Political ignorance is both rational and radical

“If there is one thing in this world

that you can depend on,

that you can bet your last dollar on,

it’s the ignorance of the American people.”

-Pase Rock

Abstract

It is commonly held that political ignorance is rational, a response to the high costs and low benefits of acquiring political information. But many recent critics of the claim that political ignorance is rational instead urge that it is a simple consequence of agents not concerning themselves with the acquisition of political information whatsoever. According to such critics, political ignorance is inadvertent radical ignorance rather than a rational response to the incentives faced by agents in democracies. And since political ignorance is not a response to incentives, these critics urge, it cannot be ameliorated by incentivizing the acquisition of political information. This paper has two goals. First, I show that these seemingly competing accounts of political ignorance are in fact complementary, together explaining much political ignorance. Indeed, there is a sense in which political ignorance can be both rational and radical at the same time. Second, I more closely examine the relationship between incentives, kinds of political ignorance, and the acquisition of political information. On the one hand, from the fact that political ignorance is rational it does not follow that it can be overcome by incentivizing the acquisition of information. On the other hand, from the fact that political ignorance is radical it does not follow that it cannot be overcome by incentivizing the acquisition of information. Lastly, the complexity of the information in question is more relevant to determining whether ignorance can be overcome than whether such ignorance is rational or radical.

1 Introduction

Decades of empirical research has shown that the typical member of the public is ignorant of most politically relevant information (Brennan, 2016; Caplan, 2007; Converse, 1964; Delli Carpini & Keeter, 1996). Unsurprisingly, the average person is ignorant of fields such as economics, political science, sociology, and the like, including relatively uncontroversial findings within such fields. Perhaps more worryingly, though, this ignorance even extends to basic, easily acquired political facts concerning such matters as the identity of elected representatives, the policy proposals of political candidates, the structure and function of important political institutions, and more (Somin, 2013: pp. 17–37; Gibbons, 2021).

There is little disagreement about the fact of such ignorance.¹ As Richard Lau and David Redlawsk put it, “[the] widespread ignorance of the general public about all but the most highly salient political events and actors is one of the best-documented facts in all of the social sciences” (Lau & Redlawsk, 2001: p. 951). There is less agreement, however, concerning the causes of such ignorance. One influential account of political ignorance, stemming from the seminal work of Anthony Downs in An Economic Theory of Democracy, suggests that part of the explanation for such pervasive political ignorance is that it is rational. Given the high costs of acquiring enough political information to vote in a well-informed manner, together with knowledge of the incredibly low probability that any single vote will be decisive, the benefits of becoming politically well-informed are swamped by the costs of acquiring political information. Accordingly, rational agents deliberately remain ignorant.²

However, recent critics of the view that political ignorance is rational allege that this account presupposes some empirically dubious assumptions (Bennett & Friedman, 2008; Evans & Friedman, 2011; Friedman, 2019; Gunn, 2015; Hoffman, 2015). Most agents, these critics urge, have no beliefs whatsoever about their likelihood of casting a decisive vote, and so no belief—let alone knowledge—that they are unlikely to cast a decisive vote features in their decision to refrain from acquiring political information. Going further, these critics also claim that political ignorance is typically not a deliberate choice at all. Far from deliberately forgoing the acquisition of political information, agents are usually ignorant of their ignorance. This higher-order radical ignorance indicates a lack of will or purposiveness rather than the cognitively sophisticated cost–benefit analyses often imputed to agents wherein political ignorance is a deliberate choice.

Determining which of these accounts of political ignorance is correct is naturally of interest to those eager to know why political ignorance is often so profound, but the dispute between proponents of these two competing models is, at least at first glance, not without practical significance. If widespread political ignorance is rational, then perhaps such ignorance can be overcome by providing the appropriate incentives for agents to acquire political information. But if such ignorance is not a deliberate, rational response to the incentives agents in democracies face, then perhaps it cannot be overcome by providing incentives to acquire political information (Evans & Friedman, 2011: pp. 78–9; Gunn, 2015: p. 292). If so, significant amounts of time, energy, and resources may be devoted to practical efforts to improve levels of political information among the public that have no real prospect of success.

This paper has two goals. First, I argue that these seemingly competing accounts of political ignorance are in fact complementary. On the one hand, even if we grant the claim that rational ignorance must be a deliberate and conscious choice, it is nonetheless true that much political ignorance is rational. Models of rational ignorance need not appeal to the empirically dubious assumptions which proponents of radical ignorance reject, and more psychologically realistic models of rational ignorance plausibly capture the actual decision-making of many agents.³ Such models can operate alongside models of radical ignorance, together accounting for much political ignorance. On the other hand, if we discharge the assumption that rational ignorance must be a deliberate and conscious choice, then it becomes clear that political ignorance can be both rational and radical at the same time, thus dissolving the apparent conflict between these two accounts. As we will see, radically ignorant agents can also be rationally ignorant insofar as their ignorance is an unconscious response to the incentives they face.

Second, I argue that the relationship between whether political ignorance is rational or radical and whether it can be overcome by incentivizing the acquisition of information is more complicated than commonly assumed. From the fact that political ignorance is rational it does not follow that it can be overcome by incentivizing the acquisition of political information. However, contrary to proponents of the claim that political ignorance is typically radical, even if political ignorance is not a response to incentives, it can in principle be overcome by providing the right incentives. Ultimately, I suggest that the complexity of the political information in question is more relevant to determining whether we can incentivize agents to acquire it than whether the underlying ignorance is rational or radical. Thus, some practical interventions to improve levels of political information may succeed, but their chances of doing so decrease as the complexity of the relevant information increases.

I proceed as follows. In Sect. 2 I outline in more detail traditional models of rational ignorance, before discussing recent criticisms of such models from scholars who claim that political ignorance is radical rather than rational. In Sect. 3 I push back against these criticisms, arguing that both accounts of political ignorance are complementary rather than conflicting. Section 4 contains extended discussion of the practical ramifications of the fact that political ignorance is both rational and radical, focusing on the relationship between incentives and the acquisition of political information. Section 5 concludes.

2 The traditional account of rational political ignorance

The claim that political ignorance is rational has been held by many economists, philosophers, and political scientists (Brennan, 2016; Caplan, 2007; Downs, 1957; Gilens, 2001; Riker & Ordeshook, 1968; Somin, 2013; Stigler, 1976; Tullock, 1972; Zaller, 1992).⁴ According to proponents of this account, pervasive political ignorance is a consequence of rational agents recognizing the fact that the benefits of acquiring political information are greatly outweighed by the associated costs.

First, consider the costs. As already mentioned, the costs of acquiring relevant political information are quite steep (Somin, 2008: pp. 258–9). Agents in democracies arguably need to acquire information about, inter alia, the platforms of competing candidates and political parties, the likely effects of different policy proposals, whether (and how) certain political agents are responsible for outcomes of concern to them, which problems we can reasonably expect government officials to address, and so on (Downs, 1957: p. 208; Somin, 2013: pp. 40–2).⁵ Additionally, to make good use of this information, agents need to both analyze and evaluate it (Downs, 1957: p. 210). Naturally, this merely increases the associated costs of acquiring political information in order to vote in a well-informed manner. Such activities take significant time and effort, much more than most are willing to allocate towards such ends. Thus, the opportunity costs of acquiring political information are extremely high.⁶

Next, consider the benefits of acquiring (and subsequently analyzing and evaluating) political information. Insofar as agents desire to influence the political process by casting decisive votes for their preferred candidates or political parties, the instrumental benefits of becoming politically well-informed are extraordinarily slight. Agents are assumed to know that the probability of casting a decisive vote in any given election is extremely low (though nonzero).⁷ Consequently, they come to believe that the eventual result of the relevant election will be the same no matter how they cast their ballot, and the instrumental benefits of acquiring the information needed to cast that vote in a well-informed manner are thereby diminished (Downs, 1957: p. 244). If agents knew that their votes were more likely to be decisive, the (subjective) instrumental benefits of acquiring the information needed to vote well would increase. But as things stand, the lower the likelihood of casting decisive votes, the lower the instrumental benefits of acquiring political information. Overall, then, when the low instrumental benefits of acquiring political information are weighed against its steep costs, rational agents deliberately refrain from it; and it is this sort of rational deliberation that is purported to explain the widespread political ignorance we find in modern democracies.⁸

One might object to the preceding account of rational ignorance on several grounds. First, one might think that it exaggerates the costs involved in acquiring the political information needed to vote in a well-informed manner. Rather than needing to procure all the relevant information themselves, rational agents can instead rely upon various heuristics which greatly reduce the costs of acquisition (Lupia & McCubbins, 1998; Oppenheimer & Edwards, 2012; Popkin, 1991; Sinhababu, 2016; Wittman, 1995). In short, by relying upon information disseminated by opinion leaders, political parties, media outlets, and more, agents can become well-informed without the need to engage in the painstaking and time-consuming task of acquiring political information themselves.

Second, one might think that it understates the benefits of information acquisition. For one thing, we need not assume that agents acquire political information only for the purposes of influencing the outcome of the political process. As Downs himself notes, “[some] people obtain information as an end in itself” (Downs, 1957: p. 214). Some people simply want to acquire political information for its own sake, whether out of a sense of curiosity or a sense of duty, and for such individuals there are benefits to information acquisition independently of its utility in influencing political outcomes. Others may want to acquire political information in ways they find entertaining (Somin, 2013: pp. 78–81).⁹ Whatever the case, ignoring these benefits may lead us to underestimate the total expected benefits of information acquisition. For another, several philosophers and social scientists have recently argued that the expected value of casting a vote in certain elections—especially elections in swing-states—is high enough that voting becomes rational (Edlin et al., 2008; MacAskill, 2015: p. 86; Barnett, 2020). And if that’s true, one might further think that taking the time to acquire the information to make good use of that vote is also a sound use of one’s time and energy. Again, the result is that the expected benefits of acquiring political information are understated by the traditional account of rational political ignorance.

However, neither sort of objection threatens the claim that political ignorance is rational. While some heuristics may be useful for some agents in certain contexts, others may in fact mislead them, with their efficacy easy to overstate (Lau & Redlawsk, 2001). And determining which of the many available heuristics are in fact reliable—which of the many heuristics in fact reduce the costs of acquiring political information—is itself a time-consuming and demanding endeavor. Accordingly, the costs for acquiring the relevant information reemerge to a great extent (Somin, 2008: p. 260; Somin, 2013: pp. 90–118). Similarly, information acquired from entertainment sources will not suffice to furnish agents with all the information necessary to be well-informed, even assuming it is accurate information.¹⁰ The costs of acquiring much information therefore remain to some extent. Finally, Jason Brennan and Christopher Freiman have persuasively argued that the case for rational voting only applies to a small subset of epistemically privileged agents (Brennan & , 2023). For most, then, even if voting is rational, the costs of information acquisition remain steep enough that information acquisition is not rational.

More importantly, both sorts of challenges take for granted the claim that awareness of the costs and benefits of acquiring political information informs the decision-making of agents who elect to remain ignorant. One might think that traditional models of rational ignorance overestimate the costs of information acquisition, but this is consistent with thinking that if the costs were sufficiently steep, rational agents would deliberately refrain from acquiring political information. One might think that traditional models of rational ignorance underestimate the probability of casting a decisive vote, but this is consistent with thinking that if the probability of doing so were sufficiently low, rational agents would deliberately refrain from acquiring political information. However, more fundamental criticisms of traditional models of rational ignorance take issue with the very idea that awareness of such costs and benefits plays a role in the decision-making of agents whatsoever.

Firstly, one might reject the claim that typical agents are aware—even roughly or intuitively—of the low probability of casting a decisive vote in most elections. After all, it isn’t obvious that people are able to intuit facts related to complicated cost–benefit analyses involving, inter alia, millions of other people (Friedman, 2019: p. 279). Moreover, further indirect evidence for the claim that agents are unaware of the low probability of casting decisive votes comes from the fact that most who decline to register to vote do not cite the fact that their vote will not make a difference (Friedman, 2019: p. 278). Perhaps this just means that those who do not register to vote (or who ultimately do not vote) refrain from doing so on grounds unrelated to the low probability of casting a decisive vote, despite their awareness of this probability. But given the available evidence, we should be open to the claim that, for many agents, belief or knowledge of their vote’s inefficacy does not play a role in their decision to refrain from acquiring political information. As Jeffrey Friedman writes, “if people have not thought about the odds that their votes are likely to tip the outcome [of an election], then they cannot possibly have decided that because the odds are low, they should not bother to inform themselves about public affairs” (Friedman, 2019: p. 280).¹¹

Second, one might deny the claim that political ignorance is a deliberate choice whatsoever. For example, Tom Hoffman, drawing upon the work of Joseph Schumpeter in his classic Capitalism, Socialism, and Democracy, argues that pervasive political ignorance stems from a “lack of will or purposiveness on the part of the public” rather than consideration of the overall costs and benefits of acquiring information (Hoffman, 2015: p. 301). Oversimplifying somewhat, typical agents are more naturally viewed as being concerned with a relatively narrow range of things that directly concern them (such as family, work, and so forth), with politics much less frequently occupying their attention (Schumpeter, 1942: pp. 258–9; Hoffman, 2015: p. 311). Consequently, they remain politically ignorant not because they have considered the prospect of acquiring political information to be too costly and therefore chosen to remain ignorant, but because they are not concerned with the process of acquiring political information at all (with occasional exceptions).

On this account, agents are “inadvertently” and “unwittingly” politically ignorant; and because this political ignorance has not been chosen, it is neither rational nor irrational (Bennett & Friedman, 2008: p. 198; Evans & Friedman, 2011: p. 75). Unaware of their own ignorance, such an account characterizes much of the public as radically ignorant (Bennett & Friedman, 2008; Friedman, 2005, 2019; Ikeda, 2003).¹²

To sum up, we are faced with two seemingly incompatible accounts of public political ignorance. Accounts of rational ignorance characterize political ignorance as a rational response to the incentives present in modern democracies. Rational agents, being aware of the high costs and low benefits of acquiring political information, deliberately forgo it. Meanwhile, accounts of radical ignorance characterize political ignorance as inadvertent, a simple consequence of most agents concerning themselves primarily with activities and interests unrelated to politics. In short, these accounts offer very different explanations for pervasive political ignorance. But they also plausibly have importantly different practical implications for efforts to improve levels of political information among the public. When viewed as a problem stemming from rational agents electing to remain ignorant in response to the overwhelming costs and minuscule benefits of information acquisition, public political ignorance seems like a problem that could be overcome by providing the appropriate incentives for agents to become well-informed. By making it beneficial for people to become well-informed (or, alternatively, by making it costly to be ill-informed), we could incentivize the acquisition of information. But if political ignorance is not the result of purposive behavior whatsoever, if it is instead completely inadvertent, providing incentives to become well-informed may be futile. As Paul Gunn writes, “[if] voters are radically ignorant—if they do not know precisely what it is they would benefit by knowing—then their ignorance will persist even if they have every incentive to know what they need to know” (Gunn, 2015: p. 292). Determining which account of political ignorance is correct is thus seemingly important from both an explanatory perspective and a practical perspective.

3 Rethinking rational political ignorance

Traditional models of rational ignorance encounter sharp criticism on the grounds that they rely upon empirically dubious assumptions. The first assumption is that rational agents are aware of the low probability of casting a decisive vote in any given election, and that this awareness plays a crucial role in their decision to refrain from information acquisition. The second assumption is that rational agents engage in a deliberate process of assessing the overall costs and benefits of information acquisition whatsoever. Both, as we have seen, may well be false: it is not clear whether typical agents have any beliefs about their likelihood of casting decisive votes, and much political ignorance may be unchosen radical ignorance rather than the result of some process of deliberation.

At the core of this critique lies the assumption that being consciously and deliberately chosen is a necessary (but insufficient) condition for ignorance to be rational. Towards the end of this section, we will consider whether to reject this assumption—that is, we will consider the possibility that political ignorance can be rational without at the same time being consciously and deliberately chosen. For now, though, let’s grant the assumption that rational ignorance must be deliberately chosen. Even so, I argue that at least some political ignorance is rational in this sense. More psychologically realistic models of rational ignorance can jettison the empirically dubious assumptions outlined earlier, and political ignorance can be rational even if most agents neither have any beliefs about the decisiveness of their votes nor any propensity to engage in complex assessments of the overall costs and benefits of information acquisition.¹³

To see how political ignorance could be rational even if these assumptions are discharged, first consider the assumptions that typically lie at the core of rational choice models. Towards the beginning of An Economic Theory of Democracy, Downs outlines the following:

A rational man is one who behaves as follows: (1) he can always make a decision when confronted with a range of alternatives; (2) he ranks all the alternatives facing him in order of his preference in such a way that each is either preferred to, indifferent to, or inferior to each other; (3) his preference ranking is transitive; (4) he always chooses from among the possible alternatives that which ranks highest in his preference ordering; and (5) he always makes the same decision each time he is confronted with the same alternatives (Downs, 1957: p. 6).¹⁴

Although some of these assumptions may be empirically questionable in their own right, note that they do not commit one to making any assumptions about what it is in virtue of which agents rank the alternatives before them. Further assumptions about what makes agents prefer one thing to another should be considered as auxiliary assumptions, the empirical adequacy of which should be ascertained on independent grounds. Consequently, those seeking to model political ignorance as rational need not commit themselves to the claim that agents choose to remain ignorant on the grounds that their votes are too unlikely to be decisive, nor do they need to commit themselves to postulating any sort of complex cost–benefit analyses whatsoever. Instead, for political ignorance to be rational, agents must be confronted with a range of options among which there are activities related to the acquisition of political information, and they must furthermore rank higher—and thereby consistently choose—those options which do not involve the acquisition of political information. For this to be the case, all sorts of considerations may influence their decision-making.

Why, then, might agents choose to avoid acquiring political information? First, they may have reasons to refrain from deliberately acquiring political information in general. For instance, something the traditional model of rational ignorance gets right is that the opportunity costs of acquiring political information are plausibly too high for many people. Setting aside information that is acquired in a happenstance manner, the acquisition of political information requires time and effort. All time allocated to such ends is time taken away from other ends which agents may prefer to allocate time towards. With the exception of people for whom the acquisition of political information is enjoyable, as well as people for whom the acquisition of political information is treated as a duty, very many people will simply prefer to spend their time doing other things. Perhaps they are politically apathetic. Perhaps engaging in politics makes them feel stressed or anxious, with deliberate information avoidance being a way to alleviate this stress or anxiety (Sweeny et al., 2010). Or perhaps they just get more enjoyment from other activities (Hardin, 2009: p. 61). Whatever the case, refraining from the acquisition of political information is rational for such people; they prefer other alternatives and consistently choose accordingly.

Second, agents may have reasons to be selective in how they acquire political information. While not avoiding the acquisition of political information entirely, they deliberately refrain from acquiring such information in certain contexts. For instance, they may prefer to acquire political information from some sources but not others. Much like engagement with politics more generally, some may find it stressful or anxiety-inducing to engage with certain sources of political information, especially if these sources reliably adopt or evince political viewpoints counter to those held by the person in question.¹⁵ Additionally, such agents may wish to avoid information that conflicts with beliefs the public expression of which they view as socially desirable in some sense (Williams, 2020a: pp. 5–6).¹⁶ Others may have reservations, whether well-founded or not, about the accuracy of the information that could be acquired from certain sources, opting to deliberately refrain from relying upon the sources they deem unreliable. Lastly, some agents, wishing to avoid information overload, may opt to cease acquiring information once they have acquired whatever amount they think suffices to make a decision at the voting booth, or to be suitably well-informed, and so on. Of course, agents may have some combination of the above reasons, as well as others not discussed here.¹⁷ Whatever the case, agents will often prefer to deliberately refrain from acquiring at least some information.

Naturally, the above is not meant to constitute a comprehensive list of ways in which agents can be rationally ignorant. Such examples are simply meant to show that agents can be rationally ignorant for a variety of reasons, even if they lack beliefs about the probability of casting decisive votes, and even if they do not engage in cognitively demanding or otherwise unrealistic cost–benefit analyses.¹⁸

Still, one might accept that agents could be rational for the reasons just outlined while doubting that many are in fact so rational. While beliefs about the decisiveness of votes or propensities to engage in complex cost–benefit analyses are optional auxiliary assumptions that need not play a role in rational choice models, assumptions about transitive preferences or assumptions about always making the same decision when confronted with the same options (and others) are not so readily dischargeable. But perhaps these too are unrealistic assumptions to make.¹⁹ If they are false of actual agents, then perhaps we should conclude that no actual agent is rationally ignorant.

However, this criticism is open to two responses. First, it is an open question whether agents in fact ever satisfy the relevant conditions.²⁰ Perhaps some fail to satisfy these conditions and so fail to be rational in the intended sense. But other may be more adequately characterized by the relevant assumptions. If these assumptions hold in the context of their decisions to refrain from the acquisition of information, then their ignorance is rational. Second, and more importantly, the conception of rationality found in standard rational choice models is not the only available conception of rationality. Even if agents who choose to remain to some extent ignorant do not satisfy all the conditions of standard rational choice models, their behavior may yet be rational in the sense that it is indicative of purposive, means-ends reasoning. In short, political ignorance may be instrumentally rational for many agents (Kolodny & Brunero, 2020). And this, of course, is precisely what proponents of models of radical ignorance deny when they claim that political ignorance indicates a lack of purposiveness.

Another criticism, though, is more difficult to dismiss. The preceding account of rational ignorance, assuming it is sound, establishes only that political ignorance can be rational even if the sort of assumptions characteristic of traditional models of rational ignorance are discharged. But even on this account, political ignorance is chosen on various grounds. Proponents of radical ignorance may complain that this is to miss the point of their account of political ignorance. According to their account, political ignorance is not rational because it is entirely unchosen, not simply because it is not chosen for the sort of reasons traditional models impute to agents. Of course, proponents of radical ignorance may insist that political ignorance is typically unchosen, and that both the traditional model of rational ignorance and the more psychologically realistic models of rational ignorance sketched above fail to explain most actual political ignorance.

At this point, the debate between these competing accounts of political ignorance seems indecisive, with proponents on each side claiming that their account best explains political ignorance in general. However, the question of which account more accurately captures the political ignorance of actual agents is an empirical one, and there is no reason to assume a priori that the pervasive political ignorance we encounter in democracies is either entirely rational or entirely radical. Perhaps the political ignorance of some is, in general, best captured by more psychologically realistic models of rational ignorance. Indeed, the political ignorance of some may be best captured by traditional models of rational ignorance, even if some of the assumptions such models make are unrealistic for most people. For example, perhaps some economists, political scientists, and philosophers—especially those influenced by the work of Downs and others who have popularized models of rational political ignorance—are accurately captured by such models.²¹ At the same time, though, the political ignorance of others is perhaps best characterized as radical rather than rational, regardless of whether we have in mind traditional accounts of rational ignorance or more psychologically realistic accounts of rational ignorance.

Furthermore, in much the same way that we shouldn’t assume a priori that either account offers the single, dominant explanation for political ignorance in general, we also should not assume that the entirety of any given agent’s political ignorance is either wholly rational or wholly radical. Instead, we should be open to the possibility that, for many agents, some of their political ignorance is best characterized as rational while some is best characterized as radical. In other words, agents consciously forgo the acquisition of some information while being unaware of their ignorance of different information. These competing accounts of political ignorance should thus be seen as potentially complementing one another, each capturing the behavior of different agents (or the same agents in different contexts), rather than as competitors which must exclude the legitimate application of the other, explaining political ignorance in toto.

Nevertheless, while acknowledging that we should not assume that either account provides the dominant explanation for political ignorance in general, one might reasonably wonder about the relative prevalence of each kind of political ignorance. And though it is true that to what extent each account can explain the behavior of different agents in different contexts is an empirical question, I offer here some tentative conjectural remarks to the effect that each kind of political ignorance prevails more heavily when different types of facts are concerned.

Generally speaking, the available empirical evidence heavily indicates that at least some people remain deliberately ignorant about certain subjects, for a variety of reasons, at least some of the time.²² And it is prima facie plausible that the reasons people have for remaining ignorant about these subjects will, on occasion, apply in cases involving the acquisition of political information. In particular, I suggest that rational political ignorance—especially psychologically realistic accounts of rational ignorance—can account for widespread ignorance of the sort of basic political facts mentioned at the beginning of this paper, as well as ignorance of the sort of political information which people are likely to hear or read about in their extended networks.²³ For instance, one might hear one’s work colleagues discussing an electoral candidate’s economic policy platform, thus becoming aware that such a platform exists (if one weren’t already). Still, one might opt to pursue the matter no further, for the sort of reasons outlined earlier in this section. In short, ignorance of basic, readily acquired political information, and information the existence of which is commonly made salient to typical agents, may be in large part rational ignorance.

At the same time, though, it seems highly plausible that at least some political ignorance is radical. All that is required for this to be the case is for some people to be politically ignorant because the option to acquire political information was not an alternative they considered at all, in the same way that they are ignorant of countless other subjects which never occupy their attention. And this, I think, is a very reasonable assumption to make about many people, especially when considering more selective forms of radical ignorance—that is, radical ignorance of certain sources of political information, or information about certain political issues, and so forth. Consider, for instance, the typical person’s relationship to the sort of complex and often obscure debates one can find in fields such as political philosophy, economics, and so on. For example, most people are not just ignorant of the academic literature on the Lockean proviso, they are radically ignorant of it.²⁴ They do not know they could acquire information on this matter while simply choosing not to; instead, they are not even aware of the existence of such information. And so it goes for many other similar disputes, though they are politically relevant in some sense. But this, of course, is paradigmatic radical ignorance.

At the very least, then, it is reasonable to conclude that our two competing accounts—rational political ignorance and radical political ignorance—are complementary insofar as both are needed to explain pervasive public political ignorance, with each better accounting for political ignorance of different kinds of information. Political ignorance is both rational and radical.

Before moving on, though, let’s return to the assumption with which this section began—namely, the assumption that political ignorance must be, at minimum, deliberately and consciously chosen in order for it to be rational. There are in fact good reasons to reject this assumption, and, once it is relinquished, it becomes clear that political ignorance can be both rational and radical at the same time.

Most notably, this assumption is rendered implausible by the fact that political ignorance may be an unconscious response to the incentives agents face. A full treatment of the way our cognitive processes—belief formation, information acquisition, information processing, and the like—can be unconsciously shaped by the incentives we encounter goes beyond the scope of this paper. But for our purposes it suffices to note that there is a large body of literature in social and political psychology where the claim that agents unconsciously respond to incentives in this manner is uncontroversial.²⁵ As Daniel Williams puts it while discussing motivated ignorance, “it is implausible that the possibility of conscious deliberation is necessary for instrumental rationality given how much decision making occurs in the absence of such deliberation” (Williams, 2020a: p. 7816). Assuming this is correct, it is easy to see how political ignorance can be both rational and radical at the same time. If conscious deliberation is not necessary for rational ignorance, then one can be rationally ignorant while being unaware of—thus radically ignorant of—one’s ignorance.

Prior treatments of the differences between these two accounts of political ignorance placed them in diametric opposition to one another. But if the preceding discussion is correct, then not only are our two competing accounts of political ignorance complementary in the sense that both may be needed to explain different kinds of political ignorance, but they are also often co-extensive, insofar as the same individual’s ignorance can be both rational and radical simultaneously. The seemingly stark conflict between both accounts has been largely dissolved.

4 On the continuing (but limited) relevance of incentives to political ignorance

If the previous section’s arguments are sound, then models of rational ignorance and radical ignorance are complementary rather than conflicting. At minimum, they are both needed to account for widespread political ignorance, with each better capturing different kinds of political ignorance. At most, if we drop the assumption that rational ignorance must be a conscious and deliberate choice, political ignorance can be both rational and radical at the same time. And this, one might think, is good news for those who seek to improve levels of political information among the general public by incentivizing the acquisition of such information. Recall that proponents of radical ignorance claim that if political ignorance is radical, then it cannot be overcome by providing incentives to acquire information. As Anthony Evans and Jeffrey Friedman write, “radical ignorance…cannot be affected by one of the main weapons in economists’ conceptual arsenal, incentives” (Evans & Friedman, 2011: p. 78). But if some political ignorance is rational, then perhaps incentives can play a role in improving levels of political information after all, at least to some degree. Specifically, rational political ignorance can be overcome by incentivizing the acquisition of information, even if some radical political ignorance—that which is not also an unconscious yet rational response to incentives—is “immune” to incentives (Ibid, 78).

However, such a conclusion oversimplifies the relationship between incentives, types of political ignorance, and the acquisition of information. Indeed, it errs in at least two important ways: (1) from the fact that ignorance is rational it does not follow that it can be overcome by incentivizing the acquisition of the relevant information, and (2) from the fact that ignorance is radical it does not follow that it cannot be overcome by incentivizing the acquisition of the relevant information. Indeed, this is true even of radical ignorance that is not also at the same time rational.

Let’s first consider how political ignorance can be overcome by incentivizing the acquisition of information. For rational political ignorance of the sort that involves conscious deliberation, the process seems quite straightforward. Agents often remain deliberately ignorant because they have subjectively compelling reasons to refrain from information acquisition. But if the process of information acquisition promises more benefits (or if the failure to acquire information yields costs of some kind), then those reasons may become substantially less compelling. One possible way to increase the benefits of information acquisition is to offer monetary incentives to acquire information. For example, one could simply pay people to take and subsequently score highly in basic civics exams (Lupia, 2016: p. 174–77; Somin, 2023: pp. 299–300). For many, once enough money is offered, the acquisition of information will become an attractive enough option that they no longer prefer to avoid it.²⁶ Similarly, agents who have heretofore unconsciously responded to the incentives they face by remaining politically ignorant may suddenly find themselves with a newfound willingness to acquire political information. Notably, since the promise of monetary reward would be salient to these agents, their previously unconscious rational ignorance may be replaced by a conscious and deliberate effort to acquire political information.

Of course, one might have various reservations about such a proposal. For instance, one might worry that the process of designing the relevant civics exams will be distorted by political bias. For another, one might worry that it will disproportionately benefit certain groups in unfair ways.²⁷ But the point here is not that this approach is the all-things-considered best way to incentivize the acquisition of information. Instead, it is merely to provide an illustrative example of a relatively straightforward and intuitively plausible way to get rationally ignorant people to begin to acquire more political information.

More importantly, this sort of simple approach could easily incentivize radically ignorant people to acquire information. Proponents of the claim that political ignorance is typically radical think that because political ignorance is not a response to incentives, it cannot be overcome by providing suitable incentives to acquire information. The underlying worry seems to be that since agents are unaware of the information of which they are ignorant, they cannot know the value of acquiring it. But without knowing the value of acquiring some information ex ante, agents have no incentive to acquire it (Evans & Friedman, 2011: p. 78). If anything, though, the offer of a precise monetary reward is a way to make salient the value of acquiring some information. And if the value of acquiring this information is made salient, agents thereby gain an incentive to acquire it. Their ignorance may begin as radical, an incidental result of not having the acquisition of political information as an option from among which they choose. But making them aware of the potential to earn money is a way to change the options they confront, whereupon they may subsequently choose to acquire the information they previously were unaware of.

Thus, even if political ignorance is radical—and even if it were true that radical political ignorance could not at the same time be unconsciously rational ignorance—it can be overcome by offering money to acquire information. In fact, Evans and Friedman concede this point, writing that the “easiest case in which one could know in advance the costs and benefits of acquiring certain information might be if one were directly paid to learn it” (Evans & Friedman, 2011: p. 80). They go on to stress that such situations are relatively infrequent (Ibid, 80). But for our purposes, this is irrelevant. What matters is that we could, in principle, incentivize the acquisition of some information by paying people to learn it, regardless of whether their initial ignorance was rational or radical.

At this point, though, various limits to our ability to incentivize the acquisition of information become clear. First, the monetary value assigned to the acquisition of information may be insufficient to induce people to acquire the relevant information. The value required will vary from person to person, with the preferences of some remaining so strong that only infeasibly large amounts of money could induce the deliberate acquisition of information. Thus, while rational ignorance can sometimes be overcome by providing monetary incentives, doing so is not guaranteed to overcome it for every rationally ignorant agent.²⁸

Naturally, of course, the same is true of radical ignorance, and some radically ignorant agents will not be induced to acquire information even if made aware of the value of doing so. But this brings us to a second limitation to our ability to incentivize the acquisition of information in radically ignorant people—namely, it is dependent not only on providing enough value to them, but also on making them aware of the value in the first place. And while this is quite straightforward for approaches that involve paying people to learn, it is much less so for other approaches to incentivizing the acquisition of information.

Consider, for instance, foot voting of the sort defended by Ilya Somin (Somin, 2020). He argues that giving people more opportunities to vote with their feet—that is, to choose which local or regional government to live under—can enable them to become better informed than they would be otherwise. As he puts it, “foot voters have strong incentives to seek out relevant information because their decisions are highly likely to matter” (Somin, 2023: p. 303). Why? Because unlike in standard electoral systems where our individual actions are unlikely to make a difference to the laws and policies we live under, the individual actions of agents are highly likely to make such a difference in systems which enable foot-voting (such as federal systems). After all, if agents discover that they would prefer to live under the laws and policies of a different local or regional government, they can simply choose to move there.²⁹

Much empirical research seemingly vindicates Somin’s optimism about the ability of foot voting to enable people to both acquire more information and make better use of what information they acquire (Somin, 2020: pp. 15–44). It empowers them to engage in open-ended searches for whatever information they deem necessary to make decisions regarding which jurisdictions to live under. At the same time, though, with foot voting the value of acquiring any given piece of information is neither precise nor obviously salient to agents. And since the value of the information is not necessarily salient, it is uncertain whether foot-voting can counteract radical ignorance since, by definition, radical ignorance involves information the existence of which agents are unaware. While agents may be induced to engage in more open-ended deliberate information acquisition, they cannot be induced to acquire information the existence of which is not salient to them. This, of course, contrasts with paying people to learn, where the relevant range of information, though much narrower than what may be acquired with open-ended searches, has a precise and predetermined value assigned to it and a greater degree of salience that can overcome even radical ignorance.

We therefore encounter a trade-off. On the one hand, incentivizing agents to engage in more open-ended searches for information can in principle lead to a wider range of information being acquired than with other methods, but the value and salience of any given piece of information is uncertain, and this renders radical ignorance more difficult to overcome, including radical ignorance of basic, uncontroversial facts. On the other hand, incentivizing agents to learn by paying them can better overcome radical ignorance by making the value of some information salient, but the range of information acquired is likely to be much narrower than with methods that allow open-ended searches. The practical significance of this trade-off is unclear, since both methods of incentivizing the acquisition of information could be implemented together. And even if they couldn’t, there may be overriding reasons to prefer one over the other, regardless of their impact on radical ignorance. Still, it is worth bearing in mind the different limitations facing different methods of incentivizing information acquisition.

In that vein, one final limitation facing efforts to improve levels of political information needs to be mentioned, one heavily emphasized by proponents of the claim that political ignorance is radical—specifically, the “inherent complexity of politics” (Friedman, 2005: p. 11).³⁰ This complexity is foregrounded since it is thought to present an alternative explanation of widespread political ignorance to more mainstream explanations found in traditional accounts of rational ignorance. For instance, Gunn claims that “we must confront the possibility that [political ignorance] results from the unnoticed complexity of political choices rather than the absence of sufficient incentives to become informed” (Gunn, 2015: p. 287). Elsewhere, while criticizing Somin’s attempts to root political ignorance in a lack of incentives to acquire information, he writes that “missing from Somin’s analysis…is the sheer difficulty of accurately and holistically interpreting the complex social world” (Ibid, 24).

Political systems are massively complex, dynamical systems that are difficult to describe or predict (Page, 2010). Given this complexity—given the sheer number of entities and processes to consider—it is often incredibly difficult to acquire facts regarding, inter alia, the consequences of implementing different laws and policies, the downstream results of institutional reform, and so forth.³¹ Accordingly, the claim that the complexity of the social and political world constitutes a fundamental impediment to our ability to acquire political information should be granted. However, there are two problems with Gunn’s analysis of the relationship between political ignorance and complexity.

First, not all political ignorance is of complex, epistemically inaccessible facts. As we noted at the beginning of this paper, agents are often ignorant of even basic political facts. One can be either rationally or radically ignorant of such facts (or both): rationally ignorant agents may simply prefer to not take the time to acquire them despite knowing that they easily could, while radically ignorant agents do not consciously confront the option to acquire such information whatsoever. Given the availability of more psychologically realistic models of rational ignorance, together with the possibility of unconsciously rational ignorance, it is highly plausible that accounts of rational political ignorance still have a role to play in explaining some political ignorance of basic political facts, even if much political ignorance instead results from the complexity of politics.

Second, such complexity presents a barrier to attaining knowledge even if people are incentivized to acquire the relevant facts. Consider an analogy: one can be incentivized, by the offer of massive sums of money, to play the drums as well as Tomas Haake, but still fail to ever become that skilled. Likewise, one can be incentivized to predict the outcome of interventions into a massively complex, dynamical system but still fail to make an accurate prediction because of the epistemic hurdles one encounters. Thus, one need not appeal to accounts of radical ignorance in order to explain the epistemic limitations presented by complexity.

In a sense, the complexity of putative political facts is more relevant to determining whether we can incentivize people to acquire them than whether people’s ignorance is rational or radical. For simple facts that can be easily ascertained—facts of the sort that appear in civics exams, for instance—we can in principle incentivize people to acquire them regardless of whether their initial ignorance is rational or radical. Paying agents to learn, as we have already seen, can overcome both kinds of political ignorance. But for complex facts, providing incentives may be insufficient given the steep epistemic limitations involved, even if people are aware of their ignorance and deeply motivated to overcome it. This doesn’t mean that the provision of incentives is irrelevant. Even if it is incredibly difficult to overcome our ignorance, it may still be worthwhile to incentivize people to attempt do so, both because doing this plausibly increases the chances that we successfully overcome it and because there may be incidental benefits of having people conduct research, engage in open-ended reflection on the bounds of our knowledge, and so on, regardless of whether we answer the questions we set out to answer in the first place.

Still, we should be realistic about what such efforts to incentivize the acquisition of information can achieve. While we may be able to overcome ignorance of simple political facts, our chances of overcoming ignorance decrease as the complexity of the relevant facts of which we are ignorant increases. Consequently, we need to manage our expectations regarding the prospects of practical efforts to incentivize the acquisition of such information.

5 Conclusion

There is a longstanding debate within economics, philosophy, and political science regarding the best explanation of public political ignorance. Some argue that such ignorance is rational, a consequence of agents recognizing the fact that the steep costs of becoming well-informed far outweigh the slight benefits conferred by a single vote with a miniscule probability of decisiveness. Others, meanwhile, argue that models of rational ignorance are empirically dubious, instead claiming that public political ignorance is radical, a consequence of agents not considering the option to acquire political information whatsoever. In addition to offering competing explanations of public political ignorance, these accounts are thought to have importantly different practical implications for efforts to improve levels of political information among members of the general public: if political ignorance is rational, then perhaps it can be overcome by providing agents with the right incentives; but if political ignorance is radical, then it is immune to incentives.

As we have seen, though, this debate rests on several mistakes. First, models of rational ignorance can discharge the relevant empirically dubious assumptions. Alternative, more psychologically realistic models of rational ignorance can play a role in explaining some political ignorance. Second, and relatedly, we should view these seemingly competing accounts as complementary, both needed to account for the political ignorance so pervasive in democracies. Indeed, once we drop the assumption that rational ignorance must be deliberately and consciously chosen, we can even see that political ignorance can be both rational and radical at the same time. Third, the relationship between incentives and types of political ignorance is not as straightforward as standardly assumed. Even if political ignorance is rational, providing incentives may not be enough to overcome it. But in a more positive vein, providing incentives to acquire political information can, in principle, overcome even radical ignorance. More relevant than whether political ignorance is rational or radical, though, is the complexity of the information of which we are ignorant. While ignorance of simple facts may be overcome by incentivizing their acquisition, we may remain ignorant of complex facts even if highly incentivized and motivated to do so. Thus, practical efforts to improve levels of political information may prove useful, but their efficacy diminishes as the complexity of the information in question increases.³²