The case for epistocratic republicanism

Epistocracy

David Estlund (2003: 53) introduced the term ‘epistocracy’ to describe rule by the knowledgeable or the wise into the lexicon of contemporary political thought. However, the roots of epistocracy run deep with Plato and John Stuart Mill held up as its most important forerunners (Estlund, 2003: 55–57; Klocksiem, 2019: 1; Lippert-Rasmussen, 2012: 242; Mulligan, 2015: 461–462). Estlund (2009: 30) defines the argument for epistocracy as resting on three ‘tenets’:

Estlund accepts the truth and knowledge tenets but is not convinced that superior knowledge can justify political authority. The core of his argument is the ‘demographic objection’, which asserts that the educated population of a political community may have damaging epistemic biases that cannot be controlled for in a sample but lurk undetected. These biases do not have to be driven by malevolence; there just has to be a reasonable chance that they could be present but undetected in the minds of the educated (Estlund, 2009: 215–217). This makes even moderate epistocracy fail the test of ‘no invidious comparisons’, which requires any group that claims expertise needs to pass a general acceptability test (Estlund, 2009: 36). A reasonable person might worry about the biases, subconscious or otherwise, that a privileged group may possess.

Brennan, however, attempts to sidestep this critique by denying the authority tenet. In Estlund’s argument, the third tenet is that superior knowledge about politics warrants political authority; it is a positive assertion. Brennan reframes the third tenet in negative terms. The focus is not on the fact that there are citizens with greater knowledge, but rather that there are citizens who are dangerously ignorant. This produces the ‘antiauthority tenet’, where the ignorance of some citizens justifies their exclusion (Brennan, 2016: 17). This argument for epistocracy is based on the ignorance of some, not the wisdom of others. The basis for restricting the franchise is the potential for innocent people to be harmed. To make this point, Brennan employs a critique of what has elsewhere been called a ‘reductive strategy’ which relies on an analogy between political and interpersonal ethics (Rodin, 2002: 127–28). The liberal tradition of political thought has defended the idea that individuals ought to be allowed to live in ways which might be harmful to themselves, so long as they do not harm other people. There is an argument that democracies have a comparable right to make bad decisions in certain conditions, such as pursuing reckless economic policies that result in mass unemployment. However, for Brennan, this analogy does not work because there is no unified moral personality in the state as there is with individual human beings. Politics imposes decisions on other people. Ignorant citizens do not just harm themselves, they harm other people (Brennan, 2016: 9). Consequently, the state ought to be structured in a way that prevents the ignorant harming the innocent.

As with any complex system of political thought, there are numerous variations of how epistocracy could be realised. Brennan (2018: 55–56) has listed several prominent ways:

There are also what might be considered ‘sci-fi’ examples of epistocratic government in which a ‘simulated oracle’ produces the decisions that a fully informed democracy would choose or that an algorithm can be used to pick out which voters choose the wisest policies and give them greater weight (Brennan, 2016; Mulligan, 2018). Although distinct, each of these forms of government is epistocratic insofar as voices of the knowledgeable are amplified, whether through direct exclusion, uneven weighting, or programming the parameters of ideal information.

Republicanism

Like epistocracy, republicanism is an old idea that has enjoyed a revival in recent years. Republicanism is built upon a negative conception of liberty that is understood as freedom from domination. Brennan (2016: 95) interprets domination in the following way: an agent dominates another when two conditions are obtained:

This accords with how domination has been conceptualised by republicans (Lovett and Pettit, 2009: 12; Skinner, 2008: 85). It also seems to accord with the condition of the slave, which is used as the paradigmatic example of domination (Blunt, 2015: 6). The slave is in a condition where all their choices are the pleasure of their master. This does not require the slave–owner to act with malevolence or indeed to act at all. All that is required is the capacity to act with impunity. The slave is within her master’s power regardless of the latter’s intentions or psychological disposition. It is a structural relationship based on the distribution of power between agents or agents and institutions.

Republican political theory is dedicated to ensuring that social institutions are characterised by non-domination. This has been expressed by a commitment to the ‘free state’, one in which citizens are secure from domination by the state (imperium) and by private citizens (dominium) (Pettit, 1997: 36, Skinner, 1998: 66–77). In contemporary political thought, republicans have been vocal defenders of democracy as a check against domination (Bellamy, 2007: 210–221, Pettit, 2014: 109–149). By providing all persons with equal political power, people can be free from arbitrary interference by political elites. Republicans have stressed deliberative and contestatory models to enhance existing democratic practices (Pettit, 2012: 227–229, 267–269). Democratic states with the rule of law and strong egalitarian social provisions are characteristic of mainstream republicanism.

However, the history of republicanism is not a paean to democracy. Cicero (1991: 57–59, §1.150) thought that unskilled labourers were not fit to be free men, because they were essentially in a contract of servitude with their master. Francesco Guicciardini argued that the virtu necessary to sustain republican government was found only in the aristocracy and that the role of the masses was to provide a structure or audience to prevent the degeneration of the elite (Pocock, 2016: 127–129). James Madison (2012: 40–46) clearly distinguished democratic and republican forms of government. The former being unacceptable because it is inseparable from factionalism and the tyranny of the majority. Consequently, Brennan’s fears about the ignorant and misinformed are not alien to the republican tradition.

The case for epistocratic republicanism

Brennan conceptualises the problems of democracy very much in the vein of freedom from interference. It is the harmful interference of the ignorant that is of concern, but this argument can be reinterpreted in the language of freedom from domination. This may, indeed, strengthen Brennan’s argument. Democracy allows an uninformed majority to arbitrarily harm innocent persons. It is not that democracy causes harm to innocent persons, but that democracy provides the capacity for arbitrary interference to the ignorant or misinformed. A democratic polity may never harm innocent persons by provoking a financial crisis or waging a war of aggression at the behest of a manipulative demagogue, but that they possess the capacity to do so is what would alarm republicans. This may lead minorities to be in a constant state of supplication and flattery to the uneducated in order to keep them ‘sweet’. These citizens will not securely possess their freedom, because it can be snatched away by the fickle majority. The model of epistocratic republicanism mooted by Brennan is based on the exclusion from some people from the franchise if they are unable or unwilling to pass a test of their political knowledge. Consequently, the cases for and against will focus on this particular form of epistocratic republicanism (Brennan, 2016: 98–99).

Democracy, here, is not a necessary condition for freedom from domination. He agrees with Pettit that empowering citizens and encouraging civic engagement is a way to protect against domination, but this does not necessitate that every person has a vote. He argues that this might be plausible in the case of groups of people rather than individuals. The disenfranchisement of all black people, he argues, would produce a domination complaint because it would expose all black people to arbitrary interference along racial lines, but this does not mean that all black people need to be enfranchised. There is no distinction between all black people holding the franchise and all black people but one holding the franchise. Political liberties empower groups, but they do not empower individuals (Brennan, 2016: 97–98). If some members of a group of citizens are enfranchised, this would be sufficient to prevent unenfranchised members of this group from being dominated. One is protected so long as some ‘people like me’ have the vote, because they will vote to protect themselves. The ‘Jamie Lannister’ objection is to say what if there are no ‘people like me’? This is dismissed by saying that, as people have numerous overlapping identities, no person is unique in this sense. Moreover, democracy is only one of many institutional means used by republicans to check domination. Deliberative and contestatory forums, publicly known laws that are impartially enforced, limitations on campaign finance, and robust welfare provisions, civic education, and solidarity all of these are compatible with epistocracy. If these are sufficient to check domination in democracies, then they should be able to do the same in an epistocracy (Brennan, 2016: 99).

The argument that democracy and republicanism are not intrinsically connected has been more precisely developed by Frank Lovett. Democracy, he argues, may be an effective check against domination but it is not essential. Freedom from domination is essentially procedural in nature. So long as there are sufficient external checks on the exercise of power, there can be freedom from domination. It does not need to be checked by the agent who is subjected to it. This is because in the presence of a third-party invigilator, power would no longer rest on the will of the empowered agent. It would be constrained and therefore non-arbitrary (Lovett, 2010: 115). If a person knows her rights and they are impartially enforced, then she is not dominated. This becomes evident when you compare her condition with that of a slave. A slave lives in a permanent condition of uncertainty. Any choice she makes is subjected to the veto of her owner. She is unable to exercise her autonomy with anything close to certainty. Compare this with the unfranchised citizen of an epistocratic republic. She might not be able to vote, but she knows her rights and they are impartially enforced. If an epistocrat attempts to arbitrarily interfere in her choices, she will have recourse to the law to check this attempt. She is not vulnerable to arbitrary interference because of the protection of third-party invigilation. If this were not the case, the root injustice would be failures in the rule of law rather than epistocracy.

The matter of unequal enfranchisement may still be a problem. A citizen who lacks voting rights might not be able to pass the republican ‘eyeball test’; Pettit (2012: 84–86) claims that republican freedom enables people to look each other in the eye, they ‘can walk tall and assume the public status, objective and subjective, of being equal in this regard with the best’. Can a person without equal voting rights the eyeball test or would it be a point of shame? Brennan (2016: 99) does not engage with this problem as, for him, this is a question about the ‘expressive meaning of unequal political rights’ rather than a problem related to freedom or power. It is a symbolic concern rather than a substantive concern about arbitrary power. Under this argument, they might be able to pass the eyeball test as they still have a set of publicly known and impartially enforced rights, which under Brennan’s scheme protects fundamental interests. They are not supplicants.

The case for epistocratic republicanism is robust. Indeed, by framing the critique of democracy as a concern about domination, it may strengthen Brennan’s argument. Although it may be uncomfortable to republicans with strong democratic commitments, it seems that they must concede that epistocracy can secure freedom from domination. If there is a case against it, it must unpick the argument of this section or show that democracy is a surer check against domination than epistocracy.

The case against epistocratic republicanism

The previous section showed that it is possible to conceive a state where citizens are not enfranchised, but thanks to publicly known and impartially enforced laws they are not liable to arbitrary interference. They have a zone in which they might pursue their plans with a degree of confidence necessary for minimal autonomy. This matches Brennan’s (2016: 10, 94–99) epistocratic state where all citizens possess civil liberties, such as freedom of expression, but not all citizens have the political liberty to vote. This indicates that epistocratic republicanism is a coherent position; there is nothing incompatible between unequally distributing the franchise and freedom from domination. If people are not vulnerable to arbitrary interference, then they seem to be living under conditions of republican freedom even if they are denied the vote. This conclusion, however, is only available if one subscribes to a simplistic version of domination. Brennan successfully captures a mode of domination in his argument, but it is not a holistic engagement. This section will argue that epistocratic regimes are characterised by systemic domination.

In recent years, many republican philosophers have embraced conceptions of domination that go beyond interactional or ‘agent-centred’ approaches (Azmanova, 2016: 471–472; Bohman, 2012: 182; Laborde, 2010: 57, Pettit, 2012: 35–44, 2014: 53). We can look to the modes of domination to clarify the sources and sites of arbitrary power (Blunt, 2015: 19). There are personally generated and socially generated sources of domination, but what matters in the case of epistocracy is the site, which he divides into interactional and systemic. Interactional domination occurs between at least two agents. This tends to be the way in which slavery is described in a lot of the republican literature. A master possesses the capacity to interfere with impunity in the choices available to his slaves. However, this does not capture the complexity of slavery as a social institution. She is not only subjected to arbitrary interference from her master but has had her place in the basic structure of society arbitrarily determined in a way she cannot contest. Slaves are subjected to two modes of domination: they experience arbitrary interference on an interactional level from their owners and various surrogate agents, but they have also had their role in the social institution arbitrarily defined in a way that they cannot contest or challenge without engaging in resistance. For example, during his time as a slave, Isaac Johnson befriended another ‘slave’ named Bob. Bob was a freeman from Canada who was wrongfully arrested while travelling in the South and was declared a fugitive slave. He had no way to assert his status as a free man. He later was brutally murdered in an attempt to escape (Johnson, 1901: 26–31). The problem with Lovett’s argument is that he neglects the systemic elements of domination; he presents an idealised apartheid example where the majority of citizens possess fewer rights based on their race. The laws in this society are publicly known and impartially enforced. A white police officer who arbitrarily interfered in the rights of black citizen would face sanctions. Be that as it may, the black majority would still suffer from domination because their place in society has been arbitrarily circumscribed. They may not suffer from interactional domination, but they are systemically dominated. An admissions officer, for example, who refuses the application of an otherwise qualified black applicant to a whites-only university does not interactionally dominated the applicant as she is following the rules, but the applicant is dominated because her status has been arbitrarily determined (Blunt, 2015: 15–17).

Epistocracy might escape this objection. Lovett gives no impression that there is a way to become enfranchised in his example as the discriminated majority cannot change their race. Epistocracy is different, because there is often a pathway to the franchise or to increase the weight of their vote (though not in the case of the enfranchisement lottery or the simulated oracle). A person may find themselves without voting rights or with minimal voting rights, but not necessarily in perpetuity. An unfranchised person can gain the requisite political knowledge to be granted the right to vote. We can assume that an epistocratic republic would include a publicly known and impartially enforced way for citizens to gain the vote. This would give all citizens the power to contest their position within the state. If they have this, it seems as though they cannot suffer from systemic domination. However, a perennial question lingers: quis custodiet ipsos custodes? In the case of epistocracy, David Runciman (2018: 184) has provided an appropriate reinterpretation: ‘who gets to set the exam?’

This domination complaint is prefigured in the ‘demographic objection’ to epistocracy which was mentioned earlier in the article: education is closely associated with privilege and may bring with it a number of known and unknown biases that negate its benefits. Estlund (2009: 36, 215–217) makes the provision that when assessing potential epistocrats, there ought to be ‘no invidious comparisons’; the concern that there could be empirically latent prejudices which negate the benefits of education is sufficient to reject all potential candidates or groups of candidates. Several critics of epistocracy have reinforced this point that educated people are no less prone to cognitive biases and groupthink as anyone else (Bhatia, 2018: 12–13; Klocksiem, 2019: 14–16; Runciman, 2018: 183–184). However, the republican concern with systemic domination is more parsimonious. The demographic objection implies that if one could track all harmful prejudices and deselect would-be epistocrats as a result, there would be no issue with epistocracy. However, this would not be sufficient to dismiss the republican concern about systemic domination. The problem is not with prejudice, known or otherwise, but with the capacity to exercise arbitrary power. A problem still exists with the unprejudiced epistocrat because they can exercise systemic domination over the status of other citizens. It may be exercised benevolently, but that is beside the point. The issue is the structure of power between agents and institutions. One can imagine a benevolent slave–owner who treats the people she owns very well and never interferes in their lives, but her slaves would still be slaves. They would still be subject to her whims. The ability of the epistocrats to ‘set the exam’ gives them arbitrary power over the status of their fellow citizens; it does not matter if they abuse it. The psychological disposition of the epistocrats is immaterial to a republican.

Anne Jeffrey (2018: 422) offers a potential reply to this problem when she suggests that the power wielded by epistocrats would not be arbitrary because the standards are independent; they are not based on their whims, but on ‘reason’. She suggests that their status gives them pro tanto authority that is not absolute, but can only be challenged by people using acceptable evidence to challenge their authority. She provides the example of medical staff working in Liberia during the Ebola outbreak. They found that despite the efforts of public health specialists, the disease continued to spread. One of the factors causing this was the non-compliance of ordinary people due to cultural and religious beliefs about the disease. These citizens did not have proper reasons to discredit the ‘epistemic authority’ of the medical staff, and there was no obligation to listen to such people (Jeffrey, 2018: 423). The problem with this argument is that it depends on internal restraints. The epistocrats are judges in their own case. They determine what is acceptable evidence and what is not, as well as who is an acceptable speaker and who is not. There is no third-party invigilation. This is not compatible with freedom from domination because the determination of what is appropriate knowledge rests on their will. Jeffery’s example seems unproblematic, but it can be contrasted with particular scandals in humanitarian assistance. In 2018, The Times alleged that senior Oxfam staff in Haiti had paid earthquake survivors for sex and an internal report had described a ‘culture of impunity’ among aid workers in the country (O’Neill, 2018). This is not an isolated incidence. In 2002, the United Nations Office for the High Commissioner for Refugees and Save the Children (UK) discovered widespread sexual exploitation of women and girls in East African refugee camps by aid workers (Reyes, 2009: 215–216). These men were in a position to use their control over vital goods and employment opportunities to exercise arbitrary power over their victims. This is not to say that humanitarian aid is rife with sexual exploitation of vulnerable people, the point is that placing one’s faith in goodwill, professionalism, or rationality is not sufficient to check domination and it is a dangerous gamble.

Brennan’s response to the demographic objection is comparable to the previous argument. He points out that licenced professions, such as doctors, often are over populated by socially advantaged groups. This does not make them intrinsically unjust or necessitate their abolition; it only means that we ought to remedy the underlying injustices that cause such imbalances (Brennan, 2018: 60–61). This argument does not have much weight against concerns of systemic domination. The problem is with the regulation of the regulators, and this becomes more pressing when considering political power. In most circumstances, my relationship with my doctor is a voluntary one. If I think she is wrong, I can get a second opinion. State membership tends to be involuntary in a meaningful way. If I disagree with a law and break it, I can be sanctioned by the state. There are, of course, ways to exit the state, but they tend to have a much higher cost than changing one’s physician, but in order for a social relationship or institution to be dominating, the asymmetry of power must be sufficient to produce dependency, which is characterised by unreasonably high exit costs (Lovett, 2010: 38–41, Blunt, 2015: 5, 11). The epistocrats are self-licenced in a relationship with high exit costs. The republican concern about this is captured by James Harrington’s (1992: 22) claim that non-domination requires a division of powers. He gives the example of two children deciding how to divide a cake. The fair way to do this is let one child cut the cake and the other child choose the first piece. The epistocrat wants to cut the cake and choose the first slice. The democrat is not in the position to cut and choose as the epistocrat. At least in a de jure sense everyone gets an equal slice. It may be true that there is a difference between the de jure and the de facto operations of democracy, but at this point, it seems wise to adopt the ‘good faith’ view advocated by Brennan for democracy as well as epistocracy. If democracy, constrained by republican institutions, operates according to plan, then it does not seem vulnerable to systemic domination. The same cannot be said of epistocracy because the necessary checks against the control over the pathway to the franchise must be set and enforced by the epistocrats. If it were otherwise, then it would be difficult to call this epistocracy since the knowledgeable would not be in power.

Judging the cases

The case for epistocratic republicanism argues that democracy leaves citizens vulnerable to the arbitrary interference of ignorant or misinformed compatriots, while the case against claims that epistocracy exposes some citizens to systemic domination. This section examines which approach provides a better path to freedom from domination. Brennan (2016: 10–14) takes an instrumentalist approach judging democracy and epistocracy; we should judge political systems as means to an end, not ends in themselves. His analysis relies on how well systems of government track ‘the truth’ or produce ‘good decisions’, but republicans cannot be so general (Brennan, 2016: 13). Republicans tend to be instrumentalist in their approach to government. A value, such as freedom, should act as a goal for an agent or institution when the purpose of the agent or institution is to promote that value. In the case of republicanism, the state’s goal is to minimise domination (Lovett, 2010: 170–179). However, this cannot be a blank cheque. A value can also act as side-constraint when the purpose of that agent or institution is to honour the value in question (Pettit, 1997: 97–98). Contemporary republicans would find it difficult to support a republican system of government that minimised overall domination by brutally dominating a very small number of citizens. Honouring the value means that we must recognise that freedom from domination is equally valuable to all persons. Diverging from an egalitarian baseline would require special justification (Pettit, 1997: 102). Consequently, we should look at which approach, democratic or epistocratic, best minimises domination while also respecting its value to all persons.

The case for epistocratic republicanism rests on the concern that uneducated, ignorant, or misinformed persons have the capacity to exercise arbitrary power over their compatriots. If the unfranchised citizens are then systemically dominated under epistocratic republicanism, then it is a price worth paying for minimising domination in society. These passive citizens will enjoy a full set of civil liberties (Brennan, 2016: 8–10). They will not be prone to interactional arbitrary interference; they will be able to plan their lives and pursue these plans with security. It is a price that can be paid, but is it a price that must be paid? The republican tradition has a long-standing fear of ‘mob rule’, and democratic republicans have a number of policy options designed to minimise this risk. They have stressed the importance of constitutional checks and balances, the rule of law, and individual rights (Pettit, 2012: 211–218). Most pertinently, in the case of epistocracy, is the tradition of civic education in the republicanism. Civic education is one of the pillars of the republican project. Children and young people are supposed to not just learn the formal curriculum of how the state works and the duties of citizens but also the ‘hidden curriculum’ of civility and respect (Dagger, 1997: 120–122). The aim is to produce ‘contestatory’ citizens who are able to participate in political life in an informed and civil manner (Pettit, 2012: 225–229). This is not to say that a state education system aiming to produce citizens who embody the values of republican government would be perfect. There will always be people who might not be up to scratch. If excluding them had minimal costs, then it might be a justifiable deviation from the value of non-domination.

The problem is that the systemic domination that accompanies epistocracy carries with it significant risks to those who lack the franchise. Several critics of epistocracy have pointed to the use of literacy tests during the Jim Crow era (Estlund, 2009: 215–216, Klocksiem, 2019: 26, Umbers, 2019: 291). Between 1956 and 1968, states in the Old South excluded citizens who were not able to read or write. At a conceptual level, this should not trouble someone like Brennan; basic literacy would be a plausible requirement for any epistocratic test. The rule was not impartially enforced but used in general to exclude black people. Yet, even if it were impartially enforced, black citizens would have been more likely to be excluded due to historical and ongoing injustices associated with white supremacy. Brennan has argued that this is an unfair criticism of epistocracy because it is placing blame on it for underlying injustices (Brennan, 2016: 223–224, 2018: 60–61). The idea that citizens must meet a minimum level of knowledge to vote is not what damaged black citizens, it was white supremacy. If we wish to make the case against epistocracy, he argues, then we need to assume the good faith of the people running it. This argument is not effective against the republican concern regarding systemic domination, because the issue is not in the background but in epistocratic institutions qua epistocratic institutions. Republicans have no interest in whether power is used malevolently or benevolently. All that matters are the constraints placed on power. If power has no external constraints and rests on the will of an agent, then it is by definition dominating. In the case of the epistocratic tests during Jim Crow, the problem is not only background injustices, but that black citizens were the subjects of systemic domination because they were not in a position to contest their exclusion from the franchise.

The epistocratic republican might concede this point but claim that the risks from systemic domination are less than the tyranny of the uneducated. It may be true that they are unable to have equal influence or standing, but on the whole the exercise of arbitrary power would be reduced. This would fail to appreciate the epistemic injustice associated with systemic domination and epistocracy. Miranda Fricker’s conception of epistemic injustice shows that control over the production, legitimation, and articulation of knowledge raises its own problems of justice. She examines how social power is used to shape identity in a way that can produce injustices in discursive exchanges. She has categorised two types of epistemic injustice and both a relevant in the case of epistocracy. The first is ‘testimonial injustice’ where someone’s inferior social status causes their testimony to be ignored or treated without due credit. They are ‘wronged in their capacity as a knower’ (Fricker, 2007: 20). Hermeneutical injustice is having a significant area of social experience obscured or occluded from collective understanding because of ‘structural identity prejudice in the collective hermeneutical resource’ (Fricker, 2007: 155). This conception of epistemic justice seeks to cultivate virtues as solution to prejudice. One way is the development of testimonial sensitivity as a path to testimonial justice. This will identify and correct prejudices in the hearer’s judgement of the credibility of a speaker (Fricker, 2007: 86–98). Regarding hermeneutical justice, hearers must develop a sensitivity to reduced intelligibility as this may be the result of marginalisation from the collective hermeneutical resource (Fricker, 2007: 169–175). These are important steps and they might go some way to assuaging concerns about epistocracy, but republicans need more. James Bohman (2012: 181–184) has developed a republican account of epistemic injustice which is apposite in the case of epistocratic republicanism, because it focusses on systemic domination. Epistemic injustice is not only the lack of sensitivity in the hearer but the fact that the speaker does not have a voice that commands respect. It is not sufficient that the hearer is sensitive or virtuous if he still possesses the ability to dismiss the person speaking to him. This is because the speaker will be keenly aware that her voice does not need to be listened to; she may have to adopt the position of the supplicant to be heard or moderate her demands. She cannot look him in the eye.

This is precisely the problem with epistocratic republicanism. The epistocrat acting in good faith and with the cultivated virtue of listening to the marginalised can still perpetrate epistemic injustice simply by virtue of his power. Brennan (2016: 212–213) makes an attempt to rebut concerns about ‘who sets the test’ by arguing that it can be rendered ‘objective and non-ideological’ if it is kept to basic facts about civics which align with the opinions of social scientists like economists. This makes a rather controversial move of depoliticising such knowledge without much of a defence. Paul Gunn (2019: 31) has pointed out that the move basically moves to enfranchise the citizens who vote the way social scientists want them to vote. It brings to light the problem of testimonial injustice and systemic domination shown in Charles W. Mills’ work on ‘white ignorance’ and academia. Mills, drawing on C.A.J. Coady’s work, also emphasises the importance of testimony in generating a social epistemology. If a group is discredited in advance, our knowledge is shaped accordingly. Their testimony is not listened to because they do not have the power to command that they are heard (Mills, 2017: 67). He casts doubt on the aspirations of Enlightenment rationality, which often privileged white perspectives and testimony under the guise of universal truth. The status of a privileged knower allowed white people to dismiss the testimony of non-whites. They were ‘primitive’, ‘backwards’, or racially inferior and so not credible speakers. This enables someone like Thomas Jefferson to declare that all men are created equal while simultaneously owning slaves and decrying the savagery of Native Americans as colonists drive them from their land (Mills, 2017: 62). Udit Bhatia (2018: 13–14) gives the example of ‘the dictators game’ to show how people avoid information when it does not suit them. The game gives its players two options: the first gives the player a very high pay off and another recipient a very low pay off; the second gives the player a slightly lower payoff than the first option, but a much higher payoff to the recipient. It was found that players tend to support the second option. However, an alternative game masks the outcomes for the recipient but gave the player the option of revealing these outcomes at no cost. Players in this scenario are more likely to choose the first option and most players who do so do not seek the further information. Those who conducted the experiment concluded that there was only an ‘illusory preference for fairness’ Dana et al. (2007: 67). Bhatia (2018: 14–16) mobilises this in his critique of epistocracy to explore how epistocrats lack the motivation to engage with structural injustices because of epistemic avoidance. It is a compelling argument that is valuable to republicans, because it is not only that they lack the motivation, due to internal cognitive biases, but that there is no power to compel motivation. It is no coincidence that this is called the ‘dictator’s game’ because one of the relevant agents possesses arbitrary power of what information is heard. The problems of epistemic avoidance ultimately reduce to epistemic domination; if people cannot make you listen, then you will only hear what you want to hear.

This also drives hermeneutical injustices where an agent cannot be heard because they have not had access to the collective hermeneutic resources. They cannot speak the language or rather cannot speak with the ‘received pronunciation’. This aligns with Mills’ critique of ideal theory in political philosophy. The dominance of this approach has helped to occlude lived-injustices by dismissing them as somewhat epiphenomenal. The assumptions of ideal theory cannot be detached from the historical context which produced them, such as white supremacy. Indeed, they enable political philosophers to evade uncomfortable connections between liberalism and racial oppression. This methodology blinkers academic philosophy. Students are trained by academics to think this way, they become academics who think this way, and who cite academics who were trained in the same fashion and think in the same way. It creates a closed shop where non-ideal concerns can be dismissed as secondary or are unintelligible (Mills, 2017: 73–80). This is not malicious. This does not require consciously held racism but can produce a sort of ‘consensual hallucination’ in the privileged group whereby they cannot understand the injustices produced by the political system they have made (Mills, 1999: 18–19). In an epistocracy, we can imagine epistocrats simply being unable to understand the concerns of the marginalised unfranchised citizen because they speak in ‘regional dialect’ rather than ‘received pronunciation’ or in an entirely alien language. The epistocrat might dismiss them out of frustration and they do not have recourse. This shows that simply declaring that a test can be ‘non-ideological’ because it aligns with received opinions carries with it hazards that Brennan does not adequately address.

It might be objected that this argument against epistocratic republicanism is based on, perhaps, the most rudimentary form of epistocracy, a restricted franchise. To this two things can be said: the first is that this is the form presented by Brennan (2016: 94–99) in his compatibility argument so it is reasonable that it should be the primary subject of criticism; second, this problem does seem to be intrinsic to epistocracy. There is not sufficient time to go through every variant of epistocracy. However, in the case of plural voting, the enfranchisement lottery, and the epistocratic veto, power is distributed based on passing a political knowledge test (Bell, 2016: 162–168; Brennan, 2011, 2016: 211–220; López-Guerra, 2014: 24–26; Mill, 2010: 174–176). Consequently, the problems about epistemic and systemic domination remain. The argument that these tests would be ‘objective’ and based on social scientific fact requires a rather naïve approach to epistemology and epistemic injustice that expose the excluded to domination. A similar problem afflicts ‘values only voting’; it relies on the knowledge of experts to determine the means, but on what basis is their knowledge established (Christiano, 2018: 207–243)? If we take Mills seriously, then it is unclear why technical expertise makes one immune to bias, conscious or otherwise. The voters may want a certain end, but the means selected by the experts may disproportionately benefit or burden certain parts of the citizenry in a way that they cannot dispute. The sci-fi solutions of ‘simulated oracle’ and algorithmic vetting cannot rest on the assumption that because these are machines that can find the ‘right’ or ‘best’ answer, they cannot be vehicles of domination (Brennan, 2016: 220–222, Mulligan, 2018: 300–302). The biases of programmers has become the subject of concern in recent years as experiments in artificial intelligence and other complex programming have shown themselves to be less than impartial (Crawford, 2016; Garcia, 2017). Claiming that programmers have bias depends on underlying injustices mistakes the objection; the problem for a republican is not that bias is present but the mere presence of the capacity to arbitrarily set the terms of social cooperation. So, while the problems of systemic and epistemic domination are clearest in the case of the restrictive franchise, it is not limited to that particular form of epistocracy. It is a problem so long as a group may arbitrarily determine what counts as correct knowledge and use that power to determine the political status of others. This is an intrinsic part of epistocracy.

Epistemic injustice shows the risks of systemic domination for plausible epistocratic constitutions. This is not something that can be easily checked by other republican institutions. If non-epistocrats were given control over epistemic power through checks and balances, then it would cease to be an epistocracy because the learned would no longer be in power. Epistocracy cannot accommodate third-party invigilation from non-epistocrats. Even if this was not the case, the problem of systemic domination would persist. There is no reason to assume that a third-party invigilator would not also be the source of epistemic injustice. The unfranchised citizen would be in the same relationship with them as they are with the epistocrats. Given that there are options that can be applied to reduce and check ignorance in a political community, such as individual rights and rigorous civic education, the divergence from non-domination as a value and the potential loss of overall freedom caused by systemic domination is unjustified. Republicans ought to be democrats.