Now that Bleeding Heart Libertarians has closed up shop, the BHL diaspora has begun. Jason Brennan, Chris Frieman, and Jess Flanigan – all friends – have started up a new blog with a different approach – 200-proof liberals. I recommend subscribing to it, though it is meant to have a very different rhetorical style than the BHL project.
J has posted a summary of a piece he has written criticizing public reason views for failing to grapple with what we know about political psychology, especially that people engage in post-hoc tribal reasoning in politics. But public reason liberalism, the dominant approach to legitimacy in political philosophy, supposes that people reason independently of their tribe, at least to some extent, such that clear rational commitments can be identified, and justifications for laws and policies can then be crafted in terms of those fairly stable reasons. So public reason liberalism is wrong because it depends on people having more stable rational commitments than they actually do.
Obviously I disagree with this argument, which Brennan has constructed by drawing on the new Achen and Bartels book in particular. Fortunately, I think the argument is good enough that I devote some time to it in my forthcoming book, Trust in a Polarized Age. Here I’m going to reproduce much of what I say there. But before I do, let me just say this. I think Americans these days tend to be more impressed by the irrationality of political behavior than Europeans. Many Europeans countries aren’t especially polarized, and some of these countries are very trusting, like the Nordic countries, and so I don’t think the trends J is pointing to are universal features of democratic polities. People are politically tribal to varying degrees, and the US is in a particularly but peculiarly bad place right now relative to some other democracies. So if we take an international perspective, I think the idea that politics is necessarily war looks less plausible even from 30,000 feet. OK, with that, here’s what I have to say about the main line of argument.
I summarize the Achen-Bartels view J appeals to in section I. So if you know the view, skip to section II.
I. Achen and Bartels’s Challenge.
The thesis of Achen and Bartels’s Democracy for Realists is that elections don’t produce responsive government. Instead, “voters, even the most informed voters, typically make choices not on the basis of policy preferences or ideology, but on the basis of who they are—their social identities.” The worry is that “if voting behavior primarily reflects and reinforces voters’ social loyalties, it is a mistake to suppose that elections result in popular control of public policy.” Instead, election outcomes are simply “erratic reflections of the current balance of partisan loyalties in a given political system.” I have already addressed the possibility of using small-scale deliberative bodies to reveal the preferences and reasons of citizens in chapter 7, but Achen and Bartels argue that empirical studies of these small political bodies “seem to be less relevant for understanding democratic politics on a national scale.” For the purposes of this section, I will assume that microcosmic deliberation cannot serve as a substitute for national elections.
Part of Achen and Bartels’s argument is that citizens lack enough information and motivation to engage in retrospective voting, voting based on the past performance of political parties and candidates. In general, Achen and Bartels claim to cast “considerable doubt on the view that citizens can reliably form and act upon sensible retrospective judgments at election time.” Retrospective voting is complicated by three factors: low political knowledge, mistaken assessments of one’s own well-being, and a limited time horizon for economic accountability where citizens only attend to their economic condition in proximity to a national election.
Achen and Bartels embrace a group theory of politics. This means that “the primary sources of partisan loyalties and voting behavior … are social identities, group attachments, and myopic retrospections, not policy preferences or ideological principles.” This is so in part because extra information doesn’t make citizens more accurate; instead, “party identification colors the perceptions of the most politically informed citizens far more than the relatively less informed citizens.” Citizens are not generally responsive to information about political parties, candidates, and the like, but rather filter the information to empower their group and disempower out-groups.
Achen and Bartels think their research has profound implications for democratic theory. It leaves democratic theory “in a shambles” because all “the conventional defenses of democratic government are at odds with demonstrable, centrally important facts of life,” specifically human limitations in information and motivation.
II. Preliminary Difficulties
I think Achen and Bartels overstate their case. One reply is that voting based on group membership can be an informational shortcut insofar as one’s group allegiance leads one to support the candidates one’s reasons would lead one to support. For instance, if black Americans vote based on how leading black legislators tell them to vote, that may indeed be a reasonable heuristic that helps protect them from serious harms by more powerful social groups. Thus, voting based on group membership, while sometimes problematic, may not be irrational. We can only show that it is irrational if we can show that following the views of the leaders of one’s group leads one to embrace more objectionable policies than one might otherwise accept.
Second, Achen and Bartels need to argue that because we determine our policy preferences by group membership, those preferences are unjustified. But what makes our beliefs rational is not how they came about (the causes of the beliefs) but whether the beliefs are justified by good reasons at present (the sustaining justifications of the beliefs). Thus, to show that our political preferences are problematically arbitrary, it is not enough to show that the beliefs are caused by group membership; it has to be shown that the influence of group membership shows that political preferences are irrational. After all, most of our beliefs are heavily affected by all kinds of nonrational factors, but this fact does not by itself undermine the rationality of our beliefs. Achen and Bartels can reply that policy preferences and related beliefs change so erratically and quickly that it is hard to imagine that these beliefs could be justified while they’re held. Rational beliefs should be stabler, since the reasons to support various policies do not change erratically and quickly. However, if it turns out that making political judgments is very difficult even for highly rational agents, then the evidence we have might not force us to hold one position or even limit us to a handful. If we move between reasonable, incompatible beliefs, even unpredictably and quickly, then that may merely mean the issue at hand is hard to resolve, not that the public’s preferences are irrational. The point is that while the causes of a belief or position might be arbitrary and tribal, the belief or position might still be a reasonable one.
But the best way to answer Achen and Bartels is to examine how their arguments challenge particular approaches to justifying democracy. So, I will now address how their views might undermine my public-justification approach. The basic defense I offer is that Achen and Bartels at best show either that most people have bad reasons to support, or that most people have no reasons to support, whatever policies they support—that most people do not have genuinely justifying reasons. But neither claim is a good reason to abandon a public-justification approach.
III. Achen and Bartels vs. Public Reason – The Tracking Objection
The data that Achen and Bartels cite does not actually challenge my approach to the public justification of democracy as a whole. Instead, they at best can be used to argue that a right to participate in elections will not tend to produce publicly justified laws, but there might be other public justifications for democratic elections.
So, the refined challenge is that exposing law making to elections may yield fewer publicly justified laws and policies than rules that limit the range of laws and policies that elections affect. This suggests that a publicly justified polity should rely more heavily on epistocratic mechanisms to insulate the selection of political officials from direct citizen input.
Let us call laws and policies that depend in some direct way on electoral outcomes democratically exposed. Laws might be democratically exposed through referenda or in directly elected legislative bodies. Achen and Bartels’s arguments can be used to challenge my claim that democratically exposed laws and policies will track the reasons of moderately idealized agents (note to blog readers: I think the reasons that figure into public justification can be identified by asking what a moderately idealized version of persons would endorse, save with enough time and information to adequately reflect). This is because, given extreme group loyalty, democratic outcomes will be sensitive to arbitrary factors that undermine any connection to what moderately idealized agents support. Call this the tracking objection. What’s worse, the data suggests that many citizens may not have justificatory reasons at the right level of idealization because they have no settled views about what to support or oppose. Thus, democratically exposed laws and policies will tend to reflect arbitrary and biased preferences and group loyalties. Call this the bad reason objection.
The tracking objection can be answered by conceding that citizens’ political preferences and behaviors can be affected by their group while insisting that political parties may nonetheless be good proxies for what moderately idealized versions of their members would support. One reason for this is fairly obvious: if citizens are committed to valuing what their in-group values, and their in-group is their political party, then what the parties favor is directly tied to what citizens favor. Thus, insofar as moderately idealized citizens are committed to their party’s success and effectiveness, their good reasons will be determined by what the parties favor. Compare a citizen’s commitment to the scientific community’s findings. If a person has the reason “Generally believe what the scientific community believes, because scientists are rational people, and I’m a rational person too” then that person has sufficient reason to defer to the findings of the scientific community. In the same way, if a person has the reason “Generally believe what my party believes,” then that person has sufficient reason to track what the party believes.
One might respond that moderately idealized agents will give less weight to the opinions of their group vis-à-vis good arguments for choosing policies and political officials. Moderately idealized agents will endorse laws and policies on the basis of decent information and arguments, and those determinations are unlikely to mirror what their group affirms. And yet, while this is so, many political issues are sufficiently complicated that boundedly rational agents will not be able to form opinions on all issues, so they may nonetheless want to use party or political group support as a heuristic. Along the same lines, parties are not always biased in favor of power and a desire for oppression. In many democratic societies, parties track political ideologies and the moral values of large swaths of the electorate. They also articulate platforms and arguments that are loosely responsive to evidence. So citizens may not act irrationally in embracing party platforms, and their endorsements may well track their justificatory reasons. Their motivation for adopting the party platform may be largely tribal, but they may yet have good reason to endorse that platform all the same.
A second objection is that, in societies with significant partisan divergence, especially when parties are more divergent than the general public, parties may be subject to greater biases than ordinary people. In that case, the heuristic “Generally believe what my party believes” may lead people to epistemically worse beliefs. Consequently, the heuristic may not help people’s opinions track their justificatory reasons but, rather, make them less likely to recognize what they have good reason to believe even by their own lights. But this leads into the bad reason objection.
IV. Achen and Bartels vs. Public Reason – The Bad Reason Objection
The bad reason objection says that, since people typically have tribal preferences, moderately idealized versions of tribal people will endorse not good reasons but reasons that we ordinarily would not want to have figure into public justification. They may, for instance, change their views on the appropriate foreign policy with respect to Russia based on the fact that President Trump says good things about Vladimir Putin. This implies that we should not base laws on the reasons that moderately idealized persons will typically, normally endorse.
I have two replies to this objection. First, even if many citizens have merely tribal reasons to support a coercive policy, that merely means at worst that we cannot have that policy. And if the law is not coercive, or if persons regard the process for selecting the law as legitimate, then we can permissibly impose the law on them because they are too flighty to have sound objections. We must of course ensure that the law is publicly justified for persons who do embrace good reasons. But the mere fact that most members of the public lack good reasons to favor or oppose many laws and policies does not make a public justification standard inappropriate. It just means that the public justification of the law depends primarily on the reasons of that minority of the public whose reasons should figure into public justifications.
Second, it is implausible to think that persons will have no good reasons at a moderate level of idealization simply because they exhibit arbitrary and easily altered preferences in the real world. Deliberative polls, for instance, often yield stable judgments, which even Jason Brennan acknowledges have promise. And deliberative polls are arguably proxies for moderate idealization since they subject opinions to critical scrutiny and proffer better information.
Moreover, plenty of empirical work in psychology demonstrates “value stability,” the empirical finding that the values that persons embrace are remarkably stable over the course of their lives. Thus, insofar as citizens can trace inferential routes from their stable values to their support of or opposition to particular laws and policies, there is a basis for ascribing good reasons to them because laws and policies can clearly advance or undermine their stable moral values.
For this reason, I conclude that a a public reason liberal case for democracy is not in shambles, despite Achen and Bartels’s important work. Their work does challenge the position I defend, but their objections can be answered. Perhaps in future work, Achen and Bartels, or other democratic theorists, will tease out the way in which their findings challenge public reason liberalism, but for now we can proceed without too much worry.