Narendra Modi is the Prime Minister of India, recently on a speaking tour in the US and elsewhere, and promoting his nationalist agenda for India wherever he goes. He’s very much the talk of the town, which is a great irony, considering that less than a decade ago, he was in effect barred from entering town: he was denied a visa to enter the United States for his failure to stop the Gujarat riots/massacre of 2002, and boycotted in Europe for much the same reason. Back then, he was a relatively obscure figure, at least by international standards. Now, he’s the Prime Minister of a major world power. Evidently, the principle involved here is that if you play a presumptively culpable role in a massacre, you’re to be treated with contempt until you assume the trappings of power. Once you do, the passage of time and victory in a democratic election jointly wipe the slate clean, and you’re to be treated with respect and admiration, bygones being bygones–even if some of the bygones include a few thousand corpses for which you were plausibly thought to bear responsibility. Wait and win, and you’re out of the moral doghouse. It’s an interesting lesson–about India, about democratic politics, about historical memory, and about justice. Worth remembering for other contexts.
I bring the issue up not (merely) to moralize about Modi (whom I admittedly despise) but as a counter-example to the discussion of character-based voting in Jason Brennan’s much-praised and much-discussed book, The Ethics of Voting. I generally agree with the thesis of Brennan’s book, at least as I understand it, which is that epistemically incompetent voters ought not to vote. But having edited a symposium on the book in Reason Papers last year, and having read some of the discussion of it there and elsewhere, I’ve been surprised at how much of what is contestable about Brennan’s argument has gone entirely undiscussed and uncontested by his peers, peer-reviewed and otherwise. Brennan’s discussion of character-based voting is one such example (EV, pp. 84-85), and the case of Narendra Modi conveniently serves to focus the issues.
The discussion of character-based voting comes up in chapter 3 of the book, which is devoted to explication of the concept of “wrongful voting.” “Unexcused harmful voting occurs when a person votes, without epistemic justification, for harmful policies or for candidates likely to enact harmful policies” (EV, p. 68). Brennan’s point is that we ought to refrain from engaging in unexcused harmful voting. It follows that citizens should vote only if their beliefs about the prospective harm or welfare-conductivity of policies are epistemically justified–probably a small minority of actual voters. The two crucial concepts here are epistemic justification and harmful policies. Brennan doesn’t explicate either, but assumes that on some version of both, his thesis turns out to be correct.
One objection to Brennan’s view is that we might vote for or against someone on the basis of character, not knowledge about the welfare-conducivity of the policies they intend to enact. Take some political candidate, X. My beliefs about X’s policies may be unjustified, epistemically speaking; they may be vague, vacuous, or based on very little evidence. But my beliefs about his character may be perfectly on target. Suppose I correctly regard X as (very) immoral, correctly regard Y as morally decent, and vote for Y because I regard X as so immoral that a vote for Y is preferable to one for X even if I’m generally (though not completely) ignorant of the relative policy implications of voting for X versus Y. Assume that Y’s policies will predictably be worse than X’s (though not egregiously so), but X’s past is egregiously unjust whereas Y’s is perfectly decent.
According to Brennan, unless I treat my judgments of X’s past immorality as proxies for predictions about X’s future policies, my vote is a case of unexcused wrongful voting no matter what X might have done in the past. It’s not clear why, however, and he doesn’t make it clear. What if, as a voter, I weight character over policy as a criterion for voting, at least in cases as egregious as those like X’s? Why is that unexcused wrongful voting? Or is it excused wrongful voting?
Let X be Narendra Modi, and ex hypothesi assume the worst about Modi’s role in Gujarat. Further (also ex hypothesi), assume the worst about Gujarat. If I were an Indian citizen voting in the last election, I would have voted against Modi simply because his role in the Gujarat affair put him beyond the pale for holding the position of Prime Minister of India. The issue is not so much that I expect Modi to repeat his past behavior; in fact, it’s unlikely that he will. Given the scrutiny he’s gotten over the years–and the visa denials and boycotts, etc.–Modi is likely to be more careful about how he expresses his nationalist sentiments, and is likely to overcompensate for his past sins, at least in policy contexts. Nor is it that I think that Modi’s policies are likely to be inferior to those of his political rivals. They may actually end up being a bit better, at least from a free market perspective. The relevant point is simply that Modi ought not to be rewarded for his past behavior, and voting him into office is a reward (a huge one), one that evades the moral significance of that past behavior. Justice demands that we not grant the unearned, and Modi’s past behavior disqualifies him from earning title to the office he now holds.
For purposes of this post, I don’t want to go into the factual details of Modi’s actual (past) behavior. Doing so is unnecessary, since my aim here is to contest Brennan’s discussion of character-based voting, and to that end, Modi simply draws attention to a relevant possibility–viz., the evil political candidate who is likely to enact better policies than the merely decent one. So if you’re not familiar with Modi’s past, or you disagree with my interpretation of it, we can simply imagine a Modi-like figure and use that as a point of departure for thinking about Brennan’s argument.
So here are my stipulations: Imagine a Modi-like candidate for office who has behaved disgracefully in an affair like the Gujarat massacres of 2002 (or worse). Imagine that he showed callous disregard for the lives of his fellow citizens when he had the responsibility to protect them. Imagine that thousands of innocents died as a result. Imagine that there is good evidence that his own nationalist political agenda explains the animus for those massacred, motivated those who killed them, and rationalized the killings (using “rationalized” in the colloquial, not Davidsonian sense).
Now fast forward about a decade. Imagine our Modi-like candidate going up for the highest office in the land. Imagine that he claims to have cleaned up his act, having done an about-face from his bad old days a decade ago. Imagine that there is no serious question that his forthcoming policies will revert to the ways of his bad old days. Assume that his policies will actually be an improvement on what the country currently has, and what the other candidate has to offer. Now imagine a rival anti-Modi candidate who is morally decent (and in particular, critical of Modi’s behavior in Gujarat). Let his policies be relatively indeterminate–not great, but not of the sort that might lead to any policy disaster. (Take the party you dislike most and imagine them enacting a safe and pragmatic version of politics as usual.) Let them just end up being worse at the margins than those of the Modi-like immoral candidate. Brennan’s thesis implies that we ought to vote for the immoral candidate, not the decent one. But why?
Here’s what he says:
To a significant degree, voting for character is voting for wrong reasons. When we elect someone, we give him power. That power can be used for good or bad. The office of the presidency is not an honorific meant to show [that] we respect that person’s character. Giving someone the presidency is not bestowing a medal or a certification of commendation but giving him (some) control of the state, an institution that makes rules, and forces innocent people to comply with these rules using violence and threats of violence. We need to be sure he will do a good job controlling it.
The first sentence refers the reader in a footnote to Russell Hardin’s How Do You Know? The Economics of Ordinary Knowledge and Bernard Manin’s The Principles of Representative Government. I haven’t read either book, so if my objection is answered there, I’d concede my case. But for present purposes my point is that my objection is not answered by Brennan in The Ethics of Voting: the passage above does nothing to answer what I regard as the most obvious objection, posed by Modi-like cases. It begs the question and offers an ignoratio elenchi.
Brennan starts by telling us that voting for character is “to a significant degree” voting for the wrong reasons. But I’d respond that it’s a principle of justice that we ought not to reward wrongdoing. Modi-like people have, to put mildly, engaged in serious wrongdoing. So justice entails that we ought not to reward them. The issue then turns on whether a vote for high office like Prime Minister of India is a reward. I think it obviously is: you reward someone when you give them what they want, and that in turn gives them power, prestige, income, a place in history, an opportunity to enact their values on a wide scale, and other perks they couldn’t otherwise have gotten.
Brennan’s discussion bypasses this last fact. The office of the presidency, he tells us, is not an honorific. Why not? He gives no argument for its not being an honorific, and doesn’t consider the possibility that it might constitute a reward without being primarily an honorific. To the extent that he thinks he has an argument, he simply adduces a different fact about the presidency that supposedly rebuts the claim that the presidency is in part an honorific–that the presidency involves the exercise of power. But this fact, though true, doesn’t really help his case. Political offices could both confer rewards on office-holders and give people control of the state. The one function is perfectly compatible with the other, especially when control is a reward.
Yes, we need to make sure that those who hold power control their use of it. But we can do this as long as we have very vague ideas about a person’s policies. As long as I know that Modi’s rivals aren’t going to enact crazy policies that will take India to perdition, and I believe that they’re decent people lacking an outright lust for power, I know they will manage to control power somehow. (Brennan basically concedes this point in his paper “The Right to a Competent Electorate,” when he says: “We should not overestimate the damage bad voting can do…Even in the US or the UK disastrous candidates rarely have a chance of winning….” p. 707). The wrongs they do via bad policies may be bad, but I may find them more tolerable than the wrong done by rewarding someone responsible for mass murder with the perks of high office. I may be ignorant about the details of the relative merits of Modi’s versus his rival’s policies. But as long as I know that Modi is immoral or evil, and that his rival is tolerable, I may know all I need to know to vote against him.
In other words: If I am more committed to the backward-looking principle of not rewarding past evil than I am (within limits) to forward-looking considerations about good or bad policies, it seems to me I can justifiably vote for the policy-suboptimal candidate so as to avoid voting for the evil candidate even when the evil candidate is more likely to enact better policies. When I do so, I’ll have some vague thoughts about policy considerations, but I’ll have epistemically justified thoughts about character, and the thoughts I have need not be proxies for future policies but might instead be motivated by considerations about not rewarding past immorality (assuming that that immorality crosses a certain threshold, e.g., culpable involvement in mass death rather than, say, adultery or smoking pot). I don’t see that Brennan even considers this possibility, much less rebuts it.
A digression: Brennan writes of “the presidency,” but I assume that he means any electoral office, including everything from municipal judge to American presidency to prime ministership, etc., including offices whose perks are enormously large. It’s an odd feature of Brennan’s book that while its topic is “the ethics of voting” as such, his focus is almost exclusively American, as though American voting were the paradigm of the phenomenon of voting as such, and as though empirical work on American voting generalized to voting everywhere. I raised this issue in an editorial context with one of the contributors to the Reason Papers symposium (not Brennan), who told me that as a social scientist it was his view that the default position is that empirical work on American politics ought to generalize to politics everywhere unless it can specifically be proven that it didn’t generalize to some particular context elsewhere. In other words, if Jones cites a study on American voting patterns, Jones can, absent contrary evidence, assume that the American voting patterns generalize to Pakistanis, Indians, Italians, or Palestinians. Generalization from the American case is the default rule. That strikes me as a pretty bizarre methodological assumption, but I’ll let the bona fide social scientists fight over it.
Brennan devotes one more paragraph to the topic, but as far as I can see, it merely elaborates on the conclusion Brennan thinks he’s established in the preceding passage.
So character-based voting is acceptable only insofar as it is a proxy to the quality of the governance a candidate is likely to produce. To what degree good character and good policies are correlated is largely an empirical question. If someone is morally corrupt, there is a pretty good chance he will use the power of the state for personal benefit rather than to promote the common good. Yet a virtuous politician with a powerful sense of justice might still be deeply misguided and committed to all sorts of counterproductive, harmful policies. Having the right values is not sufficient for making good policy, because it requires social-scientific knowledge to know whether any given set of policies is likely to achieve those values…If there is good evidence that a politician is likely to enact harmful policies, one should not vote for her (without sufficient reason) even if she is a good person. Voting on the moral virtue of a candidate counts as good voting only to the extent that the candidate’s moral virtue is evidence that she will enact good policies.
(1) My first and most basic comment on this passage is that it doesn’t address the objection I’ve raised, and doesn’t address what seems to me the most obvious objection that could be raised.
(2) My second comment is that it doesn’t cohere very well with the preceding passage. The first passage told us that voting for character was “to a significant degree” voting for the wrong reasons. The second passage tells us that voting for character can be voting for the right reasons under certain circumstances, and it’s an open empirical question to what degree good character and good policies correlate.
So which is it? If it’s an open question whether good character and good policies correlate, they might well correlate, and there are (as Brennan himself seems to admit) commonsense reasons for thinking that they do correlate. If so, why is voting for character “to a significant degree” wrong? Why is it wrong at all? And how can we know to what degree it’s wrong if it might turn out to be right?
Perhaps Brennan means to say that we don’t know whether it’s right, and if we don’t, it can’t typically be epistemically justified to use character as a basis for voting since no one has the social scientific evidence in hand to demonstrate the relevant correlations. But since he himself admits that there is plausibility in the idea that character and policy are correlated, it’s not clear that the claim about epistemic justification follows. Is p only epistemically justified in political contexts if we have a peer reviewed study (or set of them) showing us that p is the case? If I infer that a habitual liar and promise-breaker will be an unreliable implementer of good policies, is that inference epistemically unjustified? Why isn’t banking on a “pretty good chance” good enough?
I don’t mean that there’s an outright contradiction or inconsistency here; I mean that there’s a failure of exposition that leads to an obvious and unresolved puzzle in the reader’s mind about what Brennan is saying. The failure of exposition arises from the legalistic character of Brennan’s writing: we’re told that something is wrong “to a significant degree,” but not told what that phrase means; later we’re told that the issue is empirically undetermined, but we’re told one contestable claim is highly plausible. From one perspective this looks like very careful, rigorous writing, but from another it looks like a confusing way of covering all the bases so as to avoid being held to any particular claim; it also seems to put the author in the position of dialectical victory no matter what objection is made, simply because the claims in the text are so elliptical that they can be made to say anything that the author wants them to say–without saying anything a reader can pin down.
(3) Third comment: there is such a thing as epistemic virtue. Why not consider the possibility that we ought to vote not on the basis of predictable policies per se, nor on the basis of character minus epistemic virtue, but on candidates’ moral plus epistemic character? I’d be curious to know the state of the social science literature on this subject. What evidence is more easily and effectively available to voters–information about a candidate’s epistemic virtue, or information about the predicted outcomes of the policies he can be predicted to enact during x years of a term when the issues he faces are themselves partly unpredictable? Having the right values may be sufficient (or as close to sufficient as matters) for making good policy (or at least for predicting good policy) if the values in question are both moral and epistemic.
(4) If there is good evidence that a politician is likely to enact harmful policies, perhaps one should vote for her as long as she is a good person, the other candidate is an evil person, and the harmful policies are not that harmful. The preceding claim seems incompatible with Brennan’s thesis, which entails that we ought not to vote for politicians who will enact harmful policies simply because the other candidate has a bad character. Brennan adds the parenthentical “without sufficient reason” in the penultimate sentence of the passage as though to include the Modi-like case I have in mind–in which case my Modi-like case wouldn’t be a counterexample to his thesis, but something he’d already thought of, and carefully baked into the thesis ab initio.
But Brennan’s parenthetical seems inconsistent and ad hoc. Either Brennan’s point is that character is relevant to voting (only)** when it is a proxy for future policy, or not. If the first disjunct is the case, Modi-like cases are a counterexample to Brennan’s view. On the other hand, if we go by the second disjunct, i.e., if there can be sufficient reason to vote for character when character is not a proxy for future policy, it is unclear what Brennan has been saying in this section of the book, or why he thinks what he has said is a response to the objection under discussion in the section. If “voting on the moral virtue of a candidate counts as good voting only to the extent that the candidate’s moral virtue is evidence that she will enact good policies,” the “only” implies that there cannot be a reason for voting on the moral virtue of a candidate when there is no evidence of a connection between character and expected policies. [And if there cannot be a reason, there cannot be sufficient reason. Hence the reference to “sufficient reason” is incoherent.]*
On the whole, though I generally agree with Brennan’s thesis in The Ethics of Voting, and regard it as an important contribution to the literature, I’m not crazy about the way in which he deals with objections in the book. The issue of character-based voting is merely a case in point, but in my view a clear one. Though one blurb for Brennan’s book describes it as “beautifully clear and eminently readable,” this particular section is neither. I don’t think the failure is mine as a reader but his as its author.
*I added the bracketed sentences a few hours after I originally posted this.
**Added for clarification’s sake a day after the post went up.
Postscript, October 9, 2014: A belated afterthought: doesn’t Brennan’s view entail that voter disenfranchisement of convicted felons is only justified to the extent that being-convicted-of-a-felony is a proxy for high-likelihood-of-wrongful-voting by the felon? After all, Brennan’s view is that the relevant issue as regards the right to vote is always the voter’s epistemic justifiedness or competence (on a rather narrow understanding of competence that is operationalizable and excludes, e.g., “softer” moral considerations considerations of empathy, etc.) That is the motivation for Brennan’s rejection of character-based voting. But there is nothing about being a murderer, rapist, or robber that a priori excludes competence or epistemic justifiedness in the relevant sense. So it seems to follow on his view that felons ought not to be disenfranchised qua felons. They ought to be enfranchised, regardless of their crimes, and we ought then to give them a chance to become competent voters. Since felons currently lack the right to vote, they haven’t had practice either at voting or at acquiring the necessary skills for it. But dispositionally, they might be fantastically competence under the right conditions. If they can achieve Brennan-competence (once we arrange the remedial conditions), they ought to be allowed to vote. Right? Practically speaking, that would be a bit of a headache, but such considerations don’t otherwise faze Brennan (consider the practical headaches of administering a nationalized poll test, though to be fair, we do have precedents to work from, e.g., the NAEP Civics Assessment) so why not?
Postscript, December 8, 2014: Here’s an excellent background essay, by William Dalrymple, on the Modi phenomenon, written for Britain’s New Statesman in May 2014, well before the elections that brought Modi to power. It’s journalism at its best, and has a richness that no thought-experiment could hope to have.
Postscript, December 9, 2014: This article in The New York Times, “Modi’s Campaign Stop in Kashmir Is Notable for Lack of Unrest,” provides some useful specification of the point I’m making in the post. For one thing, note that the real Modi’s behavior resembles that of the hypothetical Modi I describe: he’s not been alarmingly nationalistic or anti-Muslim, and he’s promised (perhaps credibly promised) policies that advance the economic prospects of Kashmiris, including the poor. And yet, the article ends, unsurprisingly, with this:
But as the crowd filed out afterward, a knot of well-dressed men stood nearby and watched with smoldering eyes.
“I didn’t go, because that man is a criminal,” Ahtisham Shah, a 40-year-old manager at the local office of a telecommunications company, said of Mr. Modi. “He still has to answer for the massacre in 2002.”
Asked if any of them would vote in the next round on Sunday, all five men shook their heads.
“To hell with India and to hell with Pakistan,” said Basharat Ahmad. “Kashmir is an independent country.”
The guys with the smoldering eyes aren’t voting because there’s no credible anti-Modi candidate to vote for, and, as they see it, any vote is a vote for India anyway. But ignore both things for the moment. Imagine that there was an anti-Modi candidate to vote for, and imagine that the smoldering-eye guys put aside their “to hell with India and Pakistan” attitude long enough to vote. (I’m not giving advice; I’m just imagining the possibility.) Should they, as per Brennan’s suggestion, not vote on character, even if the candidate in question “has to answer for the massacre in 2002”? Better yet: should we, as per Brennan’s suggestion, regard them as wrongful or incompetent voters if they do? Should they be required to pass a Brennanite competence test before being regarded as competent to vote for or against Modi? Isn’t there something nearly obscene about the suggestion that they should be disenfranchised for their failure to pass a Brennanite poll test?
My answers: they should vote on character; we should not regard them as wrongful or incompetent for doing so; they should not be required to pass a Brennanite competence test as a necessary condition of being considered competent; there is something obscene about Brennan’s suggestion (even if, as he likes to brag, he managed to make the suggestion in a well-known “peer reviewed” philosophy journal: I have to wonder how many of Brennan’s “peers” at Phil Quarterly were Indian or Kashmiri Muslims).
If Brennan didn’t intend any of the preceding conclusions in The Ethics of Voting, he might want to make that explicit, but nothing about the text of the book or any of the articles of his that I’ve read excludes my interpretation or even considers the possibilities I’ve raised.
Postscript, March 28, 2015: Reports are now coming out of India that suggest I may have been too charitable in my predictions about the situation of Muslims in Modi‘s India, not that that affects my argument at all.
Postscript, March 30, 2015: OK, so I guess I was being really over-charitable to Modi.
Postscript, April 27, 2015: I find it depressing that Barack Obama has chosen to write a laudatory essay on Modi for Time’s “100 Most Influential People in the World” issue (April 27, 2015). Frankly, I find George Wallace’s moral rehabilitation more impressive than Modi’s. The rehabilitation doesn’t stop Wikipedia for telling us (accurately) that Wallace is “remembered” for his segregationist stance (not his subsequent rehabilitation). Meanwhile, this is Obama on “Narendra”:
When he came to Washington, Narendra and I visited the memorial to Dr. Martin Luther King Jr. We reflected on the teachings of King and Gandhi and how the diversity of backgrounds and faiths in our countries is a strength we have to protect. Prime Minister Modi recognizes that more than 1 billion Indians living and succeeding together can be an inspiring model for the world.
Yes, they can.
Postscript, November 27, 2015: The controversy over Donald Trump and the post 9/11 celebration rumors induces me to revisit the topic of Jason Brennan’s critique of character-based voting (see here as well). Also relevant (though perhaps less clearly so) is some of the commentary on the recent elections in the Bihar province of India, where Narendra Modi’s BJP was unseated by its rivals. In thinking about both controversies, it belatedly occurs to me that Brennan is operating with either an overly narrow or an equivocal conception of “governance.” Go back to the last passage of Brennan’s that I excerpted. It begins like this:
So character-based voting is acceptable only insofar as it is a proxy to the quality of the governance a candidate is likely to produce. To what degree good character and good policies are correlated is largely an empirical question.
The second sentence seems to presuppose that good governance is reducible to the enactment of good policies, so that if a judgment about character is indeterminate with respect to the enactment of good policies, it’s irrelevant to judgments about whether to vote for the candidate in question. I’m not sure how Brennan understands the term “policy,” but on the ordinary understanding of the term, good governance isn’t reducible to the enactment of policies. It certainly can’t be the case that “good governance” can be understood by this formula:
S engages in good governance if and only if, for every policy P that S enacts (or proposes or broadly speaking facilitates), P is a good policy.
What about the good policies that S fails to enact? Is the failure to enact or propose (etc.) a policy a failure of policy on Brennan’s view? Charitably read, I think he’d say “yes.” At any rate, his view doesn’t prevent him from saying “yes,” so I’ll give him that.
But what about actions unrelated to policy? Here, it seems to me, he faces a real problem. His view seems to be that either a politician is in the business of enacting/refusing to enact policies, or he’s not on the job at all. But this strikes me as a reductive and oversimplified conception of politics.
Arguably, some of what political leaders do is essentially discursive. They talk to us, and often this talk has little or nothing to do with policy–or at least need not have much to do with policy. In their discursive capacities, politicians play (or as I see it, ought to play) the role of public intellectuals: they comment authoritatively on matters of public concern in a responsible way. Arguably, in doing so, political leaders also serve as models of civic virtue: they don’t just comment authoritatively on matters of public concern, but deal with those matters in ways that self-consciously exemplify a concern for truth and justice in public affairs. They give speeches, they answer questions at press conferences, they visit distressed places under their jurisdiction, they interact with people in a face to face way, and they’re recorded as interacting with people in that way. This isn’t just PR or show business. It’s a form of interaction that’s essential to leadership. In other words, a successful politician isn’t just a technician or policy wonk, but a leader–a public figure who functions as a moral exemplar, at least in certain limited respects. And good governance isn’t just policy wonkery; it’s leadership.
So character is bound to be relevant to governance whether or not it’s relevant to policy. The details of a person’s sex life may or may not be relevant to being, say, President of the United States or Prime Minister of India. But a person’s attitude toward (say) race relations certainly is relevant to both offices, whether or not those attitudes are proxies for any policy that the relevant individual proposes, enacts, or doesn’t propose or enact. The next president of the United States might well conclude that no new policies of any kind regarding race need to be enacted during his term. Suppose ex hypothesi that this is the correct decision as a matter of policy. Suppose ex hypothesi that the president successfully plays the political game so that he gets his way, and no policies are passed. That is ex hypothesi the optimal policy outcome, but it’s not the end of the story: it certainly matters how he pulls it off, e.g., how he defends his decisions, and how he deals with critics.
Once we cross a certain threshold, an asshole is not a good leader, and not a practitioner of good governance, even if he enacts the greatest policies in the world. (I’ll grant that we have to tolerate some degree of assholishness in almost any political leader, but even in politics, there’s such a thing as crossing the Asshole Rubicon, and once we do, all bets are off. Bright lines may be hard to draw here, but I think it’s obvious that Modi and Trump crossed the Rubicon a long time ago.) This is a subtly different point from the one I had originally made. My original point was backward-looking: having crossed a certain threshold, we shouldn’t reward wrongdoers with the perks of political office given their past misdeeds. My present point is present- and forward-oriented: we shouldn’t regard character merely as a proxy variable for predictions regarding policy-enactment, narrowly understood. We need to employ a broader conception of governance than that.
The topic of assholes brings me to Donald Trump (and Narendra Modi). Put it this way: imagine that Donald Trump becomes president, but that (miraculously) while in office, he changes his tune and enacts perfectly reasonable policies, even with respect to Arab and Muslim Americans. But imagine that he continues to comport himself as he currently is doing. Would he make a good president? No. His current comportment would undercut his claims to good governance even if he was enacting the right policies, and declining to enact the wrong ones. My point is not that ill comportment would undermine the policies per se, but that demeanor is an autonomous desideratum in a political leader, and that judgments of character are, in an obvious way, a proxy for it. You can’t be a good leader if you systematically disrespect and insult the people you govern, even if you enact the right policies in the process.
Incidentally, in saying that good character and good policy-enactment is an empirical matter, Brennan seems to be implying (as he often does) that identification of the correlation is a matter of consulting double- or triple-blind peer review social science studies. But if that’s what he means, he needs to deal with some obvious but unacknowledged questions.
First of all, there are many, many situations in life in which we have pre-scientific beliefs but no scientific studies to consult on the matter. Is his view that in every such case, we should simply ditch our pre-scientific beliefs on grounds of unreliability? Or is it sometimes permissible to use the pre-scientific beliefs as a guide to action? The first claim is really implausible, but the second claim sits uneasily with his rejection of character-based voting. He himself admits without consulting “the social scientific literature” that “[i]f someone is morally corrupt, there is a pretty good chance he will use the power of the state for personal benefit rather than to promote the common good.” Well, yes, that’s a matter of pre-scientific common sense. But what social scientific literature proves that it’s true? What social science literature has ever taken the population of “someones” as its sample?
Second, we know that a great deal of social science is unreliable. (Much of it is trivial as well.) To what degree, then, can we assume a priori that social scientific findings are, regardless of subject matter, more reliable than pre-scientific beliefs? I don’t see any reason to think that social scientists have the inside track on the nature of moral virtue. If they don’t, I don’t see any reason to think that their findings are always more reliable than pre-scientific beliefs on questions related to virtue.
Third, Brennan exaggerates the univocality of social scientific findings. Social scientists disagree with one another in both interdisciplinary and intradisciplinary fashion. Unless Brennan can show us that one branch or sub-branch of “social science” has a monopoly on the truth about politics, the appeal to social science strikes me as a bit of dogmatism and a bit of intimidatory hand-waving.
Since Brennan admits that character is something of a proxy for judgments about good governance, but also admits that it has variable weight, and can’t exclude the possibility that it sometimes might be highly weighted, and has no principled reason for preferring global reliance on social science over pre-scientific beliefs, and has an overly narrow conception of governance, I conclude that he lacks a coherent objection to judicious character-based voting. (Proviso: Brennan has a new book that’s either forthcoming or just recently out from Princeton University Press, called Against Politics. For obvious reasons, I haven’t read it, so I don’t know whether he discusses any of what I’ve said in this post there.)
Postscript, December 10, 2015: This critique of Trump is useful because it very clearly enumerates moral defects of Trump’s that aren’t necessarily proxy variables for predictions about policy.
In the article, Hillyer argues that Trump has a long record of undermining or abusing those weaker than him who get in his way. To the extent that Trump hasn’t broken any laws, that gives us a clear inference to the conclusion that Trump is immoral in his personal/business dealings, but yields no clear predictions about any particular policy conclusion. Hillyer goes on to argue that Trump mistreated his workers, but even if we grant that, we can’t infer that Trump’s economic policies would necessarily slight workers; after all, Trump is sufficiently opportunistic to change his tune and plump for pro-worker legislation.
Suppose you’re against eminent domain. Can we infer from Trump’s reliance on eminent domain that Trump would favor the expansion of eminent domain as a matter of policy? No, not really. Given his opportunism, Trump could just as easily restrict eminent domain if he found that desirable.
Maybe all of this proves that we shouldn’t vote for Trump because his impulsiveness and opportunism would make for bad policy. Maybe, but to be consistent, a defender of Brennan’s thesis would need to adduce social scientific evidence to suggest not only that impulsiveness and opportunism make for bad policy, but that they make for worse policy than intentionally aiming at bad policies (cf. Hillary Clinton, at least as viewed from a libertarian free market perspective). Note that this social science couldn’t just leave its findings at some set of generalities; it would have to be sufficiently tailored to the Trump vs. Clinton option to allow us to decide between them. I’m skeptical that any existing social scientific literature can yield such a conclusion.
The real reason why we shouldn’t vote for someone with a Trump-like track record is that electing someone to political office confers a reward on the person, and justice forbids rewarding someone’s past malfeasance whether or not the past malfeasance is a proxy variable for the person’s enacting bad policies in the future. In short, we shouldn’t vote for a Trump-like or Modi-like candidate simply because they don’t deserve our votes.
Postscript, December 12, 2015: More on the same theme. Consider Megyn Kelly’s now-famous misogyny question of Donald Trump back in August. The latter half of her question had an indirect sort of connection to policy (roughly, “how do you answer the charge that you are part of the war on women?”), but the first part did not. After quoting some of Trump’s remarks, she asked, “Does that sound [to] you like the temperament of the man we should elect president….”? (my emphasis).
As I see it, the first half of the question would have been entirely fair and appropriate on its own, even without the tie back to policy. Arguably, an avowed, explicit, egregious misogynist does lack the temperament of “the man we should elect president” even if he promises do great things for women (as Trump did, in answer to Kelly). Other things equal, misogyny of Trump’s variety ought to be a reason for voting against him (or not voting for him), whether or not the misogyny predicts any particular policy position he might take as president.
Suppose that other things aren’t equal, however. It’s an interesting question what a voter should do if faced with an out-and-out misogynist whose policy positions are, all things considered, appreciably better than the non-misogynist. To keep things relatively simple, imagine a pro-choice misogynist running against an anti-abortion non-misogynist (and suppose ex hypothesi that abortion ought to be legal). Unless you take avowal of an anti-abortion position to be prima facie evidence of misogyny on its own, I’d be inclined to say that the pro-choice policy position trumps the misogynistic defect of character, so to speak. Of course, if you regard an anti-abortion position as evidence of misogyny, then the choice here is between two misogynists, so that the dilemma is resolved from the outset. But though I’m pro-choice on abortion, that approach seems implausible to me.
Postscript, December 26, 2015: So Modi “surprises” us again, though this move comes as a less of a surprise to me than some of stuff I’ve described in the postscripts above. Given Modi’s character, however, it’s hard to interpret: is it a sincere step forward, or just play-acting? It seems obvious to me that one can’t easily disentangle the policy-related issue involved here (discussed near the end of the article) from an issue of moral character: is Modi honest or trustworthy? If he is, his trip to Lahore seems like a step forward in Indo-Pak relations. If he isn’t, there’s no way to know where he stands on policy, because we can’t trust anything he says or does.
Brennan might claim that the preceding observation is consistent with the letter of his claim in The Ethics of Voting (as it is), but I would say that it contradicts the spirit of his claim: if judgments of intellectual character are relevant to virtually every prediction we make about a candidate’s prospective policies, it makes good sense to vote for intellectually virtuous (honest, intellectually responsible, non-demagogic, etc.) candidates, and makes good sense to figure out whether a given candidate is in fact intellectually virtuous (honest, responsible, non-demagogic, etc.).Other things being equal, we should vote for the intellectually virtuous candidate, the more the virtuous the better.
If so, it makes no sense to come out against character-based voting, pointlessly adding the proviso that character-based voting is OK as long as it’s relevant to policy. How could a candidate’s honesty, candor, probity, conscientiousness, accuracy, trustworthiness etc. be irrelevant to policy? We might (accurately but in this context a little tendentiously) call those policy-relevant traits of character, and think of them as the policy-makers’ analogue to the intellectual virtues required to do good science. If policy-relevant traits of character are always relevant to policy, then a ban on character-based voting that allows for them is either toothless or misleading or both.
Thanks to Faisal Jilani and Aftab Khawaja for the discussion on Modi that inspired this post, and to Fawad Zakariya for driving home to me the moral significance of what happened in Gujarat. The usual caveat applies.