What is Inequality?

I’ve made a post on the question of what equality is, but what is its opposite? The obvious answer would be: the opposite of everything in the equality post. But there are more nuances to it than that.

To me there seem to be a two primary ways we can think about inequality:

  1. Vertical Inequality (Hierarchies)
  2. Lateral Inequality (Diversity)

These two ways of breaking down inequality can also be categorized as either inequality of opportunity and inequality of outcome, each of which can then be broken down into natural inequalities and artificial inequalities (look at that, I’m making a hierarchy).

Opportunity and Outcome inequalities could almost be thought of like functions that take in the natural and artificial inequalities as inputs and output a given situation. Say we call Opportunity inequalities the function Pn(x) and Outcome inequalities the function Tn(x) for person n. What, then, is x? It would contain things like

Natural Inequalities:

  • IQ/social/emotional intelligence
  • Height, facial symmetry, general attractiveness
  • Race/sex/gender/orientation
  • Genetic variations in preferences/proclivities/dispositions and genetic ailments

Artificial Inequalities:

  • Upbringing (parenting style with which one is brought up; parent’s socioeconomic status; diet; religion; values; education)
  • Environmental (physical environment; cultural environment; technological environment)
  • Structural (political; economic; laws and the justice system)

There could be more, and certainly there is crossover (e.g., one’s score on either Pn(x) and Tn(x) are not completely independent, nor are all the inputs listed above independent of each other), but the idea is that we can conceptualize one’s inequalities of opportunity and of outcome as functions that, say, take in x (which is the above bulleted lists) and outputs a sort of score (say, between 1 and 100) that ranks a person’s place within the inequality. A person with a score of, say, Pn(x) = 1 would be someone who has had everything stacked against them opportunity-wise. We might think of such an unfortunate person as being a child born diseased and in abject poverty in a war-torn country. Then someone who has a score of Pn(x) = 100 is someone who has every opportunity imaginable open to them.

We could thus say that the closer people within some community or society fall within that ranking, the less inequality there is. Indeed, we could take an average and find the standard deviation, which would be a measure of the inequality in that community/society:

We should then add the further constraint that Pn(x) and Tn(x) can be different for whether we are discussing vertical or horizontal inequalities. We might then instead have Pn(x, V) and Tn(x, V) or Pn(x, H) and Tn(x, H) depending on which one we are talking about. However, I think it is probably easier to conceptualize this in terms of vertical (Hierarchical) inequalities than horizontal (Diversity) types of inequalities. For instance, a very diverse society can be one in which the standard deviation (in either Pn(x, H) or Tn(x, H) or both) is very small, but a very hierarchical society is one that is unlikely to have a small standard deviation in inequality scores (although one might conceive of a sort of feudal society where a very tiny fraction of the population are fabulously wealthy while the vast, vast majority are destitute peasants, which would have a smaller standard deviation while still being quite hierarchical).

It also needs to be noted that having a small standard deviation is not sufficient for a just, functioning, and/or flourishing society. A society could have a very small standard deviation, but only because most or all the people in that society fall within some range between 10 and 20, while another could have an equally small standard deviation while falling between 80 and 90. Likewise, a society might have a larger standard deviation, but that is because they fall between 40 and 90 compared to a society with a small standard deviation while falling between 40 and 50 (i.e., both societies have the same low, but one has a higher high). The point being, while having a smaller standard deviation might turn out to be a necessary condition for an optimal society (in terms of a society that promotes human flourishing in all aspects, not just something like economic success), it is almost certainly not a sufficient condition for an optimal society.

So far I’ve only discussed inequality in terms of extant humans. Taking animal rights into account, or future people into account, could potentially alter this calculus dramatically. For the purposes of this post, unless stated otherwise, I will primarily be taking extant humans (and very near-future humans) into account. I’ll touch on the two main kinds (Hierarchy and Diversity) and point out where the other breakdowns occur as I go.

Hierarchies (Vertical Inequality)

A hierarchy is a means of organizing a system such that some subset(s) of the system are subsumed under or classified lower than other subset(s), usually based on some set of criteria. In this there is a ranking in which some subset(s) are “above” some other subset(s). Often this is applied to people, such as in the government, a business, a military, a club, a team, a family, and so on. The chief executive ranks above their deputies or lieutenants, who rank above some other mid-level persons, and so on down to the lowest level. Humans and our fellow primates tend to think hierarchically and organize ourselves in a hierarchical manner.

Language itself is hierarchical. Propositions are formulated in terms of subsuming the subject under the predicate: S is P. In this, we are saying that P is some more general thing or concept under which S is a more specific or particular instance. It could then be that P is also subsumed under something else, with that something else being subsumed again under yet something else, and so on up the hierarchy. This is easily observed in biological taxonomy:

We also observe this in our love of ranked lists of things:

Tell me you’ve never watched a video or read an article where someone used the tier list

We humans seem hardwired for hierarchies. As discussed, our language and thought is hierarchical. But also, our relatives in the animal kingdom are hierarchical; pre-verbal infants recognize dominance hierarchies; humans can perceive dominance expressions and postures in subliminal images.

Humans are often most concerned about hierarchies within human organizations. These social structures are so important to us that our position in them has affects on health and well-being. What’s interesting is that, according to studies on baboon hierarchies, from second rank on down the amount of stress a male baboon suffers increases as the individual’s place in the hierarchy goes down. But, it seems, being on top is also quite stressful:

Glucocorticoids are an important stress hormone and indicate stress levels. (Source)

Yet people tend to have differing opinions on the ethical status of hierarchies. Opinions that tend to track with political identity, where those on the political left (at least ostensibly) dislike hierarchies, while those on the political right favor hierarchies. The former tend to see hierarchies as innately oppressive while the latter view hierarchies as the key to maintaining law and order. This seems to parallel one’s moral foundations:

Where in the above “liberal” is meant in the U.S. sense of being left-leaning. What the data show is that the political left strongly endorse the moral foundations of fairness/cheating and care/harm while more weakly endorsing authority/subversion and loyalty/betrayal (ingroup). Those who endorse the former (fairness and care/harm) would be more concerned with an equitable distribution of resources and power, while those who endorse the latter (authority and ingroup) would likely favor hierarchies (the authority-subordinate and ingroup-outgroup dichotomies being a way to rank people and groups of people).  A measure of one’s support for social hierarchy and the desire for their ingroup be superior to outgroups is a person’s Social Dominance Orientation (SDO).

It’s here where we can bring in our notions of inequality of opportunity and inequality of outcome (equity).

In the vertical sense, inequality of opportunity would mean that people higher up in the hierarchy have more opportunities open to them than people lower down in the hierarchy. This could be due to our natural or our artificial inequalities. For instance, a natural inequality might be that someone who is physically attractive is likely to have more opportunity open to them that someone who is physically unattractive.  An artificial inequality might be that someone with an abusive parent would has reduced opportunities simply because they are not given the right skill set to compete. Or, someone brought up in a totalitarian state (e.g., North Korea) is going to have many opportunities denied them compared to someone brought up in a more open and free society.

We can also think of vertical inequality at a society-wide scale. Wealth inequality is probably the first to come to mind. Within a society, there is an obvious inequality between person A who makes $30,000 USD a year, person B who makes $130,000 USD a year, person C who makes $300,000 USD a year,and person D who makes $3,000,000 USD a year. Person A likely has very little political clout, very few opportunities for novel experiences, and limited options for living conditions (i.e., the domiciles which they can afford, where they actually live, what kinds of things they own, what kinds of food they eat, and so on). Meanwhile, person D wants for nothing, with even expensive items like vehicles and vacations and expensive foods being tantamount to free to them (they won’t have to save up for months or even years to buy a car or go on a vacation, but can just do it with little or no thought for their finances). Person D can also likely exercise much more political influence as well. Indeed, it may even be the case that person D employs person A, giving person D direct power and influence over person A.

Wealth inequality is an artificial kind of inequality. We might say that a natural inequality at the societal level is something like intelligence, however one wants to define it. Intelligence is a predictor of success, as are other traits not of our own choosing, such as personality (and even extremely non-meritocratic traits such as one’s height). The point being, it is not just the success or wealth of the situation into which one is born that predicts a person’s success as an adult, but there are less mutable predictors of success at play. Ideally, meritocracy is supposed to sort individuals into levels of hierarchy commensurate with their abilities (as well as motivation and willingness to do the work) rather than by accidents of birth.

Of course, what a lot of critics of hierarchies (and meritocracy) contend is that what we are defining as “success” is determined by these traits. In other words, we see people with higher intelligence or possessing certain personality traits succeed because our social, political, and economic structures are so constituted as to reward these traits and punish those who reject or deviate from them. Something like intelligence is simply defined as the ability to navigate the current social milieu; it is not some natural hierarchy wherein those with more intelligence are innately “better” than those lacking it. Additionally, these traits are valued due to the historical contingencies that found white, heterosexual, cisgender men from Europe dominate and oppress other cultures. Thus, a more equitable society, in this thinking, would be one with a culture in which people possessing certain traits would not be rewarded by the mere virtue that they possess the traits unjustly defined as successful or good.

While this criticism may be true, at least to some extent, those who are proponents of hierarchies might retort that this is not a criticism of hierarchies in general. Hierarchies, it may be argued, are unavoidable, and if we remove the current hierarchies around which society is organized they will simply be replaced with another that may be more or less similar to what we already have, and is in all likelihood to be worse. Few, if any, societies have ever made it work to have no hierarchies. Especially a society with more than just a few hundred people, i.e., to have a large-scale, even global society function will require some sort of hierarchies, and the ones we inhabit now have a proven track-record of lifting more people from poverty, reducing violence, disease, and hunger, and extending rights to the greatest number. Additionally, even without permanent hierarchical structures, there will always be temporary ones, such as when people go to someone for their expertise on a matter, thereby temporarily subordinating themselves to the person’s superior knowledge or skill on the matter. In other words, to have a society free of hierarchies, all people would need to possess all the relevant knowledge on all the relevant matters; absent this, whether temporary, long-term, or permanent, some people possessing the relevant knowledge or skill on a matter will need to take charge.

We therefore have (at least) two axes on which to plot hierarchies: inevitability and desirability. The former is a measure of how likely it is that the certain hierarchy will come about, even if it is not imposed or consciously constructed. The latter is a measure of how functional, useful, or perhaps even moral, a hierarchy is. The two axes are, of course, not mutually independent, as we might say that a hierarchy is more likely to arise because it confers a great amount of utility. It is also difficult to say what criteria ought to adjudicate inevitability or desirability: is a hierarchy inevitable because it needs to exist (perhaps for reasons of functionality), or is it because the kinds of people who attain power will always impose it for their own benefit (regardless of who ends up in power)? Is a hierarchy desirable because it functions most efficiently? Or because it functions to more people’s benefit? Because certain hierarchies are just intrinsically good, even if they ultimately lead to worse outcomes?

Diversity (Horizontal Inequality)

No two human individuals are exactly alike. Even identical twins are not interchangeable. There is a great variety of different people, with different personalities, desires, interests, dispositions, proclivities, idiosyncrasies, hopes, dreams, regrets, and so on. Although the space of possible humans contains clusters, of what we would call cultures, no two people completely overlap within this space.

Thus, in this horizontal sense we would say that no two humans are equal. While our diversity is not free of judgements in the vertical sense (some individuals and cultures are deemed by other individuals and cultures as better or worse than others in certain regards), as a description it seems uncontroversial to say that no one individual is exactly equal to another in terms of personality. Of course, in western society we say that all people are equal, but this is meant in moral and legal terms. What I mean here, with horizontal inequality, is that no two humans are fungible or interchangeable, each one bringing their own uniqueness to every interaction.

But to say that everyone is unique is practically a cliche. When people talk about things like diversity, what they are usually interested in are those clusters of people. Race, sex, gender, orientation, nationality, ethnicity, culture, religion, socioeconomic status, political affiliation, disabilities, various subcultures, education, occupation, and so on. Yet the criteria for clustering is often in dispute. For instance, should natal women be considered a cluster, or should transgender women be included in this cluster? Does intersectionality entail that black women are a separate cluster from black men? Should clusters of a certain variety (or distance from each other in the space of all possible humans) suggest a need for segregating or integrating certain clusters? Is there a vertical (Hierarchical) inequality implied by certain clusters? Should cluster averages justify differential treatments of individuals within the cluster? Should past injustices against certain clusters be remediated through special treatments? Should we maintain a policy of tolerance toward all clusters, or are some more worthy of respect than others? And where and how do we even draw the lines demarcating these clusters (e.g., necessary and sufficient conditions, homeostatic property clusters, natural kinds, prototypes, etc.)?

All these questions can be categorized under two basic questions: what are the clusters? And what ought to be the clusters (if any of them are worth having at all)? The first is descriptive while the second is normative. One of the problems with the first one is, as the links at the end of the previous paragraph indicate, categorizing even fairly stable and well-characterized things in the world can be difficult and not without disagreement. Trying to do this with human beings is even more fraught. Especially since, in many such cases, we end up having to speak in generalities, using averages with fairly large standard deviations. Things get even messier when humans seemed hardwired to use heuristic thinking, make oversimplifications, issue snap value judgements, and subconsciously arrange things hierarchically. In other words, to partition humans into two or more groups, a person will immediately identify with whatever partition in which they find themselves and then make heuristic and biased judgements about others (e.g., Jane Elliott’s famous eye-color experiments).

This inability to discern the margins of these clusters empirically (or a priori) then leads into the second question: what should be the clusters (and, as a corollary, what importance should we give them)? If such demarcations are not to be discovered, then they will have to be engineered (whether consciously or unconsciously). As I said above, it is practically a cliche that all individuals are unique, and it is largely uncontroversial that clustering exists (e.g., cultures). But we are left doing a sort of regression analysis to attempt to determine what the clusters are. Unable to make the distinctions purely by empirical criteria, we are left with the responsibility that the borders be drawn based on other criteria (e.g., normative criteria, pragmatic criteria, etc.).

We can try to think about this by seeing what lies between the extremes. On one extreme would be the notion that all clustering be completely ignored, or at least that it be resolved to such a degree that each individual becomes a cluster unto themselves. I’ll call this anarchism, which says that humans should be anti-nationalistic, anarchistic, completely free to decide our own values, religion, sex, gender, race, and so on. This obviously runs into problems, such as: what if the way one (or more) person(s) decide they identify with some organized religion, state, ethnic group, race, or whatever? At the anarchic pole, this would not be forbidden, but it would thereby refute the anarchic position. And what about the fact that some things are just impossible not to notice about other people, and that clusters of people have some trait in common while others do not have it? Not to mention that humans do not possess doxastic voluntarism. Put simply: this kind of anarchy just is not going to happen (even if we did assume it was desirable).

At another extreme would be that the clusters ought to be incredibly important and well-defined. I’ll call this segregationism. This leads to notions like cultural appropriation and standpoint epistemology, where cultural ques should not be displayed (or perhaps even acknowledged) by out-group people, and where a person’s “way of knowing” is largely determined by which cluster(s) they inhabit. This is also where things like race separatism arises, i.e., the notion that the races ought not mix. Despite the Enlightenment project attempting to dismantle this way of thinking, extremes on both the left and the right have begun to take up these ideas again. It is not likely to lead to anything good, and as such, even if we assumed it was workable (which it almost certainly would not be in practice), it would not be desirable.

Another extreme would be to try reducing the resolution on the clusters as much as possible and attempt to make a sort of single, all-inclusive group that covers all of humankind. I’ll call this universalism. If this sounds like rainbows and unicorns, that’s because it is. It runs into some of the same issues as the anarchic approach: many people just are not going to buy into it, and again it is difficult not to notice differences in people that are shared among one subset of the population but not another. In other words, even if there is no empirical criteria to make uncontroversial adjudications concerning clusters, it is clear to the vast majority of people that such clusters exist at some resolution.

The fourth extreme would be to make the clusters as fluid as possible, i.e., to say they exist, but that they are always and forever open to revision, integrating, re-partitioning, and rearranging. I’ll call this fluidism. While in some sense this is already the case (which is why it is difficult to reach agreements about the clusters, and why they tend to evolve over time), if we take it to its extreme it causes a sort of breakdown that both makes cluster identification unhelpful (because they are so mercurial and transient), and in the end just brings things back to a sort of anarchism (each individual being allowed to arbitrarily draw the boundaries wherever they wish is as useful as having no boundaries at all).

In this model we therefore have a two-axis space of approaches to mapping clusters of people based on horizontal inequality. Things like critical race theory and race separatism lie somewhere in the upper left quadrant; some forms of new-ageism and communism lie somewhere in the upper right quadrant; certain forms of anarchism and libertarianism lie somewhere in the lower left quadrant; postmodernism and queer theory both lie somewhere in the bottom right quadrant. Less extreme approaches to mapping the clusters will lie in the plane somewhere closer to the origin.

My own position is probably somewhere not far below the horizontal axis on the border between anarchism and fluidism (slightly blurry boundaries, intermediate resolution). I think it clearly just is the case that most boundaries allow for some amounts of wiggle room, which is why I tend toward the blurred boundary part of the vertical axis. I’m generally for individual liberty and autonomy, and so the well-defined boundaries do not really appeal to me. But, I’m not radical enough in my individualism to side too strongly with anarchism; yet I also think that some resolution in cluster demarcation is both inevitable and practical, and so I cannot side too strongly with fluidism.

Concluding Remarks

Inequality is a hyperobject. Both vertical (Hierarchical) and horizontal (Diversity) forms of inequality are not likely going away anytime soon. Even if we made an effort to flatten the hierarchy or narrow the diversity, such a program is almost certainly doomed to fail. But more than that, it would not be desirable. Fascism sought to narrow horizontal inequality and communism sought to flatten hierarchical inequality. Both have yielded catastrophic results.

Inequalities are not categorically bad (or good). The trick is to determine which ones are inevitable so as to find way of working with and around them, and to determine which ones are good in order to promote them. The catch-22 is that, because of our (horizontal) inequality, it is unlikely that most people will ever agree about the inevitability/desirability of most inequalities. Just another source of my unrelenting pessimism.