Which Elections? A Dilemma for Proponents of the Duty to Vote

Abstract

Proponents of the duty to vote (DTV) argue that in normal circumstances, citizens have the moral duty to vote in political elections. Discussions about DTV analyze what the duty is, who has this duty, when they have it, and why they have it. Missing are answers to the Specification Question: to which elections does DTV apply? A dilemma arises for some supporters of DTV—in this paper, I focus on Julia Maskivker’s work—because either answer is problematic. First, I argue that it is implausible that DTV applies to all elections because this makes the duty too costly for the voter. Second, I argue that there are no good reasons why under normal circumstances DTV applies only to some elections. I consider objections but conclude that the dilemma is successful and therefore the case for DTV is incomplete.

Introduction

Voting in political elections is often seen as an indispensable part of democratic institutions. Through voting, citizens exercise their collective power to install and remove leaders, voice their approval or disapproval of policies, and contribute—for better or for worse—to the shape of the political landscape. If democracy is valuable, then so is voting. Some philosophers and political scientists think that voting is so important that citizens have the duty to vote (DTV).¹

On DTV, voting is essential for a successful democracy. Political scientist Lisa Hill explains:

Voting is not just any duty; it is a special duty because the existence and proper functioning of representative democracy depends on its performance. So too do our welfare and rights. (Brennan and Hill 2014, p. 171, emphasis original)

So, voting is the political responsibility of citizens and therefore it is not optional; in other words, you are obligated to vote.² We can describe the duty to vote a bit more precisely:

DTV: In normal circumstances, eligible citizens have the moral duty to vote well in political elections.³

Suppose on Election Day, I deliberate whether to go to the polls but eventually decide to stay home and binge-watch The Office. According to DTV, I have done something morally blameworthy—specifically, I abstained from voting, which is morally impermissible.

In this paper, I will argue that the case for DTV falls short. I do this by offering an argument, by way of a dilemma, which shows that DTV is implausible. The dilemma arises when we ask this crucial question, which I call the Specification Question: to which elections does DTV apply? Well, we either have this duty in all elections or only some elections, but both answers are problematic. Or so I will argue.

Before I present the argument, let me be explicit about my assumptions. First, my argument does not show that DTV is necessarily false. This is because there are cases in which plausibly I would accrue such a duty, as when I promise a colleague to vote in the upcoming election, an assassin threatens to kill my family unless I vote a certain way, or I know that my vote will be decisive to remove a corrupt president. I set these cases aside because they fall outside the scope of ‘normal circumstances’ as DTV stipulates. So, the paper does not seek to establish that we could never be obligated to vote. This is an important point because the concession that in special circumstances some people are obligated to vote is compatible with one’s rejection of DTV, which is typically understood as a general duty of citizens. Second, my argument targets DTV in a democracy relevantly similar—perhaps primarily in population size—to the United States. The focus on the US is purely practical in nature, and not because the country represents the best form of democracy or because it is the standard for other democracies.⁴

Here is the outline of the paper: in ‘The Specification Argument’ section, I present my argument with minimal commentary. Then, I defend a crucial premise in ‘Supporting Premise 3’ and respond to criticisms in ‘Objections to Premise 3’. Next, I support the other crucial premise in ‘Supporting Premise 6’ and consider objections in ‘A Strong Objection to Premise 6’ before concluding.

The Specification Argument

Discussions about DTV analyze what the duty is, who has this duty, when they have it, and why they have it. Missing are answers to the Specification Question: to which elections does DTV apply? This is a relevant question that is mostly overlooked in the debate on the ethics of voting, and I will argue that absent a good answer, the allure of DTV will be diminished. This is because no matter how proponents of DTV answer the Specification Question, trouble arises for their position. Here is my argument:

• 1. If DTV is true, then either (a) DTV applies to all elections or (b) DTV applies only to some elections.

• 2. If (a) is true, then you must vote in all federal, state, and local elections.

• 3. But it’s not plausible that you must vote in all federal, state, and local elections.

• 4. Therefore, not (a). (From 2, 3)

• 5. If (b) is true, then there are good reasons why you must vote only in some elections.

• 6. But there are no good reasons why under normal circumstances you must vote only in some elections.

• 7. Therefore, not (b). (From 5, 6)

• 8. Therefore, it is not plausible that DTV is true. (From 1, 4, 7)

Since propositions 4, 7, and 8 are conclusions, and premises 1, 2, and 5 should be relatively uncontroversial, that leaves premises 3 and 6 to do the heavy lifting in the argument. The argument is deductively valid, and thus in order to escape the dilemma, proponents of DTV must attack either 3 or 6. I will make a few comments about the other premises before defending 3 and 6.

Premise 1 is uncontroversial because it exhausts logical possibilities; it is a necessary truth that if you have DTV, that duty applies either to all elections or not all elections, and the latter is logically equivalent to saying DTV applies only to some elections if any at all. Premise 2 clarifies clause (a) in the first premise. It states that if there is DTV in all elections, then you must vote for candidates running for office on all levels of the government. For example, in the 2020 general election, my state had the following offices on the ballot: president and vice president, senator, representative, governor, lieutenant governor, attorney general, auditor, commissioners of agriculture, insurance, and labor, secretary of state, superintendent of public instruction, treasurer, and lastly, several state supreme court and court of appeals seats. Then there are local races, such as those for mayor, city council, register of deeds, and sheriff.⁵

Of course, the ‘all’ in the premise needs to be qualified because there are elections in which you are ineligible to cast a vote, e.g., in the gubernatorial race in a neighboring state. Accordingly, ‘all elections’ should be read as all elections in which you are eligible to vote. Still, the number of elections in which you are permitted to vote is not small. DTV implies that you must vote in all these races. But this is implausible for at least two reasons, as I show next.

Supporting Premise 3

First, it is implausible that citizens have DTV in all federal, state, and local elections because the duty is too costly (demanding) for the voter. Julia Maskivker has offered what is perhaps the strongest defense of DTV, and thus mainly I will interact with her position. In her book The Duty to Vote (and other work), Maskivker defends what may be called political Samaritanism. Samaritan duties of aid require us to help others when doing so would not be too costly for us; even if we will be late for work, we should help the wounded victim at least by doing the bare minimum, such as calling for help. Similarly, citizens have DTV when doing so would not cost them much. However, voters must vote with care, i.e., ‘with information and a sense of the public interest’ (2019, p. 4). Voting with care will require some sacrifices, such as time spent gathering information, but these sacrifices are not unacceptably demanding given the benefits for society that could be brought about via elections; voting has a cost, but it’s not unduly costly (Ibid., p. 78). Therefore, ‘under the right circumstances, casting a careful vote is morally obligatory because there is no excuse that may help us escape this requirement’ (Ibid., p. 206). I will now argue that if DTV applies to all federal, state, and local elections, then the costs of voting exceed even the low threshold set by political Samaritanism.

Voting is not costless. Time spent voting is time spent not doing other things, such as going on a date with your spouse, playing with your kids, or enjoying a cup of coffee with a friend. (Or watching The Office.) So, voting has an opportunity cost. Numerous thinkers have already made these points in the relevant literature, and I rely on their work to support premise 3. I am afraid that those unmoved by the arguments of Brennan, Caplan, Somin, and others will not find a novel argument from me to support 3.

Nonetheless, let us think a bit more about the opportunity costs of voting. It takes time to register, drive to the polls, wait in line, vote, and then drive home.⁶ Suppose that takes an hour. But now consider what that hour spent voting is worth. Here is how Christopher Freiman puts it:

If you spent that hour earning somewhere near the average American wage, you’d make enough to donate the money needed to save about four months of life for a child dying of malaria. So the expected value of your vote is significantly lower than the expected value of alternative uses of your time. (2020, p. 134)

One might think that even an hour spent voting is too costly because you could have done something more valuable. But let us grant that the hour used on voting is not unreasonably costly. The cost rises when we begin to consider the frequency of elections. That is, my argument will not be that voting in a single election is unduly costly but rather, voting in multiple elections, year after year, is unduly costly.

There are the well-known quadrennial general elections but also biennial mid-term elections. Some local elections are held separately from all federal and state elections. According to L. Sandy Maisel, American voters are asked to vote more often than citizens of other countries (2007, p. 137).⁷ Consider the number of times you are asked to cast a ballot. Depending on where you live, answers will vary, but throughout your life, if you fulfill DTV in all federal, state, and local elections, you will sacrifice many hours of your time on political participation. The initial hour of voting in a single election quickly multiplies to the point that having DTV in all elections accrues a cost that is no longer insignificant.

For the sake of the argument, let us suppose that the government figures out a way to drastically minimize the costs of voting mentioned above. Let us suppose that you can quickly and reliably vote from the comfort of your home; on this scenario, voting costs are drastically decreased. Nevertheless, I will argue that the cost of voting is still high if there are epistemic requirements on how we vote. Here is the second way to support premise 3: it is implausible that citizens have DTV in all federal, state, and local elections if also they have the duty to vote well or the duty not to vote badly.

Of course, some advocates of DTV will deny that the ethics of voting entail that the voter must vote well.⁸ On these views, what is really important is not how citizens vote, but that they vote (e.g., Arvan 2010; Maring 2016). While this is a popular view that has its able defenders (e.g., Elliot 2023), I think that the view only reinforces the skepticism of those who reject DTV as implausible. How can one have a duty to do something which, collectively, can bring about harm (such as the installment of corrupt politicians)? Thus, to make the strongest case possible, I think advocates of DTV should maintain that voting well is what is required to fulfill one’s duty to vote.⁹ At any rate, this assumption cannot be defended here and thus I proceed to examine DTV on which voting well is not optional. Those convinced otherwise will not find my arguments convincing.

Maskivker has recently defended this version of DTV at length. She argues that incompetent and prejudiced voters can be ‘agents of harm’ if they elect (in conjunction with other bad voters) bad and unjust governments, and thus if we really care about democratic outcomes, voters should vote with ‘information and a sense of justice’ (2023, p. 2). According to Maskivker, the duty to vote well ‘demands that we expend a minimum of effort so that our votes reflect certain standards of rationality, rightness, and knowledge’ (2016, p. 224). Furthermore, voting with care ‘requires sufficient information about facts necessary to assess the superiority of one electoral option over others (or the equal lack of appeal thereof)’ (2019, p. 80). I now argue that meeting these conditions is costly if these conditions must be met when evaluating all of the candidates (and their policies) on the ballot in all federal, state, and local elections in which you could cast a vote.

Consider what it takes to cast an informed vote. First, you must know who is running for office. This might seem simple enough, but as Ilya Somin reveals in his Democracy and Political Ignorance, only 15% of Americans could identify one candidate running for the House of Representatives in their own district and only 4% could identify two (2013, p. 32).¹⁰ Next, you must know about the candidate’s policies, e.g., what is their position on healthcare, immigration, taxes, trade, gun control, social issues, education, the military, etc.? Acquiring knowledge about a candidate’s policies is no small task. Now rinse and repeat for all other candidates. Even if the voter need not know all of the policies of every candidate, it would seem that to cast an informed vote they would need to know the policies of at least the major contenders. And as I will argue later, voting merely based on political allegiance is insufficient to count as a well-informed vote.

Next, you must evaluate whether the proposed policies are any good. Proper evaluation requires a basic grasp of economics, foreign relations, philosophy, sociology, political science, and other sciences; the time you will spend on reaching basic competence is not a small amount. Moreover, judging which consequences a policy is likely to produce will add to the cost. The problem here is that the average citizen lacks basic political competence to vote well: in the words of Larry Bartels, ‘The political ignorance of the American voter is one of the best-documented features of contemporary politics’ (Bartels 1996, p. 194).¹¹ Some policies are ethical in nature and thus deeply controversial, e.g., the permissibility of capital punishment, abortion, income redistribution, drug decriminalization, and military intervention in other countries. Having justified beliefs about these issues is no small task. But suppose you finally have justified beliefs about a candidate’s policies and their expected effectiveness. Now you must evaluate the policies of all the other candidates and come to justified beliefs about which policies will be most beneficial for the common good. Given the current widespread political ignorance, it will take a lot of effort for citizens truly to be well-informed.¹² This effort will be costly. The resources they will spend on voting well in all elections seem so demanding that DTV is no longer like Samaritan duties of easy aid that have small costs.

Maskivker thinks that to vote minimally well requires that voters must do better than voting at random, and this level of competence is easily satisfied if voters acquire the relevant information and put some thought into their decisions (2016, pp. 230–231). But Bryan Caplan has argued that when it comes to voters’ beliefs about economics, voters are systematically in error—and Caplan thinks that these errors extend to beliefs about other subjects as well (2007, pp. 9–10). As Jason Brennan puts it:

They [voters] are not merely ignorant know-nothings; they are misinformed and know less than nothing. For a great many issues, we would get better government performance by deciding questions on the basis of coin flips than by asking voters. (Brennan 2011, p. 171)

If these conclusions are true, then it looks like Maskivker’s minimal requirement will be unsatisfied in many cases. Voting well requires voting better than voting at random, and to satisfy this criterion, voters must acquire a certain satisfactory level of political knowledge. Moreover, they must correctly apply this knowledge. I am assuming here that there is some level of objectivity in voting well, such that there are standards (above and beyond the voters’ own preferences) by which the vote could be evaluated. For example, in voting for a candidate with harmful policies or policies that are not good in some sense, the voter votes badly.

Maskivker argues that ‘Voting with information and a sense of responsibility for society could be seen as an acceptable cost to undergo in order to contribute to an important collective good such as good governance’ (2019, p. 41). But based on the discussion so far, I think that ‘voting with information’ in tens (and eventually hundreds) of political elections is where the costs begin to pile up.¹³ Let us consider three objections to my line of reasoning.

Objections to Premise 3

Objection 1: the cost of voting is minimal because citizens can use cognitive heuristics (short cuts) in their decision-making. For example, one common heuristic used by voters is to vote according to a candidate’s party affiliation; if you are a Republican, then simply choose all Republicans on the ballot.

But the problem with this proposal is that if the only reason you vote for a candidate is that they are from your party, this reason seems to fail the requirement for voting well. The candidate can be corrupt, propose immoral policies, or otherwise be unfit for office. But on Maskivker’s own view, voting well requires that the voter possess ‘facts necessary to assess the superiority of one electoral option over others (or the equal lack of appeal thereof)’ (2019, p. 80). However, voting along party lines does not automatically ensure that the voter is choosing the superior candidate, or even that the voter meets the minimum level of competence by voting, as Maskivker requires, better than random. Even if voting for a candidate from your own party rather than from the other party satisfies the criterion of voting better than random—because you now have a reason to prefer one candidate over the other—voting this way will not help the voter determine which candidate from one’s party is best. Consequently, voting based on party affiliation fails to meet the minimum requirements of voting well (or fails to avoid voting badly). Of course, voters could use multiple heuristics to help them decide on a candidate. But again, it will take time and effort to utilize various heuristics as they sort through all of the candidates. Furthermore, if the average voter is as politically ignorant as evidence suggests (Caplan 2007; Somin 2013), then, plausibly, voters lack even minimal levels of competence to effectively use heuristics in their voting methodology.¹⁴

A version of this objection states that the cost of voting is minimal because citizens could be single-issue voters. Suppose you are pro-life and you believe that this is the most important issue at stake. Then, you would simply vote for all candidates who support pro-life policies. However, things are not so simple. That is because a candidate’s pro-life stance is not sufficient to justify voting for her. Imagine that the pro-life candidate vows to declare war on North Korea to exterminate the communist threat once and for all; experts predict a global nuclear war. The single-issue approach might be used to quickly narrow the list of all candidates for whom you think it is permissible to vote, but you still need more information, e.g., you need to know that the candidate is not supporting a different harmful policy. Voting on a single issue without considering the candidate’s other issues would not be voting well. So for example, if a candidate is not pro-life, that might be sufficient for you not to vote for her, but if a candidate is pro-life, that is not sufficient to vote for her because she might have other bad and unjust policies, in which case you would be voting badly. For this reason, this objection fails.

Objection 2: the cost of voting is minimal because citizens can vote for character instead of policy. Instead of having to sift through all of the policies offered by all the candidates, let us just figure out what kind of person they are. This minimizes the cost of voting. But as Brennan argues, character-based voting is bad voting because it largely means voting for the wrong reason and because character is not a reliable guide to a candidate’s success. According to Brennan,

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. (2011, p. 84)

Consequently, voting for character is a bad strategy to implement because either you would not be voting well or be breaking a duty to avoid voting badly.

Objection 3: the cost of voting is minimal because citizens need not vote well; what matters is that they vote, not how they vote. Brennan has argued that this position is mistaken. If voters decide to vote, then they have a duty not to vote badly, which means that they must be rational, unbiased, just, and informed; voting badly is voting without a sufficient (or justified) reason (Brennan 2009, pp. 536–537). Bad voting occurs when citizens vote from immoral beliefs, ignorance, or epistemic irrationality and bias. Since we have a duty not to engage in collectively harmful activities—an activity that is not harmful in itself but becomes harmful when many people engage in it—and bad voting is collectively harmful, citizens are obligated not to pollute the polls, i.e., vote badly (Ibid., pp. 538–539).¹⁵

In a reply to Brennan, Marcus Arvan (2010) argues that citizens do not have the duty to avoid voting badly; this is because the cost of not voting badly is high. For example, voters would need to be well-educated if they are to vote well, and obtaining a university-level education is very costly. Therefore, it is permissible to vote badly. I have two replies. First, Arvan actually helps my case because his reply to Brennan is based on the thesis that voting well is costly. I wholeheartedly agree. Second, Arvan is mistaken to conclude that if the cost of voting well is too high, it is permissible to vote badly. Instead, the voter should simply abstain and not contribute to collective harm that can occur when bad voters install bad governments; such contributions are ‘morally condemnable’ (Maskivker 2023, p. 5). Arvan neither establishes the position that voting badly satisfies the duty to vote, nor does he consider that option of abstaining, which, all things considered, might be what morality requires. At any rate, Arvan’s argument will not help proponents of DTV such as Maskivker escape my dilemma because they agree that voters must vote well.

These objections are inadequate to show that the costs involved with voting well in all federal, state, and local elections are minimal. My conclusion is that even if the cost of voting in a single election is not unduly costly, voting in numerous elections multiplies the cost, and it is this aggregate cost of voting that is burdensome. Consequently, if DTV applies to all elections, then DTV is implausible; this duty is too demanding—contrary to what supporters of DTV say—and thus not a Samaritan-like duty. Premise 3 stands undefeated, and I now turn to the other horn of the dilemma.

Supporting Premise 6

Suppose that I have convinced you that the duty to cast a ballot in all federal, state, and local elections is implausible. This leaves just one way out of the dilemma, and that is to affirm that DTV applies only to some elections. Premise 5 claims that if DTV does not extend to all elections, there must be some plausible reason or explanation for this fact. Furthermore, such reasons or explanations would be good, i.e., they will not be ad hoc, arbitrary, or trivial.

Recall the Specification Question: to which elections does DTV apply? Proponents of DTV have not provided an answer, at least not one that makes clear why DTV applies to this election but not others. For example, Maskivker writes that ‘Episodic voting does not have to require constant or even frequent political participation, although it does require attention to issues of concern as important elections draw closer’ (2019, p. 11) and that ‘voting every so often at important elections is not that burdensome’ (Ibid., p. 78). But which elections are important? Important for whom? Important how? Presumably, important elections are the ones in which you should vote, and now it is just a matter of figuring out which elections are significant enough to generate DTV.

As I see it, the burden of proof falls on advocates of DTV to provide compelling reasons why, under normal circumstances, DTV applies to some elections but not others. Next, I will consider two reasons which they might offer.

Reason A: DTV applies only to those elections in which your vote might make a difference to the outcome. Your vote can make a difference if it is either decisive or causally efficacious. A vote is decisive if it breaks a tie (and thus Candidate A wins), or if it creates a tie (and thus prevents Candidate B from winning). A vote is causally efficacious if the vote is part of the subset of all cast votes required to win the election, although the individual vote is neither decisive nor necessary (Tuck 2008; Goldman 1999). I now offer replies.

First, the chance of your vote being decisive is extremely small. How small? That depends on the number of voters, but all else being equal, increasing voters lowers the probability of your vote being decisive (Lomasky and Brennan 2000, p. 66). In a presidential election, that probability can be as high as 1 in 10 million or as low as 1 in a billion (Freiman 2020, p. 44). For this to happen, not only must you vote in a state in which the candidate must win the Electoral College, but your vote has to make or break a tie, and the odds of these events occurring are small (Ibid.)¹⁶

Second, if it is sufficient that your vote could be causally efficacious for you to have DTV in a particular election, then this reason is likewise sufficient to show that you have DTV in all elections because it is possible that it be causally efficacious as well. Unless we are aware of a principled way to determine in which elections our vote might be causally efficacious, we are still in the dark about in which elections you ought to vote. In normal circumstances, voters do not know whether their vote will be decisive or causally efficacious, and thus it would be irrational for them to appeal to Reason A as justification for why they ought to vote in this election but not the other.¹⁷

Lastly, Reason A is unavailable for proponents of DTV such as Maskivker because she concedes that your vote will not make a difference in most cases, and thus she seeks to establish the value of voting on other grounds besides consequential ones. Your vote is derivatively valuable ‘because it contributes, however marginally, to a larger valuable project’, a project which, via elections, furthers goods such as justice, freedom, and fair governance (2019, p. 67; cf. 2018, p. 414). But Maskivker’s reasoning is equally applicable to all elections because your vote contributes to the valuable project and might make a difference, and thus she has not shown why DTV applies only to certain elections.

Brennan (2016) thinks that the most plausible case for DTV will not rely on the claim that voters make a difference in changing the outcome of an election, but on other considerations why voters should vote anyway. He lists three arguments: (i) The Generalization/Public Goods/Debt to Society Argument, which argues that if you abstain from voting, you free ride by enjoying the benefits of good government or fail to pay your debt to society; (ii) The Civic Virtue Argument, which argues that you should vote because you have the duty to exercise civic virtue; and (iii) The Complicity Argument, which argues that you should vote to avoid being complicit in governmental injustice. But notice that these three arguments could be applied to all federal, state, and local elections. That is, proponents of DTV cannot appeal to, say, argument (iii) as a reason for why you should vote in the presidential election—thus answering the Specification Question—because citizens should not be complicit in governmental injustices at all levels of government, not just the highest ones. If you are obligated to cast a vote in hopes of removing a corrupt president, then it seems like you are also obligated to cast a vote in a local election in hopes of removing, e.g., a corrupt mayor. Likewise, the free-riding and civic virtue arguments appear equally applicable to all kinds of elections and thus cannot be used to support the claim that DTV applies only to some elections. Now, let us consider another reason.

Reason B: DTV applies only to significant elections, even if your vote will not make a difference. On this approach, it is acknowledged that voting to affect the outcome of an election is not the point of voting. Rather, you should vote to express yourself or make a statement. So, by voting for a president whose policies will economically benefit minorities or voting to remove an unjust president whose policies restrict religious freedom, you signal that you care about these issues. Although I think Reason B is much more plausible than Reason A, I have two replies.

First, we will need clarification on what makes elections significant. As Maskivker claims, ‘Voting with a sense of responsibility... does require attention to issues of concern as important elections draw closer’ (2018, p. 412). But this claim does not shed light on what makes an election important. Is it the policies involved? But which policies are significant, and what makes them significant? Significant morally, economically, socially, etc.? Significant for whom? All Americans, or a specific demographic? And should I vote even if the election is not significant for me, but is significant for others?

Perhaps a hint toward an answer is found in the following passage:

Voting with a sense of responsibility for society at presidential elections does not amount to becoming a homos politicus. . . But at certain points we ought to get involved. When? When elections afford us the opportunity to assist society by way of choosing governments that we expect to rule minimally fairly—or more fairly than all the other realistic alternatives at play. (Maskivker 2019, p. 40)

So Maskivker focuses on (or at least emphasizes) presidential elections, but why? Reasonably, it is because their outcomes affect more people. But why think the mere quantity of people affected by a candidate’s policy generates DTV which otherwise would be absent? Suppose a presidential candidate’s policy will raise annual taxes for all Americans by $1. Also, suppose a mayoral candidate plans to defund the city’s police department by 80%. If you think that the outcome of the mayoral race will objectively be more harmful to the community than raised taxes, it seems you ought to vote in the mayoral race. But then our initial explanation of why DTV applies only to presidential elections fails.

It is true that issues of national interest are more weighty than local issues, and that presidential races (as well as congressional and senate races) are more significant than local races. Still, advocates of DTV need to provide an argument for why the duty to vote arises only when national interests are on the line. That is, just because the local mayoral race affects fewer people does not mean that the impact of the local election is less significant for that community; possibly the voter thinks—perhaps mistakenly—that the mayoral race is more important than the presidential election. It seems that DTV implies that the voter should participate in the local election. If DTV does not imply this, why not? If the answer is that the local race is not of national interest, again, we need an argument for the conclusion that DTV should be restricted only to elections that have a national impact. All of these issues must be dealt with by fans of DTV in order for Reason B to be explanatorily adequate; until then, the reason is vague and thus difficult to access.

But there is a second problem with Reason B, and it is this: if it is the voters’ responsibility to decide which elections are significant, then many of my previous points about the cost of voting could be employed here to show that this is too demanding on the voter. To have a justified belief about which election is significant, voters would need to acquire and analyze information about candidates and policies, and this will take a lot of time and energy.

I conclude that so far, premise 6 stands undefeated because Reason A and Reason B turn out to be not good reasons at all. However, proponents of DTV might have other strategies to rebut my argument and in the next section, I consider what I take to be the strongest line of reasoning against my argument.

A Strong Objection to Premise 6

Premise 6 states that there are no good reasons why under normal circumstances you must vote only in some elections. If DTV applies only to some elections, then it could be interpreted as an imperfect duty. The Kantian distinction between perfect and imperfect duties is roughly this: if you have a perfect duty to do X, then this duty admits of no exceptions, but if you have an imperfect duty to X, then there are cases in which it is morally permissible to refrain from doing X. In normal situations, assisting others is an imperfect duty and thus only sometimes, and to some extent, are we required to do so. But if DTV is an imperfect duty, the critic contends that my argument proves too much and thus leads to counterintuitive consequences. To see this, consider this parody of my argument:

1*. If there is a duty to help others, then whenever it is possible to help someone, we must either (i) help every time or (ii) help some of the time.

2*. Not (i): helping in all cases would be too demanding.

3*. Not (ii): there is no clear principle specifying in which cases to help.

4*. Therefore, there is no duty to help others.

I take it that most people would reject the conclusion of this parody argument, but then for the same reason, they should reject my argument. The idea is that just as the duty to help others is imperfect and thus it is up to the individual’s discretion when and where to discharge it, similarly, it is up to citizens to figure out in which elections they have the imperfect duty to vote. Both of these imperfect duties require making a judgment call, and one might argue that both duties could be discharged with about the same level of discretion. So, if lacking a clear principle informing me in which cases I must assist someone does not abolish the imperfect duty to help, then plausibly (and by analogy), lacking a clear principle specifying to which elections DTV applies does not abolish DTV.

Although this is an interesting objection, some relevant disanalogies weaken the objection. It will be easier to see the disanalogies if we think about specific cases. For example, imagine that on your hike in the woods you come across Vic laying on the path, blood gushing from his neck. Immediately, you call 911, thereby discharging your duty to assist. There are two factors to consider here. First, calling for help was not unreasonably costly. Second, and perhaps more importantly, there was a high probability that your calling for help would lead to the desired outcome: medical professionals (and perhaps the police) arriving on the scene. So in this case, you judged that your actions will be effective. If you had no phone, you might apply direct pressure to the wound and yell for help in hopes of another hiker hearing you. Or you might load Vic into your car and drive him to the hospital. Whatever course of action you take, barring improbable or unusual circumstances, your actions will be efficacious even if Vic dies before help arrives: this is, there is a high probability of success that your call will reach emergency services, or that pressure on the wound will decrease blood loss, or that your car will start and you will be able to drive to the hospital.

When it comes to voting, things are not so straightforward. If the earlier arguments of this paper are successful, then when we vote, we can almost be sure that our vote will not make a difference in any given election; when you vote, you cannot reasonably expect that your action will bring about (or even contribute to) the desired outcome; Maskivker (2018, p. 414), for one, agrees.

Another disanalogy between helping others and voting is that we can aid the person in need in multiple ways. We could help Vic by calling for help, putting pressure on the wound, bandaging the wound, driving him to the hospital, etc. But with voting, there is only one thing you could do: vote. Why is this a relevant difference? Because it is plausible to think that our judgment about whether or not the duty to help applies in a particular case is in part based on our judgment about the possible courses of action we could take and the expected outcome of those actions. If I decide to help Vic by calling 911, it is reasonable to expect that—assuming the call goes through—help will be on the way. If I see a toddler drowning in the pool, I know that the success of rescuing the child will be extremely high; even if I cannot swim, I can get the lifeguard’s attention or throw a life vest. Helping a person in a specific circumstance usually allows for numerous ways in which that person could be helped, and at least some of those ways will have a good chance of bringing about the desired outcome. I can discharge the duty to help others in various ways, and pick the way that will be most effective, but I can only discharge the duty to vote in one way, and that is by voting. But as I have mentioned earlier, voting is not an effective way to bring about the desired result; if that is the case, then the effectiveness of your vote cannot be a guiding principle that informs you in which elections to vote because in most elections, your vote will not make a difference. Consequently, the parody argument fails twice over because when you make a judgment about whether to assist a needy person, the judgment is partly based on considerations about the possible actions available to you and the expected outcome of those actions; not so with voting. I am not claiming that there are not other considerations that must be entertained in your deliberation about whether to assist, only that this consideration provides at least some guidance in your deliberation; and since this consideration is unavailable in deliberations about in which elections to vote, the duty to help others is not analogous to DTV.

Maskivker argues that ‘at certain points we ought to get [politically] involved’, and those points are ‘[w]hen elections afford us the opportunity to assist society by way of choosing governments that we expect to rule minimally fairly…’ (2019, p. 40). But as I have argued, the opportunity to assist society via voting is not analogous to the opportunity to help someone in need because when I help the needy person, there is a high probability that my actions—aimed at providing aid—will be causally efficacious and this is not the case with voting. Because this is a relevant disanalogy between my argument and the parody argument, this strategy is unsuccessful.

Interestingly, Maskivker denies that DTV should be seen as an imperfect duty (2019, p. 135) and pursues a different strategy. Instead, she thinks that DTV is a ‘duty of time and place’—and these duties are not optional (Ibid., p. 137). In special circumstances, the opportunity to assist a needy person arises and given that you are in a good position to help then and there, you are obligated to do so even if, say, you have already helped someone earlier that day. Maskivker calls this idea ‘the Principle of Moral Inescapability’ and it states that ‘given certain confluence of factors, aiding is the right and obligatory thing to do’ (p. 136). The principle does not imply that you are obligated to help in every conceivable situation, only that in some specific circumstances, given the time, place, and ability, you ought to assist. Maskivker then extends this logic to voting:

I propose to think of the duty to vote with care as a duty of time and place, to borrow Locke’s language. The duty becomes stringent when elections offer us an easy way to contribute to aiding society by way of choosing decent governments or ousting indecent ones. Failing to act when this circumstance unfolds would be morally problematic, other things equal. Locke’s duties of “time and place,” one could say, are morally inescapable when the right circumstances come to emerge that justify action. (p. 138)

So if you are well-situated to vote and could do so at minimal costs to yourself, you should vote. I have a few responses. First, it is not clear to me how these so-called duties of time and place differ from imperfect duties. When I come across bleeding Vic and I could easily help by calling 911, the Principle of Moral Inescapability requires my assistance. But surely, I ought to help even if the duty to help others is imperfect; if the theory of imperfect duties implied otherwise, we should reject the theory. The reason why I am obligated to help Vic is that I am confident that my actions will be effective, and there is a high probability that they will causally contribute to bringing about the desired results, e.g., Vic’s medical treatment. But when I come across a car that has run out of gas on the road and I see ten people already pushing it to the gas station around the corner, I know my assistance will be of little or no help and thus I am morally permitted to keep driving. But given these circumstances, it seems that the Principle of Moral Inescapability would not require me to help, and thus it leads to the same result as the imperfect duty to help, i.e., under no obligation to assist. So what is the relevant difference?

But Maskivker denies that the duty to assist (or vote) goes away if your contribution will be minimal or even unnecessary. She thinks that in some cases, the duty to help can be discharged only through a collective effort of assistance even if your contribution is technically not required to reach the goal (2019, p. 142). For her, DTV is ‘a Samaritan duty to contribute, however marginally, to a collective activity that will or can have a discernible (worthy) effect’ (2018, p. 414.) Maskivker argues that abstaining from voting would be wrong given the circumstances that emerge around elections. Specifically, elections offer citizens an easy way to contribute to the betterment of society; they do not require much effort and only minimal political knowledge. Therefore, voting is morally inescapable given the right circumstances.

But here, another dilemma arises. Are the circumstances always right or not? If they are, then you have the duty of time and place to vote in all elections. As I have already argued, this is implausible. But if the circumstances are not always right, then how can we tell whether they are right around a particular election? The Principle of Moral Inescapability provides no guidance. At least when you stop to help Vic, given those circumstances and given that your actions will be effective, you must offer help. But when it comes to voting, knowing that your vote most likely will play no role, what other factors must be present for you to conclude that in this election, you must vote? Maskivker suggests that we ought to vote when ‘it is punctually necessary that we do so’, for example, in removing bad or inefficient leaders (2023, p. 11). But this answer just raises another question: in which elections is it ‘punctually necessary’ to vote? Thus, it seems that we still do not have a good answer to the Specification Question.

But perhaps by focusing on the collective nature of DTV the Specification Question could be answered. For Maskivker, there is a collective dimension to voting: an individual vote will almost certainly not make a difference, but the collective activity of voting does make a difference in bringing about certain political goods, such as fair governance, and thus DTV should be seen as ‘a duty to cooperate with others in bringing about justice’ (Maskivker 2019, p. 11, emphasis original.) On this view,

the duty to vote is a Samaritan duty to contribute to a collective activity that can have a discernible (worthy) effect. It is, ultimately, a duty of common pursuit. It calls us to cooperate with others in the attainment of a worthy goal, to wit, just governance (or less unjust governance). (Ibid., p. 46)

Let us suppose that voting duties are collective duties: citizens ought to vote because through their collective contribution via voting, they are able to promote and reach morally significant ends. Does this give us guidance as to which elections DTV applies? No. Maskivker’s reasoning gives us no reason to privilege presidential elections over local elections. Sure, voting can be seen as a duty of common pursuit, but common for whom? For all people in the country or to one’s immediate community? Unless Maskivker is simply assuming that the common interests are those only at the national level (perhaps because they affect the largest quantity of people), seeing DTV as a duty of common pursuit equally applies to local elections as well. If DTV is a duty to cooperate with others to bring about justice, why privilege pursuing justice at the national level if, for example, the injustice experienced in a local community is much more harmful? If it is not so privileged, then citizens are obligated to engage in the collective activity of bringing about political goods in all elections, which is to fall on the first horn of the dilemma. Maskivker focuses on presidential elections. But again, missing is an argument that shows us that DTV, explicated as a collective duty, applies only to some elections. But if that is the case, then my dilemma remains intact.

Now, we could simply stipulate that I have sufficient political knowledge, have nothing better to do on election day, and the technology is in place to vote online from the comfort of my own home; in this scenario, it seems like I should cast a ballot. However, even if I concede that some people have DTV on some occasions in some elections, this will be a far weaker claim that proponents of DTV make, namely, that the obligation to vote is a general duty of eligible citizens. This is because, on their view, citizenship entails voting, and more generally, DTV is ‘one among many instantiations of a natural duty to promote and support justice’ (Maskivker 2016, p. 225).¹⁸ Brennan and Freiman argue that if Maskivker grounds DTV in the duty to help others, then if there is an alternative available to you that is at least as helpful as casting a vote, you will discharge your duty by doing that alternative (Brennan and Freiman 2021, p. 8). It will not be difficult to think of overriding reasons in favor of not voting because you always have reasons to invest your time and energy in other things such as visiting family, exercising, or cultivating friendships.

Perhaps proponents of DTV have something like this in mind: DTV should be seen as a duty to vote in some elections but not in any particular election. On this view, DTV does not aim to provide an objective answer to the Specification Question and allows citizens to decide for themselves when to vote. Suppose a citizen judges that in a period of a decade, there was only one election that was worth voting in. A more extreme scenario: the citizen votes only once in several decades. Surely, in these scenarios, it would be a stretch to think that the citizen fulfilled DTV. Of course, advocates of DTV might say that the citizen has fulfilled her duties. However, if voting once in a few decades satisfies DTV, this understanding of DTV escapes my dilemma but at the price of going against the spirit of DTV according to which citizens should regularly vote to install just leaders and remove corrupt ones. Again, conceding the point that DTV applies in some circumstances (such as once a year or decade) does not threaten my main argument. So, it looks like we are still left without a good reason why DTV—understood as an imperfect duty or a duty of time and place or a collective duty—applies only to certain elections, and until those reasons are revealed, supporters of DTV have not avoided the second horn of the dilemma.

Conclusion

Advocates of DTV must answer the Specification Question: to which elections does DTV apply? This question has two answers, and both are problematic. I have argued that it is implausible that DTV applies to all elections, but also that there is no good reason to think that under normal circumstances it applies only to some elections. Those attracted to DTV have some explaining to do.

However, even if my argument is sound, voting might be a praiseworthy and even virtuous activity that promotes certain goods for society. My argument does not imply that there are not good reasons in favor of voting, and thus one can consistently reject DTV and nevertheless be an active political participant. But I think that my argument does show that as it stands DTV is implausible. One way to raise its plausibility is to provide a compelling answer to the Specification Question, but until that happens, the case for DTV is incomplete.