It’s been a while since I’ve beaten up on Jason Brennan’s “argument” against character-based voting, but I’m feeling the urge again, so here I am, hot to go.(1) The crux of Brennan’s argument is that it’s wrong to vote for political candidates on the basis of their traits of character, except when character is a predictive proxy for the policies they can be expected to enact once in office. In a formula (Brennan’s formula, made in discussion here on PoT): “policy > character.” Taken literally, the argument proscribes voting against any candidate, no matter how evil, if the evil he exemplifies is policy-irrelevant. My aim here is to add yet another counter-example to my ever-growing list of counter-examples to Brennan’s thesis, partly for the understanding it affords, and partly for the fun of it.

Brennan’s rejection of character-based voting arises in the context of his extended case against universal suffrage. Brennan takes himself to have proven that the average voter is epistemically incompetent to vote; such voters (he continues) lack the knowledge, temperament, and capacity for reasoning that make for electoral competence. Once proven incompetent, for instance by a “voter qualification exam,” they should not vote. They should in fact be disenfranchised.(2) 

A critic might accept the premises of Brennan’s argument, but look for a way to bypass the conclusion. Yes, such a critic might say, the average voter is epistemically incompetent to vote, at least in the sense of being incompetent to vote on policy. But perhaps such voters can be held competent in a lesser or at least different way: they can vote on character. Judgments of character (the critic continues) are less epistemically demanding than judgments of policy. Held to this less demanding standard, perhaps the average voter still ends up satisfying an acceptable standard of epistemic competence. Such voters need not be disenfranchised. Indeed, it would be unfair to them to do so.

Contrary to the way Brennan runs the objection in The Ethics of Voting, the relevant issue is not that “voters tend to vote for character, not for policies,” but that they legitimately can.(3) Put more precisely: the character-based objection holds (or should hold) that there are electoral contexts such that the average voter can, in those contexts, competently vote on the basis of character rather than policy. In those contexts, disenfranchisement would both be unnecessary and unjust. Unless Brennan has a mechanism for filtering the relevant contexts and voters out ahead of time, his policy of disenfranchising all policy-incompetent voters is wrongheaded. If competence is multiply realizable, then policy-competence is too narrow a basis for disenfranchisement. 

Brennan insists in response that character is fundamentally irrelevant to electoral competence–with one (apparently minor) exception. The exception is that character is relevant to voting only insofar as character is a predictive proxy for policy. Otherwise, not. Brennan offers no clear or direct answer to the question of whether or to what extent character is predictive of policy. Indeed, what he says across the breadth of his writing on this topic strikes me as a mess of ad hoc-ery, inconsistency, and confusion.

But I’ll leave that discussion for another occasion. For now, it’s enough to point out that Brennan does nothing to exclude the possibility that the exception he allows swallows the rule he seems to want to defend. If character turns out to be a good proxy for policy, then character-based voting is justified. It doesn’t matter whether character-based voters know anything about the proxy relationship or not. The relationship could be entirely opaque to them. Yet if character is a proxy for policy, then if they vote correctly on character, they vote correctly. But if they vote correctly, and the correctness of their voting is non-fortuitous, they cannot be said to vote wrongfully or incompetently. So the character-based objection to Brennan was right all along. Rhetorical appearances aside, Brennan has “answered” it by conceding it. 

But that’s not my main point, either. I want to insist that character-based voting is justified whether it’s a proxy for policy or not. What Brennan’s critic is saying is that people lacking sophisticated policy knowledge are not as epistemically bereft as he makes them out to be. They can, on occasion–assuming it’s the right occasion–compensate for their lack of policy knowledge by voting on character. This is clearly the case when the following four conditions are met:

(a) The vote they cast functions as a veto against a candidate whose character falls beneath a certain threshold;

(b) The character-inferior candidate is running against someone whose character is above that threshold; 

(c) The two candidates are, on policy grounds, either roughly on par, or else the character-superior candidate is not markedly inferior to the character-inferior one; and

(d) The voter isn’t doing (a)-(c) fortuitously or indiscriminately, but is instead choosing in a counterfactually stable way to vote per (a)-(c) when it’s appropriate to do so, and not voting that way when it’s not. 

Conditions (a)-(d) are sufficient for the justifiability of character-based voting (though probably not necessary). It’s fairly easy to generate (or find) examples that fit this schema, and I think it’s clear that any example that does so is a counter-example to Brennan’s argument. Though we strictly speaking don’t need to adduce actual examples that do so, I have, over the years, adduced lots of examples that do. What I’m doing here is piling on yet another. 

You might wonder what epistemic gain there is in piling up examples of this sort. The answer, I’d say, is that each example gives us a slightly better understanding of the role of character in politics, a topic that’s been butchered by ideologues and political scientists alike. That’s interesting enough in its own right. Further, each example has slightly different practical implications, which are well worth noting. And each example (I’d say) confirms my suggestion that Brennan’s argument fails, helps explain why it fails, and shows something about the stakes involved in the failure.

Imagine that an American politician, call her Smith, takes federal office. While in office, Smith commits genocide. Let’s assume ex hypothesi that she really did commit genocide, and that she met all of the relevant conditions of moral responsibility for doing so. She was a primary actor; in other words, the genocide was in some sense primarily her doing, not something she was told to do. She knew exactly what she was doing; she wasn’t ignorant. She did it freely; she wasn’t coerced. She carried on with it for years, so it wasn’t some one-off genocidal accident. Evidence of the genocidal character of her actions was repeatedly brought before her, but she evaded, minimized, or dismissed it all in a craven, culpable manner. Confronted with the corpses of the dead and the suffering of the living, she insouciantly dismissed them all as irrelevant or even deserving, and proceeded unapologetically to committing genocide. Indeed, faced with particularly loud and adamant critics, she did what she could to have them silenced and punished. In short, to whatever extent she was ignorant, she made herself so. The world, Smith insisted, had to be unpeopled.(4)  And so it was, at her hands.

It doesn’t matter to my argument whether you find this a plausible portrait of an American politician or official, or of American foreign policy, or of the state of the world. If you do, no problem. If you don’t, go back and re-read the first word of the preceding paragraph, and take it to heart. 

Suppose that Smith finishes out her term in federal office. Now suppose that she runs for a lower-level office where foreign policy or military policy are ex hypothesi not involved–mayor, town council, school board, school superintendent, county clerk, even state governor, or university president. This kind of thing happens all the time, by the way. I filled an early draft of this post with examples of high-level political officials seeking and getting low-level electoral offices, but ended up deleting them. There’s no shortage of examples, but nothing in my argument turns on how many examples I pile up, or even whether any exist. All I need is a politically plausible possibility–not a bare conceivability, but something that really could happen in political life. And this could. 

Whatever the office, stipulate two things about it. First stipulate that military-type policy is irrelevant to the office. In other words, stipulate that nothing about the lower-level position involves powers that could be used to effect military policy, much less another genocide (or anything remotely like it). The powers and requirements of the office are constrained in such a way that you couldn’t carry out another genocide (or anything like it) even if you wanted to.

Second, stipulate that Smith’s prior commission of genocide is irrelevant to predicting her behavior in the lower-level office. Stipulate, in other words, that the vices she exhibited during her commission of the genocide were in some sense, specific to the occasion and entirely localizable. Yes, what she did was evil, but there’s no reason to think that she’ll repeat that particular evil (or anything like it) here. Yes, her commission of genocide involved the expression of deep moral and epistemic vices, but those vices were prompted by the specific situation that inspired the genocide, a situation vanishingly unlikely to recur here. Given the first stipulation (and as many others as we need), Smith’s vices are not predictive of genocidal-ish behavior as mayor, or on the town council, or as local judge, university president, animal control officer, etc. That, at any rate, is my stipulation for present purposes. 

If you find this implausible (and there’s no reason you should), just remind yourself of the human capacity for compartmentalization. A genocidaire in one context can be a nice guy in the next. You couldn’t legitimately infer that because Smith was a genocidaire, she was, let’s say, necessarily a domestic abuser, an animal torturer, an embezzler, or a litter bug. Those inferences simply don’t fly. She could be a genocidaire, but also a great wife and mom, a vegan, an assiduous respecter of property rights, and a neat freak. If you’re really having trouble with this (and you really shouldn’t), just imagine that the genocide Smith committed targeted a specific group, X, that’s completely absent from the locality where Smith is to take office. If Smith’s maxim is “genocide if and only if X,” then where there’s no X, there will be no genocide. Nothing to worry about–by stipulation.(5)

In short, I’m assuming that genocide is policy-irrelevant to the lower-level office in this case. Genocide has no bearing on the policies Smith can enact in that office, and Smith’s prior commission of genocide has no bearing on what she’ll do in office. As far as policy is concerned, the genocide is something we can put behind us.

Suppose it’s Election Day, and Smith is on the ballot for this lower-level office. It strikes me as intuitively obvious that you could “penalize” Smith by voting against her because of her prior commission of genocide. I bracket discussion of all the things you might justifiably do to Smith that go well beyond voting against her, like putting her on trial or killing her. Let those be topics for another time. For now, the question is restricted to just this one: can a voter use Smith’s commission of genocide in federal office against her when Smith is seeking political office where genocide is policy-irrelevant? I would say “yes,” obviously so.

There are, I think, two ways of justifying this “yes.” Either one works as far as I’m concerned, and each poses a problem for Brennan, as do both conjointly. I call the first “moral disqualification,” and the second “moral bankruptcy,” and regard them as distinct and irreducible considerations against Smith’s holding office.

Hind Rajab at her kindergarten graduation (photo credit: Maktoob Media)

Moral disqualification. We could say that Smith’s commission of genocide in federal office reveals a corruption of character incompatible with holding lower-level office, despite the policy-irrelevance of genocide to that office. In this case, as far as the voter is concerned, character directly trumps policy. Contrary to Brennan, office-holding is, in a weak sense at least, an honorific that presupposes that the office-holder is morally worthy of holding the office.(6) If a person falls desperately short of the threshold, she doesn’t deserve to hold office, and it’s permissible to treat that as a strike against her that has its own independent normative force. Prior commission of genocide is a large strike against a candidate. Since Smith has ex hypothesi committed genocide, there’s no office she’s fit to hold. She’s unfit herself from public life as such, regardless of whatever else we do with her.  

Moral bankruptcy. Alternatively, we could say that Smith’s commission of genocide demands a response in terms of rectificatory justice that’s incompatible with permitting her to hold office. In this case, character indirectly trumps policy by way of the requirements of punishment. By committing genocide, Smith stands in society’s debt, not the other way around. She’s the one who owes us; we don’t owe her.(7)  A fortiori, we don’t owe her our votes. What she deserves is a punitive response, not a reward (and you don’t have to be a retributivist to say that). Voting for her would violate justice, that is, would flout justice by grantly her the wildly undeserved.

Though desert is not always an overriding consideration in moral affairs, its normative force is proportional to its basis. The greater an infraction, the more deserving the perpetrator is of punishment (or the greater the debt she’s incurred). The more deserving of punishment, the greater the imperative to punish. The greater the imperative to punish, and the greater the punishment, the more incompatible with permitting the perpetrator to take political office. You can’t be mayor of Peoria while standing trial at the Hague. You can’t be mayor of Peoria while serving out your sentence for genocide, either. 

The moral bankruptcy argument is one from rectificatory justice rather than character as such, but it’s still a problem for Brennan. Brennan’s explicit thesis is that policy trumps character, but that implication follows from the more general assumption that policy is the trumping consideration in electoral contexts, so that policy trumps non-policy. In the example under discussion, rectificatory justice is policy-irrelevant to lower level office. It’s a case of non-policy relative to that office. In other words, the reasons why Smith has incurred a debt to society, and should be barred from office because of it, have nothing to do with the prospective policies she would enact in lower-level office. So while character doesn’t directly trump policy in this case, a policy-irrelevant consideration of justice ends up doing so. 

What this reveals is that there are such things as policy-irrelevant considerations of justice, and that these considerations can in principle trump policy in electoral politics. In short, policy is not the self-evident trump card that Brennan takes it to be. Brennan ignores this possibility, but that’s a defect in his argument. 

Obviously, the preceding two considerations are compatible with each other. You could say that Smith doesn’t deserve office, full stop, and that she deserves to be arrested, tried, convicted, and punished, which is ipso facto incompatible with taking office. Whether taken individually or jointly, we reach the same conclusion: Smith has no business being in office, and it’s wrong to put her there. 

If genocide is the crime of crimes, then the commission of genocide is the trump of all trumps. That means that if you’ve committed genocide, you should simply be ejected from politics, full stop. That may be an understatement, but understatement or not, it’s true. De-genocidation trumps policy.

It’s worth pausing to consider some of the practical implications of the preceding claim. The basic implication is that de-genocidation hugely simplifies political deliberation by reducing it to a single issue. If you’re a genocidaire, and up for local office, we don’t have to sit around parsing your views on budgetary matters, or whether to install the new HVAC system in the middle school, or whether there should be tampons in the boys’ bathrooms or trans athletes on the swim team, or whether a stop sign should be installed at the corner of Maple and Spruce, or whether the Gen Ed Curriculum should give way to a STEM-based one. We don’t have to care what you think about immigration, or the minimum wage, or health care, or the crime rate (!), or anything else. It doesn’t matter that you’re a great mom, a vegan, a stickler for property rights, and/or assiduous about not littering, either. All that matters is that because you committed genocide, you need to be taken out of politics, whatever your other offsetting virtues. Full stop.

The only real complication here is how proximate to the genocide a given candidate was. Primary or secondary? Actively engaged or merely complicit? Acting voluntarily or under some duress? Fully knowledgeable or non-culpably ignorant? Etc. But as far as Smith-like actors are concerned, the relevant deliberation isn’t that hard. Being a primary actor in a genocide entails your wholesale ejection from public life. So we can hit eject without further deliberation or apology. 

Notice also that while I’ve used the move from high-level to low-level office as an expository convenience, the practical implication of my claim applies to all cases, not just that one. I used the high-to-low office case as a convenient way of illustrating the fact that the policy implications of past genocidal behavior are irrelevant to a decision about whether to vote for a genocidaire. So I cooked up a case in which the past commission of genocide had no policy implications for a given office. Since the conclusion of the argument is that genocidaires should never hold office, however, the practical implication for other cases is as clear as the implication for the case I happened to discuss. We should vote against genocidaires wherever they are, whether they can repeat the past genocidal behavior or not. It makes things worse for them if they can, but it doesn’t change things if they can’t. 

Some philosophers might be inclined to call this a deontic conception of politics, but I prefer to regard it as a justice-based rejection of consequentialism. It’s just a bedrock truth (not necessarily a deontic one) that no legitimate government or institution can be run by genocidaires, no matter who they are, or what promises they make, or what excuses they invoke for what they did. Just as evil may not be done that good may come, so evil may not be rewarded that good may come. Rewarding evil just is an instance of doing it.

All of this might sound excessively Kantian to some ears, and it’s perfectly reasonable to think that Kant took retribution too far.(8) But nothing in my argument turns on any attachment to Kantian deontology, or to Kant’s theory of punishment, or even to retributivism generally. My concerns lie elsewhere. The problem with our political culture, as I see it, is that it’s addicted to a form of amnesia. “What happened, happened. Let it go.” That’s what I’m rejecting. Kant may have taken punishment too far, but we’ve taken “forgive and forget” too far. Some things aren’t forgivable. Those things should never be rewarded. Not everybody deserves a prize. 

Like the rest of his epistocratic brethren, Brennan is a pessimist about the epistemic competence of ordinary people. Democracy will never work, these episto-bros complain, because politics corrupts our capacity for judgment, and turns us into mindless hooligans. I have mixed feelings about this verdict. There are contexts in which it’s clearly true, but Brennan strikes me as systematically insensitive to the contexts in which it’s clearly false. In any case, it seems to me that the issue of genocide greatly simplifies things. What if voters had a really simple lesson to learn? What if we merely educated them, or expected them to educate themselves, on whether or not their leaders had committed genocide or something like it? We teach school-aged kids about Hitler, Anne Frank, and the Holocaust. We teach them about slavery and Jim Crow, too. Since those lessons have been learned, it seems plausible enough to think that lessons like them could be learned. What if, when the facts warranted it, we taught people or expected them to learn, that they were ruled by a bunch of Nazis and Klansmen who’d made victims of another generation of Anne Franks and Emmett Tills?

I mean, if it was true, it would be hard to see the excuse for not teaching it. Whatever the specifics of the lesson plan, students (whether as children or adults) would be taught never, ever to trust or reward those to whom the lesson applied. It would eventually become common knowledge that genocidaires are to be treated as moral pariahs–despised, reviled, and excluded from public life. You might say that this is an awfully recriminatory lesson for people to learn. I agree. On the bright side, recrimination is something easily taught and learned, and has clear, determinate implications for action. If there’s some other objection to it, let me know.

Imagine a society in which commission of or complicity in genocide was widespread. Commission of genocide, in such a society, would not be a matter of a Smith here and a Jones there to be excluded from public life, but of the moral disqualification and/or bankrupticy of whole classes of people who’d made a culpable contribution to a large-scale project of extermination. However terrible this situation might be, it would greatly simplify electoral politics. Voters would simply be tasked with identifying the culpable and voting against them, ignoring all other considerations. Electoral politics would then become straightforward naysaying, a perpetual vote of no confidence in a largely guilty ruling class. I don’t claim that this is an inspiring or cheerful prospect. I simply claim that it’s easy enough to pull off. There’s no reason to worry about the intricacies of policy, or to wring our hands over the electorate’s inability to learn microeconomics. There would just be the imperative to hunt down the guilty, and hold them at bay until all of them were good and dead.

I know that epistocrats like Brennan have all kinds of supposedly rigorous arguments intended to show that people are too “rationally irrational” to learn anything significant about politics. I’ve read them all, but none of them impress me. It’s only too obvious that the social science these episto-bros invoke (and often mischaracterize) is inadequate to capturing relevant social realities. Social reality outstrips and outpaces anything you’re going to read in a peer reviewed journal. By the time it gets there, assuming it’s replicable at all, it’s likely to be old hat you already knew. And the world is just bigger, more varied, and more complex than anything you can call up on JSTOR. 

I spent twenty-six years teaching philosophy to underprepared college students, seven of them working for the National Assessment of Education Progress, a good chunk of that working on the Civics Assessment, a modest chunk of that as an item-writer and field-level scoring supervisor for the assessment. I’ve spent time as an educator in the Occupied West Bank, in Nicaraguan slums, in urban minority-majority universities, and obscure community colleges. Some of the students I taught were functionally illiterate and innumerate. Some were functionally homeless. Some were brutalized. Some were chronically hungry. Yet I had pedagogical successes in the least likely and least promising places. There is no social science that even begins to capture the qualitative experience on either end of that exchange. You could pore through hundreds or thousands of pages of work on “transfer of learning” or trend-line assessments of educational outcomes, and not scratch the surface of what happens when instructor meets students in the classroom. Educators live for outliers. So do students. But try to discover this truth in the literature on “educational outcomes.” 

At any rate, I’m optimistic enough to think that the lesson I have in mind here can be learned by anyone motivated to learn it–not just in the abstract, but as something to be weaponized against the “experts” who have yet to learn it. The issue is not cognitive capacity but motivation.

And if you find the right audience, the motivation is there. “Disempower the enemies of humanity” is a lesson I could take into Deheishe refugee camp, into the slums of Managua, or into the most fucked-up neighborhoods of the United States–Camden, Newark, Pine Ridge–with the full expectation of being fully understood. As an activist for Resistencia en Acción, a migrant-defense group, I work every day with people who, despite lacking very much formal education, show a better grasp of these issues than the academic philosophers I knew when I was an academic, or the talking head “experts” who hold forth at the School of Public and International Affairs a few blocks away.

Despite their lack of formal education, the average migrant has intellectual virtues that the average “expert” conspicuously lacks. Victims of systemic injustice have many cognitive faults, to be sure, but they tend to understand systemic injustice in ways that privileged, cossetted experts often don’t. They understand, as well, that the perpetrators of systemic injustice have to be dealt with, not tolerated, rewarded, or evaded. They understand, finally, that those who reward evil are complicit in it, and that the supposedly sophisticated rationalizations for rewarding evil are essentially noise intended to distract attention from this fact. For present purposes, that’s all they need to understand. Could I raise an army or a sizable voting bloc of such people? I don’t know. But I don’t need to raise an army or win an election to have a good counterexample, which is all I claim to have.

Endnotes

1. I’d like to dedicate this post to Chappell Roan, who got the issue half-right. I try not to care about the part she got half-wrong, but I have to admit: it hurts my feelings. 
2. See Jason Brennan, The Ethics of Voting (2011), and Against Democracy (2017). The disenfranchisement proposal, described as an “experiment,” is laid out in chapter 8 of Against Democracy, pp. 211-214. The same chapter contains a proposal for “universal suffrage with epistocratic veto,” in which an epistocratic council with a power to “unmake law” stands in judgment of the decisions of the general electorate. Unsurprisingly, members of this council are required to pass “rigorous competency exams,” but for reasons Brennan doesn’t explain, membership “may also require some sort of character check” (Against Democracy, p. 216). Unless Brennan can show that these character checks are predictive proxies for rectifications of bad policy, he isn’t entitled to them. And he makes no effort to show this. Indeed, he evinces no recognition that anything of the sort has to be shown. 
3. The Ethics of Voting, p. 84. 
4. To paraphrase Benedick in Shakespeare’s Much Ado about Nothing, Act 2, Scene 3, line 244: “The world must be peopled.” 
5. Granted, this is a stipulation. I don’t mean to be denying that outside of the stipulation, a genocidaire could find a way to commit genocide in a lower-level office, or that genocidal traits could corrupt the genocidaire’s policy enactments in such an office. Nor am I passing judgment as to the probability of any of that happening. I’m simply asserting and relying on the reverse possibilities for purposes of the example. Real world violations of the stipulation would not adversely affect my argument. 
6. The Ethics of Voting, p. 84. 
7. Cf. Daniel McDermott, “The Permissibility of Punishment,” Law and Philosophy 20 (2001), pp. 403-32. Though McDermott describes his view as “debt-based retributivism,” it seems to me to straddle the line between a retributivist and compensatory conception of punishment. In any case, McDermott nicely formulates the general idea I mean to convey here: a rights-violator incurs a series of debts both to the victim and to society as such in the act of violating rights, and in committing the offense, incurs the obligation to pay that debt to both. In saying “we don’t owe her,” I don’t mean that we literally owe her nothing in the way of respect for her rights, but rather that Smith’s debts to society far exceed society’s to her.
8. Immanuel Kant, The Metaphysics of Morals, II.1.E, Ak. 331-337.

Thanks to Susan Gordon for the constant stream of Facebook posts that informed this post, as well as to Hilary Persky for some helpful conversation. The usual caveat applies.