Mandevillian Virtues

Studies in collective intelligence have shown that suboptimal cognitive traits of individuals can lead a group to succeed in a collective cognitive task, in recent literature this is called mandevillian intelligence. Analogically, as Mandeville has suggested, the moral vices of individuals can sometimes also lead to collective good. I suggest that this mandevillian morality can happen in many ways in collaborative activities. Mandevillian morality presents a challenge for normative virtue theories in ethics. The core of the problem is that mandevillian morality implies that individual vice is, in some cases, valuable. However, normative virtue theories generally see vice as disvaluable. A consequence of this is that virtue theories struggle to account for the good that can emerge in a collective. I argue that normative virtue theories can in fact accommodate for mandevillian emergent good. I put forward three distinctive features that allow a virtue theory to do so: a distinction between individual and group virtues, a distinction between motivational and teleological virtues, and an acknowledgement of the normativity of “vicious” roles in groups.


Introduction
Virtue ethics is traditionally concerned with the character traits and flourishing of individual persons. However, an important part of moral life is the way individuals can contribute to the collective flourishing of groups or communities. Is it possible that traits that help a community to flourish can sometimes differ from the traits that help an individual flourish? Recent work on epistemic traits describes how epistemic vices of individuals can sometimes lead a group to flourish epistemically.
Paul Smart (2018a) argues that the suboptimal cognitive performance of individuals within a group can actually benefit the group in carrying out epistemic tasks. Smart calls this mandevillian intelligence. He refers to studies involving groups of people and computersimulated nodes. The experiments show that some tendencies make it less likely for an individual node to answer a question correctly, but increase the likelihood that the network they are a part of convenes on the right answer faster. Smart describes the phenomenon as an instance of collective epistemic virtue arising out of individual epistemic vice (Smart 2018b). Other authors have also described in recent years the way individual epistemic vices contributing to emergent epistemic goods (Levy and Alfano 2019;Morton 2012). In some cases, knowledge arising out of bad epistemic character can be a lucky fluke. However, in other cases the dynamic of a group can be such that an individual who consistently displays vicious behaviour, consistently leads the group to good epistemic behaviour. Stable, long-term, epistemically disvaluable traits, or epistemic vices, if embedded in the right way within a collective or group, can lead to the collective or group to behave in an epistemically virtuous manner (Smart 2018b).
I suggest that mandevillian intelligence has a counterpart in ethics, which I shall call mandevillian morality. The spirit of mandevillian morality is closer to Mandeville's message in the work The Fable of the Bees or Private Vices Publick Benefits, where he likens a capitalist economy to a beehive. Each bee acts out of self interest but this has the systemic effect that the hive functions well and keeps all the bees alive (Mandeville 1714). My definition for mandevillian morality is however more specific. I mean instances where the moral vices of individuals contribute to virtuous behaviour at the collective level. Mandevillian morality refers to stable, long-term character traits of individuals, which cause stable, long term behaviour in groups. Mandevillian morality presents a challenge for most virtue ethical theories because it has the counterintuitive consequence that vice can sometimes be valuable or morally good. I characterize Mandevillian morality as the virtuous behaviour of groups, as a result of vicious behaviour of individual agents within the group.
Mandevillian morality can arise in different ways. Particularly in non-ideal situations where there is a prevalence of vice, the prevalence of the vice at the opposite extreme can make the group function at the golden middle. For example, in a group of cowardly collaborators, an overly confident member can push the group to behave more courageously. Human resource managers often use opposing strengths and weaknesses to create a well functioning team. In the words of Belbin, famous for developing the Belbin team roles, "What is needed is not well balanced individuals, but individuals who balance well with each other" (P. 319, Bjarne 2012; Belbin 2012).
Many virtue theorists adopt what is called a pure virtue theory. A pure virtue theory is a virtue theory which sees virtues as the only fundamental source of moral good (p.79, Zagzebski 1996). If one accepts a pure virtue theory, then one necessarily maintains that all moral good boils down to individual virtue. In cases of mandevillian morality collective morally valuable behaviour does not boil down to individual virtues only, but also, importantly, to individual vice. This, I argue, presents a challenge for pure virtue theories. A pure virtue theorist will struggle to account for the moral good of mandevillian morality, because it cannot be explained through a connection to virtue only.
How, then, should pure virtue ethics account for mandevillian morality? I suggest that recent work in virtue epistemology can provide guidance. However, recent work in epistemology has expanded the notion of epistemic virtue in such a way that Mandevillian intelligence is more easily accounted for. Various authors have proposed theories about shared-belief formation, collective epistemic virtues and the social extension of epistemic virtues (Gilbert 1987;Fricker 2010;Alfano and Skorburg 2016.;Levy and Alfano 2019;Palermos 2016;Pritchard 2018) Such developments allow epistemic virtue theories to account for emergent epistemic goods, even when they arise out of vice. However, moral virtue theories have not seen this kind of developments. Hence, this paper is an exploration into the possibility of similar expansions to virtue ethics, so that mandevillian morality can also be accounted for.
Accounting for mandevillian morality would make virtue ethics, especially pure virtue theories, stronger as a moral theory, especially with regard to group, community and collaborative contexts. If a trait that is clearly a vice leads a group or community to flourish, then pure virtue ethics comes to a trilemma: either the trait is not a vice, the group is not really flourishing or that vices can sometimes lead to group flourishing. All of these are counterintuitive in true cases of mandevillian morality.
Mandevillian morality can arise in business, governance, innovation, science and other collaborative areas of life. There are plenty of true stories about of flawed or devious characters leading groups and organizations to success. Therefore, if a virtue ethical response to mandevillian morality can be formulated, then this may prove to be useful in fields of applied ethics concerning such contexts. It would also make pure virtue theories stronger against possible objections to their applicability to collective life.
I propose that virtue ethics can account for mandevillian morality if a multitude of types of virtue is recognized, much as recent virtue epistemological work has done. I put forward an account of moral virtue, with the following three distinctive components: 1. a distinction between collective and individual virtues 2. a distinction between motivational and purely teleological virtues 3. a strong role-based virtue theory which acknowledges "vicious" roles.
The three above distinctions provide three possible ways to explain how a vice, which is fundamentally bad, contributes to morally virtuous group behaviour, which is fundamentally good.
In section 2 of this article, I give an extensive definition of mandevillian morality. In section 3 I explain what is meant by pure virtue theories. In section 4 I argue that mandevillian morality presents a challenge to pure virtue theories. Section 5 presents a structure for an ethical virtue theory with the three features above that allow it to answer the challenges of mandevillian morality, sometimes drawing inspiration from the field of virtue epistemology. The concluding remarks in section 6 elaborate in how far the three features align with many common intuitions about virtues and present some questions for future research.

What Is Mandevillian Morality?
To begin, my definition of mandevillian morality is inspired by Smart (2018a)'s definition of mandevillian intelligence, which is: Cognitive and epistemic properties that are typically seen as shortcomings, limitations or biases at the individual level can, on occasion, play a positive functional role in supporting the emergence of intelligent behavior at the collective level (Smart 2018a).

I define mandevillian morality as:
A specific form of stable and long-term, ethical group agent behaviour in which character traits that are typically seen as moral vices at the individual level play a structural role in constituting this ethical behaviour at the group-agent level.
In what follows, I shall explain some of the terms in this definition. By "ethical group agent behaviour" I mean morally valuable behaviour, such as courage or reflectiveness, manifested by a group agent. This ethical behaviour can be a description of how the group behaves towards itself, like the members respecting each other appropriately or ethical behaviour regarding the shared goal of the group, like the group taking on a supererogatory courageous challenge.
In order to count as mandevillian morality the group agent behaviour must be stable and longterm. One-off good consequences arising "accidentally" out of vicious actions are to be excluded. I have chosen to limit the definition of mandevillian morality to applying to "group agents" rather than collectives, as Smart does with mandevillian intelligence. I see a "group agent" through List and Pettit's definition as a collection of people which collectively possesses three characteristics. The group possesses representational states, motivational states and the capacity to process representational and motivational states, to intervene in the environment, so that the environment is brought to match the motivational state (List and Pettit 2011).
By "moral vices at the individual level" I mean disvaluable, or morally bad, traits of individuals, which in the broader context of a person's life inhibit their flourishing. In the Aristotelian tradition vices are generally seen as excesses or deficiencies, like foolhardiness or cowardice. In between the two extremes is a virtue, like courage. If such a trait structurally contributes to stable, long-term ethical behaviour of a group agents, then it is a vice which causes mandevillian morality.
What examples are there of mandevillian morality? Individual vices can create ethical group behaviour in many ways. One example is third-party punishment of norm violation, which can be beneficial for the establishment and maintenance of social norms (Fehr and Fischbacher 2004). This could mean that in some cases a disposition to spite can be helpful in maintaining moral norms in a community, especially if others in the community are less disposed to punish third parties. Imagine, for instance, that an individual within a group possesses the vice of wrath. She may therefore be particularly disposed to engaging in thirdparty punishment. Hence, the wrathful can systematically and consistently keep the group behaving more ethically than they would if she were not there.
The addition of a spiteful individual to a group of benevolent and non-spiteful people, can make the group more just. Vices can also add to the deliberative capacities of a group. A bigot in a group is annoying and possibly destructive to ethical group behaviour. However, if embedded in the right kind of group, a bigot might cause more dialogue. This can make a group more reflective as a whole because the bigot raises discussions about morality and probably some basic assumptions that everyone in the group accepts or conforms to. Vice can sometimes make us even more conscious of the morals we hold, by disrupting them and confronting us with their opposite. As Swierstra puts it: "'cold' morality turns into 'hot' ethics: invisible, solid, moral routines become fluid in ethics" (Swierstra 2016). There is undoubtedly something valuable about this event, where people are confronted with each other's differences and engage explicitly in moral thinking. While intuitively convincing, it has also been argued more systematically by Fabienne Peter that deliberation about ethical values is not merely instrumentally valuable, but also has procedural value. Having procedural value means that it is not merely valuable for the purpose of achieving some valuable end, like consensus or a sense of community, but rather that the deliberation itself has some intrinsic value which makes it good regardless of the consequences (Peter 2013). Behaviour resulting from vices reliably initiates these valuable events. This would be a prime example of mandevillian morality.
Sometimes epistemic emergent good also arises from individual moral vice, or moral emergent good from epistemic vice. Adam Morton (2014) discusses a phenomenon, which I would interpret as a hybrid form of mandevillean morality and mandevillian intelligence. Namely, when individual moral vices lead to better epistemic group behaviour. One example that Morton discusses is that epistemic self-indulgence can, if embedded in a group in the right way, make the group flourish more epistemically. Epistemic self-indulgence is when people seek confidence or respect by deviating into fields of knowledge where they feel less challenged by epistemic authorities. For example, teenagers are known to seek knowledge about things that do not interest adults, to be in a space where they can be more of an epistemic authority than elsewhere. The presence of such individuals in an epistemic community often means more epistemic rivalry, conflict and hence more better supported views and theories as a consequence of having to defend one's position against diverse attacks (Morton 2014).
Another example of moral vices leading to epistemic benefits is the systemic benefits of dogmatic group cognition in democracies. Citizens are sometimes overly dogmatic regarding political issues that touch their own social group. Dogmatism can be seen as both an epistemic and a moral vice. Lepoutre discusses the example of city dwellers who worry about the environment and underestimate the effect of environmental measures on job creation (Lepoutre 2020). Rural populations who see a lack of jobs in their environment may overprioritize the issue of job creation over environmental protection. Both opinions may be coddled in negative perceptions about the other social group, i.e. that city dwellers are arrogant. This "dogmatic group cognition" can be beneficial systemically. In his words: "Rural inhabitants who overprioritize job creation will be more disposed to uncover evidence of how environmental regulation harms rural jobs… …Conversely, environmentalists may, because of their contrasting epistemic standards, be more effective at finding flaws in evidence that environmental regulation harms jobs and more likely to generate new evidence of climate change" (Lepoutre 2020).
I emphasize that mandevillian morality is stable and long-term and involves a structural contribution of individual vice. This is what distinguishes it from other moral-luck-like phenomena. Older debates on moral luck and other areas have for long discussed the possibility of good things arising out of bad actions. A politician might lie in order to influence others to get a life-saving bill passed. Waltzer states in a 1973 article on the problem of dirty hands "No one succeeds in politics without getting one's hands dirty" Subsequently, the author asks "But one's hands get dirty from doing what it is wrong to do. And how can it be wrong to do what is right? Or, how can we get our hands dirty by doing what we ought to do?" (Walzer 1973). A single instance of a politician doing a good action with dirty hands, is not mandevillian morality. However, if this politician's stable tendency to lie contributes to the parliament collectively taking greater efforts to fact-check, or helps the parliament avoid harmful stagnation, then it is mandevillian morality.

Pure Virtue-Ethical Theories
Mandevillian morality is problematic if one adopts a "pure" form of virtue ethics. In order to show this, I will first define virtue ethics and what such a "pure" virtue theory entails, beginning with the definition of virtue.
Virtue theories are normative theories that place the good in the character traits of people. Moral virtue theories focus on the moral character of people, which lead them to flourish as human beings (Hursthouse and Pettigrove 2018).
The central entity in virtue ethics is the virtue, or the morally good character trait, as well as its opposite, the vice. Zagzebski characterizes a virtue by listing the distinctive properties of virtues: 1. a virtue is "an acquired excellence of a person in a deep and lasting sense" (p.135, Zagzebski 1996). 2. "a virtue is acquired by a process that involves a certain amount of time and effort" (p.135, Zagzebski 1996) 3. A virtue is not just a skill. A skill is distinguished from a virtue in the following ways: a. a skill need not be exercised, but a virtue must be exercised to be possessed b. The difficulty of skills is due to technical difficulties of the actions themselves, the difficulty of virtues arises generally from a lack of sufficient passion c. A virtue, unlike a skill, cannot be compatible with voluntarily giving it up. It is characteristic of having a virtue that one would never want to give it up. d. "The behaviour consisting in the exercise of a skill is not essentially connected to anything valuable" (p. 113 Zagzebski 1996) A skill is good through the objects or acts it produces, whereas a virtue is good to have in itself. e. A skill is not affected if distractions or persuasion prevents a person from exercising it, but a virtue is affected by this. f. a skill has no contrary, or opposite, whereas virtues have vices as opposites. 1 4. A virtue has a motivational component. The person possessing the virtue has the desired emotional disposition. i.e. if a person is benevolent, they are motivated to act benevolently by feelings of duty or empathy, rather than merely by the prospect of improving one's reputation. 5. A virtue has a success component. This means that a person who has a virtue is also generally reliable in bringing about the desired outcomes of a virtue. I.e. if someone is benevolent, they also tend to succeed in their efforts to help others.
Zagzebski and others also make a distinction between moral and epistemic virtues. Epistemic virtue theories focus on the epistemic character of people that leads them to flourish as human beings, or at least as knowers (Zagzebski 1996;Code 1984). Similarly, moral virtue theories focus on character traits like honesty, generosity and justice. These traits are developed through practice and help one to lead a good life. Moral and epistemic virtue theories are also popular in application to collaborative contexts like science, business and innovation (Demetriou 2013;Hicks and Stapleford 2016;Steen 2013;Sand 2018). A pure virtue theory is described by Zagzebski as "a theory which makes the concept of a right act derivative from the concept of a virtue or some inner state of a person that is component of virtue" (p. 79 Zagzebski 1996). Pure virtue theories are held, notably, by Anscombe and MacIntyre. Anscombe famously argued that deontological theories are vacuous in the absence of a divine legislator, and hence untenable in a secular worldview. Consequentialist theories, cannot be moral theories because they permit obviously immoral acts. Virtue theories are the only way in which a moral notion of the good can reasonably be approached 1 For a more detailed discussion on the difference between skills and virtues, see section 2.4 in Zagzebski (1996). (Anscombe 1958). MacIntyre defended the fundamentality of virtues in a similar way, arguing that other moral theories collapse into vacuous statements without the content of virtues and vices (MacIntyre 1986).
Of course, some see virtues as additional to other moral theories. Nussbaum, for instance, argues that virtue theories are not rival theories to utilitarianism and deontology, but rather something additional to both (Nussbaum 1999). However, the pure virtue theorist rejects utilitarian and deontological notions of good completely.

Why Mandevillian Morality Challenges Pure Virtue-Ethical Theories
My claim is that mandevillian morality is widespread and important in human moral life, and virtue ethics should be able to explain and account for it. However, mandevillian morality has some consequences that seem to misalign with a typical picture of virtue. If an individual vice plays a systematic role in morally good group behavior, then it is reasonable to say that that vice is valuable. However, in virtue ethics vices are supposed to be disvaluable, the opposite of valuable. In the case of mandevillian morality, a person might reasonably be praised or rewarded for their vice. This is counterintuitive from the perspective of virtue ethics. This counterintuitiveness stems from the fact that a virtue ethicist upholding pure virtue theories will struggle to explain what is good about good group behaviour, like collective reflectiveness, in a case of mandevillian morality.
Within a pure virtue theory especially, vices being systematically valuable is problematic. Because if one sees virtues as the foundation of morality, then one explains every instance of moral good with reference to the virtues. If confronted with mandevillian morality, a pure virtue theorist would be forced to either admit that the good group behavior is not morally good or that the vice in question which plays a role in this behavior is actually morally valuable.
This problem does not necessarily apply to virtue theorists who see the virtues as a supplementary concept to other moral theories. If one argues, with Nussbaum, that virtue ethics is not a "third category" but an addition, then one can use virtue ethics to account for the good of individual character and another theory to account for collective phenomena (Nussbaum 1999). In this article, however, I aim to investigate in how far virtue ethics itself can be modified to account for these collective and individual phenomena simultaneously.
Human moral life happens interpersonally. Much of our life is spent collaborating and contributing to groups and communities. Vicious character, whether it is epistemic or moral, can have an important role in such a group or community. If we uphold a pure virtue theory, and accept that moral virtue is the source of moral good, then an explanation is needed for how vices that lead to group flourishing can still be vices. If pure virtue theories can provide such an explanation, then they are better suited for prescribing ethical conduct for group and communal behaviour.

Three Features of a Virtue Theory that Can Account for Mandevillian Morality
The examples in the previous section show that mandevillian morality is irreducible to individual virtue only. Sometimes individual vices or non-virtues play an essential role in creating emergent good. As I have argued, it is hard to explain mandevillian morality in terms of traditional virtue ethics, and virtue ethics would benefit from being able to account for it.
It seems to me that a virtue-theoretic response can be formulated. There is at least one way in which a virtue theory could account for emergent good. As I shall argue, we can do so by accepting a type of virtue theory which has three essential components, all of which combine into a theory which accounts for the role of vice, or the virtue of vice. The three features, which work synergistically in both virtue ethics and epistemology, are: 1. a distinction between individuals and groups as possessors of virtues 2. a distinction between motivational and teleological virtues and 3. recognizing role-based virtues that can attach to "vicious" roles The first feature accepts that groups can also possess virtues and vices in the same way individuals can. A group can develop mandevillian virtues which are possessed by the group as a collective agent. The second feature entails accepting that there are two types of virtue: teleological and motivational virtues. Teleological virtues are good because of their leading to group flourishing or furthering the groups progress towards a moral telos. Motivational virtues are those which are good mainly because of virtuous motivations.
The third feature calls for attention into the specific virtues and vices connected to each role that people can take in collaborative activities. A group may require a role for "the dogmatist," "the egoist," or "the wrathful" in order to expand group creativity or maintain social order.
The three features of the view I explore make cross-cutting distinctions among epistemic or moral virtue theories. While each feature is in itself sufficient in accounting for emergent good, accepting all three allows for dealing with various different kinds of mandevillian morality. The last 2 features should be seen as a re-classification of what traits count as virtues. They are categories by which a trait may be classified as a virtue. For instance, a trait which is a teleological virtue can be a procedural vice and a role-based virtue. I will show how each of these distinctions still fits virtue within the confines of Zagzebski's characterization of virtue.

A Distinction between Individuals and Groups as Possessors of Virtues
Many accounts of epistemic virtues already entail this feature. Theories in virtue ethics and epistemology that attribute virtues to groups are often referred to as collectivist. Collectivism is a metaethical position which states that groups can have moral responsibility in the same way that individuals can. The position also has virtue-theoretic followers in responsibilist epistemology and ethics. A root of collectivist thinking is in most cases the plural subject-thesis in the work of Margaret Gilbert. Gilbert (1992) argues that groups or dyads of subjects, when both convinced of a belief held by a group they belong to, compose a plural subject, a single knower or single agent. Therefore, groups can have moral agency as a unit and carry moral responsibility for the actions that the group takes as a unit.
Fricker states that if all members of a group have some virtuous motive individually and are reliable in bringing about its object, then the group can be said to possess a virtue. She also argues that it is not necessary for members in the group to possess that virtue individually. Members of a group may also "go along with" a motivation, because they see it as appropriate to the role they take in a certain institution. In this case the group can still be described as having a virtue or a vice. She contends that both Aristotelian virtues, which require a motivational component, and Platonic/Stoic virtues where the virtue is simply an excellence, can be possessed by collectives (Fricker 2010). She does not go as far as claiming that mandevillian intelligence or morality can constitute a group virtue. She discusses the example of a jury where individual, but differing, prejudices cancel each other out, creating fair-minded decisions. This, she states, is a matter of luck rather than virtue, because the jury cannot be held creditworthy (Fricker 2010).
Fricker's example of a diversely biased jury is indeed not an example of collective virtue, nor mandevillian morality. But this is because the jury in question is a temporary group, not a group agent. A jury is generally assembled for a short period of time and hence does not have time to develop a "group personality." 2 It should be noted that virtues are not merely inclinations, but long-term and stable character traits. A collective virtue, like the fair-mindedness of a diversely biased jury, can also be a long-term stable character trait, let us imagine that the jury is assembled to meet regularly over the course of a year, for example. The group dynamic, in which the dogmatism and biases of individuals play a constitutive role, can evolve to a state where the diverse biases cancel each other out in the making of decisions, which may be helped by the biases of the participants also being long-term and stable on the individual level.
On the individual level, developing a "group character" consists of developing appropriate ways of communicating with each member, developing a sense of group identity and perhaps even the individuals learning to take pleasure in certain modes of communication. All of these individual developments give rise to individual motivations that respond to the traits of other members, and keep the group in balance. They might take forms like "Robert talks too much, so I better shut him up before everyone gets bored" or "this group is full of fragile snowflakes, I better assert my opinion louder to toughen them up." If these micro motivations become systematic in the group dynamic, and contribute to group reflectiveness, then they can constitute the motivational component of the group virtue in the same way as Fricker's "willingness to go along with"-motivation can. What makes these micro-motivations count for a virtuous group motivation, is that they are in some sense aimed at group virtue, reflectiveness in this case.
It may thus be possible to account for mandevillian intelligence or morality within virtue theories by acknowledging their value and seeing them as collective mandevillian virtues. Then one can reasonably say that a single bigot in a group has a vice, but the group as a single collective agent has the virtue of being reflective, as a result of the bigot. One can still explain the fact that bigotry is a vice, but a group with a bigot that functions as a reflection mechanism has a collective virtue at the same time. A distinction can be made between the individual person and the collective person, which is a collective agent that can possess virtues and vices. This distinction allows us to talk about two different agents, the individual with the vice and the group with the virtue.
If we take Zagzebski's list of properties of virtues, we see that collective virtues can also fit this description and thus that we can meaningfully speak about collective virtues. 1. A collective virtue can be "an acquired excellence of a person in a deep and lasting sense." In this case, the "person" is a collective agent. These ways of operating can also be stable and long-lasting.
2. A collective virtue can be "acquired by a process that involves a certain amount of time and effort." Groups take time, and collective effort to develop fruitful dynamics and ways of operating.

A virtue is not just a skill. A skill is distinguished from virtues in various different ways.
A collective virtue, like reflectiveness fall into the category of virtues rather than skills. Group reflectiveness, for instance, must be exercised in order to be had. It does not make sense to speak of a reflective group, if the group does not take part in deliberation among members. The good of group reflectiveness is also not merely good because it leads to more informed decisions, better strategies and such. Such a "group trait" is also often described as having intrinsic value (Peter 2013). 4. As explained earlier, group virtue also has a motivational component. This consists of the virtuous motivations of individuals, and/or a "willingness to go along with a virtuous motivation" or systematic motivations towards improving or upholding a good group dynamic, even if these stem partially from vice. 5. Group virtues can also be successful in hitting their desired targets, while possessing the abovementioned qualities. For instance, a target of reflectiveness would be the ability to form new intentional states based on deliberation. In the case of group reflectiveness, these intentional states can be intentional states held by the group agent.

A Distinction between Motivational and Teleological Virtues
In addition to distinguishing between collectives and individuals as possessors of virtues, I propose distinguishing between two categories of virtue. What are referred to as "teleological virtues" in this paper are often called "consequential virtues" in literature. The word "teleological" is preferable over "consequential" to avoid associations with utilitarianism. Making the distinction between motivational and teleological virtues means that a case of mandevillian intelligence or morality could be evaluated in the following way: the individual has a motivational vice, but this vice is also a teleological virtue. The teleological and motivational component are two different ways in which a trait can be blameworthy or praiseworthy. Motivational vices are traits that are inherently or categorically wrong. Teleological vices would be vices which are mostly bad due to the negative consequences. Motivational virtues are traits like benevolence or honesty, which are inherently good. Teleological virtues are those which are good mostly because they facilitate the end goal of flourishing. No virtue theoretical account explicitly uses this distinction to postulate two kinds of virtues. However, the distinction seems to be implicit in many accounts when describing the aspects that constitute a virtue. Bradley, who defends virtue consequentialism, also notes that prominent virtue theorists like Philippa Foot, Linda Zagzebski and Julia Driver all embrace virtue consequentialism in some way (Bradley 2005). However, these authors (with the exception of Driver, who favors virtue consequentialism over other accounts (Driver 1989)) all also acknowledge that there is another component to virtue which has nothing to do with the consequences of one's virtuous behaviour. 3 Virtues are sometimes described as having a motivational requirement, as well as the requirement of reliable bringing about of valuable consequences. It seems that it is at least plausible that there are virtues where the teleological component is more important or pronounced and others where the motivational component is more important or pronounced. It also seems plausible that there are people possessing a virtuous trait that can be described as more teleological or more motivational in the specific instance of that particular person. Because of this, it is also plausible that some virtues can be described as completely motivational, referred to as motivational from here on, or as completely teleological.
To illustrate the difference, think of two versions of the virtue of honesty. Someone may have acquired a strong motivation to tell others the truth and to practice honesty whenever they have the opportunity. However, the same person may be bad at communicating, due to extreme shyness or inability to articulate what one means or judge how much another understands. This person is clearly honest, but their honesty does not "hit its target" which is to make others aware of what one thinks is true. This person is motivationally honest, but not, or at least less, teleologically honest. Think of another person, who possesses a different kind of honesty. They are not necessarily motivated to tell others the truth, but they have acquired the habit of reliably doing so, and communicating their perceptions very accurately to others, perhaps because it saves them trouble. However, even if confronted with a situation where it does not save them trouble to be honest, they reliably act honestly, because it is a part of their character and identity. Such a person also possesses the virtue of honesty. However, their type of honesty is a more teleological, and therefore "target-hitting," type of virtue and admirable mostly for that, rather than for the motivations behind it.
In the case of a bigoted individual who nevertheless makes the group more reflective as a whole: one could say that the individual has a motivational vice. They may be motivationally arrogant or insouciant. This vice can however still be a teleological virtue, like intellectual courage, if it is an ingrained habit which tends to "hit its target" which is challenging unspoken assumptions and leading to fruitful dialogue.
Let us reflect on whether making this distinction aligns with Zagzebski's list of properties of virtues. Both teleological and motivational virtues must be acquired excellences of the character in a deep and lasting sense as well as involve time and effort to acquire. While virtues traditionally have both a motivational component and a success component, I suggest that some virtues can have more of one than the other. This challenges the idea that having a virtue means doing the right thing, at the right time, for the right reason. However, it does seem to make virtue something more achievable for most people, answering the common critique that virtue theories are elitist.
An opponent may argue against the conception of teleological virtues by arguing that teleological virtues are merely skills. However, teleological virtue can in my view be distinguished from a skill in various ways. Firstly, a teleological virtue, like teleological honesty, must be practiced to be had. No one will be called honest if they do not show honesty. Communication skills, which may benefit one in being honest, do not need to be practiced to be had. Secondly, teleological honesty would be affected if the person was persuaded not to practice it or was too distracted to practice honesty.

Recognizing Role-Based Virtues that Can Attach to "Vicious" Roles
Lastly, I propose a third feature for virtue theories to deal with mandevillian morality: recognition of role-based virtues that can attach to "vicious" roles. Fields of applied ethics for professional contexts sometimes use a role-based ethics. This means that the right traits or actions are dependent on the role that a person has in a particular circumstance. Virtue approaches are also sometimes connected with role-based ethics. In such accounts, what is virtuous is dependent on one's role and there can be role-specific virtues (Nuyen 2007;Swanton 2016). Levine and Cox (2016), for instance, argue, using MacIntyre's approach, that virtues can be understood through social practices. The authors argue that due to the specific practices that are involved in academic work, we can distinguish academia-specific virtues, relating to the role of the academic as a researcher and as an educator. Similar accounts can also be imagined for virtue epistemology. For instance, the virtue of "impartiality" may be more relevant for a judge than for a defense lawyer. One's role as a judge may therefore impose a greater normative demand on developing this virtue. Similarly, one could imagine that a good defense lawyer is precisely someone who lacks this virtue.
An account that can deal with emergent good would benefit from such a role-based conception of the virtues. In a case of mandevillian intelligence or morality, the context is an essential part of what makes the virtue. Egoism or stubbornness can only become a teleological or group-level virtue once it is embedded in the right social context. While one egoist might be essential for a group virtue, ten egoists are perhaps too many. There may be "a role" for an egoist in a specific context, but once that role is occupied, another egoist may simply be destructive to the group virtue. Importantly, an account that can deal with mandevillian intelligence and morality should acknowledge the existence of "evil roles" such as the role of the "dogmatist" or "reflection generating bigot." Let us consider Fricker's example of the jury that consists of diversely biased individuals, which functions in a fairminded way. If one imagines that the jury was compiled, on purpose, to be as diverse as possible, precisely so that it would come to fairminded conclusions, then each member takes a certain role. If each has "being biased" as a part of their role description, then acting in accordance with that bias can be considered virtuous.
Accepting the possibility that virtues can be relative to roles does not expand the notion of virtue in any way that would risk making it incompatible with the definition of virtues I have been using. A role-based virtue can be "an acquired excellence of a person in a deep and lasting sense" (p.135, Zagzebski 1996) if that role is maintained by a person for a longer period of time. Such a virtue can also be "acquired by a process that involves a certain amount of time and effort" (p.135, Zagzebski 1996) as one takes time and effort to adjust to a role. The motivational component of such a virtue can be an identification with the role, or knowledge of its significance and a passion to fulfill this significance. A "success component" would be the ability of the person to bring about the intended positive effects of having the role in the group or community. Such a role-based virtue can also be distinguished from a skill.

Conclusion
Traditional virtue theories have often revolved around a notion of balance or harmony between the virtues within the individual soul as well as between different people within society. Plato's virtues are distributed unevenly in society. The soldier lacks the wisdom of the philosopher but makes up in courage. Together the different members of the state can form a harmonious whole (441ad, Plato 2010). In this paper, the central claim has been that virtue theories would benefit from acknowledging the way that vices too are distributed unevenly in a community, and how they can also contribute to a harmonious whole. If ones sees virtues as the right middle between two vices, then this golden mean can also be achieved between two vicious people. A group where all the vices are represented can develop a collective way of working wherein all the vices are balanced into all the virtues, so that the collective agent is virtuous.
There may be a case where a bigot causes a group to be more reflective. We may blame the bigot for her vice, but we may praise the group, including her, for having the virtue of deliberation. Imagine that her bigotry does not result in better group morality, but rather, to the fact that the group manages, partially thanks to her bigotry, to innovate a tremendously valuable innovation. Then we may blame her for her bigotry but at the same time praise her for her teleological virtue of having that same trait which pushes the group forward towards creative directions. We may understand this teleologically beneficial trait as intrinsically good by virtue of the role of "representing evil," "expanding the collective imagination" or "triggering action" that she occupies in the group. These roles can, just like group virtues, be acquired and developed over a period of time.
An opponent may worry that accepting collective, teleological and evil-role-based virtues is a danger to the importance of individual character development. One might argue that if an account of virtue integrates vice, that this allows people to explain away their vices as virtues and no improvement is incentivized. A twofold response can be given: Firstly, the opponent's question could be asked the other way around too: whether rejecting these kinds of virtues is not a danger to the importance of virtues arising from group dynamics. Humans live and collaborate in groups, so why should a virtue theory prioritize the importance of the individual as an agent over the group as an agent? Secondly, the advantage of the account I put forward is precisely that individual virtues remain valid and important, while emergent virtues are also acknowledged. If anything, this allows for a wider spectrum of appreciation for the valuable kinds of character development and group development that people undertake.
The approach to virtue that I recommend makes distinctions between different types of virtue, different roles and different kinds of possessors of virtues. While it still leaves many questions open, it is a plausible answer to mandevillian morality, mandevillian intelligence and similar phenomena that pose a challenge to virtue ethics. Other possible responses also merit investigation.