(Un)doing gender (and citation)

Judith Butler stans might appreciate this, by Candice West and Don Zimmerman (2009, pp. 112-113):

‘The initial ideas for “Doing Gender” came in 1975 and 1976 […]. We presented “Doing Gender” at a meeting of the American Sociological Association in 1977; we spent the next ten years trying to get it into print.

‘Between 1977 and 1987, this work was rejected by some of the most respected journals in our field (including Signs: Journal of Women in Culture and Society and Social Problems). In fact, Erving Goffman, one reader of an early draft, passed away in the time it took to get the paper published. During those ten years, we continued to circulate prepublication versions to friends and colleagues, and we continued to refine and polish the paper in response to their remarks. We were more than gratified to see “Doing Gender” finally published in 1987 […].

‘Today, “doing gender” often appears in print without acknowledgment of its source, and some scholars (such as Judith Butler) play on our wording (Undoing Gender, Butler 2004) without ever citing our work.’


West, C., & Zimmerman, D. H. (2009). Accounting for doing gender. Gender and Society, 23, 112–122.

The L’Oréal principle

“Starting from Disney animations that we watch as young children telling us that if we believe in ourselves, we can achieve anything, we are bombarded with the message that individuals, and they alone, are responsible for what they get in their lives. We are persuaded to accept what I call the L’Oréal principle – if some people are paid tens of millions of pounds per year, it must be because they are ‘worth it’. The implication is that, if people are poor, it must be because they are either not good enough or not trying hard enough.”
– Ha-Joon Chang (2014). Economics: The User’s Guide.

What is a social fact? The grounding-anchoring model

Society depends on social facts, which are quite different to facts about, say, physics and chemistry. Statements we may wish to evaluate to determine whether they are indeed facts include:

  • The gang ran down the street.
  • This person’s name is Alex.
  • Alex is a woman.
  • Spain won the World Cup.
  • This elaborate polymer sheet is a valid Bank of England banknote.

Brian Epstein’s (2015) grounding-anchoring model aims to help us. According to Epstein, there are two kinds of social fact which represent extremes along a continuum:

Type 1 facts are relatively straightforwardly defined in terms of facts about individual people. For instance, a social fact, “the gang ran down the street”, depends on facts about each person in the gang running down the street.

Type 2 facts cannot be defined in this reductive way. For instance, I go to the corner shop for emergency chocolate supplies and exchange an elaborate polymer sheet for the chocolate. The social fact “this elaborate polymer sheet is a valid Bank of England banknote” cannot be reduced to facts about physically smaller units of anything.

As Epstein explains, it’s more complicated than this. The Type 1 facts need something more than only a description of the individual people running. We also need some sort of agreed understanding of what constitutes a gang. Consider different groups of people running down a street. How do they know each other? Are they running away from danger or to try to catch a bus? What age and gender are they? How menacing do they appear? Consider the following dictionary definition:

“a group of young people, especially young men, who spend time together, often fighting with other groups and behaving badly”

Note the “especially” and the “often” in this. Who decides what constitutes bad behaviour?

Ingredients of a social fact

I like Epstein’s approach because it neatly organises the problem of pondering social facts by drawing out different elements involved, making it easier to see what the arbitrary choices are.

The banknote fact, “this elaborate polymer sheet is a valid Bank of England banknote”, depends on other facts. For instance, the note has to look a particular way and be made of a particular material. But it wouldn’t suffice for me to print it this way in my flat. The Forgery and Counterfeiting Act 1981 would land me in trouble if I were caught doing so. A valid polymer banknote should be produced by a special print works near Debden tube station in Essex.

Epstein likes to give his social entities names, so consider Poly the polymer bank note. Poly is a valid banknote if (a) Poly is made of the correct material and (b) Poly was produced at Debden Printing Works. These facts about Poly ground the social fact that Poly is a valid banknote. They are said to be grounding conditions.

The more general principle, of which this fact about Poly is an instance, is the frame principle. For all b, if b is made of the correct material and produced at Debden Printing Works, then b is a valid Bank of England banknote.

Now the interesting question is why does this arbitrary sounding frame principle work? This leads to the final ingredient: the social facts that anchor the frame principle (anchoring facts). In the case of banknotes, laws are important. There is also a more general anchoring fact that seems to apply for many non-legal social facts: “we collectively take it to be the case that…”. Note that anchoring facts do not explain how or why the frame principle came to be, just what justifies it now.

The diagram below (Epstein, 2015, p. 84) illustrates these ingredients:

So, a lot was going on when I handed Poly the polymer banknote to the shop keeper and received a bar of chocolate (and handful of metal) in return. Something else to ponder is how rapidly transactions like these happen. Much is taken on trust, e.g., the shopkeeper didn’t run through a checklist of half a dozen features to verify that Poly was valid.

Unenrolled deed polls provide another intriguing example. You print a short incantation saying things like:

“BY THIS DEED OF CHANGE OF NAME made by myself the undersigned … I ABSOLUTELY and entirely renounce, relinquish and abandon the use of my said former name…”

Get it signed in wet ink with two witnesses and, as if by magic, thereafter you have a new name and a new social fact. You have also entirely renounced, relinquished and abandoned another social fact and call on others to do the same. This signed sheet of paper can be presented at your bank and lead to them issuing you with a card under your new name, even though – since it is unenrolled – the signed document is not officially filed anywhere. Insert that card in a bank machine, and out come more sheets of polymer. Exercise to the reader: think through what the frame principle, grounding conditions, and anchoring facts are for this example.


Epstein defines social facts in terms of frame principles, which are recipes for grounding conditions concerning specific entities. Another set of anchoring facts justify the frame principle: they anchor the frame principle and hence grounding conditions.


Epstein, B. (2015). The ant trap: Rebuilding the foundations of the social sciences. Oxford University Press.

Mildly revised on 30 August 2023

Conventional versus non-conventional facets of reality

One of the slipperiest concepts in social theorising is that of social construction. Another potentially more productive way to look at social reality is in terms of what properties of the world are conventional rather than non-conventional. A great example comes from the physics of time. I’ll briefly introduce this and then sketch out parallels with theorising sex/gender.

The duration of a second is conventional. One definition is that a second is

‘the duration of 9192631770 periods of radiation produced by transition of an electron between two hyperfine levels of a Caesium-133 atom.’

It is conventional in the sense that it could have been otherwise and in fact the definition of a second is changing. As Bradley Dowden explains:

‘conventions are up to us to freely adopt and are not objective features of the external world that we are forced to accept if we seek the truth.’

A non-conventional feature of time is the transitivity of events in the same frame of reference. That means if an event

A precedes B, and
B precedes C, then
A precedes C.

Let’s give this distinction a go for sex/gender. Phenotypic sex category at birth has a clear definition in terms of genitals which, for most people, coincides with sex chromosomes. The category is highly reliable and is predictive of much, including sexual reproductive roles – but only for cisgendered and heterosexual people, and even then it is imperfect (due to menopause, for instance). Sex category loses its predictive power for LGBTQ+ people.

The male/female dichotomy could have been otherwise. For instance it is a fact that many adults have a gender identity that differs from the phenotypic sex category they were assigned at birth and some people are nonbinary, i.e., neither a man nor a woman. In terms of social roles, there is much greater variation than two categories (see, e.g., twinks and bears). A binary category tethered to sexual reproductive role is not the only way to carve nature at its joints.

However, there are non-conventional facts about reproductive biology that determine, for instance, whether someone is able to become pregnant. Katrina Karkazis argues that sex/gender category should be detached from the specifics of what bodily organs we have:

‘It is long overdue that we understand sex not as an essential property of individuals but as a set of biological traits and social factors that become important only in specific contexts, such as medicine, and even then complexity persists. If we are concerned with certain cancers, for example, knowing whether someone has a prostate or ovaries is what’s important, not their “sex” per se. If reproduction is the interest, what matters is whether one produces sperm or eggs, whether one has a uterus, a vaginal opening, and so on.’

Opportunity hoarding / social closure

“In order for certain jobs to confer high income and special advantages, it is important for their incumbents to have various means of excluding others from access to them. This is also sometimes referred to as a process of social closure, in which access to a position becomes restricted. One way of doing this is by creating requirements that are very costly for people to fulfill.”

– Wright, E. O. (2009, p. 104). [Understanding Class: Towards an Integrated Analytical Approach. New Left Review, 60, 101–116.]

Algorithmically infused societies

“We are witnessing the emergence of algorithmically infused societies who are shaped by deeply entangled algorithmic and human processes and behaviour […]. Algorithms—and the tools, services and platforms they power—mediate social, economic and political processes by shaping a wide range of activities and decision-making practices across many areas. Algorithmic influence can be observed in how people consume information or cultural artefacts and how they interact with others. This influence is also increasingly tangible in many high-stakes areas, such as healthcare, credit scoring, law enforcement and employment. Although algorithms can provide many benefits, including efficiency, objectivity, auditability, fairness and social good, they can also amplify—either inadvertently or purposefully—existing inequalities and biases in society, or might even introduce new ones.”

– Wagner, C., Strohmaier, M., Olteanu, A., Kıcıman, E., Contractor, N., & Eliassi-Rad, T. (2021) [Measuring algorithmically infused societies. Nature, 595(7866), 197–204.]

Three useful terms

Amatonormativity: “the widespread assumption that everyone is better off in an exclusive, romantic, long-term coupled relationship, and that everyone is seeking such a relationship” (Elizabeth Brake)

Cisnormativity (also cissexism): “the belief or assumption that cis people’s gender identities, expressions, and embodiments are more natural and legitimate than those of trans people” (Julia Serano)

Heteronormativity (also heterosexism): “the belief or assumption that heterosexual attractions and relationships are more natural and legitimate than their same-sex counterparts” (Julia Serano)