested waters instead will not be strictly circular, but we can see that it approaches circularity – if they won't do one, they almost certainly won't do the other.

Circular recommendations look to be poor solutions to problems, and thus certain ‘people won't’ objections should count against understandings of problem-solving concepts. However, I wonder if ‘people won't’ objections are not always objections to a posited understanding of a problem-solving concept. When people ask us for advice, we generally try to take that person's character into account. If my partner asks me for advice on how to solve a problem, I will not advise her to do something I know that she definitely won't (unless I think that the act in question is morally required – in which case, my advice is based not on problem-solving considerations alone, but on deontic considerations as well. In the case of Estlund's democratic theory, the QAR provides a direct moral imperative, and therefore that part of the theory is arguably not subject to the same realism constraints as the problem-solving part. Moral imperatives provide outer limits for acceptable solutions). But given a range of solutions that are morally permissible, it just does not seem as if you have understood the role of problem solver to posit something that in actual fact you know will never happen. It does not look as if you have thought about your audience – the person who has the problem and requires guidance. When we try to solve problems (for ourselves or others) we must take seriously the kind of thing it is that has the problem.

I certainly think there is an especially strong case for taking ‘people won't’ objections into account when we are trying to solve problems for groups. As we have seen, ‘people can't’ objections certainly count against understandings of problem-solving concepts. When we try to solve problems faced by groups, I claim, won't-based objections can rise to the level of can't-based objections. This is because individual won'ts lead (in some sense) to collective can'ts. To explain: consider our troops who need to cross the river. Imagine that we find a solution that involves all of them performing a particular task – there are 100 of them, and we know that all 100 must do something in order for the solution to work. However, we know that, due to human nature, at least three people in any given 100 won't perform the necessary task. It seems to me legitimate for the soldier talking to the political philosopher to say ‘we know that some of us (i.e., the three) won't, so we can't’. The individual won'ts of the three lead, in an important sense, to a collective can't for the 100. And can'ts defeat proposed solutions.

All this is relevant to our discussion of democracy because democracy can be seen, as I have argued Estlund sees it, as a solution to a problem. I have argued that theories of concepts that receive their value as solutions to problems should be significantly more embedded in social and physical reality than evaluative/telic and first-order normative/deontic concepts. Since on this kind of view democracy has no inherent value, then it seems strange to say that an idealistic picture of democracy paints things as they should be – democracy is not a picture of how things should be, things should be just. Democracy's value is to be found in helping us get there. Therefore, it would seem sensible to take into account what we are like. In addition, democracy is a solution directed toward a collective and in such solutions individual won'ts that become collective can'ts might be taken to defeat proposed solutions.

When thinking about democracy conceived of as a problem-solving concept, we begin our theorising deep within non-ideal theory – we are not to assume strict compliance, for it is the absence of strict compliance that generates, in part, the need for coercive institutions in the first place. And we are not to assume ideal reasoning capacities for, again, those form part of the problem. Since the problem arises from certain aspects of human nature, any solution would need to take human nature into account in a way that theories of justice or fairness may not need to. Any theory of democracy that does not take human nature seriously may run the risk of providing circular recommendations, and circular recommendations are not good solutions to problems.

What Estlund gives us is, to use his own metaphor, a beautiful village some miles off that looks great but that we seriously doubt we can or will get to (pp. 269–70). I am not sure this has done us much good, because we started out looking at a beautiful village we doubted we would get to – Justice – and all we have been given is a different one. Imagine we want to set out from our village, Reality, and make it to a village on the hill – Justice. We ask a political philosopher to help us get there. If she comes back with a new, far-off, village called Ideal Democracy, which she admits is OK, but not as good as Justice, has she helped us with our problem? Taking our eyes off that beautiful village and pointing us to a slightly closer but inferior one does not seem worth it, for two reasons: we have not been given any way to solve our problem that takes seriously what we are like, we have just replaced one ideal with another; and by taking our eyes off Justice we might lose sight of that most ultimate prize, in search of the acknowledged inferior, which only drew value from its ability to approximate Justice. If we want ideals to dream of – we have one. If we want solutions, another ideal is not going to cut it.

So instead of building an idealised vision of democracy, which becomes a standard for living up to more ultimate standards, why do we not take those external, ultimate standards (justice and justifiability) and build a theory of democracy that is a real usable tool for the people who exist here and now to use in approximating them? Cut out the middle man.

I have not had the space here to say anything about Estlund's particular version of deliberative democracy – and it is less ‘utopian’ than many other versions. But he does want to defend unrealistic conceptions of democracy by making a defence of unrealistic political theory. My contention is that we should look to the particular nature of a political concept, and its relationship to other concepts, before deciding whether or not we can defend unrealistic conceptions of it. And if democracy is a solution to a problem, then there may be strong reasons to begin our theorising about democracy from the perspective of the people and societies that have that problem.