This post is part of a series on identity on the modern Left:

Critique of Norm Criticism

2022-01-29 • 16 min read

My own instinct is that in the feminist utopia, there will be no gender, or there will be so many genders as to be fairly meaningless. I would like a totally arbitrary relation between people’s bodies and how they perform: how they dress, how they act, with whom they partner, how they choose to have families, and so on.[1]

– Amia Srinivasan

Here is an extract from a guide written in 2014 (and revised in 2016) by the Swedish National Office for Gender Studies (Nationella sekretariatet för genusforskning) and directed to Swedish government agencies:

Norms can be described as commonly shared behavioural rules and expectations. Sometimes these rules and expectations are unspoken and violations surface through subtle expressions like surprised or concerned expressions, sometimes they are even written down in law and norm violations can be punished by imprisonment.

In society, there are norms around for example gender, gender identity or expression, age, ethnicity, religion, sexual orientation, disability and class, as a consequence of which people are sorted into categories. Norms linked to these categorisations give people different structural living conditions. If a person is categorised as a woman, she will statistically speaking have a lower wage, spend more time on parental leave and take more responsibility for unpaid home and care labour than if that person were categorised as a man. If a person is categorised with another ethnicity than Swedish, they will statistically speaking have a harder time finding a house or apartment, be more likely to be subjected to workplace discrimination or hate crime in public than if the person were categorised as ethnically Swedish.

The different group identities that people are given through this categorisation have different amounts of power. This is because power structures give different groups different amounts of power. When we say that there are power structures we are describing structural, that is general and societal patterns to which there are individual exceptions. On a structural level men, for example, have higher wages than women, but there are individual women who have higher wages than individual men.

The term intersectionality comes from the English word intersection which means crossing. When an intersectional analysis is made it is the crossings or intersections between different power structures that is analysed. This in order to create an understanding for how different categories interact and when power imbalances are created.[2]

In Sweden, feminism, gender studies and intersectionality are closely associated with normkritik, or norm criticism. Norm criticism is, roughly speaking, a method for analysing the effects that norms have on individuals who cannot or do not wish to conform to them. The extract above is, I think, a fusion of norm criticism and intersectionality. It seems reasonable to me, with a few exceptions. But it doesn’t say anything about norms outside their ability to harm. Is that all norms do? I think Chesterton’s Fence applies here:

Chesterton’s Fence is the principle that reforms should not be made until the reasoning behind the existing state of affairs is understood.

The principle is relevant because human culture and institutions are often the products of years of accumulated knowledge, experience and (cultural) evolution. Do norm critics generally reason clearly about the norms they criticise? I want to say, maybe? If I squint, I can kind of see it there in the passage we started out with, though in an understandably condensed form. There is, for instance, a causal explanation of how norms can become oppressive for people. On the other hand, I’ve often had the impression that the two chief activities that norm critics engage in are describing and condemning bad norms. But I don’t know if this is really true.

Norms – How Do They Work and What Are They Good for? #

Here is how I see norms.

Norms are a solution to coordination problems. Take stealing, for example. One person benefits by stealing if they are the only thief in town, but if everyone steals, that’s bad for everyone – the best thing for the village as a whole is that no one steals. So how do you prevent people from stealing? You do it with a norm against stealing. You make stealing shameful and taboo.

Norms are sustained by people’s monitoring and punishing other members of their community for deviations.[3] In small communities, where everyone knows each other, norms are enforced via reproach, gossip, ostracism and stuff like that. That kind of enforcement doesn’t scale, so once some communities started growing larger, large enough that not everyone knew everyone else, they encoded their norms in laws and endowed specially appointed people with the right to enforce them, e.g. cops.

Norms affect people’s behaviour – obviously, because that is why they exist. They affect the decisions that we make, and they do this in at least two ways. The first is by making the expected value of some outcomes much worse (e.g. “if you cheat on your spouse you will go to hell”).[3:1] The second is via the aforementioned punishment (e.g. “if you cheat on your spouse, we will shun you and talk about you behind your back”).[3:2] And of course often the mere thought of such consequences is enough to deter people from norm-deviating behaviour. They may not even realise that it is because of a norm that they are averse to the behaviour. That is what people mean when they say that norms are internalised.

It’s worth pointing out that, even when norms aren’t enshrined in law, the punishment for violating them can very bad. Shunning, ostracism, bullying – these work really well and even the threat of them can be oppressive. They can and often do make people feel terrible. They can cause people bodily harm, or even death. So not only should we work for norms that cause desirable behaviour, but we should also work to dismantle norms that do not. This, of course, is what norm critics have set out to do.

Norm critics, though they focus their energy on harmful norms, surely grant that there are excellent norms, too. But that makes me wonder what norm critics do that is so special? Isn’t that what the culture war is – a massive, prolonged, entrenched war over norms? When a conservative is denouncing no platforming and other free speech violations, they are fighting over norms. When a devout Christian opposes abortion rights or premarital sex, they are fighting over norms. Are they doing norm criticism? I suspect norm critics don’t think so. There has to be something more to norm criticism. Maybe that something is power.

Norms and Identity #

Here is another obvious point: norms are different in different communities. For example, in many Muslim-majority communities, drinking alcohol is frowned upon, whereas in Northern European communities people are often expected to drink alcohol in social settings, to our own detriment. But there are also different norms applied to different people within communities. These can be temporary, like smoking or drinking alcohol being taboo for pregnant women. Or they can be tied to more permanent markers like identities, for example men being expected to be courageous risk-takers and that sort of thing.

I think there is a twofold interaction between norms and identity. The first is that, so long as we have norms associated with certain identities, knowing someone’s identity can give us an idea of what norms they adhere to. I will talk more about this in the next section.

The other is that powerful identity groups can use norms as tools to achieve their goals, possibly without taking others’ interests into account, and can moreover use norms to retain their power. Until recently in European history, the Christian church has been a supremely dominant power, and has used norms both to fortify itself (e.g. taboos against doubting or disbelieving) and to get what it wants (e.g. norms enforcing the paying of tithes or attending church or treating priests with reverence). In these societies an individual might find it hard to leave their given identity – that of a Christian, say – behind, due to norms harming people of other identities, such as that it is acceptable or even good not to do business with people of other faiths, or to apply the law more strictly to heretics, etc.

(Sometimes norms are harmful for certain groups for purely accidental reasons. For example, among the Fore of the Papua New Guinean highlands, women and children were more likely to get kuru, a “rare, incurable and fatal neurodegenerative disorder”. That was because kuru was caused by the spread of prions that occurred when the Fore ate the remains of dead family members. Prions especially accumulate in the brain, and tradition held that it was the brain that women and children should eat. But of course none of the Fore knew any of this, and so it must have been pure bad luck that made this norm of funerary cannibalism especially harmful for Fore women and children.)

So when the National Office for Gender Studies writes the following –

In society, there are norms around for example gender, gender identity or expression, age, ethnicity, religion, sexual orientation, disability and class, as a consequence of which people are sorted into categories.[2:1]

– I am unsure what they are actually saying. They are not saying that people are forced to behave in a certain way because that is expected of people of their identity, although they probably think so. Are they telling us that different norms apply to different identities, and that it is this that sorts people into categories? Are they telling us that there are norms that themselves tell us to sort people into categories? The meaning is surprisingly slippery.

I for one don’t think people “are sorted” into categories; or not merely sorted, I should say. I think that people both are sorted, and sort themselves into categories. I have two reasons for thinking so. The first is that many people join and identify with various subcultures, often acting against majority norms, seemingly of their own volition. The second is that there is evidence for innate gender differences in preferences; for instance, boys and girls show large differences in which toys they choose to play with starting at a very young age, and even as infants.[4] That makes it plausible that, with gender at least, norms don’t do all the job of sorting us into categories, though that is not to say that they don’t matter.

Still, if one wants a good world, where people are free to choose their own ends to pursue, it is enough to know that it does happen – if only sometimes – that people are sorted into categories which they would not have chosen for themselves for us to want to put an end to it, because now those people are stuck with a set of norms to which they do not want to adhere, and with a threat of punishment should they deviate. I see four possible ways out of this problem. We could make punishment for norm deviations laxer. We could tear down all norms that apply only to a part of our community. We could stop sorting people into categories and let them choose their own identities when they are old enough to do so. Or we could kill the norm that says people can’t move between categories once they have already been assigned. I am not sure which of these the median norm critic prefers.

But when the National Office for Gender Studies writes the following –

The different group identities that people are given through this categorisation have different amounts of power. This is because power structures give different groups different amounts of power.[2:2]

– I think I get it. It’s not that these groups meet and conspire to rid their adversaries of all rights and power. It’s that they evolve norms – often unwittingly, I think, via cultural-evolutionary processes – that allow them to achieve their goals and enforce their values, and that these norms don’t always take into account the interests of those who aren’t members of these groups. The fact that a group is in power is evidence that it has developed power-retaining norms. The norms might even be a reason that it took power in the first place.

(I’ll admit, however, to not quite understanding Sekretariatet’s use of the phrase “power structures” here. As I understand the phrase, it describes the relation in power between different individuals or groups in some community. But to say that this is what gives different groups different amounts of power seems like a tautology.)

Identities and Prediction #

I wrote earlier that one way I see norms and identity interact is that, so long as some norms are associated with some identities, knowing somebody’s identity gives us an idea of what norms that person adheres to. Why does that matter?

It matters because we all want to achieve our goals. To achieve our goals, we need to make good decisions. We are more likely to make good decisions if we can correctly predict the future, including how others will react to what we do or to what happens. Of course, we know something about humans, so we can make guesses about how other humans will act and react – guesses that will be right more often than not, probably. But different people sometimes act differently. And because, as we said earlier, different communities have different norms, knowing that someone is a member of a certain community has some predictive power.

Humans are hardwired to separate things into categories.[3:3] This has been evolutionarily adaptive for us because it allows us to make predictions based on class similarity.[3:4] For instance, when we see that such-and-such a plant belongs to such-and-such a species, we know that it likely possesses the traits that plants of that species have generally, and similarly for animals, landscapes, weather patterns and so on.[3:5] It seems likely that this sort of predictive processing happens with humans too, for example by categorising somebody as of a personality type, or a tribe, or a gender.

Obviously the predictive information is the same for some community of people no matter which individual is under consideration. If I think that Georgians (of the Caucasus) are more likely to be heavy-hearted poets than other peoples, and all I know about Mr. Chavchavadze and Mr. Guramishvili is that they are Georgians, that might be (weak) evidence for the possibility that they are such poets, but, importantly, equally so for both of them. Until I get to know them as individuals, all I know about them is what I know about Georgians generally. We call this phenomenon stereotyping.

Stereotypes are often quite accurate.[5][6] That is, they often capture average differences between groups.[7] So the stereotype that Irish people are red-haired doesn’t imply that more than 50% of Irish people are red-haired, but that Ireland has a higher proportion of redheads than do other countries, which, with the exception of Scotland, is true. If we don’t have access to good information about an individual, these heuristics are generally better than nothing at predicting attributes or behaviour. That is not to say that they are perfect. Even if an Irishman is more likely to have red hair than a Frenchman, the majority of Irishmen are not red-haired, and there will be Frenchmen who are. And there are stereotypes which almost certainly don’t capture any real group differences, such as that red-haired women are wild and hot-tempered.

These stereotyping heuristics do not seem bad to me except when they are used on beings with moral standing, like humans. You might think, “That is a major caveat!” I agree. The reason I think this is because the moral philosophy that makes most sense to me is Korsgaardian Kantianism, and, quoting Korsgaard:

Even if members of genders [or any other community] do tend to be alike, there are individuals who are different. If the genders are treated differently, these individuals will be treated wrongly. By Kantian standards the argument is over: it does not matter whether these individuals are many or few.[8]

The point is that we should act with respect for other people as individuals – as people who can make their own choices about what is good for them.[9]

Utopia of Unpredictability? #

Let’s return to the epigraph:

While I think it’s true that there are lots of women, cis and trans, who don’t experience their identity as a woman to be oppressive, I think nonetheless they exist under a system in which gender itself is structured hierarchically. My own instinct is that in the feminist utopia, there will be no gender, or there will be so many genders as to be fairly meaningless. I would like a totally arbitrary relation between people’s bodies and how they perform: how they dress, how they act, with whom they partner, how they choose to have families, and so on.[1:1]

The way I interpret her – and it seems likely to me that my interpretation is wrong in some way – it’s not that she wants people to not make inferences about others based on their identity (or other marker of community), rather she wants there to be no correlation at all between people’s identity (or other markers of community) and their behaviour. The ideal relation is “totally arbitrary”.[10]

This is weird for two reasons. The first is that it means it’s not the offending behaviour that should change, but that we should reconfigure society such that the offending behaviour is no longer possible. For example, it does not require the jingoist to stop making stereotyped assumptions about foreigners. Instead, it wants all of us to act in a way, and to allow others to act in a way, that does not correlate with our nationality. That way it does not matter if the jingoist wants to make inferences about people based on their nationality, because there is no information to be gained by doing so. And because these norm-based predictions are no longer useful, they will become extinct. (I’m sure this is not how Amia Srinivasan would put it, but it is how I put it.)

I empathise with this perspective, because it reminds me of how I see as the likeliest path forward for farmed animals not a moral awakening on an unprecedented scale but the development of affordable, healthy and delicious plant-based and cultured alternatives.

The second reason it’s weird follows from the first, and it is that this seems to demand that a fragile, unstable equilibrium be maintained forever. Take Srinivasan’s example of gender and how people dress. To realise Srinivasan’s utopia, it is not enough that five percent or ten percent of women (anyone who identifies as a woman) dress differently than the norm. Even if only a narrow majority dresses according to the norm, that is still enough for a woman’s identity to provide (some, weak) evidence about her behaviour (in this case, making it slightly more likely that she wears dresses or skirts or blouses or slim-fitting pants or tops or whatever). A “totally arbitrary relation” demands an extraordinarily disciplined commitment to diversity. How could such a state of affairs possibly be maintained?

I don’t think it could. I think that, even if Srinivasan’s utopia did arrive, it would vanish in an instant. Take, for example, subcultures. People constantly form new subcultures around shared interests and attitudes. But the fact that they share these interests and attitudes, and that this marks them out as different from the general population, means that they have norms. Knowing that someone identifies as a member of the subculture means one is better positioned to make inferences about their interests and attitudes, and therefore also about their behaviour. Can Srinivasan’s utopia withstand the human tendency to form subcultures and, for that matter, distinct cultures? Seems unlikely, even for a utopia.

(Hmm, maybe I am being a bit unfair here. Srinivasan isn’t talking about the relation between identity and behaviour, but of that between behaviour and body. People’s bodies are often correlated with their identities, but there are many identities that have nothing to do with people’s bodies, and it’s possible that Srinivasan only cares about those identities that are, or have traditionally been, linked to bodies, e.g. gender, ethnicity, sexual orientation, disability, age, etc. Maybe Srinivasan’s view is that only these identities – because they are often not chosen, but assigned at birth – should be dissolved, and that all other identities may be permitted. If so, that takes some of the force out of the subculture argument.)

Conclusion #

Norm critics analyse norms, especially those that have to do with gender, sexuality and ethnicity, and advocate for their removal or alteration. Norms are a social technique for enforcing a desired behaviour across a community; the community sustains the norms by monitoring and, if necessary, punishing norm violations.

There are good norms, like norms against lying, stealing and murdering, and there are bad norms, like norms against believing in the wrong faith, being a vegan and having gay sex. (Many norms are more ambiguous than that.) Good norms prescribe good behaviour. Bad norms either prescribe bad behaviour or forbid behaviour that isn’t bad; the latter sort is harmful because the punishment for norm violations is often harsh, and we don’t want to punish people without good reason.

Because norms differ across different communities, knowing someone’s community membership can help us predict how that person will behave. However, actually treating someone in a certain way due to their community membership is, in my opinion, wrong, because it involves not fully taking that person’s interests into consideration. I don’t think we have any hope of getting rid of norms, or of differences of norms between different communities. Instead, I think we should work to improve norms, revamping or removing the bad ones and protecting and nourishing the good ones, in our communities.

Footnotes #

  1. This is from an interview with Amia Srinivasan by Regan Penaluna. ↩︎ ↩︎

  2. Translated by myself from the original Swedish:

    "Normer kan beskrivas som allmänt delade regler och förväntningar på beteenden. Ibland är dessa regler och förväntningar outtalade och överträdelser märks genom subtila uttryck som förvånade eller besvärade blickar, ibland är de till och med nedtecknade i lag och normöverträdelser kan bestraffas med fängelse.

    I samhället finns normer kring till exempel kön, könsidentitet eller könsuttryck, ålder, etnicitet, religion, sexuell läggning, funktionalitet och klass, som gör att människor sorteras in i kategorier. Normer som är kopplade till kategoriseringarna ger människor olika strukturella levnadsvillkor. Kategoriseras en person som kvinna kommer hon statistiskt sett att ha lägre lön, vara föräldraledig längre och ta ett större ansvar för obetalt hem- och omsorgsarbete än om personen kategoriseras som man. Kategoriseras en person som varandes av annan etnisk tillhörighet än svensk, kommer personen statistiskt sett ha svårare att få en bostad, i högre utsträckning bli utsatt för diskriminering på arbetet eller hatbrott i det offentliga rummet än om personen kategoriseras som etnisk svensk.

    De olika grupptillhörigheter som personer genom denna kategorisering tillskrivs, har olika mycket makt. Det beror på att maktordningar ger olika grupper olika mycket makt. När vi talar om att det finns maktordningar beskriver vi strukturella, det vill säga generella samhälleliga mönster från vilka det finns individuella undantag. På strukturell nivå har män exempelvis högre lön än kvinnor, men det finns enskilda kvinnor som har högre lön än enskilda män.

    Begreppet intersektionalitet kommer från engelskans intersection som betyder korsning. När en intersektionell analys görs är det korsningarna eller skärningspunkterna mellan olika maktordningar som analyseras. Detta för att skapa en förståelse av hur olika kategorier samverkar och när över- eller underordning skapas." ↩︎ ↩︎ ↩︎

  3. Henrich, J. (2015). The secret of our success. Princeton University Press. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  4. Davis, J. T., & Hines, M. (2020). How large are gender differences in toy preferences? A systematic review and meta-analysis of toy preference research. Archives of Sexual Behavior, /49/(2), 373-394. ↩︎

  5. Jussim, L., Crawford, J. T., Anglin, S. M., Chambers, J. R., Stevens, S. T., & Cohen, F. (2016). Stereotype accuracy: One of the largest and most replicable effects in all of social psychology. ↩︎

  6. In Thinking and Deciding, Jonathan Baron describes an early study pointing in this direction:

    “McCauley and Stitt (1978) examined white Americans’ stereotypes of black Americans[.] The subjects answered questions like ‘What percent of adult Americans have completed high school?’ and ‘What percent of black Americans have completed high school?’ The ratio of these two answers serves as a rough measure of correlation. A ratio of one means no correlation. When the study was done, the ratio of these questions was .65: Blacks were less likely to have completed high school. Mean ratios from the responses of different groups of white Americans ranged from .60 (social work students) to .74 (social workers). For this question and most of the other questions (illegitimate children, unemployment, victim of crime, on welfare, four or more children, female head of family), the average ratios of the respondents were in the same direction as the true ratios based on census data and somewhat less extreme – closer to 1.0. In other words, this kind of ‘stereotype’ is fairly accurate, and people judge the correlation correctly between (for example) being black and having completed high school.” ↩︎

  7. Bordalo, P., Coffman, K., Gennaioli, N., & Shleifer, A. (2016). Stereotypes. The Quarterly Journal of Economics, /131/(4), 1753-1794. ↩︎

  8. Korsgaard, C. M. (1995). A note on the value of gender-identification. Women, Culture, and Development: A Study of Human Capabilities, 1, 401. ↩︎

  9. I leave it open whether there are situations where stereotyping is permissible because the intended action does not involve or affect the stereotyped person or persons directly, though I have not thought that through in detail. ↩︎

  10. It’s possible that Srinivasan is dreaming for something slightly different, namely that we should keep norms, and keep identities that are connected with them, but disconnect those identities from physical markers of community. So we can let a thousand genders bloom so long as none are more likely to appear than others for any one cluster of body types. This doesn’t seem like it would get rid of “structured hierarchi[es]”, though. And I don’t see how vast numbers of identities will prevent oppressive and power-gaining norms to form. In fact, the more identities (and therefore communities), the more opportunities for such norms to form and spread. ↩︎