Two Concepts of Representation

The idea of representation has been enjoying a boom of late. Of course, our modern democracies are representative ones, and the idea that our governments and parliaments should be representative has been around for a while. Increasingly, though, other groups of people are expected to be representative too, from teaching staff at universities to the boards of corporations. There’s even a feeling that the fictional worlds presented to us in books and films should be representative of the real world that we inhabit.

These two concepts of representation are actually quite different, however. The first has been used in formal political institutions since ancient times in an effort to make sure that it really is the people—or a good sample of the people—that is exercising political power. The second is more focused on the idea that specific groups—not just political institutions, but also companies, clubs and other organizations—should reflect the demographic make-up of society as a whole. Only the first conception is truly democratic; the second has taken on some of the prestige and legitimacy of democracy, but has very little to do with it.

The first conception of representation is grounded in the core idea of democracy: that the people (dēmos) should exercise power (kratos). What exactly this means is, of course, open to interpretation: there are a range of different democratic institutions and procedures, each with a fair claim to embody the power of the people. Whether they incorporate referenda, proportional representation or other measures, all these institutions and procedures are designed to create as accurate a picture as possible of what people actually want.

In this, they resemble a particular kind of applied statistics. Now, an essential part of arriving at sound statistical conclusions about a population is obtaining a sample that is truly representative. Just as a good social scientist strives to ensure that her sample will allow her to drawn reliable inferences about the population she’s studying, so a good political system will produce a more or less accurate picture of the balance of opinion in society as a whole.

There are various ways of doing this, and, through the course of history, more than a few have been tried. The citizens of ancient Greek cities were often divided into subdivisions, which were then used as a basis for selection to central political institutions. In Athens, for example, the citizens were subdivided into ten tribes; each year, fifty men randomly selected from each of the tribes would make up the city’s Council of Five Hundred, one of the democracy’s central institutions. The men selected wouldn’t necessarily be expected to speak on behalf of their tribe, but they did provide it with a presence in one of the city’s main organs of state. The fact that they only served one-year terms ensured that a fair number of the members of each tribe would, at some point, have a seat on the council, and take part in its proceedings. These tribes were artificial subdivisions of the citizen body. They served, essentially, as a kind of built-in mechanism to ensure that members of the central council came from all across the Athenian citizenry. It was a low-cost but effective way of coming some way towards a representative sample of the Athenian dēmos.

Nowadays, of course, we do representation differently. In our much larger states, significant numbers of citizens cannot turn up to central institutions in person. Instead, we have them elect representatives to speak on their behalf in a central parliament or congress. There’s a long-running controversy about what exactly this should entail: eighteenth-century statesman and writer Edmund Burke, for example, argues that MPs should represent their electors’ best interests, not necessarily what the electors think are their best interests. But this is a pretty rare view nowadays: most people insist that their representatives make some effort to reflect the beliefs and preferences of their constituents. How do representatives get an accurate sense of the opinions of the people who have elected them? This, of course, has long been part of the craft of politics. Electoral success (or failure) can give politicians an idea of what the people want. So can being deeply embedded in the norms and networks of particular communities. Over the years, the craft of finding out what constituents want has become a kind of science, with teams of pollsters on standby to measure voters’ feelings on any number of issues.

The ancient Athenian system and our modern ones, then, can be seen as different approaches to the problem of how to make sure that the preferences of the people as a whole are communicated to the major institutions of the state. But why is it so important that these institutions are representative? First and foremost, because major state institutions wield the considerable powers of the state. They have, as Weber puts it, a monopoly on the legitimate use of violence; they also make the decisions that will shape the future of the communities we live in, for good or ill. But these bodies are also supposed to be democratic institutions, and one of their major functions is to manifest the will of the nation in microcosm. In fact, their function as embodied opinion polls and their function as organs of the state go hand in hand, since, in a democratic order, it’s primarily those decisions that reflect the will of the whole people that are seen as legitimate.

In short, then, it’s important for democratic political institutions to be representative because they’re democratic political institutions: their purpose is to exercise the specific and extraordinary prerogatives of state power, in a democratic way. This brings us to the non-political realm and to our second concept of representation. I’ll call this demographic mirroring, since the ideal seems to be that any particular group of people—from the attendees at an academic conference to the characters in a film—should approximate the demographic proportions of society at large as closely as possible.

People are found together in lots of different agglomerations and very few of these, if you take them separately, display exactly the same distribution of attributes that you’d find in the population at large. How could they? Human beings don’t acquire their various attributes randomly: they are the result of a combination of genetics and of the things that have influenced them during their lives. Neither of those things will be exactly the same irrespective of where you happen to be—which means that there will be many differences between groups in different parts of the world or even within a single country. However, whereas democratic political institutions have an obligation to be representative of the balance of opinion in society, there’s no reason for other groups to bear the same burden.

It also makes little sense to impose such an obligation on fictional groups of people, such as the characters in a film. Many films are set in the distant past, in places that were significantly less ethnically diverse than many modern nations. Some novels and films even depict groups of people who have never existed, but might at some time in the future. Should they have to reflect our current demographic balance too? Parliaments should be representative in some sense, but why should the cast of characters in a work of fiction? They don’t even exist.

This second concept of representation is also sometimes applied to formal political institutions. For example, Justin Trudeau’s first cabinet was designed to “look like Canada” in that it reflected the country’s demographic balance. That, of course, was meant to make it representative. But in what sense? Since this was a formal political institution, it would seem that our first concept of representation must have been somewhere at the back of people’s minds. It was a good thing that the members of Trudeau’s cabinet were of different ethnicities (the assumption seemed to be) because they’d be able to speak on behalf of the groups they represented.

But the idea that an ethnically Indian politician would be able to speak on behalf of other Canadians who happen to be ethnically Indian is a curious one. The ethnicity with which you identify may correlate with some things (for example, which religion you espouse) to some extent, but it is not a good way of predicting your views on a range of discrete political issues in any detail. Ethnicity might be a proxy for certain attitudes, but even then it’s a very rough one, and if we really want to find out people’s political views there are ways of doing that: face-to-face meetings, polls, elections, etc. The democratic procedures we have are already designed to make sure that our representatives have some knowledge of what our views are, and so can re-present them in our central democratic institutions. We don’t need to have recourse to crude guesswork based on what people look like. In political institutions, then, though a demand for representation is clearly more legitimate than it is in a rugby club or a TV series, it can be focused on the wrong thing—on race or sex rather than on people’s political opinions and preferences. The point of democratic political processes, after all, is to get a more or less accurate sense of the distribution of views, not of the distribution of skin colour or other physical characteristics.

So why is representation invoked so often, even in instances where its focus might be wrong, or where it obviously has no place? Maybe people are simply confused about where the concept should apply, and where it shouldn’t. But it’s also possible that the concept of representation has been co-opted because of its prestige. Saying a film or a university department needs to be more representative has a legalistic, almost constitutional ring to it: the claim that a film should have more Canadians in it (say) is suddenly elevated to the same level as the idea that a parliament should reflect the views of the population at large.

In reality, though, a wide gulf separates the first concept of representation, which aims to ensure that the spread of a nation’s views is reproduced in its most powerful decision-making bodies, and the second, which reflects a desire to have a country’s demographic make-up echoed by a whole range of smaller groups. There may be other considerations in favour of that desire than the ones I’ve looked at here. But to describe it as a striving for representation just looks like an attempt to draw on that concept’s connotations of democratic legitimacy for purposes that have very little to do with democracy.

If you enjoy our articles, be a part of our growth and help us produce more writing for you:


  1. Thanks for reading, Nathaniel. I’m not sure what you meant by what you say about trading off two interests against each other. Can you say more about that?
    Could you also say more about what lessons you think your personal experiences have produced, and how they might affect what I’ve written in this article?
    In your final sentence, you seem to be saying that women and ethnic minorities should be represented in order to represent the interests of those groups (women and the various ethnic minorities). But 1) ethnicity isn’t an entirely stable thing. The same person might define himself differently over time, or depending on what groups he’s living in (there are studies that suggest that’s indeed the case). For example, someone might identify as Latino in one neighbourhood or time period and as black in another. That means it’s hard to say which individuals a Latino or black representative may be representing at any particular time. A good political system will simply table the question of what ethnicity people identify as and just say, ‘Whoever supports X and Y policies, or Z vision for the country’s future, can get behind this candidate as their representative, no matter what ethnicity they identify as.’ This brings us to 2) I think we should be careful about the idea that someone of the same ethnicity (whatever that is) will necessarily represent my needs and political views. At the base of this there’s an assumption that all Asian people (say) will think more or less the same. But there are many contexts in which we find that assumption pretty objectionable. For example, if someone said at a party, ‘You’re Asian, so you’re bound to support more money for STEM subjects!’ wouldn’t you find that strange? If you would, why don’t you find it strange when the same assumption’s made in the context of representation? Now, it may be that some assumptions we make about people are largely correct (e.g. Canadians are more likely to like watching hockey). But there are always lots of exceptions, and, with the political system we have, we don’t need to make any assumptions of this sort. We can, again, just let people be represented by the politicians that best support their views on a range of issues. That may turn out to correlate with various markers of identity to some extent, but that’s not something we want to assume at the start of the process. Focussing on issues rather than identities actually protects the desires and preferences of atypical people: the few Canadians who don’t like hockey can vote for someone who also doesn’t like hockey, and the non-Canadians who do like hockey can all rally behind a hockey-loving candidate.

    1. “I’m not sure what you meant by what you say about trading off two interests against each other. Can you say more about that?”
      Take the hobby lobby supreme court case in the US. Providing birth control for women is an important thing. Not making people do things that violates their religion is an important thing. How important are each of those? That’s a super subjective question and one that I don’t even feel like the law tries to make less subjective. The law, the way it was written in that case, was that in order to violate religious freedom, there must be a “compelling state interest”. People can and do disagree about what that means. I think that if it affects you or people you identify with (which there are lots of ways to identify with someone, women and ethnic minority groups are just my lazy examples), you rate it as more important.

      I think your points in your comment all make sense, I just don’t think they map on to the real world in a useful way. Part of this is because a person is so many different things in so many different contexts that it’s hard to say, this person shares my views so I want them to represent me. They could share your views on wealth redistribution and climate policy but be very different on religious freedom or sexual harassment.

      And what happens if an issue comes up in the middle of a term that wasn’t even talked about during a campaign? The court case about discrimination in acceptance against Asians at Harvard could be an example of that soon. How important is the unfairness of Asians getting passed up for lower scoring students versus the university’s desire to have a diverse student body? That’s another very subjective question. And hypothetically, even if someone identifies as Filipino versus broadly Asian, if it becomes clear that an antagonist groups them together when making their decisions, that could make them identify with the bigger group since they are experiencing the same problem. I think that having a representative amount of Asians in a governing body helps with this problem.

      Clearly, there is no great solution if the group you need representation for is not even formed when you pick your representatives, but in today’s world there are reliably enough issues that will come up that fall along gender and ethnic group lines that it’s worth it to use that metric.

      1. Thanks – I don’t doubt that the questions you raise are real issues. Politics is often complicated. But this is exactly why, in the current system, people decide who represents them based on who they think should represent them, not based on top-down judgments (from the PM, say) about what the right balance of physical characteristic should be (and about what the relevant physical characteristics are). That doesn’t mean identity can’t come into people’s voting decisions. If you think someone with the same skin colour as you really will represent your views better than anyone else because of that fact, there’s nothing stopping you voting for them for that reason. But the current system lets every voter make that choice for themselves. It lets them weigh up identity and policy, and even consider how good a proxy identity might be for policy, and then vote according to what they’ve decided. The problem with pushing a political body towards demographic mirroring from above is that it might effectively pre-empt that democratic process (or deny it its proper weight). Voters are basically told, ‘Your decisions, when counted, may not have decided that race is that important, but we’re going to fix that for you!’ I find that idea of insisting on racial identity as a factor in making up a political body, even if the people themselves have rejected it, puzzling for a number of reasons. Most are liberal democratic (people should be free to choose; their votes should be fully respected); but one is humanist and anti-racist: why divide people according to race when it isn’t necessary, and hasn’t even been requested?

        Granted, that’s not quite what Mr. Trudeau did with his cabinet, but there did seem to be an assumption that (say) a French Canadian will better ‘represent’ another French Canadian than an English one, and vice-versa. But I don’t know if all French Canadians have the same views as Maxime Bernier (for example), or if no English Canadians do share his views. You’re right to some extent that there are some issues that fall along gender and ethnic issues (though there are also lots that don’t). But ethnicity only predicts policy to a certain extent. It’s hard to think of an issue that absolutely everyone in an ethnic group feels the same about, and that means there are always going to be a few people of a given ethnicity who don’t feel represented by a leader chosen with ethnicity in mind. So you could go with ethnic representation as a proxy, since ethnicity might well predict 80% or something of people’s views on a given issue. And that might not sound that terrible. The thing is, though, that we’ve got a system that will do much better than that, which involves doing away with the proxy and just letting people choose whichever candidate they want, regardless of their race. That system – focusing on representation of people’s own preferences, not the elite’s current obsession with demographic mirroring – that system also has the added advantage of not classifying people on the basis of unchangeable characteristics, something that we know people can find inappropriate or demeaning.

  2. There’s no absolutes here. One problem with political representation is that if there’s too much focus on ‘representing’ people’s view rather than just them and their interests it easily tips over in a kind of majoritarianism that can run roughshod over minorities. It’s obviously something Madison identified. Burke too was right at least in part right on the role of representatives.
    The flip side is a kind of elite capture that leads to disillusionment with the process by significant numbers of people. There’s no simple answer to the balance between these two things.
    Same goes for repsentation by demographic. Clearly diversity is important, but as was noted you can’t assume some ‘represents’ some identity group simply by having that label attached to them.
    Again there’s no simple formula for how that should work. Quotas are by their nature reductive as they assume someone is representative by virtue of sharing certain immutable characteristics. That’s not to saying aiming for something close to gender parity or ensuring ethnic groups have close to proportional representation can’t be a laudable aim, it just can’t be seen as the only way of looking at representation.

    1. Thanks Andrew. Why do you think that focussing on representing people’s views rather than ‘them and their interests’ will tip over into majoritarian tyranny, whereas representing people’s interests won’t? I mean, it strikes me that there have been examples of tyranny in which leaders who claimed to intuit the needs of the majority, or even to have a sense of the ‘people’s soul’ or some such, did great harm to minorities.
      Also, where do you think Madison touched on that particular issue? (I find Madison an interesting thinker, so I’m interested.)
      What I was trying to do in this piece was make a pretty narrow point: that the word ‘representation’ is often used in support of what I call demographic mirroring, in a way which makes it sound like the kind of representation that we rightly want as a feature of our political institutions. But it’s actually a different concept of representation, either because a) it’s aimed at groups that don’t have anything to do with politics (sports clubs, fictional characters) or b) it focusses on the wrong thing (race or sex rather than viewpoints or preferences).
      As I try to make clear at the end, that’s not to say that there are no other arguments that might be made in favour of demographic mirroring in some groups. Certainly, I don’t deal with all of the possible arguments for that here. All I’ve really argued here that where there’s an implication that ‘representation’ has some of the legitimacy of democratic representation, that may well not be the case.

  3. I’m disagreeing with this one pretty hard. The idea that sex or skin color doesn’t need to be representative as long as you have representative views is fine in theory but when you have a lot of decisions that are trading off 2 interests against each other, people act the selfishly. Not even in a bad way necessarily, but when you are engaging with a problem in near mode (this problem affects me and my family on a day to day basis), issues are completely different than if you just think about them abstractly. That makes representation of women and ethnic minorities very important to representing the interests of a group.


Leave a Reply