Sex and the Scottish Census

Harry Josephine Giles
Mar 26 · 36 min read
Image by Ted Ayton on Flickr, licensed under Creative Commons. A person with mid-length crown hair and sunglasses, wearing a white bandanna and a black t-shirt reading “All Genders Matter”, is waving a trans pride flag.

If I can’t be counted then I don’t count for anything. I am a nonbinary person in Scotland — someone who’s neither male nor female — and I have no legal status. I don’t know how many of me there are. When I want to access healthcare, there’s no funding and no statistical basis to support it, there’s no legal status to back up my right to that healthcare, and there’s a considerable lack of social and medical understanding — three tremendous barriers which, added to the fact that I’ve now been on an NHS waiting list for 15 months before getting a single appointment, makes my sex-related healthcare effectively non-existent.

When I learned that the 2021 Scottish Census was probably going to include a question on nonbinary people, I was pleased and hopeful, because this would give the strongest possible statistical foundation to our appeals for legal recognition and healthcare. That, to me, was all it meant: it wasn’t about the government recognising me, and it wasn’t about any politics or philosophy of sex and gender; it was just a simple matter of being counted.

Now, however, I’ve become a political and philosophical football. The Stage 1 Report on the Census (Amendment) (Scotland) Bill issued by the Culture, Tourism, Europe and External Affairs Committee of the Scottish Parliament (now there’s a mouthful — I’ll just call it the Census Bill and the Committee from now on) surprised statisticians and rights campaigners by disputing many of the initial recommendations made to the Parliament, and was accompanied by a video and Twitter thread from Joan McAlpine MSP that made it clear that the Committee had been strongly influenced by a large input of evidence from various campaigns organised to oppose transgender rights. My feeling was that quite a few people seemed to think that I didn’t count, and didn’t really want me to be counted — and that a relatively simple matter of counting me well had got caught up in an immense social, political and philosophical argument about sex and gender.

I’m faking naivety a bit. I do, of course, have a number of strong social, political and philosophical opinions about sex and gender. For the sake of this article, and for the sake of the census, what I want to do here is to concern myself only with what makes for good counting. (Because I can’t help myself, though, I’ve saved some trans politics for the end. More crucially, this whole project also needs a critique of what it means to be counted.) To do this, I want to address directly points made by McAlpine and by the campaigners she has cited in this and other public communications. First, I’m going to discuss what sex and gender are in UK law, and then I’m going to consider what makes for good census questions. This article deals with very technical matters, but I have tried to write it for a general audience. It is addressed to trans rights campaigners to help them understand these issues, and it is addressed to the Parliamentarians and civil servants legislating on these issues and delivering the results. I will send it personally to the Committee members, to my constituency MSPs, and to MSPs who express an interest.

In all this, I am not a legal expert or a professional campaigner. I am a layperson who is engaged in a deep, personal, reflective way, as a feminist and a trans person, with the fight for women’s and gender minorities’ liberation. I believe, however, that my research into the legal basis for my arguments is at least as strong as that of those who gave evidence to the Committee, the majority of whom are also, for these purposes, laypeople. It’s important for anyone engaging with this debate to realise that the vast majority of us producing articles like this are doing so without institutional backing, without pay, but with considerable expertise by experience — even when you see credentials (and I’ll be a Doctor soon, but a Doctor of Poems), they’re often unrelated to the matters at hand. In any case, I welcome the perspectives of legal expertise on my arguments, and I particularly welcome additional case and statutory legal analysis that has a bearing on these issues.

A summary of what follows for the short of time:

  • There is no general definition of sex in UK law.
  • Gender identity does not exist in UK law.
  • Trans people access the vast majority of sexed services and legal statuses on the basis of their lived sex.
  • The Equality Act protects trans people on the basis of their lived sex.
  • There is no legal grounding for terms like “legal sex” and “biological sex”: this a philosophical position only.
  • The census should be based on what is established law and practice, not on a single philosophical position in an ongoing debate.
  • Established census practice is for trans people to answer a sex question by their lived sex, not by their birth certificate.
  • Nonbinary people need a third write-in option in the sex question.
  • The fears of data inconsistency with previous censuses if more trans-inclusive questions are asked are ill-grounded, and suggested alternatives risk the data far more.
  • The census should have one question on sex, with a third write-in option, without a “biological” or birth certificate definition of sex, and one voluntary question on trans status.
  • The census process must consult with intersex-led advocacy organisations and intersex people to establish better practices for their monitoring in the census.
  • We need to be very wary of separating sex and gender identity in law, theory and practice.
Photo by Stephen Yu on Flickr, licensed under Creative Commons. A large crowd at Pride Scotia 2014: lots of balloons and a Nae Nazis sign are visible.

What is “legal sex”?

Here I’m going to consider sex and gender in the Equality Act, and then in the case law of Corbett v Corbett and others, before concluding with thoughts on the practicality of sex operating in law.

Sex and Gender in Statutory Law

Sex has no general definition in UK law. The Equality Act 2010 defines sex as involving “a reference to a man or to a woman”, and that’s it: it doesn’t tell us what men and women are. I have found no statutory law which references any biological or social factors as definitional for sex, and only one piece of case law, discussed below. When McAlpine says that the Equality Act “gives women protections based on biological sex”, this is untrue: there is no biological language in the Act. All people have protections based on sex, whatever that is, and there are some limited exemptions to the protections all people have based on gender reassignment, but there are no “biological sex” protections.

When people say “legal sex”, they are generally referring to the sex that is recorded on someone’s birth certificate. This is assigned at birth, and can be amended through a Gender Recognition Certificate (GRC) thanks to the Gender Recognition Act 2004. The reason for not calling this “legal sex” is that sex operates differently in different areas of law. For example, you do not need a GRC to change your passport: your sex at the border is not defined by birth certificate. The same goes for drivers’ licenses. You do not need a Gender Recognition Certificate to access the bathroom of your lived sex, as just thinking about it for a moment will indicate: there is not (thankfully) a customs agent at every bathroom door checking people’s sex, though panic about trans people in bathrooms has made life harder for all gender-nonconforming people in public loos, such as butch women. If you think about it, nobody checks your sex for the majority of things in life where sex is relevant, and those conventions and expectations (and how they are shaped by case law) define the legal grounding for sex in those situations. Some of the areas where birth certificate sex does matter are pensions, inheritance and marriage rights, which is why there’s a campaign to reform the GRA to improve these situations for trans people.

Gender, meanwhile, has no meaning separate from sex in UK law. I have found neither statutory nor case law which is relevant. Indeed, “gender” and “sex” are used interchangeably in many relevant laws, such as the Equality Act, which offers the delicious semi-tautology that gender reassignment means reassigning sex: “A person has the protected characteristic of gender reassignment if the person is proposing to undergo, is undergoing or has undergone a process (or part of a process) for the purpose of reassigning the person’s sex by changing physiological or other attributes of sex.” This strongly implies that for the Equality Act gender and sex are the same thing. “Gender identity” is not used anywhere in UK law.

This means that when McAlpine and Committee evidence-givers say things like “sex and identity are not the same thing and require separate questions”, this has no legal basis. Some of this arises from the fact that sex and gender do have separate but interconnected meanings in sociology and politics: sex tends to refer to biological characteristics and gender tends to refer to social characteristics, but these two are fundamentally intertwined. Components of both include: genitals; chromosomes; gonads; secondary sex characteristics such as breasts, bone structure and hair; hormone levels and hormone sensitivity; how people see you and interact with you in the world; what social roles you’re asked or choose to play; race, economic class, disability and other social and class systems; how you experience yourself and what your identity is. Different theorists may make dividing lines between sex and gender in different places between these categories, and many theorists will say that that is not possible. For these reasons, I never use “gender” separately from “sex” in this essay, except in quotation and in legal terminology, usually in compounds like “gender reassignment”. When I use “gender identity”, I refer only to the factors of “how you experience yourself and what your identity is”, and I use this very sparingly and usually in quotation.

So: how to work with the categories of sex and gender identity is an ongoing debate, but for the purposes of the census and in UK law there is no general separation. When Professors Rosemary Auchmuty and Rosa Freedman, in their evidence, state that “Conflating sex and gender identity will undermine sex as a separate category protected by law,” their argument proceeds on this factually incorrect basis.

Sex and Gender in Case Law

[Content Note: the fifth/penultimate paragraph briefly discusses the law around rape and trans women, with necessary description that may be difficult to read. It, and this whole section, can be skipped if needed as it is summarised in the conclusion.]

The foundational case law is Corbett v Corbett, in which Captain Arthur Cameron Corbett, 3rd Baron Rowallan, sought an anullment to his marriage to trans woman April Ashley on the grounds of her trans status. This is the law that Freedman cites as the basis of her argument, when she says that “The law clearly sets out in that case that sex is biological, and that transsexualism (what we would now term transgender) is psychological. As a result, when we use the word ‘sex’ in law, we are referring to biological sex.” This, however, is legally wrong.

In the decision of Lord Justice Ormrod in the case, he states directly that the question is “what is meant by the word ‘woman’ in the context of a marriage, for I am not concerned to determine the ‘legal sex’ of the respondent at large.” That is, Ormrod explicitly states that he is not providing a legal definition of sex, but Freedman says that he is. He even puts inverted commas around “legal sex”, as I do, in acknowledgement that this is not a legal term. He is, rather, determining what a woman is in the single legal context of marriage. He goes on to say, “Having regard to the essentially heterosexual character of the relationship which is called marriage, the criteria must, in my judgment, be biological, for even the most extreme degree of transsexualism in a male […] cannot reproduce a person who is naturally capable of performing the essential role of a woman in marriage.” That is, his definition of marriage is heterosexual, and his definition of woman is a person who is “naturally capable” of the “essential role” — that is, making babies. This is not a definition of woman that I think any feminist would like to hold in law. Fortunately, given that the legal definition of marriage has now changed, there is some question as to whether this case law still holds. This is indicated but not determined by cases like S vs S-T (Formerly J) in which Ormrod’s ruling is “applied reluctantly”, and Bellinger v Bellinger, in which it was acknowledged that “though increasing recognition has been given to the complexities of gender identity over the years, this was an area which, if it was to be reformed, must be reformed by parliament and as a comprehensive whole, not piecemeal by judges”.

Ormrod does, however, provide the founding consideration of what sex is in law and thus some groundwork on which future definitions could be made. He says: “All the medical witnesses accept that there are, at least, four criteria for assessing the sexual condition of an individual. These are: (i) Chromosomal factors. (ii) Gonadal factors (ie presence or absence of testes or ovaries). (iii) Genital factors (including internal sex organs). (iv) Psychological factors.” That is, Ormrod, contra Freedman, states that psychology — “gender identity”, perhaps — is a medical component of sex. He rules this out when defining what a woman is for the purposes of marriage, because he thinks women are essentially baby-makers, and he explicitly declines to provide a legal definition of sex, but he does say that what we might now call gender identity could be a part of sex for legal purposes.

I have brought up these points with Professor Freedman, after she contacted me following my initial comments on the Bill, and she has not responded. Similar points were also made by Professor Sharon Cowan in the evidence, but were not responded to by Freedman or acknowledged by McAlpine in her numerous statements.

Further case law illustrates the points that (1) there is no general legal definition of sex, and that (2) the definition of sex varies depending on the area of law. Against my argument, Regina v Tan 1983 extended the Corbett ruling to also be relevant to the regulation of sex work, ruling that chromosomal, genital and gonadal factors only were relevant to whether or not someone is a man for the purposes of a conviction for living on the earnings of prostitution. (Feminists may also wish to note the homophobic judgements and language in this ruling.) For my argument, W v W 2000 argued that “on the true construction of the Matrimonial Causes Act, greater emphasis could be placed on gender rather than sex” (i.e. that psychological and social factors could be relevant to sex in the case of marriage) but did not require this to be true in order to make a ruling: the ruling concerned an intersex woman who was assigned-male-at-birth, and whose chromosomal and gonadal factors were classified as male under Corbett, but whose genital, hormonal and social factors were not. (Again, feminists should also note how attempting to extend Corbett has dangerous effects for intersex people, and that “birth certificate sex” is not sufficient.) Finally, as a further indication of the interpretability of sex for legal purposes, R v John Matthews ruled that a trans woman’s vagina is legally a vagina for the purposes of rape law, even though the Corbett ruling established (though may no longer still do so) that a trans woman’s vagina is not a vagina for the purposes of marriage law.

There will certainly be relevant case law which I have not covered here, but I am very confident that there is no case law providing a general legal definition of sex, and that Corbett does not do so, because none of these cases reference such a law and all of them repeat that Corbett is not a general definition of sex. Moreover, the closer the cases come to the present, the more reluctant are the applications of Corbett, and the more urgent the calls from the rulings for statutory law to advance. Since the rulings discussed here, the Gender Recognition Act 2004 and the Equality Act 2010 have changed the legal landscape considerably, and so I will discuss their application now.

Application of the Equality Act to Trans People

McAlpine, Auchmuty and Freedman all refer to sex-based protections in the Equality Act. They want to ensure that women are able to access law which protects them from sex-based discrimination, as do I — I just include all women in that. I also want to ensure that all trans people are able to access law which protects them from sex-based discrimination. I believe that the Equality Act provides this protection, for all the reasons stated above, but to the best of my knowledge this has not been tested in court, and nor is it likely to be. This is because the Equality Act also protects gender reassignment status, and this protection can be invoked in the majority if not all cases where sex is relevant: this stems back to P v S and Cornwall County Council at the European Court of Justice, which laid the groundwork for these protections. The Equality Act also defines gender reassignment status as referring to someone anywhere in the process of transition, which means that, as with sex, a GRC is not relevant.

I think it is worth taking a moment to consider the law of this fully, because of how it clarifies the often heated debates around the Gender Recognition Act, sex and gender. Access to bathrooms already does not depend on a GRC. Access to prisons already does not depend on a GRC. Access to refuges does not depend on a GRC. All of these accesses are already underway, already protected by the Equality Act, and the organisations responsible for campaigning for and administering women’s services support the inclusion of trans women and have not, in their considerable experience, found reason to consider trans women a threat. I do understand why women have fears and concerns about this, and am happy to talk these through with any woman interested in a respectful and evidence-based conversation, but bring it up here purely as a further illustration of the law, which does not distinguish between sex and gender, and which already allows for self-identification of lived sex in every case bar the birth certificate.

I have qualms about the term “self-identification”, which I find is often used to diminish trans people’s claims to their lived sex, by deliberately connoting a world in which people are randomly choosing their sex all the time. Notwithstanding the importance of gender identity to trans people and the contribution of analyses and practices of gender identity to transfeminist theory, the imaginary of free-flowing self-identification is not a world that exists in practice or in law: both social life and the law that regulates it depend on lived processes. That is, trans folk can access sex-based rights because they live as their sex, not because of a statutory declaration via a GRC or otherwise. I therefore use the term “lived sex” wherever possible. To be clear, this is not a clearly-defined legal term either, but is based on the use of “living in the acquired gender” in the Gender Recognition Act and the interchangeability of sex and gender in UK law. This is also an accurate sociological description of how sex operates in the world and in law. Aspects of “lived sex” include but are not limited to: social presentation, names and pronouns, physical appearance, gender identity, secondary sex characteristics however they are acquired, hormone levels however they are acquired, genitals however they are acquired, and so on. In short: if you live in the world as a woman, you are a woman for almost every legal and social purpose. The boundaries around that category are a long way beyond the scope of this essay, and I have no interest in policing them.

Consider the imagined case of a trans woman without a GRC, who lives and is perceived as a woman, and who believes she is being paid less than a male colleague for the same work. If she brings a pay discrimination case, would it have to be on the basis of the protected characteristic of gender reassignment, or could it also be on the protected characteristic of her lived sex? To my knowledge, there is no existing case law. My argument here is that the sex characteristic applies, not only for the reasons stated above, but also because the definition of the discrimination in the Equality Act includes perception: that is, for the purposes of the Equality Act, if someone thinks you’re a woman, and discriminates against you on that basis, you’re entitled to the Act’s protection. This neatly illustrates again how sex actually operates in the majority of law: as it does in daily life, on the basis of how you live it and how people see you.

[Edit 7/4/19: While the sex-based rights of trans people with a GRC are not in dispute, there is some legal conflict over what rights trans people without a GRC can prove in court. Key cases are Green v Secretary of State for Justice and Brook v Tasker, which reached potentially conflicting conclusions in different situations and levels of court. These are discussed here, with further links. In practice, though, trans people’s rights to, e.g., use toilets of their lived sex are not contested, with or without a GRC, which is one reason the case law is thin.]

To Conclude (This Bit)

Gender does not exist separately from sex in UK law, and sex has no biological definition in UK law, and when McAlpine and the relevant evidence-givers state that sex and gender identity are separated in law they are wrong. This putative separation cannot be the legal foundation for a separation in the census, which must instead consider what produces good data.

A final political point: I have serious reservations about the suggestion that we solve this problem by giving sex a rigorous legal definition. This is because, even if the technical difficulties could be overcome, when protected characteristics have this rigorous definition, that definition can be used against those whom it was intended to protect. Consider Ormrod above, who wanted to define women as “essential” baby-makers. This is not a hypothetical scenario: consider this (failed) bill in Utah which sought to limit trans rights by defining a woman as “an individual with ovaries who is confirmed before or at birth to have external anatomical characteristics that appear to have the purpose of performing the natural reproductive function of providing eggs and receiving sperm from a male donor.” Again, not very feminist. For a different sort of case, consider LGB asylum-seekers fleeing persecution on the basis of their sexuality, forced to prove their sexuality in strict legal terms, and how that leads to outrageous abuses. The weapons of our liberation can also be turned against us.

Photo by Joe Blaycock on Flickr, licensed under Creative Commons. Sheets of stickers labelled She, He and They; some red-yellow-green colour cards are also visible to the left.

What should we count in the census and how?

There are three questions facing civil servants, the Committee and MSPs considering the Census Bill: (1) Should the Census count sex or gender identity or both? (2) Should the Census count intersex and nonbinary people, and if so how? (3) How should trans status be counted? The third question is not in debate: everyone on all sides has agreed that a voluntary question asking whether or not someone is trans should be added to the Census, so I won’t discuss this here, but I will take that fact into account. I’ll consider the other two questions in turn, but first it’s worth giving an account of why some people are worried about this question.

Why Some People Are Worried

The first issue about which people are concerned is trans people’s access to sex-segregated spaces and services. This accounts for points 6–11 of Freedman and Auchmuty’s evidence, much of the other anti-inclusion evidence, and a significant degree of McAlpine’s ensuing comment. I understand that there are fears, and repeat my offer to talk these through with any woman interested in a respectful and evidence-based conversation, but they are outside the scope of this article and the Census Bill: the Census Bill is not a piece of legislation which affects the legal definition of sex and gender, only what is counted. The Census Bill cannot change the rights and protections that already exist. For now, I’d refer anyone interested in these issues to the position and work of Engender, Scottish Women’s Aid, Rape Crisis Scotland, Zero Tolerance, Equate Scotland, Close the Gap and the Women 5050 Campaign.

The second issue is whether any shifts in data gathered would affect the provision of women’s services. As Professor Kathleen Stock says, “Having accurate information about actual sex class is extremely important for tracking all sorts of related statistics to do with discrimination.” She does not say what these statistics are, and it is worth noting that Engender, Scotland’s feminist membership organisation, states that they “are not aware of any ways in which we, or our sisters in other women’s sector organisations who undertake detailed gender-analysis of policy, use the data from the census to do so”. Rather, the census plays “an important normative role in shaping how information is gathered in other more frequent or localised data gathering”—the data gathering actually used by women’s organisations. The question then is what norms the census has established, and whether there’s a justification for changing them. As discussed above, the legal and social norm for sex categorisation in the vast majority of circumstances is self-identification by lived sex; as discussed below, this was also the norm established by the 2011 census if not earlier.

Stock makes her argument on the basis that “Transitioning female transmen [sic] do not opt out of sex-based discrimination because they identify as men”. I agree that sex-based discrimination is an important area of data-gathering; however, Stock is not making an argument, but a petitio principii fallacy. She argues that trans men face discrimination as women, therefore they should not be counted in the census as men, so that we can accurately count discrimination against them as discrimination against women. The question begged is, do they face discrimination as women? Or as trans people? I want data to be able to demonstrate this one way or the other: the question cannot be decided by legislating census categories. Curiously, Stock also says “Late transitioning trans women, socialised as males, do not become subject to sex-based discrimination; what discrimination they receive has another cause.” The second petitio principii aside, it is odd that her evidence states that trans women face discrimination because they are trans, but trans men face discrimination because they are women: I am not sure on what evidential or philosophical basis she can make both claims simultaneously. But, philosophy aside, I agree entirely with her motivation, because I too want good statistics to protect women’s rights and trans rights both. The question is, how do we get these statistics? That is, what census data will give us good information about women and trans people?— inclusive of both intersex and endosex women, of both trans and cis women, and of trans men, trans women and nonbinary people. Asking a trans status question is one way to improve this information; how a nonbinary question can also improve the data is discussed in the next section.

[Edit 7/4/19. Three initial studies illustrate the complications here. Schilt and Wiswall 2008 found indicative results that trans women experienced pay discrimination as women, possibly even more so than cis women, while trans men experienced increased pay and less pay discrimination than women, though possibly not as much as cis men. Meanwhile, FORGE 2011 found that trans men and trans women both experienced sexual violence at roughly equal rates, and both at higher rates than cis women. Finally, Dhejne et al 2011 found indicative results that a pre-1989 cohort of trans men committed crimes more frequently than women, whereas the same cohort of trans women did not commit crimes less frequently — but that both effects were not present post-1989, when acceptance and social care increased, after which trans women committed crimes at the same rate as cis women. Notably, McAlpine has mis-cited the Dhejne study on several public occasions: this interview with the author helps clarify how and why. In short: how trans people experience sex-based oppression, and how this is expressed in statistics, is extremely complicated and should not be so easily simplified.]

The third issue is consistency with previous censuses, which Lucy Hunter Blackburn raises in her evidence. Such questions include: If trans men were counted in previous censuses as women, and are now counted as men, wouldn’t that create incompatibilities with datasets? If sex was previously counted as “biological” or “legal”, wouldn’t counting it on “self-identification” create incompatibilities? And if sex was previously counted on a purely binary basis, wouldn’t introducing a third option create incompatibilities? I will address the third option question below, as it is by far the most complex and important, but the first two questions are more easily dealt with.

It is crucial to understand that, in accordance with the legal and social norms recognised in the Equality Act, in the 2011 census trans people were already advised to answer on the basis of identity: “If you are transgender or transsexual, please select the option for the sex that you identify yourself as. You can select either ‘male’ or ‘female’, whichever you believe is correct, irrespective of the details recorded on your birth certificate. You do not need to have a Gender Recognition Certificate.” In other words, trans men were previously counted as men if they identified themselves as men, and trans women likewise. Therefore, it is changing this question and guidance that would create inconsistency. Oddly, this fact is already noted by Blackburn in her evidence, even though she says that “Moving explicitly to self-identification in 2021 would be expected therefore to cause a larger departure from legal/biological sex than in the previous census”. There is no move, and therefore no departure, and because legal/biological sex are not legal terms they were not used in the previous census: every element of the statement is factually incorrect.

I currently have a Freedom of Information Request underway to learn what question was asked and what guidance provided in previous censuses. My suspicion is that trans people have always answered by their self-identified sex: given the past and evolving legal situation, it is likely that no-one would or could have provided any guidance otherwise, and given the determination of all the trans people I know I would be surprised if most of them had answered otherwise. [Edit 4/7/19: This request is now complete and can be seen at the link. The 2001 census for England and Wales also advised trans people to answer by lived sex. No other census has defined sex “biologically” or “legally”, or provided specific advice to trans people. Answering sex by self-identity now has 20 years of census practice.]

So: the terms of the Census Bill cannot and should not affect Equality Act protections and terms; trans-inclusivity does not threaten data collection for women’s services but offers an opportunity to improve it; and self-identifying your sex was established in the 2011 census if not before and so does not produce inconsistencies now.

Sex and Gender Questions

As we have discussed, gender identity has no status in UK law, and no previous census has separated gender identity from sex. I do not yet know whether gender identity was explicitly considered a part of sex for census purposes prior to 2011, but I do know that Corbett v Corbett established that as a valid legal possibility. Given that the Equality Network, the Scottish Trans Alliance and Stonewall all do not want a gender identity question, and nor do the organisations which campaign for and administer women’s services, and that those giving evidence sceptical of trans rights do not think gender identity is a protected characteristic or relevant vector of discrimination and so have not advocated for the question either, I think on all these grounds we can say that a gender identity question should not be asked. The question has, however, been suggested by the Office for National Statistics for the census for England and Wales, so it is worth considering a little further, which I do in the section below on nonbinary people.

What should the sex question be? How should sex be defined in the census, given that it has no general definition in law? Any change from self-identification for trans people is categorically inconsistent with the previous census, and with the majority of social and legal convention, as well as risking the continuity of the data, and so would need a considerable justification. Every justification that I have seen is on philosophical grounds: Stock, Blackburn, Auchmuty, Freedman and McAlpine all believe that sex and gender identity are fundamentally separate and so should be administered separately. Some of them seem to think that trans people should only answer as their lived sex if they have a GRC (which, remember, would be a practice inconsistent with almost every legal area of sexed life); some of them seem to think that trans people should answer as their assigned sex at birth (which is likewise inconsistent, and could also only be a voluntary question, because forcing disclosure of gender reassignment status is unlawful under the Equality Act). I am happy to concede, as Engender do, that this is a fraught and difficult area of philosophy and politics. In so doing, I want to assert the centrality of the census as a data gathering exercise: it should gather data on the categories relevant to us now. It cannot and should not seek to legislate on philosophical and legal questions of the definition of sex; it is an utterly inappropriate forum for doing so. Nor can it pre-empt, on the basis of a political ideology — either McAlpine’s or mine — the outcome of an ongoing debate. What it can do is provide appropriate data to frame that debate.

I understand the concerns of those involved that differing social and economic histories of trans people may impact gender-disaggregated data if trans people continue to be included in their self-identified lived sex. Given the astonishing levels of poverty, sexual violence and discrimination faced by trans people, I would be very surprised if, for example, Stock’s putative “late-transitioning transwomen [sic]”, carrying all their purported male socialisation and privilege, could make the slightest dent in statistics on women’s income, rates of sexual violence and experience of discrimination. But maybe I’m wrong. Accurate data can tell us. And because the proposal is to include a “trans status” question alongside a sex question, it enables not only gender-disaggregated data but also trans-disaggregated data. That is, with this data gathering norm you will be able to separate out trans men from cis men when analysing the data. This provides a better statistical foundation for any data gathering exercise than any alternative and addresses the concern about any impact trans people may have on sex data: count us properly, and we can all see what the statistics say.

Muddying the waters with gender identity and “legal sex” would make this kind of analysis harder, not easier, because it would separate trans people into different legal groups and census categories that are harder to disaggregate from the general sex data. It would also create social absurdities: under this proposal trans men with a GRC and trans women without one would both be counted as trans and as men, and that trans women with a GRC and trans men without one would both be counted as trans and women. Moreover, the trans people more likely to access a GRC and be counted with their self-identified sex under such proposals are those more able to access economic resources — that is, those who look more like Stock’s imagined privileged trans women — and so would, if there indeed is any data distortion, distort the data more. Stock’s possible further implication that, for example, all trans men should answer as women (and, presumably, all intersex people who were assigned female at birth, regardless of any of their sex and gender characteristics or their lived sex) would produce an even greater statistical anomaly, incorporating a population who live as men for almost all legal and social purposes into the statistics for women. It would also entail making the sex question voluntary, because disclosing gender reassignment status cannot be mandatory under the Equality Act, as acknowledged by ONS’ research, and so would be the greatest risk to the data of all.

So: we should stick with prior legal and data-gathering convention by continuing to guide trans people to self-identify their lived sex, and ask an additional voluntary question on trans status so that data disaggregation is possible, enabling better data gathering norms and services for all women and trans people.

Intersex and Nonbinary Questions

Two groups not previously counted in censuses are intersex people and nonbinary people. Intersex people are “born with physical sex characteristics that don’t fit medical and social norms for female or male bodies” (Intersex Human Rights Australia). How many intersex people there are depends on the definition: McAlpine has cited 1 in 5500, which is dsdfamilies’ statistic specifically for births with genital difference, given their founding and core interest in this specific intersex category; dsdfamilies also acknowledge 1 in 60 statistic for a far broader definition; and the Intersex Society of North America provides a breakdown by different cases with some different numbers. Nonbinary people (like me) are people who are neither male nor female based on a range of sex and gender characteristics, including but not limited to gender identity and other forms of material and social experience. We do not know how many nonbinary people there are in the UK, due to the lack of a census question or fully rigorous research, but it may be around 1 in 250. A few but not all intersex people are also nonbinary (and vice versa), and many but not all nonbinary people are also trans (and vice versa), and so on. How and when should we count both of these groups?

I am not going to answer the question on how best to count intersex people, because I am not intersex and that conversation should be centred on intersex-led organisations. Trans people have a bad history of using intersex people as an example in an argument rather than centring intersex voices, and I do not want to continue that practice. I also regret that some muddy answers and poor understanding led recommendations in the Census Bill to falsely imply that all intersex people were nonbinary or trans, as pointed out by the Committee.

I will, however, speak against three things. First, in response to those slippages of language, the Committee report seems now to be implying that no intersex people are trans or nonbinary. This is false, as noted by ISNA, campaigned on by the Intersex Campaign for Equality, and discussed in this article. Second, Stock makes the extraordinary and unevidenced claim, repeated by McAlpine and others, that “Due to our best understanding of the categories, intersex females are still female; intersex males are still male”, and that these are fully determined and the only options. I do not know what that understanding is, what evidence that is based on, or who advocates for this position among intersex-led organisations. Given the primacy of “biological sex” in these comments, and that intersex people are precisely those whose “biology” is not clearly binary, I do not understand or recognise the point. Many intersex people are male, many are female, some are neither, and some of all of these categories are also trans, as discussed by all the intersex organisations referenced here. Third, Stock suggests speaking to dsdfamilies only, and this is the only organisation referenced by McAlpine and others in their commentary on the issue. While dsdfamilies is an important organisation, they are not an intersex-led advocacy or rights organisation but primarily a support and information group for families of intersex people. In their own evidence, they acknowledge that different intersex organisations and individuals of course have different approaches, and this can and should be taken into account.

As much as some trans campaigners have instrumentalised intersex people and talked over intersex advocacy, Stock and her allies in this discussion are doing the same. Everyone in this debate in Scotland so far is failing intersex people. We all need to look to diverse intersex rights and advocacy groups on this question. I have provided some resources in the links in the paragraph above; an exceptionally good resource is IHRA’s “Intersex for allies.” IHRA also offer one thorough approach to working with intersex people in forms, studies and surveys.

On nonbinary people I can and will speak. Whether nonbinary is a sex category, a gender category, a gender identity or something else is an ongoing debate, and the language is new and evolving. I believe that it is a sex category, as indicated in numerous international contexts, but I am happy to concede that the debate is continuing as our understanding of sex and gender develops and that, again, the census should not pre-empt such debates and legal statuses. We do, however, urgently need to better count nonbinary people, to make a better case for healthcare, for legal rights, and so forth. If we are not counted in this census, we seem to be in a Catch-22 situation, where we can’t get legal recognition because we don’t know how many of us there are, and we don’t know how many of us there are because we can’t get legal recognition.

There are two main options as I see it, and as discussed by both NRS and ONS. The first is to ask a separate question about whether or not the respondent is nonbinary, in the trans status question or otherwise, and the second is to include a third write-in box in the sex question, as suggested in the original recommendations to the Committee, in which nonbinary people can state their nonbinary status (including those intersex people who are nonbinary, as recommended by IHRA).

The problem with the first option, asking a separate question on nonbinary identity, is that it pre-empts the question of what nonbinary status is: it says, for both endosex and intersex nonbinary people, that nonbinary status is a gender identity but not a sex. Moreover, it would require nonbinary people to state a binary sex and that they are nonbinary. Aside from being a little insulting and pre-empting the question about what nonbinary status is, this is also likely to induce high rates of census non-compliance among nonbinary people. Worst of all, it separates gender identity from sex in a way likely to be highly prejudicial to trans men and trans women: it implies that their sex, too, is not their sex. Not only does this risk confusion, it again risks dividing the trans population into different census categories, which damages the data in the ways described above. Consider the options suggested by ONS, which require trans people to answer both what their sex is, whether their gender is the same as the sex registered at birth, and what their gender is, but not what their sex registered at birth is. Even I’m a little confused. It would take considerable guidance to ensure that trans people answer the sex question as well as the gender identity question by lived sex, and we could expect resulting confusions and uncertainty in the data. Moreover, given that the Committee and all evidence-givers has recommended not asking a gender identity question, none of ONS’s suggestions or the research into them can be applied.

The second option, a third write-in option box in the sex question, avoids these problems. It might seem that this suggestion also partially pre-empts the outcome of the debate. However, it would only imply that nonbinary status is a part of the broad category of sex, as it is currently used interchangeably in UK law with gender, rather than making any determination as to how different elements of this category could or should be treated differently socially and legally. That is, gender identity is in legal, social and census conventions treated as one element of sex, and the different aspects of sex have no general separation in UK law, so to include nonbinary status within the sex question does not pre-empt the debate as to whether it is a gender identity or something else. That said, even if there is some pre-emption in the case either way, we must choose the least prejudicial and most statistically useful option, which is to continue the established practice of self-identifying in the census.

However, the problem as noted by Blackburn is that a third option creates inconsistencies with previous censuses. This is fair, but I concur with Engender that, while nonbinary population numbers are very important to me and to our services, they are very small overall and are highly unlikely to impact broader data. I would add that, given that male and female population levels are fairly stable, it ought to be possible to disaggregate nonbinary people assigned male at birth from those assigned female at birth by an application of population levels for some statistical purposes, though this method would be limited. A related method is currently proposed by NRS, as discussed extensively in the Parliamentary debate, of distributing the non-binary population among the male and female categories by population level for some purposes; NRS recommended this as having absolutely minimal impact on the data quality, while still providing both inclusion and an overall population figure for nonbinary people. I would admit that I also have curiosity on statistical and outcome differences between nonbinary people based on sex assignment at birth, but given that no-one at all is suggesting that this question be asked, and indeed that it cannot legitimately be asked without risking privacy around gender reassignment status, I think that question would have to be left to separate, less politicised and more finely-tuned data gathering exercises. The broad point is that we are too small a population to have a significant effect on census sex data (which, remember, is not even used by women’s organisations when campaigning and providing services), but too significant a population not to count.

It is fair to say then that there are some risks to the data however we approach this problem. The decision must be made on which risks are less severe, and which are more acceptable in the interests of gathering accurate population data for social reforms. There is at present disagreement on this between the Office for National Statistics (which makes recommendations for England and Wales) and the National Records for Scotland. The former has recommended asking a separate gender identity question and a binary sex question, without making a recommendation about how trans people should answer the sex question; the latter has recommended asking a nonbinary sex question and a trans status question, but not a gender identity question. Both have conducted testing, and I hope everyone in this discussion can agree that further testing is essential. I have argued strongly in favour of the National Records for Scotland’s position on a legal basis, and on the basis of their own arguments on statistical compatibility and respondence rates, but also on the basis of Scotland’s own equalities priorities: to ensure the better treatment and recognition of trans and nonbinary people in society. I believe that the legal case for this conclusion is very strong indeed, and that the statistical case is also secure, and that all of this is underlined and strengthened by Scotland’s commitment to trans and nonbinary people.

To Conclude (This Bit)

The two questions that should be asked in the Scottish Census are sex and, on a voluntary basis, trans status. Guidance on the sex question to trans people should be consistent with previous years and the majority of legal situations, and advise answering on the basis of self-identifying sex. An optional third write-in category should be provided in the sex question, with guidance for and about nonbinary people. Better consultation with intersex people and their own organisations about appropriate intersex questions and whether or not to include them is vital.

This brings us to the end of my discussion of sex and the Scottish Census. I want to conclude with some thoughts addressed to my comrades in trans rights work, inspired by this research and argument.

Photo by Glenn Euloth on Flickr, licensed under Creative Commons. Two young people walk under a multicoloured umbrella; one has yellow hair and a black shirt; the other has a green cap and a white shirt.

On Sex and Gender Identity for Trans Rights Campaigners

My investigation of this issue has made it very clear to me that those who oppose trans rights are using, extensively, the rhetorical separation of gender identity from sex. They state, in many ways, that sex is the real thing that is assigned at birth and is the basis of all discrimination, and that gender identity is this other not-really-real thing, and that trans people might have the gender identity they say they do but that they don’t have the sex. We know that this is practically, legally, socially, philosophically and politically false, and based on a foreshortened critique of how gender and sex actually operate in the world.

Unpacking all this is well beyond the scope of this article, but I would urge trans people very strongly to read their own history and canon of radical thought, from grounding texts like Sylvia Rivera’s Christopher Street Liberation Day speech, Leslie Feinberg’s Transgender Liberation and Sandy Stone’s Posttransexual Manifesto, through the considerable catalogue of Transgender Studies Quarterly, with a particular focus on the “Decolonizing the Transgender Imaginary” and “The Trans*-ness of Blackness” issues, to contemporary radical trans thinkers with significant work freely available online including Alyson Escalante, Jules Joanne Gleeson, and Nat Raha. We spend too long reinventing the wheel, and not enough time reading the work that has already been done and thinking about it. If you have financial or other accessibility barriers to reading any of this work, please get in touch and I’ll try to help.

We also know that this clear-cut sex and gender identity distinction has no legal basis, as discussed here, but I think we can expect those opposed to our rights to begin making legal moves to try and establish this distinction in law, as a means of separating trans people from our sex-based rights, and from our protection against discrimination. This is a dangerous moment and one that should be watched carefully.

Alongside that, I would advise trans people to take great care in how they use “sex” and “gender identity” in our politics. The distinction has been important for our politics as a means of decoupling our lived experience from what we were assigned at birth, but our politics has also always been about understanding that distinction as complex and contingent. I am worried that we too are separating sex from gender identity too easily and habitually. Naming the components of sex and gender, from chromosomes to breasts, from identity to hormones, from gonads to social roles, is useful insofar as it strengthens the organisation of — to reappropriate a term from so-called “gender critical feminists”, or rather to acknowledge that gender critical feminism as it operates in the UK is a severely truncated pastiche of the plural transfeminism that has helped animate women’s movements since before women’s movements were named — women and gender minorities as a sex class. Identifying some of the differences between us should strengthen the solidarities between people as wonderfully various as nonbinary trans women, intersex cis women, transmasculine demibois assigned-female-at-birth, and so on. What we should not do is reify and police these differences in theory, practice or the law, whether it’s as Harry Benjamin Syndrome-identified transsexuals organising with those who oppose trans rights, or well-meaning cis allies proliferating endless taxonomical genderbread people across the internet. Our analysis should enliven our solidarity, not police its borders.

Sex was always a social identity, and gender was always a material situation, intimately bound up with breathing bodies, with physical violence, with economics. Gender identity shapes the distribution of wealth, gonads shape the distribution of health, presentation shapes the distribution of assault, genitals shape the distribution of care — and all of these processes are intertwined, all of sex and gender is intertwined. This materiosocial analysis has to be at the centre of our politics. Healthcare, economic freedom, transformative justice, social liberation. Ultimately, I want to be counted not so that I can feel the government says I count, but to make myself count: to access the health, the resources and the solidarity needed for all our liberation.

Closing Notes

Thanks to Christine Burns, David Allsop, Jules Joanne Gleeson and Rosa Campbell for conversations on the law that helped clarify my understanding. Thanks to Jesse McAulay for tracking down the guidance on the 2011 census. Thanks to Jane Ditchfield, B D Owens, Morvern Cunningham and Sally Evans for helpful comments on drafts of this article. All of the mistakes in this article are mine and mine alone. I am grateful for any corrections and additional information people have to offer, particularly on the case and statutory law.

I have limited time and resources, but am happy to discuss the matter of this article with people who disagree with me, however vehemently, as long as you’re willing to not throw slurs at me, not expose me to a harassing audience, and extend a basic level of respect, which I will extend to you. Discussing these matters online has drawn me a lot of violent abuse in the past, and I operate a light-fingered blocking policy. I am particularly keen, though, to talk these things through with people with whom I already have social relationships, or forms of existing political solidarity and mutual support, because I do not want to cast you out: I want us to be able to be in community together. Even if we do not come to agreement, I think we can learn together.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade