Much research demonstrates that voters are largely ignorant about the issues and institutions at stake when they vote. Some research also suggests that voters are selfish and irrational. Such poor voter performance might be thought to pose problems for an epistemic theory of democracy since if the performance is bad enough democracy may not deliver on its epistemic promise – it might fail to outperform choosing policies ‘with a roulette wheel’ (262). Estlund has two answers to the challenge posed by the reports about voters’ poor performance. The main answer consists in showing that there is in general value in theories that demand of people that they behave better than they in fact behave or are ever likely to behave.
But Estlund also offers another answer. While admitting that this is not his focus in the chapter, he offers reasons why we should not concede that voters are inevitably poor decision makers. First, it is unclear just how selfish or irrational voters really are. Second, even if voters are pretty ignorant, as I take it Estlund accepts, it is unclear if this translates into poor decision making. Parents, he notes, lack all sorts of information that would seem necessary to reach good decisions about the life of their children (for example, they either have been shown to be ignorant about or are likely to be ignorant about basic nutrition, basic first aid, the value of a college degree in promoting later happiness and success etc.), and yet we would not conclude that parents cannot make good decisions for their children or that parenting is not good for children. The final reason why we should not concede that voters are inevitably poor decision makers is that even if they are poor decision makers at present, it might not be too difficult to rectify this.
If we assume, however, that voters do make poor decisions and that this is unlikely to change, we can turn to the main argument of the chapter, which is Estlund’s general defence of theories that demand more of people (or institutions) than they are ever likely to deliver. Estlund begins his defence by categorising normative theories depending on their realism. He observes that all good normative theory must lie between two extremes of too much and too little realism. So, on the one hand, any theory must avoid complacent realism and, on the other hand, it must avoid utopianism. A theory is complacently realistic if it recommends no change from the status quo. According to Estlund ‘any theory that implies criticism of actual institutions or behavior is not as realistic as it could be’ – it is not ‘entirely realistic’ (263). At the other extreme lie theories that are factually or morally utopian. A theory is factually utopian when it ignores (relevant) non-moral facts about the world; such theories are straightforwardly false. A theory is morally utopian when it sets impossible requirements for people or institutions. Morally utopian theories are false because moral standards that are impossible for people to meet are false. Theories that lie between these two extremes of complacent realism and utopianism lie within the ‘noncomplacent nonutopian range’ (264) and Estlund adds that most political theorists would agree that this is the right range for normative political theory to occupy.
Noncomplacent nonutopian theories further divide into hopeful and hopeless theories. Since they are not utopian they only require what is possible of people. But, in contrast to hopeful theories, hopeless theories hold individuals (or institutions) to standards about which there is good reason to believe that they will never be met. Thus, for example, given what we know about people’s fallibility, a theory requiring everyone to be virtuous would be a hopeless theory since there is no chance of everyone being virtuous.
Estlund offers one more categorisation: he distinguishes between aspirational and concessive theory. A theory is aspirational when it posits standards that are currently not met but could be met. A theory is concessive if it concedes facts about how people and institutions are likely to act and guides action on the basis of this concession. If I understood him correctly, all hopeless theories are aspirational: they concede nothing and this is why they remain hopeless. Hopeful theories, on the other hand, could be aspirational or concessive (or, it seems to me, both). I gather that a hopeful theory would be aspirational, rather than concessive, if it so happened that people were likely to act as, ideally, they should. If this is correct, then the distinction between hopeful and hopeless theory concerns what is and what is not likely, while the distinction between aspirational and concessive theory concerns whether any adjustments are made to what a theory recommends for the sole purpose of increasing the likelihood of the theory’s requirements being met: aspirational theory does not make such adjustments while concessive theory does.
Estlund considers his epistemic theory of democracy to be aspirational. He defends aspirational theory by defending it even when it takes on the form of hopeless theory. Hopeless theory can be attacked on many fronts. First, Estlund defends it from those who would want to classify it as utopian. Hopeless theory would collapse into moral utopia if we could not distinguish between, on the one hand, requiring people to do what there is good reason to believe they will never do (hopeless theory) and, on the other hand, requiring people to do the impossible (moral utopia). According to the challenge Estlund considers, ‘impossible’ simply means ‘zero probability’ and so requirements that won’t be fulfilled should be classified as impossible. But, according to Estlund, such a definitional move would commit us to an implausible claim, namely that doing something might be impossible even though it would be easy to do it. He uses the following example to illustrate this point: the probability that two strangers at a café will move their hands and arms the same way for the next five minutes has probability close to zero. So, if we granted the challenge we would have to say that such synchronised movement is nearly impossible. However, it cannot be nearly impossible since it would in fact be easy for one person to intentionally copy the arms and hands movements of another. At the same time, while easy to do, the probability that it would happen remains close to zero since it would be very embarrassing for one person to copy another’s movements like that. If so, then a close to zero probability of X does not amount to the near impossibility of X and, therefore, hopeless theories do not collapse into morally utopian ones.
Second, hopeless theory can be attacked on account of what I take to be the quality of its recommendations (although note that Estlund deals with the challenges in a different order from the one I list them in here). For one, we might worry that a hopeless theory will be dangerous: by hypothesis it is likely to inspire only partial compliance with its requirements but compliance that is only partial might be worse than no change at all. A related worry is that hopeless theories simply fail to recommend any actions since what they recommend is so unlikely. Estlund’s answer here is that hopeless theories do recommend actions that are collectively prescriptive: they tell ‘all people, together, to behave differently.’ Making what one person is required to do conditional on how others behave is, as he points out, utterly familiar and does not make the requirement any less practical: ‘a moral requirement to help push a friend’s car out of a snowdrift does not say to push futilely even when no one else helps…’ (266). This is a practical requirement even if we can anticipate that no one else would try to push the car out of the snow. I take it that Estlund’s answer here (to the worry that hopeless theory offers no recommendations) also deals with the worry about hopeless theories being dangerous, since if recommendations were to be followed collectively (i.e. as they ought to be followed) they would not be dangerous. Estlund makes two further points in the chapter that bear on how we should assess the quality of recommendations of hopeless theories. First, Estlund admits that institutional recommendations of hopeless (and aspirational) theories are likely to be imprecise since such theories assume that people will act very differently to how they actually act. Since such theories cannot straightforwardly extrapolate from data on current behaviour they require ‘principled reticence about institutional recommendations’ (271). Second, it is worth noting that Estlund obviously only sets out to defend hopeless theories that are morally correct: he admits that one way in which hopeless theory might be morally incorrect is if it is unreasonably harsh – demanding of agents more than can reasonably be demanded – but he does not see this as an inevitable feature of hopeless theories.
But some might think that even if correct hopeless theories avoid the problems mentioned above, at the very least we should see them as worse than hopeful theories. We should see them as worse because their recommendations will simply not be followed. Estlund rejects this challenge as well. Hopeless theories are not worse than hopeful ones because the answer to what people should do does not depend on what they are likely to do (even though, as already mentioned, it does depend on what it is possible for them to do). As he puts it: ‘It is not the case that ought implies reasonably likely’ (265). Basically, I think that Estlund is suggesting here that we should keep two questions apart (although note that he does not put his point this way): (1) what people should do in practice given what others are likely to do (establishing this is the task of concessive theory) and (2) what people should collectively do (establishing this is the task of aspirational theory). Each of these questions is worth answering and since the correct answer to question no. 2 might turn out to be hopelessly unrealistic we should not dismiss hopeless theories as inferior.
The difficulty, of course, is to show that the second question is worth answering even if there is good reason to believe that the recommendations given in answer to that question are unlikely to be followed (i.e. even if we find ourselves constructing a hopeless theory). Estlund’s strategy for showing that the question is worth answering is to point out the two roles hopeless theory can play. It has a philosophical role in that there is intrinsic value in good philosophical inquiry and in ‘seeking the philosophically most defensible account of what political arrangements should be like’ (269). It also has a causal role: it prevents us from expecting too little of ourselves. ‘…[I]f we are not sure that it is impossible [to get to a new place], then even if we are unlikely to get there, it could be worth thinking about how we might’ (270). Given this causal role of hopeless (and in general aspirational) theory we can identify a parallel weakness in concessive theory, which is that it can steer us away from a course of action simply because an easier one is available. The upshot of the discussion is that hopeless theory is necessary to answer certain normative questions and that it is also independently useful in keeping us on track for the best reform.
Having offered a defence of why it would not count as a defect for his theory of democracy to demand good voter performance even if such performance was very unlikely, Estlund also rejects the view that voting itself is not sensible. He acknowledges that if we think of voting in terms of what difference a single vote makes, voting may seem foolish. But he suggests that we think of voting instead in terms of each vote’s contribution to what voting can bring about. Specifically, thoughtful voting by a sufficient number of people can prevent disasters and, he adds, we all have moral reason to contribute to efforts to prevent disasters even if we are pretty sure that our contribution would not make any difference. He offers the following analogy here: ‘suppose that if enough of us arrayed ourselves at the edge of town at midnight on the first day of each month, we could deter a dangerous gang from moving in…’ Even if it is ‘virtually certain that my presence will not make or break the effort…there is a clear moral reason to join this effort.’ (272).
In the chapter, Estlund also develops a vocabulary that allows him to compare the strength of his argument for democracy in relation to those offered by fairness and correctness theories. An argument is stronger (in the sense of being more efficient) than another if it can establish the same conclusion as another argument but on the basis of weaker premises (i.e. premises that are satisfied in more possible worlds). More efficient arguments may have a rhetorical/polemical advantage, notes Estlund, if weaker premises are more acceptable to the relevant audience. An argument is also stronger (in the sense of yielding more) than another if it establishes a stronger conclusion from premises that are no stronger. Estlund notes that in comparison with his epistemic theory of democracy, fairness theories have weaker premises (which would make their arguments potentially more efficient and more acceptable) but they also yield less since they ‘yield very little by way of authority or legitimacy’ (274). Correctness theories on the other hand have stronger premises than Estlund’s theory and, if Estlund’s theory is correct, are inefficient. Correctness theories are therefore also too demanding in the specific sense of demanding more than is necessary to account for authority and legitimacy.
Estlund concludes the chapter by returning to the idea of model deliberation. In light of the argument of the chapter it is clear that it should not count against model deliberation that its standards would not be met in practice: to reject model deliberation on those grounds alone would amount to inappropriate utopophobia. It may be that the standards of model deliberation are what we morally should aspire to. However, since it is very unlikely that we could realistically mirror model deliberation in public political deliberation, we should concede this and not set it as our aim but rather develop a 2nd best solution.
I found myself agreeing with all the important claims of the chapter (except, perhaps, with the answer to the question ‘why vote?’). I thought that the defence of hopeless theory was excellent. So my comments are on tangential issues:
(1) I wondered if hopeless (as opposed to just aspirational) theories really can be said to have the causal role attributed to them? If we really have a good reason (a pretty conclusive reason?) to think that some recommendations will not be followed then does the worry over settling for too little really require us to construct a theory that would issue such recommendations? Why not construct a theory whose recommendations are unlikely but we just might meet them? But maybe the point Estlund is making is simply that we don’t always know whether a theory will turn out to be hopeless and, when in doubt, we should not abandon it too quickly?
(2) Like Jonathan in week one of the reading group, I was not certain why we should reject theories that demand the impossible. Estlund simply stated that we should. But let’s say (borrowing Jonathan’s example) that we all agree that it is impossible at present, and in the foreseeable future, to compensate people adequately for being born blind (when they see their blindness as a disadvantage). Putting aside all the other well-known worries about such attempts to offer compensation, I do not see why a theory demanding compensation should be dismissed simply because it is impossible to bring about what it advocates. Such a theory could still, for instance, measure the distance between our world and a world in which no one could complain that their lives go much worse than those of others through no fault of their own. Is Estlund assuming that the purpose of normative theory exhausts itself in guiding or evaluating action rather than, for example, evaluating states of affairs? In any case, I also wondered how extensive was the set of utopian theories: would it consist of anything other than theories that demand what is conceptually impossible? All other theories, after all, could claim that while there is no good reason to think that what they demand is ever likely to be met, it is not impossible if only we invest enough in R&D.
(3) This is just a small point in relation to how Estlund draws the boundaries of the set of entirely realistic theories. Can a theory that advocates change really never be entirely realistic? It seems to me that there is a range of change that is just as realistic as the status quo if we know exactly how to bring it about. This could be the case if the change that we are aiming at falls within some approximate range and requires only a sub-set of all the people who could bring it about to comply. To use a crude example, we could propose an entirely realistic change of reducing the number of people who use a given short cut at night if we informed them that it was extremely dangerous.
(4) I found that the move to distinguish between what is impossible and what will not happen (but is possible) in terms of whether it is easy ingenious. But I thought that the claim that it would be easy to copy the other person’s arms and hands movements was controversial since it would also be very difficult to do so given how embarrassing we have to assume it to be (so that the probability of it happening remains close to zero). It seems that we are asked to assess the easiness of a given action while assuming that the relevant person is already motivated to carry it out. But if this is so then how do we assess if achieving some motivational state is very unlikely but not impossible?