Safety, fairness, and inclusion: transgender athletes and the essence of Rugby

ABSTRACT

In this paper, I link philosophical discussion of policies for trans inclusion or exclusion, to a method of policy making. I address the relationship between concerns about safety, fairness, and inclusion in policy making about the inclusion of transwomen athletes into women’s sport. I argue for an approach based on lexical priority rather than simple ‘balancing’, considering the different values in a specific order. I present justifying reasons for this approach and this lexical order, based on the special obligations of International Federations such as World Rugby. As a result, I provide a justificatory framework for the WR Guidelines that exclude transwomen from the women’s game in WR competitions. Finally, I give an account of a maximally safe, maximally fair and maximally inclusive form of sex categorisation in sport.

KEYWORDS:

• Safety
• fairness
• inclusion
• trans athletes
• World Rugby
• pluralism
• local essentialism
• policymaking
• applied ethics

Ethical regulation and trans athletes: an introduction

There is a growing literature in which philosophers grapple with the question of the inclusion of transgender athletes into sex segregated sport (Bianchi 2017, 2018; Coggon, Hammond, and Holm 2008; Devine 2018; Gleaves and Lehrbach 2016; Knox, Anderson, and Heather 2019; Koenigsberger 2017; Sailors 2013, 2020; Teetzel 2013). However, the question of how, exactly, to make policy in this area has received less attention. This paper, which draws on the author’s own involvement in policy making in this area, is an attempt to begin to fill that gap.¹ It also offers a justification of the first regulatory move specifically to exclude transgender women from a particular level of a particular competitive women’s sport: the Transgender Guideline of World Rugby (World Rugby Working Group 2020)

Through 2020 World Rugby (the International Federation (IF) governing Rugby Union) embarked on a process of investigation and policy formation, which is not yet complete. Rugby’s position is not unique, but it is special. The nature of rugby as a ‘combat’ sport, like Australian Rules football and American Football, raises safety considerations that are not present in other sports. Hence, the specific relevance of the issue for Rugby, highlighted by cases in which it was suggested that trans athletes competing in women’s rugby posed an unacceptable risk. (see, for example (Coleman-Phillips BBC Sport 22 August 2019), where it is said of the Welsh trans player Kelly Morgan that she folds up opposition players ‘like a deckchair’.)

In many ways, (and viewing it partly from the inside) the process in World Rugby seems exemplary, involving key stakeholders and experts in face to face presentation and discussion. Trans players and trans advocates were included in the discussion, as were women’s rights groups. Amongst the experts who were consulted were trans health specialists, lawyers and insurers, experts in sports science, in developmental biology, in human rights, and in ethics. The drafting of the new transgender guideline was delegated to a working group made up almost entirely by players or former players of the women’s game. The process was remarkably transparent: all the presentations were placed online, and the draft guideline was published on 9 October 2020 after its adoption by World Rugby’s ExCo. There was and is significant media interest in the policy, and this is ongoing at the time of writing. This particular policy development process is important because World Rugby is the first IF to adopt an approach that specifically and explicitly excludes transwomen from competing in elite women’s sport: in this case, rugby at a World Cup level (that is, in tournaments organised by World Rugby itself, not by national governing bodies)

Most of those who write in this area quite quickly realise that it is one in which deep values conflict. But this conflict is not always made explicit: arguments about testosterone levels kick in before basic assumptions are declared, and this can lead to a muddled debate. There are exceptions; for example, Pam Sailors makes it clear that her position comes from a mutualist perspective, and an assumption that fairness is fundamental:

I think there is a strong case to be made that fairness in physical competition is the fundamental value in sport, perhaps even a prerequisite for the existence of sport. If we think of sport as a mutual quest for excellence, then participants should have a reasonable chance of winning. Without such a chance, a contest is neither competitive nor fair to the participants who are certain to lose. (Sailors 2020, 2)

In contrast, the advocates of trans inclusion, John Gleaves and Tim Lehrbach give a radically different account of the nature of sport. They suggest that:

the rationale for inclusion of transgender and intersex athletes must move beyond the idea of fairness. … We have argued that a better rationale emphasizes that sport is about meaningful narratives and that gendered narratives constitute at least one type, and perhaps the most significant type, of sport’s meaningful narratives. (Gleaves and Lehrbach 2016, 14)

It may be difficult to see a way of reconciling these deep commitments about the nature of sport. It may be that these are just clashing paradigms, incommensurable ‘ethical conceptions of sport’ which show us when we have reached the end of rational disagreement and can only bang the table (Morgan 2020). I lack space here to enter that argument, but it is worth noting a practical upshot: this would appear to leave philosophers of sport hamstrung in their attempts to influence sport regulators – and so hamstrung in any attempt to make sport better, or, more pessimistically, to stop it getting any worse. But there may be an alternative approach based on a different understanding of sport. I attempt to illustrate this in what follows.

Balancing vs. lexical priority

A common approach to this policy issue, with some intuitive appeal, is the idea of ‘balancing’ different considerations. Indeed, this is often the default position. Proposals for trans inclusion in women’s sport seem to raise questions about safety, fairness and inclusion. So, it is fairly standard to read pleas that these values are balanced against each other. For examples of this, see John William Devine (Devine 2018, 166) who urges that:

Even if some unfairness does arise where returning dopers and trans women are deemed eligible to compete, fairness is not the only value. Fairness in competition must often be balanced against other competing values, most notably in this case, ‘inclusivity’. In addition to ensuring that competition is fair, we may wish to ensure that sport is inclusive of all gender identities. (JP italics)²

In a similar vein, Sarah Teetzel argues that:

There are several methods for evaluating the moral acceptability of rules. In setting eligibility rules for participation, sport governing bodies have to weigh the desire for fairness with other considerations, including practical considerations and economic implications. (JP italics) (Teetzel 2013)

In the same way, the Canadian Centre for Ethics in Sport evinces considerable enthusiasm for the balancing metaphor:

The Expert Working Group further sought a balance among the seven True Sport Principles. Sport, at its best, balances a series of principles to create a fair, safe, inclusive and open environment. Depending on the level of sport, these principles need to exist in different degrees and proportions. Creating good sport policy is about finding a balance between these principles while keeping all of them present in a way that makes sport a positive experience for everyone. (Kidd et al. 2016, 7)

There are three problems with this approach. First, a balancing approach might be promising if the various considerations could be expressed in the same commensurable and comparable units, and if cost-benefit analysis were thought to be the most appropriate methodology. Cost-benefit analysis can usefully illuminate trade-offs between different comparable goods. However, it seems clear that these particular goods cannot be compared in this way. What amount of ‘fairness’ ought to be sacrificed for what amount of increase in ‘inclusion’? Second, this practical problem arises from and expresses a prior assumption that decision-makers are entitled to trade off different goods – that, for example, they are entitled to trade off injury risk or unfairness against inclusivity. An alternative ethical conception – and a more compelling one – is that authorities have basic duties that cannot permissibly be traded off in this way. There are far better approaches available in applied ethics: the ideas of lexical priority, threshold deontology,³ priotarianism,⁴ the liability to harm,⁵ are all available to decision-makers who are fully informed. Third, importantly, because it cannot be concretised, this allows regulators to regulate, yet with a completely opaque process. This introduces an element of ad hocery into the process: ‘How did you decide? Oh, well we balanced up various considerations!’ Unsurprisingly, this is not explanatory, because we are faced by plural values. In this case, at least three sets of values are in play: first, the safety of competitors and whether they are confronted by unacceptable risk, second, the fairness of the sport, and third, the inclusion of all who wish to play the sport in a way in which they wish to be included. Should the existence of these plural values be a cause for concern? Michael Stocker thinks not:

My conclusions are these: plural considerations as such are not problematic. They do not force us simply to plump for one option or another. We are often enough able to form and use higher‐level synthesizing categories … I think these are undeniable and obvious facts. (Stocker 1990, 194)

For the case under analysis, we ought then to seek ‘higher level synthesizing categories’: what is needed is a method which allows us to generate policy. How can we develop these higher-level synthesizing categories, to determine how safety, fairness, and inclusivity relate?

At first sight, inclusivity is a thin concept, in the sense that it does not carry, of itself positive or negative valence (in contrast to safety and fairness, which are thick concepts). We are not obliged to include the Covid-19 virus in the diversity of nature, nor losing semi-finalists in the final of the Rugby World Cup. Nonetheless, inclusion of persons is generally taken to be a value of some sort and we ought to determine its right place.

To do this, I want to invoke the notion of ‘lexical priority’. This is an idea drawn from Rawls ([1971] 1999, 37–8). The term ‘lexical’ here comes from the idea of a dictionary, we sort words by their first letter: acorn before bacon, then by their second letter: aardvark before acorn, and so on. Sorting items lexically means organising them according to one principle, then a second, then a third, and so on. But it can also be thought of like a filter system, or a decision tree. It is this notion of lexical priority that is basic to Rawls’ criticism of utilitarianism, and which sets public policy on the right track, against the Government House Utilitarianism so effectively criticised by Williams (Williams 1985, 108). It meshes well with the idea of the priority of the Right over the Good, a theme of mainstream liberal thought since W.D. Ross ([1930] 2002). In urging International Federations and other regulatory bodies to adopt an approach of lexical priority, I urge them to get up to date with twentieth century moral and political philosophy.

Well-advised IFs should do the following: Consider three values: safety, fairness, inclusion in that order. All of them, let us say, are good to some extent. But what order should they come in? Let us first assume that these values are, analytically speaking, simples rather than complexes. Safety comes first, fairness second, and inclusion third. So, IFs ought not to adopt fair rules that are unsafe. Of the possible sets of rules, we want only those that are safe. Of the set of safe rules, we want those that are also fair. Of the set of safe and fair rules, we want those that are also inclusive.⁶

What IFs must not do, therefore, is ‘trade off’, or ‘balance’ fairness against inclusion. This general approach needs justification and modification. In the next section, I will attempt to justify it, and in the last section, I will modify it.

Justifying the lexical approach

In justifying the lexical approach as a matter of methodology, it is first important to look at who is making the policy. Sports policy is a matter for governing bodies – in particular IFs – and we ought to specify their ethical (not merely legal) duties. I take it that IFs are interested not only in being compliant with the law, but also in constructing ethically sound policy. In addressing these duties, it is useful to know about a distinction that is made in the philosophical literature between special duties and general duties (Dworkin 1986)

Suppose we think of an IF – in this case World Rugby – as a sort of person, with a role. World Rugby is a corporate person in a special role. This person finds itself in a network with lots of individuals. These special obligations arise because we are in networks – like families. Because of these networks, agents have special obligations to some persons that they do not have to all persons. World Rugby in this sense is like a family member, with special obligations to other family members. Who?

It is helpful to think of this as a matter of concentric circles around a person (so I might have special duties first to my daughters, because they are especially vulnerable as children, then to close family who are not vulnerable in this way, then to friends and relatives, and in a different way to colleagues, and students, perhaps to people in my neighbourhood and so on) (Keller 2013; Scheffler 2001)

In the same way, WR has duties to the game itself (Butcher and Schneider 1998) to the practice of playing rugby, then to individual active rugby players, and to clubs, affiliated through national federations, then WR has obligations to potential future rugby players, those who might play rugby. Perhaps it has obligations to past rugby players,⁷ to ex-players, to fans, then potential fans and so on. Last, WR has obligations to the non-rugby playing, non-rugby watching public. The image is of ripples of responsibility going out like concentric circles. Duties are less and less demanding as we move through the outer rings.

World Rugby has strong special duties to the rugby family, which are more important than its general duties Why are they more important? One reason is that if World Rugby does not put rugby first, nobody else will. World Rugby is uniquely fitted for the role of protecting rugby and rugby players: it is not uniquely placed for the role of securing a world free of discrimination against trans people. I have special obligations to my children because that is part of my role as their father, to put the interests of my children first. Equally, it is part of the role of WR to put the interests of rugby, the game, its players, and its fans first, and before its obligations to the wider public. There are a number of ways to ground these obligations: one is to say that this is how human beings are – to appeal to a Humean account of the passions: another is to incorporate that into a more sophisticated view such as Railton’s (1984) – that such particular obligations can be generative of the general good. But there is another angle which arises from the metaphysics of a particular sport. Suppose we understand rugby in terms of its essential actions, first amongst these being tackling. In normal life, the practice of running at someone, grabbing them around the torso, and pushing them to the ground is a kind of assault. If I did that on the High Street, I would be arrested, because this action is likely to cause harm to the person who I run at, grapple with, and force to the ground (hereafter ‘tackle’). The practice of rugby makes this different: tackling actions are constitutive of rugby (in most of its forms) and are essential to it. The institution governing this practice⁸ – World Rugby – is therefore, and peculiarly, a tackling-permitting institution (or TPI). It is special, and peculiar to TPIs (others might be the NFL and the AFL Commission) that they permit tackling, unlike the rest of the social world. My claim is that TPIs have a special obligation to limit, to a reasonable level, the risks that arise from tackling. It is because they allow tackling that they have a special and intense obligation to limit the risks of tackling. This helps to explain the focus and seriousness with which WR has treated the issue of concussion over the last few years, investigating, researching, amending rules and changing its approach (Tucker 2017a, 2017b; Tucker et al. 2017). But it only partially helps. My point is not that those internal to a practice are the most attuned to the ethical risks of that practice: because of the risks of institutional conservatism, they might not be. It is that they ought to be, because they have special responsibilities in this regard.

This helps us to see a little more clearly why safety risk, especially with respect to tackling and concussion, must be central to the concerns of World Rugby, trumping other values. Risk is tolerable in relation to the essence of the game – what is, and is not necessary to the game being the game it is. If some feature of a sport must be there, for the sport to be the sport it is, then the associated risks are more tolerable. If some feature is merely contingent, then the associated risks from that feature are less tolerable. Of course, rugby, with tackling as part of its core, is not a completely safe game. Its participants do not have a right to absolute safety on the field of play, because such a right is incompatible with the existence of the game. But the argumentative move that because there is some risk there need be no justification for an increased risk is surely discreditable. Rather, the opposite is true: because there is some risk, it is particularly incumbent on World Rugby to be alert to increased risk, and to oppose any increased risk that is not an ineliminable part of the essence of the game.

In this way, the lexical priority of safety over other criteria arises from an account of the nature of rugby – its (local) essence (Pike 2018). On this account, different sports have different individuating actions, so that (e.g.) Rugby is essentially a tackling game (I exclude touch rugby here). These actions ‘fill in’ – give substance to – the special obligations of IFs to their members. By analogy, World Sailing (WS) has a special obligation to protect its members from drowning by specifying safety equipment, by prescribing training courses, and so on. Drowning is a risk that follows from an essential component of sailing – that it takes place at sea. The special obligation arises from the particular corporate role of WS, and it is particularly important that it ought not to be cavalier about water safety. Likewise, World Rugby: a cavalier attitude to concussion risk would be a central abdication of its special and particular duty (whilst it can be relatively relaxed about safety on the water). As a Hohfeldian correlate of this special duty (Hohfeld 2001), rugby players have a right against their governing body that it ensures personal protection congruent with the essential actions of the sport. This is a right owed to players by their governing body because it has special – intense and focussed – obligations to players in this area.

The account above rests in part on an overall essentialist theory of sport. But I think that a theory of i) bodies first, difficult actions, and competition rather than ii) ‘narrative construction’ or iii) ‘social moral conventions,’ is right, though I do not have the space to argue that out here. A pro tem justification of the lexical approach, then, is that lexical priority gives us the way to make sense of and order the values that are implicated if we are to preserve the distinctive nature of competitive sport.

Modifying the lexical approach

This account of Rugby as a sport in which a potentially dangerous action is permitted links to a proposed modification of the lexical approach. Here is an objection, which we have already met in a different form. Rugby is not a completely safe sport, so safety is not an absolute, overriding value: if it was, no one would be allowed to play. So, we need to meet a threshold where the sport is not completely but tolerably safe.

And, perhaps it might be said, by analogy we ought to seek rules that are not completely fair, but only tolerably fair. But this is not a good parallel. It is essential to rugby that it is a combat sport. An absolute and overriding commitment to safety – in the sense of eliminating all risk – is incompatible with the essence of rugby.

Is the same true of fairness? I think not. No-one would say ‘if we try to make rugby too fair, it will cease to be the same game.’⁹ The fact of rugby being fair, and initiatives to make rugby more fair by pushing the fairness envelope do not tend to detract from or undermine the essence of the game of rugby. After safety, fairness seems to have an overriding status. This, at least, is the approach taken by the IOC in its consensus statement:

“The overriding sporting objective is and remains the guarantee of fair competition. Restrictions on participation are appropriate to the extent that they are necessary and proportionate to the achievement of that objective.” (My italics) (International Olympic Committee 2015)

The important thing to see here is the difference between an overriding aim, and balancing two values, fairness and inclusion, against each other. So, fair competition can override inclusion: we ‘do’ fair competition first, and then inclusivity within, and after that.

Having justified the lexical approach, and having discussed some suggested modifications of it, I now move on to look at operationalising it, by fitting it together with scientific and empirical findings to generate policy recommendations.

Conditionals and empirical evidence

If the approach I have sketched is right, then it points to another technique to be pulled out from the philosophical toolbox: the construction of conditionals. Ethical policy decisions are (happily) not the sole responsibility of ethicists. Rather, it is the job of ethicists to fit together pieces of a puzzle, including fitting together empirical work, and conditionals play a large role in this. The empirical work is not within the expertise of philosophers, and we should not try to do it. But we can construct conditionals which show how the upshot of empirical work might fit into the policy making process (Radcliffe Richards 2000, 87–99). In the investigation of World Rugby into transgender policy, two pieces of empirical data were particularly relevant, and fit with the two lexically prior criteria which I have justified above. The first concerns the safety of players, and the second concerns the fairness of the sport. Here are two such conditionals, which make the considerations above more concrete and operable:

The Safety Conditional: If there is an [intolerable] increased risk of injury from the inclusion of transwomen in the women’s game, then it is unsafe for transwomen to compete in the women’s category.

The Fairness Conditional: If [intolerable] male advantage remains after Hormone treatment, then it is unfair for transwomen to compete in the women’s category.

If it is unsafe or unfair for transwomen to compete in the women’s game, then this ought not to be allowed, because of the lexical priority of safety and, then, fairness. The two conditionals point us to the empirical work necessary to make a decision, because the empirical work is contained in the antecedent of the conditional. That is to say, the empirical work needs to address the questions: Is there an increased risk of injury from the inclusion of transwomen in the women’s game? Does male advantage remain after the use of cross sex hormones? Then there are two further questions: if the answer is yes, then are the empirical findings such that these outcomes are intolerable?

In the World Rugby process these were precisely the questions that were addressed, and, indeed, the science shows that the answer to the two questions is yes. It is the case that the inclusion of transwomen in the women’s game leads to increased risk, and it is the case that male advantage remains after HRT. The Tucker study, specially commissioned for the working group, concluded:

While there is overlap in variables such as mass, strength speed and the resultant kinetic and kinematic forces we have modelled to explore the risk factors, the situation where a typical player with male characteristics tackles a typical player with female characteristics creates a minimum of 20%-30% greater risk for those female players. In the event of smaller female players being exposed to that risk, of larger male players acting as opponents, the risk increases significantly, and may reach levels twice as large. (World Rugby Working Group 2020, 14)

In the Hilton and Lundberg study presented to World Rugby, the authors concluded that:

Longitudinal studies examining the effects of testosterone suppression on muscle mass and strength in transgender women consistently show very modest changes, where the loss of lean body mass, muscle area and strength typically amounts to approximately 5% after 1 year of treatment. Thus, current evidence shows that the biological advantage enjoyed by transgender women is only minimally reduced when testosterone is suppressed. Sports organizations may therefore be compelled to reassess current policies regarding participation of transgender women in the female category of sport. (Hilton and Lundberg 2020)

Here, then, we have clear answers to the two antecedents of the Safety Conditional and the Fairness Conditional, leaving aside for the moment the square bracketed evaluative claim. It is clear, on the basis of the empirical evidence presented that risk is substantially increased, and that biological advantage arising from androgenisation is only minimally reduced. It looks as if the inclusion of transwomen in women’s rugby is neither fair nor safe. But is it intolerably unsafe, or intolerably unfair?

There is much to say here, but a feature of the World Rugby process was the seriousness with which it took this question. Posing the question in this way invites two responses: first, what would make a policy intolerable and second; intolerable to who? And this qualification invites the thought that the concerned parties here are those who are to bear the cost of the increased risk and increased unfairness: current female players of the women’s game.

To address the first of these: the policy seeks to distribute harms – in this case the harms of increased risk and the harms of unfairness. If we are to consider justice in the distribution of harms we can draw on a general approach in the philosophical literature: to assess out who is morally liable to harm. This is a further shift away from the cost-benefit analysis approach associated with the ‘balancing’ metaphor. It also helps to explain why the ethical considerations around transwoman playing in the women’s game and those around transmen playing in the men’s game are asymmetrical, as the World Rugby guidelines recognises. We can ask who has acted in such a way as to make themselves a ‘legitimate target’, perhaps by acting unjustly themselves or in some way alienating their right not to be harmed. It seems clear that female rugby players have not acted in such a way that it is reasonable to visit harms or potential harms upon them (regardless of the size or intensity of those harms). But if female rugby players have done nothing to deserve the harms of unfairness or the harms of increased risk, then it cannot be reasonable or tolerable to visit those harms upon them.

The second question: ‘intolerable to who?’ points us at a similar issue. Though there have been elaborate attempts to discount the actual views of current female athletes from this debate (Ivy and Conrad 2018) those attempts unravel. The survey data presented to the World Rugby Working Group (World Rugby 2020) showed clearly that elite women players who were fully informed on the two conditionals outlined above strongly opposed trans inclusion in women’s rugby. The datum that they found it intolerable is of significance in policy making.

A different approach

The regulation and division of sport ought to reflect, and be written around bodies. Hence, it ought to recognise male advantage, and protect female sport from those with male advantage. But this regulation should be substantially uninterested in gender identity. What counts is not identity but male physiological advantage. In order to centre that advantage in regulation, there ought to be some relatively minor rejigging of categories. Neither should require an affirmation, or denial of any particular gender identity: the entry conditions for each class are independent of gender identity. First, then, I propose a ‘Protected’ or ‘Closed’ or ‘Non-Androgenised (NA)’ category of sport, which excludes everyone with male advantage, including residual male advantage. Physiological male advantage depends largely on androgenisation, so this class would exclude all those who had been through male puberty. It is optional whether those who lack male advantage compete in this category, but they are welcome to do so – for example, socially transitioning transmen. Second, I propose that there should be an ‘Open’ category, in which everyone, and anyone is able to compete (with appropriate age restrictions). This category would be open to male and female bodied athletes, to transmen and transwomen: it would be neither a gender-defined category, nor a sex-defined category. This is a matter of eligibility criteria only, and of course, not of team selection criteria, competition criteria, or anything else. Because the protected class is justified on the basis of the exclusion of those with male advantage, it will exclude those with residual male advantage. There is a clear flaw, I think in most trans inclusive arguments: they seem to rest on claims that male physiological advantage does not matter. This is either (with radical implausibility) because it is entirely socially constructed, or (falsely) that it is, or can be, eliminated by Testosterone suppression, or for some other reason – that gender identity matters more than male advantage. But if male advantage does not matter, why have sex segregated sport in the first place? That those excluded from the closed category might define themselves as women or as transwomen is beside the point, since the people are not excluded from a category on the basis of some claim about what gender they are, or identify as (or are not, or do not identify as) but on the basic physiological claim about what advantages particular persons possess, by virtue of the path of physiological development that they have taken.

In this way, rules can be designed that are maximally fair and maximally inclusive. They require adjustment of both regulations and attitudes. For example, they require that male athletes open up their categories to any transwomen, (and any women) who want to compete in an open category. Instituting these categories would be ‘maximally inclusive’: it would enable everyone who is currently able to compete, to compete in a category that provides them access to fair sport.¹⁰

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. This paper draws on material presented to the World Rugby Symposium on Transgender Athletes in Rugby, on 25–26 February 2020. I would like to thank the organisers and all the participants at that workshop for their contributions. In addition, I have discussed these matters with many colleagues including John William Devine, Emma Hilton, Miroslav Imbrišević, Tommy Lundberg, Emily Ryall, Kathleen Stock, Ross Tucker, Nicola Williams, and Christopher Yorke. I am grateful to others who have contributed to my thinking in discussion, who, understandably, but in an appalling reflection of the state of open philosophical discussion of these matters, wish to remain anonymous. I am grateful for comments received from two anonymous reviewers for the Journal of the Philosophy of Sport, and from the editor, Paul Gaffney. Of course, but also importantly, none of them are responsible in any way for the arguments presented here.

2. Whilst Devine uses the balancing metaphor, it may be that he is committed to a more nuanced approach, because of the words ‘in addition’: this might betoken a threshold or lexical approach such as the one I advocate here: viz once we have assured ourselves, that competition is fair, we can attend to inclusivity, But, if so, this is not a balancing approach.

3. That certain duties are binding, and cannot be traded off, at least up to a threshold Alexander and Michael (2016).

4. That certain interests take priority, and ought to be met before other interests come into play (Parfit 1997).

5. That if harms must be distributed, they ought to be distributed according to liability to harm, and not arbitrarily (McMahan 2009).

6. It might be argued (and indeed is, by an anonymous reviewer for this journal) that the important cut is between safety and fairness on one hand, and inclusion on the other, since both safety and fairness are thick concepts, essential to the game, whereas inclusion is both thin and contingent. Why, then, prioritise safety over fairness? I have an objection and a concession to make in response. The objection lies in the seriousness of the consequences of breaches of the value of safety and fairness respectively. Notoriously, rugby injuries – concussion and spinal injuries – can be fatal or life changing. Unfairness, which distorts, spoils and devalues the game on the pitch, does not have such consequences off the pitch. But the concessive move is to accept that we need to have rules that are both safe and fair. Whether we achieve this by selecting safe rules that are also fair, or fair rules that are also safe, this will amount to the same set of rules, and need not be controversial.

7. Such as, to maintain the archive of the sport

8. For the practice/institution distinction, see, of course, Macintyre (1981).

9. I concede that something close to this is being said about the introduction of VAR in football, but that is importantly not analogous to the case we are considering.

10. There is one current restriction here, which should be acknowledged. Exogenous testosterone is on the proscribed list. Athletes using exogenous testosterone are therefore not allowed to compete in WADA governed sport. This applies to all athletes, and is a feature of anti-doping rules rather than sex segregation rules. These rules apply except in the case of a Therapeutic Use Exemption (WADA 2019) However, there is a possible avenue through this, which arises from the qualifying conditions for a retrospective TUE in conditions that are ‘manifestly unfair’, and it could be the case that a female bodied athlete using testosterone (a transman in medical transition) could qualify into the ‘open’ category in this area, because their use of exogenous testosterone would (manifestly) not overcome their pre-existing female disadvantage, but exogenous testosterone would exclude them from the closed category. In this case, it would be possible to see the outline classification above as maximally inclusive and maximally protective of female sport.