Sex has at least five axes. Eligibility rules and reference ranges have one field.

Writing in The Conversation, a biologist argues that human sex-related biological traits are neither binary nor arranged along a single spectrum, and that describing one person accurately can require multiple sliding scales rather than one position on one line ^[1][4][10]. Strip out the culture-war framing and what is left is a claim about measurement, and measurement claims are the ones that end up inside eligibility rules, laboratory reference ranges and trial enrollment criteria.

The argument is built in steps. To call sex binary, the author writes, is to assert that humans are sexually dimorphic: that female and male distributions of sex-related traits do not overlap, so that plotting a trait such as amount of coarse body hair would produce two separate bell curves ^[2]. The counterclaim, which the author says biologists and physicians have long known, is that there is overlap in every sex-related trait, so a given person can be female-typical in one, male-typical in another and intermediate in a third ^[3][4].

The move from spectrum to multidimensional is the part that matters operationally. Genital development, the essay notes, tracks androgen exposure and receptivity continuously, producing intermediate structures at intermediate levels ^[7][8], which the author says fits a color-spectrum analogy of one axis running red to violet ^[9]. But the other primary traits that get used as sex markers - chromosomes, gonads, hormones and hormone receptors, internal reproductive ducts, external genitals - can each point in different directions in the same body ^[6][9][3]. The example given is complete androgen insensitivity with a Y chromosome: testes in the abdomen, female-typical genitals, and after puberty breasts, a high-pitched voice and no coarse body hair ^[10]. One number cannot carry that.

The definitions in working use already disagree. Biologists usually define an animal's sex by gamete size, small for sperm and large for eggs, produced by testes or ovaries; physicians assign sex at birth from genitals, which may not match gonadal sex ^[12]. Because chromosomes, gonads and genitals can disagree within one person, the essay argues it is not clear that any one of them alone should set the category ^[13]. Note the size of the resulting uncertainty: intersex variations are put at 0.2% to 6% of the population depending on which variations are counted ^[11]. That is a thirtyfold spread ^[1], and it is not instrument noise. It is case definition doing the work, which is the same failure mode as any cohort defined by an unstated inclusion rule.

The clearest evidence that this is unsettled in practice is regulatory. Regulators of elite women's athletics including the International Olympic Committee have changed their sex tests repeatedly over the past century, from genitals to chromosomes to testosterone levels, and in March 2026 the IOC returned to chromosomes ^[15][16] - at least three switches of the operative variable ^[2]. Each switch redefines the population, and therefore who is in and who is out, without anyone's biology changing.

Two cautions. The essay is a single-authored argument drawn from the author's own book, "The Binary Delusion" ^[5], and it does not address clinical trials or lab reference ranges; that extension is mine. What follows from it, if the trait-overlap claim holds ^[3], is narrow and testable: a rule or a range labelled male or female is a rule about one measured variable, and it should say which one.