What is social hierarchy?

Under which conditions are social relationships hierarchical, and under which conditions are they not? This article has three main aims. First, I will explain what this question amounts to by providing a more detailed description of the general phenomenon of social hierarchy. Second, I will provide an account of what social hierarchy is. Third, I will provide some considerations in favour of this account by discussing how it improves upon three alternative ways of thinking about social hierarchy that are sometimes explicitly endorsed and sometimes suggested or presupposed in writings in philosophy and elsewhere.


INTRODUCTION
People stand in all sorts of social relationships to others. Sometimes these social relationships are hierarchical as opposed to non-hierarchical or egalitarian. Here are two examples of the kind of phenomenon I have in mind: . . . the Untouchable was not allowed to use the public streets if a Hindu was coming along, lest he should pollute the Hindu by his shadow. The Untouchable was required to have a black thread either on his wrist or around his neck, as a sign or a mark to prevent the Hindus from getting themselves polluted by his touch by mistake. In Poona, the capital of Peshwa, the Untouchable was required to carry, strung from his waist, a broom to sweep away from behind himself the dust he trod on, lest a Hindu walking on the same dust should be polluted (Ambedkar 2014: 5).
We readily identify social relationships described in these examples as hierarchical. In addition, we share a rich vocabulary to talk about social relationships of this kind. We say that people are above, below, behind, or in front of others; people rise and fall; we are looked up to and down upon. We speak of superiority and inferiority, and of people being treated or regarded as more, less, better, or worse than others. We talk about differences in status, standing, rank, position, and respect. All of this is familiar from everyday language, but the social sciences introduce various terms as well. Henrich and Gil-White (2001), for example, write about social asymmetries; Fiske (2011) about the vertical and horizontal dimensions of social relationships; Ridgeway and Balkwell (1997: 15) about differences in the degree to which people are worthy in society. Not all of this language always fits every particular instance equally well, but it is nonetheless the case that we have a body of language to refer to the general phenomenon of social hierarchy. The central question of this article is: what is social hierarchy? Under which conditions are social relationships hierarchical, and under which conditions are they not? I have three main aims. First, I will provide a more detailed description of the general phenomenon of social hierarchy ( §2). Second, I will provide an account of what social hierarchy is ( §3). Third, I will provide some considerations in favour of this account by discussing how it improves upon three alternative ways of thinking about social hierarchy that are sometimes explicitly endorsed and sometimes suggested or presupposed in writings in philosophy and elsewhere ( §4).

THE PHENOMENON OF SOCIAL HIERARCHY
Persons can occupy various social positions. 1 We can think of pluralities of individuals and the social relationships they stand in as forming networks, with the nodes representing persons, and the connections between nodes representing (different kinds of) social relationships. At a sufficiently fine-grained level of analysis, the full complement of social relationships any given person stands in to others in any realistic social network is bound to be unique. It is often possible, however, to identify structural similarities between the sets of social relationships in which different participants in the network stand to the other participants. On a hospital ward, all of the nurses may relate in similar ways to one another, and to the other participants in the network-medical assistants, residents, attending physicians, and so on. The term 'social position' refers to such structurally similar locations in social networks. 2 Occupying one or another social position in a given social network has a range of implications for how you interact with other participants and for how they interact with you. In early modern Germany, for example, in one's social interaction with an executioner, one would avoid all physical contact with him, treat his touch as contagious, and approach him with hostility (Spierenburg 1984). Of course, how one responds to a person in a given social position will depend on one's own social position and on various other circumstances. Convicts, commoners, lords, and the executioner's wife all relate to him in different ways, and will do so differently depending on whether he is performing an execution or simply walking through town. I will refer to such implications as the incidents of a social position. The incidents of any real social position are bound to be highly complex, but it is in principle possible to give a full specification of the incidents of any given social position. 3 Since persons can be members of several collections of individuals, they can also occupy several distinct social positions at the same time. The collections of individuals at issue in this article will have a certain degree of salience. We might look at the the members of tribes, clubs, or platoons, for example, or the inhabitants of communes or villages, or the workers in a particular work place. I will speak of distinct salient sets of persons and the social relations between them as distinct social contexts. A person may occupy the position of guard in the social context of a prison, mother or father in the family, and volunteer in church. Although it is possible that persons occupy salient social positions in a social network defined over an entire political society-perhaps persons occupy one or another social class-there is no sense in which we can "aggregate" a person's various social positions so that we can assign a single overall social position to each person. This means that individuals may occupy a number of different positions, high and low, in a variety of social hierarchies, but there is no such thing as a person's overall or aggregate social standing. 4 We can now state the central question of this article in a more precise way. Suppose that we can, for a given social context, identify each distinct social position in it, and that we can provide a complete specification of the incidents of each social position. Suppose further that we can provide these specifications in terms that are in themselves neutral with regard to the presence or absence of social hierarchy. 5 Given such a description of a social context, my question is under which conditions the participants in that social context are hierarchically ordered. This indicates that one central feature of social hierarchy that an adequate account needs to explain is valence: social positions in a social hierarchy are not merely different; their occupants are situated "above" and "below" one another. 6 I identify two further facts about social hierarchy that an adequate theory needs to be able to explain. The first is that the incidents of the different social positions in social hierarchies characteristically include distinctive requirements or normative expectations. 7 People not only in fact display different patterns of attitude and behaviour towards others depending on their own and those others' social positions; they also, in a sense to be explained, should. The example from Annihilation of Caste outlines some of the requirements that apply to "Untouchables," 8 and cool kids are expected to join the clique leader in derisive behaviour towards others (Adler & Adler 1998: 66-68). These examples most explicitly concern behaviour; the following illustrates some of the attitudes that may be required of us: "Yes, Miss Elizabeth, you will have the honour of seeing Lady Catherine de Bourgh on the ensuing Sunday at church, and I need not say you will be delighted with her. She is all affability and condescension, and I doubt not but you will be honoured with some portion of her notice when service is over. I have scarcely any hesitation in saying that she will include you and my sister Maria in every invitation with which she honours us during your stay here" (Austen, 1882: 135).
Elizabeth is not merely predicted to be delighted with and to feel honoured by Lady de Bourgh; this is normatively expected of her. Likewise, people in various positions may be expected to believe or want certain things, or to be ashamed, disgusted, or in awe of themselves or others.
The second fact about social hierarchy is that people are often powerfully motivated by considerations of social hierarchy. First, where social mobility is possible, people often aspire to occupy higher rather than lower social positions. People frequently try to rise in social position-wannabe kids aim to be cool kids, ladies to be duchesses. Such elevations give rise to characteristic emotions of pride and self-esteem in those moving up; envy, admiration, and anger in those left behind. Less ambitious individuals may at least try not to fall in social position, where such lowerings are often experienced as shameful, humiliating, demeaning, or degrading. 9 Second, even where social mobility is difficult or impossible, participants in social hierarchies often care a great deal about being treated and regarded in a manner that is befitting the social position they take themselves to be in. Moreover, this concern is asymmetrical: it is a lot less painful to be treated and regarded as if one occupies a higher social position than a lower one. The former might be flattering, perhaps embarrassing; the latter is often perceived as an insult and an occasion for anger or shame. We care about the appropriate social responses not just for ourselves but also for third parties. Those who go around acting as though they belong to a higher station than they actually occupy are seen as arrogant or as not knowing their place. It is vexing when our social equals are treated and regarded by others as though they are above us. Transgressions of this kind can be met with a variety of social sanctions. These kinds of motivational phenomena are not uniform; they can vary from one person to another and from one social hierarchy to another. Still, the fact that our navigation of various social hierarchies is commonly associated with strong motivations and emotions is a fact that an account of social hierarchy should be able to explain.

Social hierarchy and social norms
In this section, I provide an account of social hierarchy that has two main parts. First, I propose that social hierarchies are constituted by sets of operative social norms. For the purpose of my discussion in this article, I will rely on Bicchieri's informal account of what social norms are: A social norm is a rule of behaviour such that individuals prefer to conform to it on condition that they believe (a) most people in their reference network conform to it (empirical expectation), and (b) that most people in their reference network believe they ought to conform to it (normative expectation) (Bicchieri 2017: 34). 10 I will call a social norm operative (as opposed to latent) in a given social context if both conditions are met for the participants in that context; that is, if the participants have the relevant empirical and normative expectations. 11 Of course, the nature of social norms is itself disputed, but my account of social hierarchy can, I believe, be made compatible with a range of different views. 12 I make one key amendment to Bicchieri's account: in my view, social norms can demand not only that individuals behave in certain ways, but also that individuals display various dispositions, emotions, feelings, or other attitudes. Social norms surrounding the death of loved ones, for example, may demand that we grieve for them, where grief is not simply a matter of displaying certain kinds of behaviour. 13 Social hierarchies are sets of operative social norms that demand clusters of attitude and behaviour from the various participants in a given social context. Occupants of different social positions are subject to different requirements, and the attitudes and behaviour demanded of them varies depending on the social position of the individuals they interact with. The social requirements that apply to "Untouchables," for example, are sharply different from those that apply to Brahmins, and "Untouchables" are required to respond very differently to Brahmins than to other "Untouchables." Similar remarks can be made about the participants in my other examples. However, mere differences in the social requirements that apply to different participants in a given social context do not entail the presence of social hierarchy. Think, for example, of team members in a workplace, each with their own professional role. They may each be socially expected to perform their distinctive tasks, but nonetheless relate to one another as equals. In other words, the proposal to think of social hierarchies as sets of operative social norms leaves the valence of hierarchy unexplained. The second part of my view explains that feature.
Before I turn to the second part, I describe three important benefits of the role of social norms in my account. First, it provides a straightforward explanation for the fact that social hierarchies characteristically involve social expectations. As my examples illustrate, participants in a variety of social hierarchies are socially expected, required, demanded, or asked to behave in certain ways and to have certain attitudes. My account explains this in terms of social norms. This explanation properly accounts for the way in which these expectations are normative. The "Untouchable" should wear a broom, the vassal should provide military service for his lord, Elizabeth should feel honoured by Lady de Bourgh's invitations in the sense that these things are required of them by operative social norms. It does not follow that these individuals have all-things-considered reason to conform, nor that they are morally required to do so or that they should in some more general normative sense.
Second, it partly explains why people are often strongly motivated by considerations of social hierarchy. It is part of the structure of social norms that people prefer to comply with them conditional on certain empirical and normative expectations. This feature of social norms explains why people are often motivated to act in conformity to the social expectations that attach to their social positions. It also explains why people care about whether others conform to the expectations associated with their roles, and why we are frequently prepared to sanction others for failures in compliance: in doing so, we participate in the enforcement of the social norms that are operative within our social context. 14 This does not explain the full range of our social hierarchy-related motivations. Most importantly, it does not explain why we might aim to rise in social position, and why we might be especially averse to degradation. The second part of my account is needed to explain these further motivational features.
Third, social norms identify relevant social contexts. As far as social hierarchies are concerned, distinct social contexts are groups of individuals covered by distinct sets of operative social norms. This helps explain how one individual can occupy several positions, high and low, in a number of distinct social hierarchies-the same person may be a high-status member of an online gaming community, say, while being an outcast at school. This is possible because different groups of individuals, or even the same group of individuals in different circumstances, can be regulated by different social norms requiring different patterns of attitude and behaviour. It also helps avoid a problem of proliferation. On some views, hierarchies of esteem or power are constituted by inequalities of esteem or power, regardless of whether those inequalities are supported by social norms. Such views tend to generate social hierarchies everywhere. We may think of hierarchies of power, for example, in terms of the overall amount of power each person has. I might then count as below the current president of Chile but above a factory worker in Japan in a single social hierarchy, provided I have less overall power than the president but more than the worker. Hierarchies of esteem can be thought of in terms of third party assessment: one musician outranks another in an esteem hierarchy, say, when the former is more highly esteemed than the latter by their listeners. Such views can produce the result that each time anyone has greater esteem for one person than another, this creates a mini-hierarchy between those two people. Of course, views of these kinds may add further constraints to limit this kind of proliferation, but my understanding of social hierarchies as sets of operative social norms does this for my view.

Valuing persons more than others
The key feature of social hierarchy that remains to be explained is valence. Under which conditions does a set of operative social norms that governs the social relationships between the occupants of different social positions hierarchically order those social positions? The second part of my proposal says that a social position A is "higher than" or "above" social position B if and only if, for the participants in the relevant social network, when they display the norm-required complexes of attitude and behaviour they thereby and to that extent value the occupants of A more than the occupants of B. To illustrate this proposal, take a prestige hierarchy among professional tennis players. My view would say that such a hierarchy is constituted by a set of social norms asking the participating players to respond to one another in various ways. Most importantly, participants are expected to display different patterns of admiration, emulation, praise, attention, and so on, towards different people with whom they interact. When participants respond to different players with different levels of praise and admiration, they thereby and to that extent value them more or less as tennis players. Players are located in higher and lower positions in the hierarchy depending on how highly they are socially required to be esteemed in this sense. One immediate objection to my proposal is that the presence of social hierarchy is clearly compatible with some participants valuing certain occupants of lower social positions more than the occupants of higher social positions. Take, for example, a hospital setting in which a group of surgeons occupy a relatively high social position, and the nursing staff a lower social position. One of the surgeons might be married to one of the nurses, such that she values her spouse more, not only compared to the other nurses, but also compared to her fellow surgeons. Cases like this do not present counterexamples to my proposal. The reason is that the social norms regulating the relationships between surgeons and nurses do not ask of the surgeon that she displays a pattern of behaviour and attitude that constitutes her valuing her spouse more than her colleagues. As a surgeon, within the social context of the hospital, she may be expected to direct and make demands of the nurses, to overrule them in cases of disagreement, and to prioritise her time over theirs, and she may be expected to do none of these things in relation to her fellow surgeons. To display such a pattern of behaviour and attitude is, taken by itself, for her to value her fellow surgeons more than the nurses, within the social context of the hospital. The applicable social norms may nonetheless permit the surgeon to have a special relationship of love, care, and concern with one of the nurses, as long as the surgeon does not violate social expectations by, for example, favouring her spouse in workplace decision-making.
What does it mean to value one person more than another? I will start with the idea of objects being valuable. My pocket knife is valuable. Setting aside the question of whether my pocket knife could be just plain good, it can be good in various ways. It might be good as a cutting tool, for example, or as a collector's item. One implication of an object being valuable in one way or another is that we have reasons to hold certain attitudes towards it and to act in certain ways in regard to it, or that certain complexes of attitude and behaviour towards the object would be fitting or appropriate. 15 Persons can also be valuable in different ways. A person might be valuable as a member of the team, as a tennis player, and perhaps also as a person. Each way in which a person is valuable entails that we have reasons to hold certain attitudes towards her and act in certain ways in regard to her, or that a certain complex of attitude and behaviour towards that person is fitting. For example, we might have reasons or it might be fitting to hire her as a tennis coach, or to praise her for her contribution to the team effort. From now on, I will speak about this in terms of fitting complexes of attitude and behaviour, rather than in terms of reasons, but nothing important hangs on this choice for my purposes here. When we display a complex of attitude and behaviour towards a person that is rendered fitting by that person's being valuable in one way or another, then we value the person in that way.
The next step is the idea of valuing one person more than another. I will not address the question of whether one person can be just plain better than another, but persons can be better than one another in various ways or respects. One person may be a better member of the team than another, say, or better as a parent or as a tennis player. The different degrees to which persons are valuable in such ways render different complexes of attitude and behaviour fitting towards those persons. As a result, we can, for example, value one person more as a parent than another by displaying a complex of attitude and behaviour towards that person that is rendered appropriate by their greater value as a parent, and that would not be fitting if they would merely be as valuable as a parent as the other person is. I have no general account of the kinds of attitudes and behaviours that can constitute valuing persons, but it would seem that we commonly do so by, for example, praising, emulating, admiring, respecting, supporting, caring for, attending to, loving, and honouring them. At the other end, common ways to respond to persons as though we disvalue them include disrespecting, humiliating, ignoring, blaming, ridiculing, ostracising, or having contempt for them.
So far, I have only looked at cases in which valuing objects more or less corresponds to differences in value. It is important for my account of social hierarchy that such correspondence need not be in place. In general, we can value things even though they lack the value that would make our response fitting. I might highly value my pocket knife as a collector's item, even though it in fact is a cheap replica. Likewise, we can highly value people as team members, parents, or tennis players even though they are poor team members, parents, and tennis players. We do so by displaying the complexes of behaviour and attitude towards them that would be fitting if they were valuable in the corresponding way. As a consequence, my view can account for cases in which the relative standing of persons in a social hierarchy does not correspond to underlying differences in value. Furthermore, we can value people, and objects generally, without believing them to be correspondingly valuable. 16 As participants in a prestige hierarchy, for example, we can, on my view, emulate, admire, or esteem a person without believing that the person is correspondingly admirable or estimable. We may value persons in these ways because this is required of us by social norms, and because we would like to avoid various social sanctions.
In sum, my proposal is that a social position A is hierarchically ordered above social position B just when it is true for the participants in the relevant social context that if they display the socially expected complexes of attitude and behaviour, they thereby and to that extent value the occupants of position A more than the occupants of position B. Persons can be valuable in different ways, so the question naturally arises: which ways of being valuable are relevant for social hierarchy?
My view is that social hierarchies can be built around any respect in which persons can be more valuable than others. In some social contexts, people are socially expected to treat different people as more or less honourable, pure, or courageous; in other contexts as better or worse tennis players, dog trainers, and so on.
This flexibility allows the account to explain the wide variety of hierarchical social structures, but it also leads to some puzzles. In some social contexts, the operative social norms demand patterns of attitude and behaviour that amount to valuing different individuals in different ways. The cardiologist is expected to be respected, appreciated, and deferred to when it comes to cardiology, the gynaecologist when it comes to gynaecology, and so on. By itself, this does not constitute a social hierarchy. We should look at the patterns of attitude and behaviour demanded by all the social norms operative in the context together and see whether they involve some individuals being valued more than others. Individuals being valued in different ways does not show this to be the case. This suggests an objection to my proposal, however. Might it not be the case that in clearly hierarchical social structures people are also expected to be valued in their different roles? In a feudal system, peasants may be valued as peasants, lords as lords; in a caste system, each individual may be expected to be valued as an occupant of their distinctive caste position. My answer is that in such cases, the full set of operative social norms in the relevant context do not just require that individuals are valued as occupants of their distinct roles, but also that the occupants of some roles are valued more than others. The full set of social norms does not just require that peasants are appreciated as peasants and lords as lords; it also requires patterns of attitude and behaviour that amount to valuing lords more than peasants. 17 One advantage of the resulting account is that it explains why people often seek to rise, or at least not to fall, in social position. On my view, in doing so, people aim to occupy social positions in which they are valued by others more rather than less. Our desire to be valued by others may in turn have different sources. First, we may value this instrumentally: being highly valued by others is often associated with various advantages that we may desire for social hierarchy-independent reasons. Second, we may value this for epistemic reasons: being highly valued by others may give us reason to believe that we are highly valuable. In this way, the attainment of rank in a social hierarchy may satisfy our desire for a kind of self-esteem. Third, it is not unreasonable to suppose that we often care intrinsically about how we are being valued by others. We may wish to be valued more, rather than less, or in keeping with how valuable we deem ourselves to be, and many of us may desire to be valued more than other people. Taken together, the social norm component and the valuing component of my account offer a promising set of resources to explain why we are often strongly motivated by considerations of social hierarchy.

Problems of underinclusion
Each part of my account of social hierarchy may seem to make it underinclusive in its own way. I have illustrated the second part of the account using an example of a prestige hierarchy, which in some respects in the easiest case for my view. It may seem difficult for me to account for social hierarchies that are largely constituted by asymmetrical de facto authority relationships or by dominance relationships. Think, first, of a military unit with a clearly defined chain of command, in which sergeants have a certain degree of de facto authority over privates, but not the other way around. Sergeants occupy a higher place in the social hierarchy internal to the military unit than the privates, but it may seem that the recognition of the relevant de facto authority relationships does not constitute a way of valuing sergeants more than privates. My response is that having legitimate practical authority is a way for persons to be valuable, such that ascribing greater authority to one person than to another is to value that person more than the other. One part of the motivation for this view is that we, as persons, are capable of practical deliberation and decision-making, and the recognition of a certain degree of authority over our own affairs is one way of valuing persons as reasoning agents-it is a way of respecting persons. We also deliberate about what other persons are to do, so that to ascribe authority to a person to determine how others are to act is another way of valuing them as practically reasoning agents. The asymmetrical ascription of authority can then plausibly count as a way of valuing some people more than others.
Second, dominance hierarchies also seem difficult to fit into my account. Whether this is the case depends on how we understand dominance relationships. In some areas of research, one individual is said to dominate another when the former attacks or threatens to attack the latter more often than the other way around, and where the latter typically does not respond with aggression. Pecking orders among hens are an example of this, but relationships of this kind are found in many species. 18 A possible example among human beings is the medieval raid, a common form of warfare in which aristocratic knights-on horseback and fully armed-would pillage large areas of farm land, burning crops and villages, and attacking unarmed and defenceless peasants (Lynn 2003: chapter 3). Those peasants would have lords of their own, and would stand in hierarchical relationships to those lords, but they would confront a foreign knight not as someone owed loyalty or deference, but simply as someone with the capacity to apply overwhelming force. My view does not consider dominance relationships, understood in this thin sense, to be instances of social hierarchy. However, I do not consider this to be a fault of my proposal: I take there to be an important difference between social hierarchy and the sheer exercise of force or physical aggression. The notion of dominance is also frequently used to describe a more socially complex phenomenon (for example, by Martin (2009: chapter 4)). On such views, dominance relationships between any two individuals are mediated by these individuals' understanding of the social network as a whole and their respective places within it. Further, both aggressive behaviour and nonaggressive responses are to a large extent ritualised. The patterns of interaction that characterise individuals in different positions in a dominance hierarchy often involve displays of dominance answered by displays of submission rather than the direct exercise of physical violence. Such displays can take a wide range of different forms, with a common association between dominance and making oneself physically taller or higher, and between submission and physically lowering oneself. Interactions of this kind are in turn frequently described in terms of respect; a failure to submit to a gang leader, for example, is understood as a form of disrespect. Such ritualised displays of dominance and submission, understood by their agents as demands for and indicators of respect, have a much stronger claim to being ways of valuing some persons more than others than sheer asymmetries in the application of force.
These considerations also go some way towards answering a different complaint of underinclusiveness. The idea that social hierarchy is a social norm-governed phenomenon seems to mistakenly exclude social relationships in which there are sharp asymmetries of power or esteem, but in which there are no social norms requiring differential responses. The first part of my answer, suggested by the previous paragraph, is that many of the inequalities in esteem, power, or other factors that we wish to count as instances of social hierarchy are in fact not mere inequalities but are part of more complex social phenomena structured by social norms. Second, counting inequalities or esteem, power, and other factors as instances of social hierarchy tout court would produce an objectionable proliferation of social hierarchies. Third, my view is able to meet the objection part way by saying that inequalities in esteem, power, and the like are very closely related to the phenomenon of social hierarchy: these inequalities are such that operative social norms requiring them are social hierarchies.

DIMENSIONS, ADVANTAGE, AND IMPORTANCE
It is hard to provide a direct argument for my account of social hierarchy. In this section, I aim to provide some considerations in favour of accepting it by indicating how it avoids some of the problems of three alternative views.

Dimensions and skepticism
The first view I will consider holds that social hierarchy has a number of distinct dimensions, aspects, or elements. Such views are especially common among philosophers writing on social or relational egalitarianism. Anderson proposes 'three types or dimensions of social hierarchy: of authority, esteem, and standing' (2017: 3); Kolodny writes that social hierarchy 'has to do with' relative power, de facto authority, and consideration (2014: 295-6); and Viehoff (2014: 357) holds that equal control over a relationship is a distinctive element of relational equality, where equal freedom, consideration, and rights are other elements. Views of this kind are also found, however, in certain areas of social psychology and sociology. Blader and Chen, for example, list social status, power, socioeconomic status, dominance, prestige, influence, and leadership as distinct dimensions of social hierarchy (2014: 74). Proposed lists vary, but four social phenomena tend to recur as the main proposed dimensions. Dominance relationships centrally involve patterns of interaction in which the aggression of some individuals is answered by the submission of others. In asymmetrical (de facto) authority relationships some persons are taken to have greater standing than others to issue obligation-generating commands. Differences in prestige are a matter of the extent to which persons are esteemed in a given social group. Finally, in systems of honour, participants are expected to display respect and deference to the dignity associated with different social positions. 19 Whatever the exact list of proposed dimensions may be, the question remains how they add up to a general account of social hierarchy. When authors provide lists of this kind, it is not always clear how they intend them to be understood. One plausible way of reading at least some of these statements is as simply providing a list of common features of social hierarchy. The subsequent argument may then focus on one of these features. Kolodny's argument for democracy, for example, proceeds primarily in terms of asymmetrical authority relationships (2014). For certain projects, this may be an entirely appropriate way of proceeding. Furthermore, I have no objection to the idea that the listed dimensions are indeed common features of social hierarchy. However, understood in this way, my question would not be answered: a list of common features of social hierarchy does not constitute a general account of the conditions under which social positions are hierarchically ordered.
Contrary to the likely intentions of most of their proponents, we could take these lists to provide necessary and sufficient conditions for the presence of social hierarchy. We could say that two given social positions are hierarchically ordered just when the incidents of these social positions include asymmetrical relationships of dominance, authority, prestige, or honour between their occupants. Such a proposal is problematic. An initial problem is that the resulting view has difficulty explaining the way in which social hierarchies typically include requirements or normative expectations. This problem could straightforwardly be addressed, however, by accepting the first part of my proposal. The list of dimensions would then only take the place of the second part of my proposed view. A more significant problem is that a list of dimensions does not specify sufficient conditions for the presence of social hierarchy. One reason for this is that distinct dimensions may be present in a single social context and point in opposite directions. For example, a central idea in Dumont's description of the varna system is that Brahmins have a kind of spiritual authority or status and that Kshatriyas have temporal authority or power (1970: 71-72). A simple list of dimensions, the presence of which suffices for the presence of social hierarchy, would yield the inconsistent result that Brahmins are both above and below Kshatriyas. Some further set of conditions specifying how the different dimensions interact with one another would be needed to support Dumont's judgment that Brahmins are above Kshatriyas despite the fact that they submit to their temporal authority (1970: 69). Relatedly, just providing a list of dimensions, elements, or aspects produces an intellectually disappointing account of social hierarchy. Even if it is true that social hierarchy is often, perhaps even always, associated with asymmetries in dominance, authority, prestige, and honour, the view would say nothing about what unifies these phenomena. Why are these, and only these, asymmetries associated with social hierarchy?
This may suggest an alternative proposal. One might doubt that if we have adequate accounts of dominance, authority, prestige, and honour, there is anything further to be understood. One might think that there is nothing interesting that these social phenomena have in common, and that there is no substantial sense in which they are instances of a genuine social kind. This skeptical view provides a more interesting answer to my question, but I think we have at least some reason to be hesitant.
One reason is that, as I mentioned in the introduction, we have an elaborate vocabulary to describe social hierarchy. In many cases, the terms we use can be applied to many different kinds of social hierarchy, whether they revolve around authority, esteem, honour, or dominance. We might, for example, describe people as superiors or inferiors, as above or below us, or as higher or lower in rank, where such things could be said by nurses, wannabe kids, "Untouchables," or vassals. This observation is not just based on my own linguistic sensibilities; in academic writing on social hierarchy, similar terms are used to describe hierarchical social structures of different kinds. Writers often switch between terms, and group terms together, without further explanation or justification. Clifford Geertz provides a striking example of this: 'in deep [games], where the amounts of money are great, much more is at stake than material gain: namely, esteem, honour, dignity, respect-in a word, though in Bali a profoundly freighted word, status ' (1993: 433). Gordon Wood writes of differences in title, rank, status, quality, and social position, all on the same page and without distinguishing between them (1993: 21). Examples of this kind are not difficult to find. Blader and Chen state that researchers in social psychology frequently equate status, power, dominance, influence, prestige, and rank (2014: 72). In short, there is a large literature with a long history spanning multiple disciplines that takes itself to be concerned with social hierarchies, such that we can discuss their causes and functions, how people move around in them, and so on. According to the skeptical view, both the academic literature and our everyday language of social hierarchy rest on a mistake: they group together social phenomena that are fundamentally distinct. This is a significant cost of the skeptical view.
Second, social hierarchies, whether they revolve around considerations of dominance, authority, prestige, or honour, occasion similar patterns of emotional response. Being lowered in rank is often experienced as humiliating, shameful, or degrading. This would be true of a person being stripped of their military rank, but also of a kid who is being kicked out of the cool group, or a lord being condemned to be hanged like a common criminal (see Whitman 2002: 160). Elevation in social position fills people with pride; deviations from proper treatment arouse indignation or anger. Those at the top are looked up to with envy and admiration; those at the bottom down on with pity, contempt, or disgust. Of course, emotional responses to social hierarchy-related phenomena vary from person to person and from social context to social context. Nonetheless, there is a general pattern of emotional responses that we find in association with social hierarchies generally. The skeptical view has no good explanation for this: if social hierarchy is nothing more than a label for a set of fundamentally distinct social phenomena, it is unclear why they should lead to a similar emotional involvement in their participants.
My account improves on both list views and skeptical views by tying valence to differential valuing. When asymmetries of different kinds are present in the same social context, my view holds that we have to look at the full sets of social norms regulating the interactions between the occupants of different social positions, and at the complete complexes of behaviour and attitude required of different participants. One position is above another just when the socially-required set of attitudes and behaviours towards their respective occupants constitutes valuing the occupants of one position more than those of the other. This is not to say that the ordering of different social positions is always clear: there may well be cases in which it is undetermined which of two positions is higher, and there may be cases in which it is difficult or impossible for us to know what the ordering is. Nonetheless, my view helps with mixed cases by tying the ordering of positions to the idea of valuing some persons more than others. On this picture, it is no surprise that we use the same language to talk about the many varieties of social hierarchy, because social hierarchy is a unified social phenomenon. It is also unsurprising that we find a certain consistency in the volitional and emotional relationships people have to different kinds of social hierarchy, because it is no surprise that people would be similarly concerned with different ways in which they are (socially expected to be) valued by others.

Advantage
The second view holds that social hierarchy is a matter of relative advantage. Suppose we can, for each social position in a given social context, determine how advantageous it is to occupy that position. A set of social positions is hierarchically ordered, we can now say, just when it is more advantageous to occupy some social positions rather than others. The more advantageous a social position is, the higher it is placed in the hierarchy. 20 This view is initially plausible. Many examples of social hierarchy fit the account. The social position of a full professor does seem advantageous compared to the social position of an untenured or adjunct professor, as does the position of Brahmin compared to the position of "Untouchable," lord to vassal, or cool kid to wannabe. In addition, the account straightforwardly explains valence; it explains why people are motivated by considerations of social hierarchy and why people would aspire to occupy higher rather than lower positions. The account does not explain the distinctive requirements and normative expectations involved in social hierarchy, but here again this problem can be addressed by building social norms into the view. Different problems arise, however, depending on how advantage is understood. A first suggestion is to think of advantage in terms of welfare: a social position is advantageous to the occupant insofar as occupying that social position is good for the occupant. What is good for a person, or what contributes to a person's welfare, can in turn be understood in a number of different ways. One proposal would be to think of welfare as preference-satisfaction. It is no surprise that high social positions are often associated with high levels of preference-satisfaction-after all, it is often possible to use one's power, authority, prestige, or honour to get what one wants. It is clearly not the case, however, that occupying a higher rather than a lower social position invariably results in higher levels of preference satisfaction as a downstream consequence. Being the head of state, for example, may make you a target of coups and assassination attempts and may leave you much worse off in terms of preference-satisfaction than those situated below you. The better version of this view ties the elevation of social positions to the extent to which the incidents of those social positions tend to satisfy their occupants' preferences. Being an assassination target is a downstream consequence but not normally an incident of the social position of head of state. However, counterexamples can be identified here as well. There is, first, the simple fact that preferences vary widely across individuals. The incidents of relatively high social positions may align well with the preferences of some individuals but not with those of others. There is no inconsistency in the idea that high social positions can come to be occupied by individuals who suffer in terms of preference-satisfaction as a result. Even if we help ourselves to some notion of the preferences of the average person, problems remain. In academic hierarchies, for example, it might well be that undergraduate students and postdocs generally fare better in terms of preference satisfaction, in virtue of their position, than higher-placed PhD students and assistant professors. In a military hierarchy, the distinctive social expectations attached to the position of lieutenant may generally produce greater preference-satisfaction than those attached to the position of colonel. Such observations would not lead us to conclude that postdocs are higher up in the academic hierarchy than assistant professors, or that lieutenants outrank colonels. Such facts may even be recognised by the participants in a given social context, so that the role of leader (or other relatively high positions) is seen as a burden to be shouldered rather than a benefit to be enjoyed.
A second way to understand welfare is in terms of a list of items that constitute well-being. 21 We can now propose that the incidents of relatively high social positions constitute a greater contribution to the well-being of the occupant than the incidents of relatively low social positions, regardless of whether the occupant has a corresponding set of preferences to be satisfied. This view, too, is mistaken. As I suggested before, it is common for asymmetrical social relationships of de facto authority, prestige, and honour to be central to different kinds of social hierarchy. It is, however, also common for such social relationships to fail to correspond to differences in the extent to which people have legitimate authority, are estimable, or are honourable. The relationships between enslaved people and their enslavers, for example, may be characterised by a sharp asymmetry in de facto authority, but those enslavers have no corresponding legitimate authority. In some social contexts, there are social norms requiring people to be esteemed in accordance with the severity of the crimes they have committed (Jankowski 1991: 141-151), but having committed serious crimes is not a genuinely estimable quality. There is some plausibility to the claim that having one's legitimate authority recognised or being esteemed for one's estimable qualities is as such good for a person or as such a contributor to their well-being. It is not true, however, that these social phenomena are good for people regardless of such correspondence. It is not as such good for a student to be esteemed for having written a plagiarised paper, for a gang member to be esteemed for having committed serious crimes, or for a slave owner to have de facto authority over an enslaved person.
A second main way to think of advantage is in terms of resources or assets. We could say that social positions are higher to the extent that they provide their occupants with greater access to various resources or assets, and we can say this without claiming that access to a greater bundle of resources invariably better satisfies people's preferences or contributes to their objective good. If the relevant resources are, in turn, narrowly understood as material resources, this view is quite clearly mistaken. The problem is that a common downstream consequence of occupying a relatively high social position is being confused for a defining feature of social hierarchy. It is true, for example, that being a writer who occupies a high place in a hierarchy of prestige and esteem will likely lead to greater sales, better book deals, and lucrative speaker invitations. It is not hard to imagine, however, that one can occupy a high place in a hierarchy of prestige without such associated differences in command over material goods. We might here also think of religious authorities bound by a vow of poverty.
Such examples suggest that the view requires a more expansive conception of the relevant assets or resources. Grusky and Weisshaar develop such a conception, distinguishing eight distinct asset groups, including social, cultural, honorific, and power assets, with examples of such assets including various kinds of honour and prestige, and power and authority over other people (2014: 2-4). 22 On this view, our highly esteemed writer and impoverished religious authority receive large amounts of honorific assets, and correctly count as occupying a high social position on this basis. The problems for the resulting view start with the observation that the suggested asset groups are similar to the dimensions, elements, or aspects listed by the views discussed in the previous subsection. If the current view simply takes the various proposed dimensions of social hierarchy and labels them assets or resources, then it faces the same objections. The current view seems better positioned to answer some of these objections, however, because it does not need to say that social hierarchy is present whenever different social positions are attached to different amounts of power, authority, prestige, or honour (or yet other assets); instead, it can say that one social position is above another just when it is attached to an overall greater bundle of assets.
An initial problem with this proposal is that it would incorrectly count any inequality in overall assets as a social hierarchy. Grusky and Weisshaar include physical and economic assets (2014: 2-4), and so two people would stand in a hierarchical relationship to one another simply in virtue of having different levels of health or wealth, even if we assume all other assets to be equal. Suppose, however, that this issue can be resolved. The question would remain under which conditions one bundle of assets is overall greater than another. We have ways of answering this question when all the relevant assets are of an economic kind, such that their value can be measured in terms of their price. The range of assets we would need to recognise in order to adequately capture the many different kinds of social hierarchy must, however, be wider than such economic assets. The cool kid has access to some bundle of economic, honorific, social, cultural, and power assets, the wannabe kid to another; the same is true for Elizabeth and Lady de Bourgh, and for vassals and lords. According to which measure do we say that the latter have a greater overall bundle of assets than the former? A purely economic measure won't do, for the reasons given above. It is tempting to answer this question in terms of the extent to which different bundles of resources satisfy the preferences of those who possess them, or in terms of the extent to which they contribute to their objective good, but this would lead us back to the objections against welfare-based views.
A final suggestion would be that some asset-bundles provide individuals with greater social standing than other asset-bundles, and that people outrank others in a social hierarchy just when they have greater standing in that hierarchy. Of course, this suggestion would beg my question: I ask what it means for people to be above and below one another in a social hierarchy; the notion of differences in social standing would be another name for this phenomenon. My account does better on this point by constructing a notion of social standing out of socially required patterns of individual valuing-responses to persons. Of course, my account leaves many questions unanswered-I'm relying on a common understanding of what it is to value something, including valuing persons and valuing some persons more than others. Ultimately, these ideas need further explanation. I cannot rule out that some further explanation of what resources are, and of which resources there are and how they combine, might account for social hierarchy. I'm skeptical, however, that we have much to gain from thinking about social phenomena in terms of the distribution of resources when the notion of resources is disconnected from ideas like welfare, preference satisfaction, well-being, and price.

Power and domination
The third view holds that social hierarchy is a matter of asymmetrical power or domination. Work in the republican tradition, and especially the work of Pettit (1997Pettit ( , 2012, has done much to clarify the notion of domination. One could propose that social positions A and B are hierarchically ordered just when the occupants of A dominate the occupants of B. I should note at the outset that such a view is not proposed by Pettit, nor by any other republican I am aware of. One reason could be that domination is not present in all social hierarchies, for example, hierarchies that largely revolve around differences in esteem and prestige. We could still maintain, however, that an important class of social hierarchies are fundamentally asymmetrical relationships of domination. Even with this restriction in mind, problems remain and it is illustrative to see why. On Pettit's view, person B's choice between several options is dominated by another person A to the extent that A has the capacity to interfere in B's choice, where that capacity is not itself exercised on terms imposed by B. Understood in this way, the domination of one person's choice by another person is a ubiquitous phenomenon. One of Pettit's own examples of domination illustrates this rather well: 'there is only one copy of the newspaper available, you and I both wish to read it and, recognising the competition between us, you grab it first, thereby frustrating me ' (2012, 40). In any realistic social relationship between two individuals, both parties will dominate a range of choices of the other person.
To account for social hierarchy, it would seem that we need an account of one person dominating another person, in addition to an account of one person dominating another's choice. Pettit provides us with such an account. Social justice requires, he argues, that each citizen is guaranteed their status as a free person in relation to their fellow citizens. All citizens must be protected against domination from their fellow citizens to such an extent that they enjoy freedom as nondomination. What does it mean, on Pettit's view, to be a person free from domination? The answer is that a non-dominated person satisfies the eyeball test: They can look others in the eye without reason for the fear or deference that a power of interference might inspire; they can walk tall and assume public status, objective and subjective, of being equal in this regard with the best (2012: 84).
I take the language Pettit appeals to here-deference, walking tall, public status, and being equal with the best-to the the language of social equality. On this view, the domination of persons is explained in terms of social hierarchy, and not the other way around. Perhaps some of these issues can be avoided if we turn to power rather than domination. Power is a more general notion and more plausibly characterises all hierarchical social relationships, not just authority or dominance hierarchies. We could propose that A is "above" B in a social hierarchy just when A has greater power than B, or just when A has greater power over B than B has over A. 23 In any realistic social relationship between two individuals, hierarchical or not, each will have power to some extent and each will have certain kinds of power over the other person. This means that the proposal needs an answer to the kind of question Pettit answers with the eyeball test: under which conditions do differences in the power A and B have, or have over one another, place A and B in a hierarchical social relationship? We need to answer this question using a conception of power that is at once general enough to account for a wide range of different kinds of social hierarchy, but specific enough to provide an informative explanation of social hierarchy. The literature on power is very large, 24 but I do not know of a conception of power that can play this role.
My proposed view of the nature of social hierarchy accounts for the role of power in two main ways. First, various ways in which individuals dominate, control, and have authority and power over other individuals is central to many kinds of social hierarchy. My account thinks of such social hierarchies as sets of social norms that put certain individuals in positions of power over others. Patterns of deference, obedience, respect, and submission on the one hand, and dominance, command, and aggression on the other, are normatively expected of individuals in a relevant social group. As I indicated, such patterns of attitude and behaviour can plausibly be seen as patterns of valuing, allowing the view to account for these kinds of authority and dominance hierarchies, but also for social hierarchies in which individuals having power over other individuals is less central. Second, since social hierarchies are constituted by sets of operative social norms, on my account, social hierarchy can itself be seen as an exercise of power by a group over its individual members. Hayward criticises the sociological literature on power for focussing exclusively on the intentional control of one agent by another, proposing to reconceptualise power as 'the network of social boundaries that delimits, for all, fields of possible action ' (2000: 27). Given a conception of power along these lines, social hierarchies would count as structures of power on my account, with the relevant "boundaries" and "possible actions" understood specifically in terms of operative social norms.

CONCLUSION
Over the course of my discussion, I have made reference to a wide range of social structures and social relationships. I have argued that when we describe these as examples of social hierarchy, we say something both substantial and true. This argument is not meant to challenge or undercut the important work that has been and continues to be done to investigate distinct kinds of social hierarchy. Instead, one aim of developing a general account of social hierarchy is to facilitate communication between different academic disciplines examining social hierarchies across different times and places. With regard to philosophy specifically, I expect that my account can provide a more solid conceptual foundation for debates about the normative assessment of social hierarchy and social equality-debates we see in the literature on relational egalitarianism and elsewherewithout prejudging these normative questions. Relatedly, there is an important current that runs through modern social and political philosophy according to which the desire to acquire a kind of standing in the eyes of others-a desire for respect, esteem, status, or glory-is a key element of the moral psychology of citizens. My hope is that my account of social hierarchy can help us understand the nature of the motivation that is being ascribed to citizens in this tradition, so that we may better understand how a political society is to respond to these desires.

N O T E S
1 The concept of a social position or location is widely used in sociology, as well as in anthropology and political science. See Wasserman and Faust (1994) for a helpful account. 2 In this article, I will largely avoid the language of structure. The main reason for this is that the concept of structure is one of the most contested concepts in the social sciences-employment of the concept of structure is Scheffler, Seidman holds that to value an object, we must see the object as valuable to ourselves (2009: 288-91), and on Scanlon's view we must take ourselves to have reasons for our response to the object given by various properties of the object (1998: 95). 17 The previous two paragraphs only touch upon a few issues in an area of much greater complexity. Ultimately, my view needs a more systematic account of what it is to value people, and to value one person more than another. In this article, I rely on our standard conception of valuing, which means that I must leave some difficulties unaddressed. 18 Pecking orders among hens were first described by Schjelderup-Ebbe (1922), see Chase and Lindquist (2009) for an overview of the dominance literature. 19 Not all views distinguish these four phenomena-some take honour to be a kind of prestige, for example, or authority a kind of dominance-and some views include additional elements. 20 On this view, social hierarchy is a matter of the relative advantageousness of occupying various social positions, not a matter of the relative advantage of the occupants. It is possible that a person who occupies a high social position is disadvantaged in various ways that are unrelated to her social position, so that she is overall less advantaged than other persons who occupy a lower social position. It is also possible for one person to occupy a high, and thus advantaged, social position in one social context, and a low social position in another. 21 On this type of view, the list does not just include desirable experiences and preference or desire-satisfaction (Crisp 2017). I do not separately discuss hedonistic views in this article because the kinds of arguments used against views of social hierarchy based on preference-satisfaction can also be employed against hedonistic views. 22 Grusky and Weisshaar distinguish these asset groups in the process of providing a definition of social stratification. The relationships between the idea of social stratification and social hierarchy are complex, but it is fair to say that the concept of stratification is generally applied to a wider range of social phenomena than are the subject of this article. The attempt to account for social hierarchy in terms of distinct asset groups sketched in this paragraph is therefore mine, not theirs. 23 As before, to avoid issues of proliferation these proposals could incorporate the first part of my account, holding that social hierarchies are sets of social norms granting A greater power than, or over, B. 24 Haugaard (2002) provides a useful overview of the "power debate" in sociology.

A C K N O W L E D G E M E N T S
Previous versions of this article were presented at McGill University, Rijksuniversiteit Groningen, Stanford University, Tel Aviv University, Temple University, University College London, University of Sussex, and University of Warwick. My thanks to audience members and organisers at these institutions. I am also grateful to many individuals for their suggestions, comments, questions, and objections. With apologies to those I will inevitably forget to mention, some names that come to mind are Showkat Ali, Euan Allison, Aberdeen Berry,