1. Introduction
There is an important debate in moral epistemology about whether there is something morally troubling about deferring to someone else on a moral question. A standard case brought up in this context is
vegetarianism. Aaron has never seriously thought about the moral status of animals. Aaron knows that Birke is very reliable on moral questions. There is no difference between them with respect to non-moral information surrounding the relevant issues – both are equally well-informed about (say) the meat industry and the conditions of factory animals. Aaron asks Birke whether it is permissible to eat meat; Birke tells him that it is not. On this basis alone, Aaron stops eating meat.Footnote 1
To many writers, it has seemed that, even if Aaron made the right choice in becoming a vegetarian, there is something morally troubling in how he became one. According to these pessimists, it is in some way morally deficient to defer to others for one's moral views (e.g., Hopkins Reference Hopkins2007; Hills Reference Hills2009; McGrath Reference McGrath2009; Crisp Reference Crisp2014).Footnote 2 A more precise statement of this view is
pessimism about moral deference. In most cases, if one acts on the basis of beliefs concerning a moral subject matter, where these beliefs are only sustained on the basis of deference to someone else, then one's action is morally deficient in this respect.
Pessimism does not claim that moral deference is always all-things-considered deficient. Given the moral badness of eating factory-farmed animals, we might think it all-things-considered best if Aaron deferred to Birke's opinion. But, the pessimist claims, insofar as something is an act of deference, it is almost always morally problematic. (I will explain the qualifier “almost” in the next section.) Optimism is the denial of pessimism (e.g., Jones Reference Jones1999; Sliwa Reference Sliwa2012; Enoch Reference Enoch2014; Groll and Decker Reference Groll and Decker2014; Lillehammer Reference Lillehammer2014; Wiland Reference Wiland2017; McShane Reference McShane2018). We can distinguish two forms: according to strong optimism, moral deference is (almost) never seriously deficient, at least not qua being an act of moral deference. On the other hand, weak optimism claims that in most cases deference is not seriously morally deficient, at least not qua being moral deference.
In this paper, I am interested in the debate between pessimists and optimists in a specific area, democratic politics. There has so far been little debate of this more specific question, with one notable exception (van Wietmarschen Reference van Wietmarschen2018). But the phenomenon of political deference is an every-day one, and how we should judge it far from obvious: If a friend tells you that Brown is the best candidate to vote for, is there something morally deficient if you vote for Brown, on the basis of that advice alone? If an economic expert tells you that ending austerity is best for Greece, would you be liable to moral criticism if you publicly criticised Greece's debtors, merely because you believed the expert? If you hear a black football player state that the NFL is racist, would it be problematic if you boycotted the NFL, on the basis of that testimony alone? My answer to these questions will generally be “no”, as I will defend weak optimism about non-ideal deference on normative political issues.
This needs some explanation. I have no interest in defending deference on moral questions in general. There might be interesting continuities between the moral and political case, but I will set those aside. More importantly, my primary focus will be on non-ideal cases of deference. This does not refer to morally inferior cases of deference, but to cases of deference which happen under less heavily idealised, reasonably realistic circumstances. The epistemological literature standardly invokes highly idealised cases like vegetarianism. I will argue that even a cursory look at empirical research from political psychology suggests that real-world political deference almost never looks anything like these cases.
To preview the contents of the paper, I argue that (i) almost all non-ideal cases of political deference are impure (in a sense to be defined), and (ii) optimism is the more convincing view for impure cases (§3). Second, I offer a principled defence of the role of deference in democratic politics (§5). Democratic politics is epistemically complex, and deference is a justified part of the epistemic division of labour. Furthermore, by deferring to others with whom we share certain identities or interests, we exercise epistemic solidarity with them. These sections are flanked by an outline of the optimism – pessimism debate (§2), a defence of my methodological approach (§4), and answers to some objections (§6).
2. Pessimism and optimism about moral deference
We can start by describing the major ingredients in moral pessimism, as defined above.
“In most cases”
First, pessimism does not claim that it is always morally troubling if one defers to others.Footnote 3 A standardly given counterexample is young children deferring to their parents (Howell Reference Howell2014: 390; Mogensen Reference Mogensen2017: 263–4). McGrath concedes cases to the pessimist in which “I am too angry, too drunk, or too personally involved to think clearly” (McGrath Reference McGrath2009: 323). Another exception is people who do not have the time or the mental capacities to deal with a complicated moral issue (Jones Reference Jones1999: 57; Nickel Reference Nickel2001: 265; Hills Reference Hills2009: 123–4; McGrath Reference McGrath2009: 332; Howell Reference Howell2014: 390; Mogensen Reference Mogensen2017: 263). Several pessimists suggest that we should defer to those who are more knowledgeable in such cases.
“Morally deficient”
Second, pessimism claims that there is something morally deficient if one defers to others. An alternative form of pessimism claims that deference is epistemically deficient.Footnote 4 Perhaps, for example, testimony on moral issues does not transmit knowledge. However, my focus will be on moral forms of optimism and pessimism, for two reasons. First, epistemic forms of pessimism are less philosophically convincing (Hopkins Reference Hopkins2007). Second, moral pessimism is the more interesting form of pessimism in the political context. I am looking for characteristically normative criticism that could be offered against those who politically defer to others.
The next question is how deferring to others is morally problematic. The strongest option would be that it is all-things-considered wrong to defer to others. This is an implausible position (Howell Reference Howell2014: 390; Lillehammer Reference Lillehammer2014: 115–16; but see van Wietmarschen Reference van Wietmarschen2018). A weaker form of pessimism claims that it is pro tanto wrong to defer to others on moral issues (Crisp Reference Crisp2014: 132). Most formulations are even broader. Gopal Sreenivasan considers the claim that moral deference is “some kind of toxic stain” (Reference Sreenivasan2015: 42), while Jones discusses whether relying on others is “something regrettable” (Reference Jones1999: 75). I have phrased pessimism in terms of what is morally deficient, a deliberately vague notion meant to encompass a variety of ways of framing the debate.
Sreenivasan has argued that you can be an optimist while accepting that it would be better if people came to understand moral views by themselves (Reference Sreenivasan2015: 49–52).Footnote 5 Some action A might be superior to some other action B, without B being morally objectionable. This strikes me as the right view to take. Thus, I will concede to the pessimists that the ideal citizen forms their views autonomously; however, those who fail to do so do not necessarily exhibit a moral failure. In other words, forming one's political beliefs autonomously is generally supererogatory – praiseworthy (highly morally good) to do, but not blameworthy (morally bad) not to do.
“Acting on the basis of”
It might be morally deficient merely to form beliefs by deferring to others. But the more interesting cases are where agents also act on those beliefs.Footnote 6 Thus, I will only consider cases in which deference to others does not merely change what we believe, but also how we act. In the political case, this could take the form of voting for a different party or candidate, or other forms of activity, such as political advocacy, lobbying, protesting, and so on.
“Beliefs sustained”
We might gain a belief on the basis of deferring to someone else, but then come to support it by independent evidence or reflection of our own (Howell Reference Howell2014: 389–90). Imagine, for example, that after several years of practicing vegetarianism, Aaron has come to see why eating meat is wrong. He has gained genuine empathy for the suffering of animals, and a deepened understanding of their moral status. We can stipulate that Aaron would have never become a vegetarian if he had not deferred to Birke previously. In that sense, Birke's testimony made a crucial difference. However, his current vegetarianism is no longer sustained by it. It provided Aaron with the impulse for critical reflection of his own. What matters, then, is not whether a belief originated in testimony, but whether it is currently sustained by the testimony of others (van Wietmarschen Reference van Wietmarschen2018: 25).
“Deference”
Lastly, I have phrased pessimism in terms of deference. In this way, I hope to avoid debates around two other concepts, testimony and expertise, which are often discussed in this context. Relying on testimony is a paradigmatic of deferring to others, but it is not the only one (Mogensen Reference Mogensen2017: 265). Alternatively, for example, you could rely on others’ advice (Sliwa Reference Sliwa2012). What is important for our context is not in which way you defer, but that you defer. I take it that you defer to someone else on the question whether p if you believe that p (or not-p) merely because the other person believes that p (or not-p). Similarly, I hope to avoid issues of expertise. I will make no assumption as to whether the people we defer to are experts, or what would make one an expert in political issues.
3. Deference in politics
We can now formulate the view I aim to defend,
weak optimism about political deference. In most cases, if one politically acts on the basis of beliefs concerning a normative political subject matter, where these beliefs are only sustained on the basis of deference to someone else, then one's action is not morally deficient in this respect.
There are only two minor changes in this formulation. First, I narrowed the view down to political actions – such as voting, protesting, or deliberating in public. Second, I focus on beliefs concerning normative political subject matters. These claims could be highly abstract (e.g., “justice requires equality”) or highly specific (e.g., “Jones is the best candidate for county commissioner”).Footnote 7
There is one feature of the definition of the optimism – pessimism contrast that I have so far omitted, because here I start to deviate from the standard framing of the debate. Most authors agree that the interesting cases of deference are those where I defer to someone else on a purely normative question. Imagine that Birke knew more non-moral facts about eating meat than Aaron – e.g., about the conditions under which factory animals are held. If Aaron defers to Birke in this scenario, it is no longer clear whether he relies on her as a purely moral guide; Aaron might defer to Birke because of her superior non-moral knowledge. To avoid this complication, I stipulated that Aaron and Birke know the relevant non-moral facts equally well.
McGrath calls Aaron's deference to Birke pure for this reason (McGrath Reference McGrath2009). Almost all authors in the literature – pessimists and optimists alike – focus on pure cases of deference, and use similar stipulations. However, I will reject this focus on pure cases of deference. I will argue that (i) almost all interesting, non-ideal cases of deference are impure, and (ii) optimism is more convincing than pessimism when we focus on impure cases.Footnote 8 In other words, the idealised cases found in the literature, like vegetarianism, are misleading: optimism provides a much better fit for our intuitions once we turn away from them.
You might object that I misunderstand the reason why epistemologists focus on idealised cases, but hold your horses: I will tackle this objection below (§4). At this point, it is useful to have an impure form of moral deference at hand which we can contrast with vegetarianism.Footnote 9 Here is one such case,
union support.Footnote 10 Matt is a 56-year-old bricklayer in Battle Creek, Michigan. For three decades, he has belonged to the Michigan Bricklayers and Allied Craftworkers Union. He is overall satisfied with how the union has represented his, and other bricklayers’, interests.
The union employs a few full-time political representatives. One of them is John, who spends significant amounts of time lobbying state politicians in Lansing. Matt and John have known each other for many years, and come to trust each other, although they rarely meet. In a particularly contentious state election, Matt is torn between voting for a more conservative candidate, Blue, and a more progressive candidate, Red. Matt's socially conservative views align with Blue's. Blue also promises tax cuts, which Matt thinks would benefit him and his family.
A few days before the election, John visits Matt on a building site. John says that a vote for Red is in the best interest of bricklayers. In particular, John claims, Blue has a “stealth agenda” meant to harm unions, while at the same time wooing their members with tax cuts. But before John can elaborate further, he has to urgently leave.
Despite having no further information, Matt ends up voting for Red.
Two things are at issue here, an action and a proposition on which this action is based. The action is voting for Red; the content of the proposition is “a vote for Red is in the best interest of bricklayers”. Matt defers to John with respect to this proposition, and it forms the basis for his action; this makes it a case of political deference as I have stipulated above.
union support supports a broadly optimist position. It seems that there is nothing seriously morally deficient in how Matt comes to vote for Red. Indeed, if we compare Matt to the median voter in contemporary democracy – about whom empirical research has few nice things to sayFootnote 11 – his method of decision-making seems positively laudable. As I conceded above, it might have been better if Matt had searched out more sources of information; but whatever his omissions in this respect, they do not constitute a major moral failure.
Some principled defences of Matt's deference can be offered (§5). But before I turn to these, I want to highlight some ways in which union support is a case of impure moral deference, how such forms of impurity can be expected in non-ideal cases, and furthermore, how that impurity supports optimist intuitions.
3.1. Purely moral and mixed deference
We can start with the dimension of purity already mentioned. Many authors make a strong distinction between testimony concerning purely moral matters and mixed moral matters (e.g., McGrath Reference McGrath2009; Enoch Reference Enoch2014: 232; Mogensen Reference Mogensen2017: 263). This is the form of purity Sarah McGrath originally had in mind. In a case of pure moral deference, she writes, I defer to you on the basis of “purely moral information that one lacks” (Reference McGrath2009: 322). By contrast, in the case of impure moral deference, I defer to you because I also believe you to have additional non-normative information that bears on the normative case (McGrath Reference McGrath2009: 322).
We can juxtapose this assumption with what we know about how political opinions are actually formed. A wide array of empirical research shows that voters tend to fall into two broad groups.Footnote 12 First, we find a small group of highly informed people who follow politics closely. Voters in this group need not be political operatives like John; reading the political news regularly, for example, is enough to put you in this group. On the other hand, the large majority of voters knows little about politics: they cannot name their representatives in parliament, have trouble assigning political positions to the different parties, and wildly misestimate even basic facts about politics, like levels of distributional inequality, the percent of the public budget spent on foreign aid, and so on.
Furthermore, we know that public opinion often forms on the basis of “elite cues”: more informed voters tend to have a strong impact on the opinion of less informed voters.Footnote 13 Thus, it is a plausible assumption that political interaction often happens between two parties with large asymmetries in factual knowledge. If so, then union support represents a much more likely occasion of political deference than a case like vegetarianism.Footnote 14
The second point we need to establish is that cases which are impure along this dimension lend greater intuitive support to optimism. We can start with the observation that pessimists normally restrict their view to testimony on moral issues, and do not advocate a general pessimism about testimony (Groll and Decker Reference Groll and Decker2014). They do so for good reason, as pessimism about all testimony – an insistence that we should be autonomous knowers on all subject matters – is deeply implausible. After considering what it would mean to practically implement such a position, Elizabeth Fricker concludes that such a “human would-be epistemic autonome … is not an ideal, but either paranoid or severely cognitively lacking, or deeply rationally incoherent” (Fricker Reference Fricker, Lackey and Sosa2006: 244). Pessimists are wise to stay away from this position.
Furthermore, we cannot neatly disentangle the degree to which Matt relies on John as a moral subject guide and the degree to which he relies on him as a technical expert (cf. §4). Matt expects John to make the right call as to what is in the interest of bricklayers like him. This is a thick concept, one with a normative component (looking towards what the normatively relevant interests of bricklayers are), as well as an empirical component (looking towards what actually promotes those interests). It is implausible to think that there is a neat separation in Matt's deference to John between these two aspects. When Matt defers to John's claim that Red is “in the best interest of bricklayers”, he defers to John on both a moral and a non-moral issue.
However, if normative and non-normative deference in this case – and by extension, in most non-ideal cases – is so closely entangled, then optimism becomes more plausible. The alternative would be to hold that deference in mixed cases like these is still morally deficient. But given how ubiquitous we can expect mixed deference to be in the political case, this runs the position dangerously close to the unappealing vision of the epistemic autonome Fricker has argued so powerfully against.
3.2. Agreement on background beliefs
A second common stipulation is that the people in cases of moral deference do not share any moral background beliefs. Imagine that A trusts B's claim that p. Following Robert Howell, we can imagine the following broad types of situations (Reference Howell2014: 391–2):
1. A and B agree on all values relevant to the situation, and their relative weight. A trusts B in being faster or more reliable at finding out which option their shared values support in the given context. (Howell calls this “derivative moral deference”.)
2. A and B agree on all values relevant to the situation, although not their relative ranking. A trusts B in being more morally attuned to adjudicating difficult trade-offs between their shared values in the given context.Footnote 15 (Howell calls this “derivative moral deference with ignorance”.)
3. A and B agree neither on which values are relevant to the given situation nor these values’ weight in the given context. A trusts B in being more reliable concerning moral questions in general. (Howell calls this “foundational moral deference”.)
Deference gets increasingly problematic along these cases. In case 1, it seems that A uses B merely as a “sort of moral calculator” (Howell Reference Howell2014: 392). This is perhaps a strange way to make use of another person, but it is not clear that we think of it as very morally troubling. The most problematic seems to be case 3, where A shares no values with B. It is cases in this last category one normally encounters in the literature, such as vegetarianism.
However, most non-ideal cases of political deference will fall in the middle category. Empirical research supports this claim. Citizens do not take cues from elites randomly: they search out like-minded sources. For example, when the same political statement is ascribed to sources with different ideological backgrounds – e.g., once it is ascribed to a Democrat, and once to a Republican – experimental subjects agree with the statement much more strongly if they share the source's party affiliation.Footnote 16 Party affiliation only weakly corresponds with shared ideological views, and it might reflect shared group identity instead (a point to which I turn below). Still, it is not a bad presumption that much non-ideal deference will not be a case of foundational moral deference, but of derivative moral deference with ignorance.
Return to union support. Matt and John share a fundamental concern for the fate of bricklayers. This is a broad value which allows for a variety of interpretations. There can be significant internal disagreement over what is best for bricklayers. When Matt defers to John, he does not defer to someone who merely calculates his own values for him, but neither does he defer to John on the basis of some abstract belief in John's reliability to get “morality” right as such.
Once again, to the degree that non-ideal deference is impure along this dimension, we are moved towards the optimist case. We are unlikely to use others merely as moral calculators, but we are also unlikely to defer to them on the basis of some abstract measure of “moral reliability”. We will often know, to some degree, of values we share, of which shared party affiliation is one expression. Aaron's vegetarianism seems so troubling because there is just nothing of him – none of his values, commitments, or critical reflection – in it. But that is not quite true of Matt's vote for Red. One of Matt's fundamental values, his support for union causes, is reflected in his choice. This makes it much harder to feel a strong pessimist reaction to his case.
3.3. Perfect strangers versus shared groups
Another common assumption found in the literature is that testifier and trustee stand in no special relationship to each other: they are not friends, or relatives, or co-workers, nor do they have any other shared identity.Footnote 17 But this is once again an epistemologist's fancy. One of the most powerful forces driving voting behaviour is group identification.Footnote 18 In politics, who we are is as important, if not more important, as what we believe. In a famous study of the voting behaviour of African-Americans, for example, Michael Dawson introduced the idea of “linked fate” (Dawson Reference Dawson1994): many African-Americans, he found, do not vote on the basis of perceived self-interest, but on the basis of perceived African-American group interest, believing that “what happens to the group as a whole affects their own life” (Reference Dawson1994: 77).
If our political lives are so intensely shaped by group identity, then it is natural to think that deference happens between members of a group with a shared identity or shared interests – especially along racial, geographic, ethnic, gender, religious, and class lines. If African-American voters believe in “linked fate”, for example, then it is natural for them to adjust their views, and voting behaviour, in view of what other African-Americans think. This leads to a second question: within a given group, who are we likely to most defer to? I suspect that groups will often have an internal separation between elites and followers, such that much (but certainly not all) political influencing is from group leaders to group followers.
Once again, I have constructed union support to reflect these empirically informed suspicions about non-ideal political deference: Matt does not defer to just anyone, but to someone with whom he knows to have a shared group interest, and whose fate he can plausibly see as being linked to his. (More precisely, Matt's fate is linked to that of other bricklayers, and deferring to John is a specific way of expressing that linkage.) Just as with the other dimensions I have outlined, this moves us towards an optimistic view. If there are shared bonds between two people, testimony starts to look less troubling. I will return to this point below (§5.2), so I will delay its defence till then.
3.4. Deference and persuasion
A further way in which cases of deference can be more or less pure is by including more or less reason-giving. I have constructed union support such that it contains very little argument given by John to Matt. But we might think that even short interactions between Matt and John contain compressed arguments, and are thus not forms of pure deference.Footnote 19 Pessimists generally accept that there is no problem with one person persuading, or otherwise influencing, another (Wolff Reference Wolff1970: 13; Hills Reference Hills2009: 123; Mogensen Reference Mogensen2017: 263). If John spends an hour explaining to Matt why Red is the better candidate, and if Matt comes to understand the relevant reasons by himself on this basis, then it seems that he does not defer to John in any interesting sense.
However, there is a gradual distinction between testifying and arguing (Jones Reference Jones1999: 63). Even in the extreme case of John arguing his case for an hour, there are some ways in which Matt defers to John. For example, he relies on John framing the issue in a reliable way. John relies on Matt not to leave important aspects of the question out, of presenting the evidence in an impartial way, and so on. This is genuine normative deference, too: what is important or unimportant is itself a normative judgment (cf. Rini Reference Rini2017).
Framing effects are known to have a strong effect on political persuasion, to the degree that it has been claimed that the “essence of public opinion formation … lies in the distillation or sorting out of frames of reference” (Chong Reference Chong1993: 870).Footnote 20 One study found, for example, that it makes a difference to listeners’ opinions whether a rally of the Ku Klux Klan was presented as a free speech issue or a question of public order (Nelson et al. Reference Nelson, Clawson and Oxley1997). So even in those cases which are more persuasion than deference, we will find at least some types of deference, especially deference to the other person's framing of an issue. The question, then, is whether deference of those subtler kinds is wrong. The pessimist, on the face of it, is committed to answer this in the affirmative, too; once again, I think, the optimist has the superior view in this respect.
4. Against continuity
To summarise, non-ideal deference will often be impure amongst one or more dimensions: it will be characterised by an asymmetry of factual knowledge, will involve some broadly shared background values, and will often happen between members of some shared identity or interest group. The crucial observation is that optimism fits our intuitions better if we look at impure cases. Thus, we can grant to the pessimist that deference in vegetarianism is morally deficient; but we can reject that it is a representative case. In the majority of non-ideal cases, like union support, deference is unproblematic.
The pessimist is likely to respond that I have misunderstood why philosophers focus on pure cases. They do so, they might argue, because they follow a philosophical version of the method of isolation.Footnote 21 If there is something troubling in pure cases of moral deference, then there is something troubling about cases of moral deference across the board. Just as you want to test gravitational theories in an artificial vacuum to exclude distorting forces like air resistance, so we want to test the moral deficiency of deference by excluding distorting factors, like the impurities I have outlined. Thus, deference is assumed by pessimists to be pure not because purity is realistic, but because it allows us to home in on the clearest cases. Someone who applies this method might implicitly assume
continuity. If there is something morally deficient about moral deference in pure cases, then moral deference qua moral deference in impure cases is equally morally deficient.
If continuity is true, and if deference is morally deficient in vegetarianism, then we have identified one way in which deference is morally deficient everywhere. Thus, we could continue, we have shown that deference in union support is morally deficient. This is compatible with admitting that there are other factors in impure cases which relevantly outweigh the moral deficiency of deference. Perhaps it is all-things-considered better for Matt to defer to John. But Matt's deference still possesses a moral stain, the pessimist would insist, one which their position correctly identifies.
However, we should reject continuity: there is no reason to think that impure cases are continuous with pure cases. There are three reasons to reject continuity. The first comes from close attention to the kinds of cases we have already considered. Focus again on union support, where I stressed the importance of group solidarity. A defender of continuity might argue that the value of solidarity outweighs the disvalue of deference in this case, but the latter is not vanquished. It is still deficient, in some respect, for Matt to defer to John. This does not strike me as the correct diagnosis of the case, however. Matt has chosen a general strategy, which is to solidarize with a particular group, the cause of bricklayers. This is a fundamental, autonomous choice of Matt's, and a morally and politically unproblematic commitment at that. In union support, this commitment requires him to trust John's views. It seems to me that this not merely outweighs the supposed stain of deference; there is no such stain in the first place. The same diagnosis applies to other impure cases. I can feel some moral unease about how Matt forms his beliefs in union support. But this unease does not rise to the level of straight-out serious moral deficiency which the pessimist suggests we should see – i.e., the kind of deficiency exhibited by Aaron in vegetarianism.
Second, there is a deeper metaphysical worry. If we are to make sense of continuity, we must be able to somehow distinguish the “deference” component (as the morally deficient part) from the “solidarity” component (as the morally valuable part) in what Matt does. But it seems doubtful that we can do so, at least in a non-gimmicky way. It is more plausible that there is one thing Matt does – trusting a fellow union member – in which the solidarity component and the deference component cannot easily be disentangled. Similar points could be made with respect to the other dimensions of impurity. We will be hard-pressed, for example, to separate out the degree to which Matt trusts John as a moral expert, and to what degree he trusts him as a non-moral expert. Matt defers to John in both ways at the same time, inseparably; there is no pure moral deference with some independent non-moral deference thrown in. But if that is so, then one of the underlying assumptions of the method of isolation seems to fall away.
A last reason to reject continuity is that similar assumptions in other areas of philosophy have come to seem increasingly dubious. For example, it might be that in the Kantian Kingdom of Ends we would never tell a lie. But that does not entail that we should tell the truth to the famous murderer at the door. Similarly, an ideal theory of justice might tell us that some distribution of benefits and burdens D is morally the best. But that does not entail that if we can make our society more equal to D in one important respect, then we should always do so (Gaus Reference Gaus2016: 61–74; Wiens Reference Wiens2016). In other words, there is no general reason to think that non-ideal cases in moral and political philosophy are increasingly paler versions of the ideal case.Footnote 22
This is not an argument against idealisation in philosophy in general; I also accept that cases like vegetarianism are interesting and fruitful if we wish to tackle more abstract philosophical questions in meta-ethics or epistemology. Furthermore, some idealisation is inevitable, and some is used in union support. But if continuity fails, no simple extrapolation from one case to all cases succeeds. This does not by itself show that pessimism is wrong, as there could be an independent argument to favour pessimism for non-ideal cases. However, as I argued in the previous section, in a wide range of cases of medium impurity, optimism carries the day.
We should also note that rejecting continuity is a double-edged sword: it speaks in favour of weak optimism, but it also cuts against strong optimism, as it also stops the optimist cases I have identified from generalising either. This implication I accept, and is part of the reason I stop short of advocating strong optimism.
5. Two principled defences of deference
As far as substance goes, my argument in §3 was an argument that optimism fits our intuitions better for a wide range of cases. In this section, I will offer two more principled arguments in favour of optimism.
5.1. The division of epistemic labour
The starting assumption of the first argument is the immense epistemic complexity of political problems. These problems are a complex interplay of normative and empirical questions. Think about the election of a local judge. Think about whether to boycott a certain textile company or not. Think about whether to support a local initiative for a wildlife reserve. Each of these issues involves a highly complex set of both facts and values. There are scientific facts and theories I need to know – e.g., about the best preservationist strategies for different wildlife species. There are complex value judgments I need to make – e.g., about the value of species survival, or about the distribution of responsibilities between different actors. There are social and legal facts relevant to the question, too – e.g., what type of preserve would be realisable within legal constraints, and which would find general acceptance. Lastly, I need to answer various questions about political strategy – e.g., which political party is most likely to tackle these issues, most likely to be effective if in power, and so on.
Each of us might be in a position to competently decide some issue based on such an interwoven thicket of questions. Most of us are knowledgeable on certain political issues. For example, you might care, and know a lot, about local education policy, or the moral and legal arguments surrounding pornography legislation, or the environmental issues concerning wildlife preserves. But while each of us might be competent in such limited ways, for most political issues we are not (Brinkmann Reference Brinkmann2018).
It is important that we cannot be faulted for these limitations in our political competence. Because political problems are so large and overwhelmingly difficult, and because there are so many of them, we cannot be expected to be anything but know-nothings for most. Indeed, given how little difference most of us make to political outcomes, it is rational for us to be ignorant on political issues (Downs Reference Downs1957; Caplan Reference Caplan2008: Ch. 5). There is, in other words, normally no moral deficiency in political ignorance itself.
To this observation of the complexity of political problems we must add the claim that political problems are high stakes. Filling out your office sweepstake bracket for March Madness is epistemically demanding, but no blame falls on you if you fill it in haphazardly; indeed, seeing the random guessers succeed is part of the fun. Political problems are not of that type. To put the point dramatically, people's lives are at stake: some political choices are matters of war and peace. Which political party is in power affects the lives of millions of people, determining their fates – the resources, rights, freedoms, and punishments they will receive or fail to receive. These are not merely questions of well-being, but justice too. The stakes in politics are high, and we have a strong moral interest in getting them right. Jason Brennan has argued on these grounds that we have a duty to vote competently (Brennan Reference Brennan2009); however, all we need for the current argument is the claim that we have a strong moral interest in voting competently.Footnote 23
Onwards to the idea of the division of epistemic labour.Footnote 24 Because political problems are complex, it would take each of us significant amounts of time and effort to become competent at solving them ourselves. Indeed, even with significant amounts of time and effort, we might not become competent at solving them, and the task might simply be super-human. But if we cannot solve an important task of our own, and if solving this task is important and not morally optional, then we have a duty (or at least a strong moral interest) to do our best to contribute to a collective effort to resolve the relevant task. For example, if I have a duty to save a drowning child, but if the child is too heavy to be saved by me alone, then I should relevantly combine my efforts with those of someone else (cf. Collins Reference Collins2013). A division, or combination, of labour is morally required where our individual powers run out.
What would it mean to engage in a genuinely epistemic division of labour? Let us start from a non-political case,
particle accelerator. A group of scientists aim to build a particle accelerator. Doing so involves combining complex knowledge from a large variety of specialised subjects – e.g., particle physics, chemistry, engineering, and so on. Amanda is a construction expert for large-scale scientific instruments. Other scientific experts, on the basis of specialised knowledge unavailable to her, provide her with various specifications for the accelerator – e.g., on the proper materials to be used, the requirements for the cooling system, the needed scanning equipment, and so on.Footnote 25
Amanda, we can assume, has a strong interest to help create the particle accelerator. But she cannot do so without deferring to the other specialists. To accumulate all the relevant knowledge to build a particle accelerator would take her a lifetime, if it is possible at all. She is not expected to gain all that knowledge herself. Furthermore, Amanda cannot simply withhold belief – for example, on what materials to use. She must trust the specialised expertise of others, or she will be incapable of doing her part. Given her situation, she should defer to others.
We can now assemble all the claims on which my first defence of political deference is built: (i) we have a strong moral interest to get political questions right, (ii) we are normally incapable of solving those questions, or parts of those questions, by ourselves, (iii) we cannot generally be blamed for this incapability due to the epistemic complexity of the problem, and (iv) we can get those questions right if we combine our epistemic efforts with others, in particular, by deferring to them. If most political problems have these features, then we have a strong moral interest, perhaps even a duty, to defer to others on important political questions. Furthermore, as facts and values are often closely intertwined in political issues, we also must often defer on moral questions.
This provides us with a principled explanation why there is nothing deficient about deference in cases like union support. Matt plays his part in a large-scale division of epistemic labour. Far from ignoring his duties as a democratic citizen or a union member, he lives up to them. Indeed, there is a sense in which not deferring to John can sometimes be the morally deficient choice. Consider an analogous situation in particle accelerator. Imagine that Amanda does not defer to some specifications given to her by a theoretical physicist, Bertram. She remembers some of her undergraduate physics, and does the relevant calculations herself. Let us assume that she reaches the same conclusions as Bertram, and the particle accelerator gets built without a hitch. Still, Amanda is open to a range of criticism by her co-workers: by not trusting Bertram, she could have done significant damage to the whole project. She is not liable to that same critique if she had simply deferred to him. Along the same lines, we might sometimes be justifiably criticised if we do not defer to others on important political questions.
There are of course dangers in blind deference which threaten the benefits of the division of labour.Footnote 26 For one, you must check whether you trust the right person. Amanda must check whether those she trusts have the right credentials, are properly authorised members of the overall project, and so on. Similarly, Matt should check John's credentials (in a wide sense): is John reliable and honest? Does he seem competent enough to “call the shots”? And so on. It is difficult, and beyond the scope of this paper, to determine the correct balance of trust and mistrust in these cases. However, one point is important: even if a general policy of blind deference would be mistaken, deference on specific subject matters to specific people is justified once you have paid appropriate critical scrutiny. Matt, for example, has exercised due diligence through a long history of interacting with John. It is on the basis of this history that he can trust John. Unless he encounters defeating reasons which should make him reconsider, he has no need to double-check John's recommendations in every new case which comes up. Thus, he has no reason to check John's specific advice in union support. A demand on people to critically consider whom to defer to is thus entirely compatible with optimism.
5.2. Epistemic solidarity
I will now turn to a second set of arguments. As these ideas have been powerfully put forward by other authors (Jones Reference Jones1999; Wiland Reference Wiland2017; McShane Reference McShane2018), we can focus on how to adjust them to the political context, and how they interact with the main argument I have given. I will start with Paddy McShane's (Reference McShane2018) suggestion that deference is a constitutive ingredient in friendship. She starts from the plausible idea that trust is one of the crucial parts of friendships; depending on one's friends is one way of expressing that trust. McShane then goes on to apply this general line of argument to relying on friends’ testimony in moral matters. She writes:
[I]f we refuse to take up friends’ moral views in response to their moral testimony, when their testimony functions as an invitation to accept what they say just based on their say-so, we risk treating them poorly and undermining our relationships in a way that we do not when our friends are not addressing their moral views to us at all. (McShane Reference McShane2018: 641)
I find McShane's argument convincing. Unfortunately, it does not immediately help us with the political case. There might be some special cases where we should politically defer to someone else because of a friendship. But this does not provide us with a general line of argument that political deference is non-deficient in most cases. It is also not quite clear why we should trust others on specifically political subject matters merely because they are our friends; the idea seems intuitively more convincing with respect to moral subject matters.
But McShane points us in the right direction, by making us aware of the networks of relations and identities people share – something which I stressed above (§3.3). In particular, Matt deferring to John's advice is a way of exercising epistemic solidarity with other bricklayers. You exercise such solidarity if you adjust your beliefs and actions to that of others with whom you share a group with common interests (such as being a bricklayer, or working class, or African-American).Footnote 27 This explanation, unlike McShane's, does not immediately focus on the relationship between Matt and John. They need not be friends, or even know each other particularly well. What is important is the group with which Matt solidarizes by deferring to John – that of bricklayers, or perhaps more narrowly, Battle Creek bricklayers. John represents that group in the formal sense, but also in a more informal sense: it is plausible for Matt to assume that John's advice represents, or approximates, the collective political view of bricklayers.
There are instrumental reasons to value epistemic solidarity. A group of voters with a shared interest can achieve more if they pool their votes rather than leaving it to each individual to determine what is in the group's interests, thus splitting their collective power (Goodin and Spiekermann Reference Goodin and Spiekermann2015). Furthermore, if you believe that crowds are wiser than their individual members (Surowiecki Reference Surowiecki2004; Landemore Reference Landemore, Landemore and Elster2012), exercising solidarity might increase the chance that you choose correctly. But beyond such instrumental benefits, there is intrinsic value in epistemic solidarity. Here we can turn to Eric Wiland's account of epistemic solidarity (Wiland Reference Wiland2017). Wiland emphasises that if we solidarize with others, we take an interest in our genuinely collective achievements. Matt stops thinking of his interests, but instead starts thinking of “we, the bricklayers”. This is a mode of democratic action which seems not only unproblematic, but positively laudable. As I observed above, there is some evidence that this happens in fact, insofar as African-Americans think of their fates as being linked, and vote on the basis of perceived group-interest rather than self-interest.Footnote 28
Wiland paints solidarity of this type as a way of making “use of each other's senses” (Reference Wiland2017: 69), and he argues that epistemic solidarity “is part and parcel of the human good” (Reference Wiland2017: 72). Wiland is best interpreted as describing an ideal. In a non-ideal context, however, we will want to dial down the idea of epistemic solidarity a bit. Real-world communities are imperfect, and the solidarity lived in them often flawed. We need not think that anything as grandiose as “making use of each other's senses” describes what happens in union support, or in African-Americans voting as a unified block. Still, even so there is something intuitive to the idea that Matt deferring to John is a way in which the former expresses his allegiance to a collective cause, and that there is something morally valuable in forming such a relationship.
The arguments from epistemic solidarity and the argument from the division of epistemic labour are mutually reinforcing, even if we can see them as independent. First, you might think that it is troubling to think of political problems as merely technocratic issues, as complex intellectual problems to be solved like Sudoku puzzles, as the first argument seems to suggest. But if we add that political deference often happens as an exercise of epistemic solidarity, some of these worries will be assuaged. An agent like John in union support is part of a complex social group engaged in solving a difficult epistemic problem – what is best for bricklayers? What motivates John is not some abstract, fetishist concern for “morality as such”, but a commitment to a concrete, identifiable social group and its political interests.
Inversely, the argument from epistemic solidarity gains in strength if we add the importance of the division of intellectual labour. If we do not face a difficult problem, or not a morally significant problem, then it is less clear why epistemic solidarity should be so pressing. It is not important, for example, that union members present a unified front on local zoning regulations. On this matter, Matt might just vote as he sees fit. Similarly, for political choices which are comparatively easy – say, voting between a mainstream candidate and a pseudo-fascist – it might not matter greatly that Matt defers to his union. But if the political problem is both significant and epistemically difficult, epistemic solidarity becomes more and more valuable.
6. Objections and responses
In this section, I turn to some objections. I will limit myself to the more unique objections to the present project – I set aside general objections to optimism, and general objections to the picture of democracy on which I have relied, as those have been usefully answered elsewhere.Footnote 29
6.1. Public opinion and the scope of optimism
First, one might object that I cherry-picked what non-ideal political deference looks like. One might object that most such deference is unlike union support, and supports a broadly pessimist position. Anecdotally, for example, we can think of the rise of “fake news”Footnote 30, social media echo chambers, or ideologically biased news sources. One can cite a whole range of other results from political psychology to paint a much nastier picture. For example, you might highlight the various ways in which voters are influenced through framing of political issues (e.g., Chong and Druckman Reference Chong and Druckman2007), or how they engage in motivated reasoning to seek supporting evidence for pre-existing views, or more generally, how people are influenced by external sources in non-deliberative and subconscious ways (e.g., Taber Reference Taber, Shapiro and Jacobs2011: 370–3).
There is no point in denying that empirical research reveals some extremely troubling ways in which political opinion is formed. Remember, however, that the optimist merely claims that deference qua deference is not morally deficient. But in many cases, there are various other factors which cloud our judgments. In those cases, however, it is not the act of deference itself that is wrong; instead, we can point to various other contextual features. Consider the following example:
racist radio. Jørg holds racist prejudices about Muslim immigrants to Denmark. His friends, however, present him with convincing alternative viewpoints on the benefits of immigration, the equal rights and respect owed to them, and so on. Still, Jørg feels an urgent hatred for Muslim immigrants, and he feels there must be evidence that supports him. Jørg comes across a right-wing extremist podcast, “Enemies of Sharia”, which presents a highly unreliable view of the facts, tinged with racism. While Jørg has initial worries about the reliability of “Enemies of Sharia”, their views psychologically resonate with him, and he quickly becomes an enthusiastic follower. The podcast also advances other views Jørg had not previously considered, like anti-Semitic conspiracy theories. Jørg accepts these claims, and their implicit normative underpinnings, uncritically.
Jørg forms his opinions in an epistemically and morally irresponsible way, and there is much to criticise. But we should not be too quick in thinking that a case like racist radio lends support to pessimism, or presents a problem for optimism. The question is whether Jørg's deference in this case is morally deficient, or whether the moral deficiency is located in other, extraneous features of the case. And extraneous features there are. For one, Jørg's deeper motive in deferring to “Enemies of Sharia” is not a desire to get a moral subject matter right, but to confirm a pre-existing, highly troubling emotional-moral attitude. We could also highlight other features: just as Jørg is not a good-faith moral agent, so we can expect the makers of “Enemies of Sharia” not to be good-faith political deliberators. We reasonably suspect that the makers of “Enemies of Sharia” intend to manipulate Jørg for nefarious purposes. And so on.
It is useful to pause to note a methodological point. My appeal that epistemologists should move away from highly idealised cases must not be misunderstood as an argument against philosophical idealisation in general. My objection is to the specific idealisations epistemologists commonly impose which move it too far away from realistic cases. union support rests on idealisations, only different ones. If we wanted to apply principles derived from cases like union support to real-world cases, we would still need to add some extra level of philosophical reasoning, as there will be additional factors in realistic cases which will merit further attention. However, my claim is that we can at least reasonably recognise real-world political agents in Matt and John, which is not generally the case with Aaron and Birke. So there is some hope that principles we derive from cases like union support have some relevance to real-world cases, while I think there is little such hope with principles derived from cases like vegetarianism. The question, in short, is not one of perfect mimicry of reality, but of coming reasonably close to reality while still having a usefully abstract model.
Let us move on to a less extreme case. Imagine that Lorna belongs to the top 1% of the income distribution. Lorna listens to hedgehog news, a right-wing television channel which supports the factually incorrect, but ideologically comforting theory that wealth “trickles down” such that progressive taxation is counter-productive. Is it seriously morally deficient if Lorna comes to believe in trickle-down economics on the basis of hedgehog news? A lot depends, I think, on how we paint in the detail. On some versions, Lorna can be criticised in a similar way as Jørg, as being a deficient moral reasoner. But the moral deficiency is in the quality of the moral reasoning which leads Lorna to defer, not in the deference itself. Our intuitions might also be confused because many will think that Jørg and Lorna support political causes which are repugnant, or at least morally problematic. But then again, the moral deficiency in such cases is located elsewhere, and not in the deference itself.
There are, in short, a number of debunking strategies available to the optimist to discount various supposed counterexamples to their view (Sliwa Reference Sliwa2012; Enoch Reference Enoch2014). The scope of cases for which these debunking strategies can be made to work is open to debate. But the weak optimist does not need to have them work everywhere. They might well accept that there are many cases of deference which are morally deficient, even qua deference. As long as deference remains unproblematic in the majority of cases, it remains an interesting position.
6.2. Advantages of independence
One might also question the ideal of the epistemic division of labour on which much of my defence of epistemic deference rests. Many defences of democracy stress that democracy's advantages lie in harnessing epistemic diversity.Footnote 31 The Condorcet Jury Theorem, for example, rests on the assumption that individuals form their views independently.Footnote 32 To the degree that individuals defer to opinion leaders, then, it seems that democracy performs less well.Footnote 33
There are two replies to this objection. First, epistemic solidarity can be an effective strategy that, under certain conditions, is rational and improves democracy's performance. This has been shown by Robert Goodin and Kai Spiekermann, who consider a series of models in which the assumption of independence is relaxed (Goodin and Spiekermann Reference Goodin and Spiekermann2015; Spiekermann and Goodin Reference Spiekermann and Goodin2018). Similarly, there are formal models in the philosophy of science which suggest that deferring to others under certain conditions can improve truth-tracking.Footnote 34 While many formal models emphasise the benefits of diversity, there is certainly room for deference and consensus-seeking, too.
Assume, however, that we are facing a worst-case scenario. Imagine, for example, that all local bricklayers defer to John's claim that Red is in the best interest of bricklayers. Let us further stipulate that, while John knows more about politics, the multitude of bricklayers, collectively, possess diverse perspectives and information that would have shed additional light on who to vote for, and which gets lost if they simply defer to John. Perhaps if all bricklayers had come together and discussed the matter, for example, they could have collectively found out that John's misgivings about Blue are misplaced, and that Blue is in fact the superior candidate.
The optimist should make two responses to this case. The first response is, simply, that the moderate optimist does not claim that deference is always unobjectionable. There will be cases, especially where there are additional moral factors, where it would be wrong to defer. Imagine, for example, that Matt held some unique piece of information that would show John's suspicions of Blue to be misplaced, or at least, information that he knows would be valuable to other bricklayers. In this case, it would likely be morally deficient if Matt simply deferred to John. But this should not trouble the weak optimist, who already acknowledges many exceptions to the general rule.
The second response is that the optimist merely claims that it is not individually morally deficient to defer. But as in many cases, the combination of individually moral or rational actions might amount to a collective outcome which is morally or rationally subpar. However, we need to distinguish the institutional question (which institutions perform morally or epistemically best?) from the individual question (for which beliefs and actions should we hold individuals morally responsible?). Granted, the two forms of assessment are not wholly independent. If everyone following a strategy of political deference was inferior to people forming opinions independently, for example, then this would certainly affect how we judge individuals who followed a strategy of deference. But there is space for optimism to remain plausible.
6.3. Autonomy
I will now try to briefly assuage objections from autonomy. The main objection is simple: moral deference conflicts with the fundamental value of autonomy. We could make this argument on general grounds – there might just be a general deficiency if one forms one's moral beliefs non-autonomously.Footnote 35 But a specific political objection can also be pursued: there is value if each citizen in democracy forms their views autonomously. If democracy centrally relies on the formation and expression of a “general will”, then to the degree that people defer to each other, it is not a general will, but the will of the few.
One strategy to answer this objection would be to go through a variety of ways how one could understand autonomy and to show that deference poses no serious threat on each specification of autonomy. I think this could be done, but it would merit a paper of its own.Footnote 36 I will follow a different strategy instead, and sketch a plausible form of autonomy that people like Matt in union support exhibit. Following Thi Nguyen (Reference Nguyen2018), I call this delegational autonomy. Nguyen considers a case similar to particle accelerator in which one scientist has to trust another scientist who is a “logical alien” to them: their forms of reasoning, argument and evidence are fundamentally different from each other, and there is no feasible way to become competent in the other scientist's field. Still, Nguyen argues, if I trust the other scientist, I exhibit a form of autonomy:
It is the capacity to invest others with intellectual trust, even when their intellectual work is not transparent to us. … I am autonomous because I have chosen and delegated responsibility to somebody out of an active sense of trust and because I treat myself as responsible for the whole. (Nguyen Reference Nguyen2018: 116)
Delegational autonomy is a form of autonomy because we still control our beliefs and actions. When Matt trusts John in union support, the particular choice he makes is non-autonomous in one important respect. But this does not mean that Matt's choice is non-autonomous in all respects. Matt has autonomously chosen a general political commitment – to further the political interests of bricklayers. His deference to Matt is part and parcel of his commitment to that larger project. Nor is Matt's choice fully out of his control. If he encountered important defeating information – e.g., evidence that John is corrupt, or disagreement from other bricklayers with John's recommendation – Matt will stop deferring to John.
The pessimist might respond that delegational autonomy is still not full autonomy. It falls importantly short of the ideal of the democratic citizen who forms their views in a critical and independent-minded way. But remember what weak optimism claims: there is nothing morally deficient in deference, even if non-deference might be better. Thus, we should not be worried by the fact that delegational autonomy is a paler version of full autonomy. What matters is that delegational autonomy is a sufficiently valued form of autonomy to lift the moral stain from deference.
6.4. Deference in morally troubling societies
We can use these observations to answer another set of counterexamples. One might think, for example, of a patriarchal society in which wives are expected to endorse the views of their husbands, especially their political views. (There might be other hierarchical relations – e.g., children might be expected to endorse the views of their parents, and so forth.) We can imagine that this convention of deference has been internalised by many women, and is perhaps even endorsed by some.Footnote 37 Wives deferring to husbands in such a society seems morally troublesome.
We must first be clear that what is at issue here is epistemic deference plus action on that basis. It is not enough for a woman to vote or politically advocate in accordance with her husband's demands for this to count as political deference in our sense. What is also necessary is that she defer to her husband's beliefs, and on the basis of that deference, act politically. If you act in accordance with someone else's wishes without sharing their beliefs, you defer to their practical, but not their epistemic, authority. The de facto practical authority of men over women in patriarchy is morally objectionable, but not our current issue.
Imagine, however, that some women have come to so deeply internalise patriarchal norms that they do not just politically act, but also believe in accordance with their husband's demands. This is clearly deeply morally troubling. Can the optimist explain why? Two observations are important. First, we might wonder about the target of moral opprobrium in this case. Do we really want to say that it is women's deference which is morally problematic? It seems harsh to judge them for giving in to the unrelenting societal pressure to follow chauvinistic social expectations. Instead, it is the social convention – to do and believe as husbands demand – and those men that uphold and enforce it which are morally offensive. In addition, we should denounce the background conditions which normalise such social expectations and make them possible. In short, even in an extreme example like this, it is not clear whether we are troubled by the deference in itself, but rather the conditions and people that demand it; this gives the optimist ample opportunity to explain what is wrong in the given case.
The second observation is that deference under patriarchy might not be an expression of delegational autonomy after all, and thus not something the optimist is willing to defend. When it comes to autonomy in choosing what to do, the mere absence of coercion and outward force is insufficient. You also need to have the capacity, in some genuine sense, to respond to the merits of the actions open to you. Similarly, we might think that the many subtle forces which move women under patriarchy to defer to their husbands ultimately undermine, or at least severely stain, the degree to which that deference counts as autonomous. If so, then a major reason to defend deference in such cases falls away.
7. Concluding remarks
Let me briefly discuss three issues in closing. First, I can imagine a response, once again from the pessimist. This respondent accepts that deference in impure cases is often unproblematic, but responds that such cases do not fall under the scope of pessimism, which is only concerned with pure cases of deference. Alternatively, they might argue, the cases which I highlighted fall under the exceptions that pessimism already acknowledges (see my discussion of the “in most cases” modifier in §2). In short, this respondent objects, I have failed to deal with the core of the pessimist case which was never intended to apply to the cases I have considered. I am happy to accept that I have said nothing against the substance of such a scope-limited pessimism. But if that is the pessimist position, it strikes me as mostly of abstract philosophical interest, relevant to debates in meta-ethics and epistemology. Political philosophers who want to say something about the multitude of epistemic interactions in real-life democracies would have to look elsewhere.
Second, I have left it open what the relationship is between moral and political deference. In general, it seems to me that the pessimist case is stronger with respect to moral deference. Two arguments I have given, from the division of epistemic labour and epistemic solidarity, seem to develop more traction with respect to political problems than moral problems. Intuitively, the demands of epistemic autonomy are also stronger in the moral case, and appeals to delegational autonomy weaker. It might be that there is an asymmetry between the two areas: perhaps we should be more pessimistic about moral deference, and more optimistic about political deference. But establishing or rejecting this hypothesis would require more work.
Third, you might think that the weak optimism I offer is too weak. After all, weak optimism is compatible with both thinking that sometimes political deference is morally deficient, and that sometimes, while not by itself deficient, it is all-things-considered morally deficient. It is, in that sense, a weak pessimism as much as a weak optimism. I accept the charge, to a degree. I know of no convincing line of argument that a pessimist verdict for some cases could never be right. Indeed, a last suspicion I will offer is that any universalist view – i.e., a view which finds political deference to be either universally lacking or universally permissible – is likely implausible. We have to face up to the possibility that there are no interesting general principles for moral deference, but instead a patchwork of cases and cross-cutting principles. The failure of continuity supports this suspicion. The interesting question would then not be whether we should be optimists or pessimists in general, but for what specific range of cases we should take which position, at which degree of strength.Footnote 38 Putting this suspicion aside, I hope to have established at least one point – namely, that in a wide range of interesting, non-ideal cases like union support, optimism carries the day.Footnote 39
