Gender metamerism

Some people like to talk about gender as a spectrum. So I’ve been thinking about this idea of gender metamerism.

Think color metamerism, not earthworm metamerism. Genders, like colors, are the outcome of perceptual processes, and only are a tiny fraction of the reality of what we see. Since these are likely two of the most misunderstood topics ever, I wanted to have some fun exploring both at the same time in a convoluted way.

Let’s go through some of the basics of color perception first, because it’s not a topic that is familiar to everyone, and color metamerism is kind of a technicality. Not much background is needed.

Color perception 101

So, you know that the retina is a sensor of light. Light is electromagnetic radiation, which for the sake of this intro we will say is a stream of photons that might eventually end up in your eyes and interact with cells in the retina. Each photon has its own wavelength, a continuous quantity. We will see that somehow the mix of wavelengths of the photons that hit the retina is the physical property behind color perception.

Although there is some variance between individuals, we mostly classify four types of sensory cells in the human retina: rods (which don’t contribute much to color vision), and cones, which come in three specializations, each sensitive to a specific range of wavelengths. Remember the number three.

These three ranges are distinct by the wavelength they are most sensitive to, but overall they have large overlaps. Altogether, the range covered by the three specialized cones (rods don’t add to it) is what we call “visible spectrum”, and defines what we call “light”. Electromagnetic radiation outside of the range is entirely invisible to the naked eye because it does not activate any of our sensory cells.

One thing about these cells is that they only respond to the amount of light activating them. The signal each one sends to the brain is “hey, I’ve been activated by some photons”, and the strength of the signal depends on the amount of activation within the last few instants. No other information is there.

The wiring is such that the brain knows each cell’s position and specialization, so it can put together the individual signals to construct shapes and colors. This is the sensory entry point to the experience of vision.

Note that a lot of information is lost in the process. We don’t perceive different wavelengths if not in a very compressed form. A ridiculous number of photons interact with your eye at all times, and each has its own wavelength (and spin, but forget about that) which is a continuous quantity. You can characterize light as the amount of photons for each visible wavelength, called “spectral distribution”.

But when light reaches some area of your retina, your perception of wavelength is at best characterized by three quantities: the amount of activation of each of the three types of cone cell. For some reason, we experience different levels of activation of these three sensor cells as different colors.

So, in order to produce light that can be recognized as having a certain color, you could mix a combination of photons with varying wavelengths, or another, or yet another. I won’t be able to tell which combination or anything of its spectral distribution. There are an infinity of such combinations for each perceived color, and no way the naked eye can tell them apart.

In this sense, color (or hue) is not a quality of light itself. It describes one aspect of our experiential relationship with light. Light has photons and wavelengths and such, but color is a thing that happens in our brain in response to the activation of cone cells. The color we see is not enough to describe the spectral distribution of light.

One more thing that will be useful to remember. Generally light has spectral distributions encompassing multiple wavelengths. But light with a single wavelength is also perceived as a color. Colors evoked from this kind of light are special enough, from a physical perspective, and are called “spectral colors”. Colors of the rainbow are spectral colors. Some colors are special in the opposite sense: there is no single-wavelength light that can evoke their hue. The majority of colors we see are non-spectral, but only some hues are. Purple hues are like that, as we only perceive them when we mix wavelengths near the opposite ends of the visible spectrum.

Color vision is quirky, much quirkier than I can fit in a brief intro, and it’s an absolutely fascinating topic that I recommend exploring. Not only I didn’t mention things like unsaturated colors (activating multiple type of cones together gets less “colorful”, not more). But also how color perception is ultimately a purely psychological phenomenon that is not yet fully understood, and depends on shapes, patterns, nearby colors, movement, adaptation, and so on. Colorimetry is the field that covers the quantitative relationship between light and color perception that I discussed. And it’s fundamental for visual technology. But the focus of colorimetry is limited, and a visual artist or designer probably has a much stronger intuition for the things that affect the experience of colors.

Color metamerism

Out of the intro, hold on to one concept: while color vision is the (only?) way we perceive electromagnetic spectra, colors are not electromagnetic spectra. One color can correspond to infinite spectral distributions.

Metamerism is a fancy word to say that what we perceive as the same color might actually be produced by very different blends of photons. Blends that are different but result in the same perceived color are called metamers.

It’s useful to know about metamerism when you want to make colorful things. You are building a display, or you creating new pigments for print, or to dye objects and clothes. Anything with colors. Since you are dealing with the physical layer, you might want to know a little bit more about the spectrum of light emitted or reflected by the material. Mainly for one reason: deciding a color is not enough, you might want to choose a metamer for that color that works under different lighting conditions (e.g. when less photons with wavelengths of a certain length are available, like on a cloudy day, or are missing altogether, like on certain types of artificial lighting).

I don’t deal much with materials, light emitters, pigments and the such. I care about the word mostly because someone decided that the concept is interesting enough to give it a name (apparently appropriating a now-obsolete word from chemistry which conveys, uh, kinda the opposite meaning). The word captures something simple but with gigantic consequences: perception is about what happens in the brain, in response but as opposed to what happens in the outside physical world.

And this is where we steer vehemently towards a new topic.

“Gender is a spectrum”

I want to deal with one pet peeve of mine: the phrase “gender is a spectrum” or, rather, the ways it’s employed.

You see, it’s easy to imagine that you have two ends of a spectrum that you call “man” or “woman” and any person’s gender falls somewhere on the line. Maybe the middle region is kinda neutral, while the closer you get to the extremes the more pronounced is their masculinity or femininity. This is a good way to imagine one’s own gender for some people, but it’s extremely limited, and very problematic in the way it suggests there might exist two “pure” points at the ends.

One’s gotta appreciate the intention, which is still clear and valid. A lot of people can relate to the general idea of representing gender as a variable quantity of “something”, rather than a blunt categorization in two or three discrete classes (man, woman, and possibly a third gender of some sort).

I think that the conceptualization of the spectrum as a male-female line might just be based on a popular misunderstanding of the phrase “gender spectrum”, maybe by cisgender people, maybe by people who were in sore need of a better description for themselves.

A while ago you could have seen the Genderbread Person model going around. I’m not sure of its origin, I first saw it on ancient social media, and educators have been using it. Later on, Sam Killerman popularized his own cute infographic rendition, which resonated massively in the larger online community. The merit of the model is to distinguish one’s gender identity from their outwards gender expression, and from their anatomical sexual characteristics. All of these things are “a spectrum”, but the kind of spectrum for each one was always a single male-neutral-female line.

Closer to the point of this post, the Genderbread Person model was later improved to represent the maleness-femaleness split on two axes which are distinct. Adding masculinity does not mean removing femininity, it seems to say, nor vice versa. Instead, your gender identity (or expression, or anatomy) could be described as two independent quantities.

I genuinely think this is a greatly improved model over the single-line one. Notwithstanding the limitations (primarily, that it’s not intended to be a canonical definition of gender identity, and it cannot be), it allows to imagine one’s identity as a composition of characteristics. Maybe someone is comfortable positioning their identity close to the Woman axis, but with a low intensity and without Man characteristics. Someone else might recognize a high degree of intensity on both axes, or on neither. It feels a lot less constraining.

Me being the kind of nerd I am, I can’t help noticing that increasing the number of dimensions to two still carries a lot of binary baggage. Why is it not three, or more? Probably because it’s hard to imagine what these extra dimensions might represent, while it’s comfortable to fall back to the two familiar categories of masc-fem.

To make this even more apparent, if you are so mathematically inclined, try switching the coordinate system from Cartesian to polar. You could have one quantity for the distance from the “neutral” origin, and another for the angle in either the masc or the fem direction. We are back to the original single-line model, with an added “intensity of gender” quantity, but still moving between two ends, one for boys and another for girls.

The diagrams are interactive (if you have JavaScript enabled) and you can play around with them. There is nothing special about the starting point. Position your current gender identity on the diagram, or experiment in any way. If you are a bit like me, you might have fun for a bit and share with friends, but also get a bit frustrated at the lack of expressivity of the model.

The color of a gender

When you pick colors to represent colors on a screen, you often have the option to use a Cartesian (or rectangular) system, such as RGB or LAB, or a cylindrical one, such as HSL or LCH. In either case, you are still moving in the three-dimensional space of perceptual colors, i.e. you are not choosing the exact spectrum that a monitor will emit for a pixel, that’s up to the monitor technology. Instead, those three coordinates are calibrated so to translate to a certain amount of activation of the three types of cone cells in the viewer’s retina. That’s where the number three comes from.

While with rectangular systems you specify each component separately (red, green, and blue each very roughly cover one of the cone cells’ types), with cylindrical coords you try to describe perceptual properties of the color. Usually you have something called “hue” which describes the qualitative aspect of a color (is it blue, cyan, purple, yellow, etc) and is arranged around a circle, so it’s represented as an angle. Then a linear quantity for how intensely colorful it is, which might be called “saturation” (or “chroma” or something else). And another for “lightness”, the amount of light to be emitted.

The Genderbread Person kinda described gender as a combination of “hue” and “intensity”, as we saw above. The gender hue can be segmented in two qualitative classes, or maybe three, depending on the interpretation of the transitional segment in the middle (it could be a perfect androgynous mix of male and female, or a third gender which is neither).

One could argue that each position on the arc could be a qualitatively distinct gender hue, just like you could segment the color wheel any way you like and name each segment. But this is not suggested by the model itself, because it’s designed around the idea of two fundamental gender classes.

Might it be that, just like pigments in the cone cells, we are hardwired to have two types of gender receptors? Like, concretely, the neural circuits for gender are two, and you can activate them in some measure. A bunch of people out there would jump to embrace this conclusion, although it’s a very distorted simplification of our perceptual reality.

Take any two people who would position themselves on the same spot in the Genderbread model, and listen to their stories about their relationship with gender, and you will notice vastly different shades. Similarly, ask someone to consider the gender they are attracted to, and to draw the corresponding region on the two-axes plane. They are likely to protest that any one area of the plane could have very different inflections.

All the genders of the rainbow

What happens if we actually try to imagine the “hue wheel” for gender to represent not just two qualitative classes, but a larger number of them? We might lack the vocabulary to name each one, but that’s a clear chicken-and-egg situation! You need to be able to see (or imagine) the colors to name the colors.

In fact, people have been using a vast vocabulary for gender-related description for a long time. The attempt is systematically mocked or politically opposed by those who don’t see (or refuse to see) the need for a variety of labels. Nevertheless, even without resorting to the “there are as many genders as there are people” wild card, you take any group of people with some freedom to express themselves, and you will see a colorful snapshot of what “gender” might mean.

I’m being deliberately handwavy because I don’t want to suggest my interpretation, and I invite you to consider yours, or that of the people you know.

While you do that, let me go back to cheap math shenanigans. So, if I manage to imagine a palette of uniquely different qualitative gender hues, and I arrange them on a wheel, I’ve removed one element to the the binariety of the Genderbread model. The “angle” does not describe a position in the man-woman dichotomy anymore, yay, binary gender is defeated by the sheer force of imagination. Can we move beyond the two-dimensional plane as well, now?

Remember that we have two dimensions because we are starting from the idea that an individual can describe their own gender identity as a composition of two values: an amount of man-ness and an amount of woman-ness. I suggested that the notion is a familiar one, which is exactly the reason why The Genderbread Person demonstrated useful for educative purposes. I’ll wager that this familiarity has strong cultural roots (or I’ll leave that to the entire feminist literature of the past hundred years).

Perceptual gender

What I mean is that, yeah, I do have two neural circuits for masc/fem detection, although they are largely the result of myself learning the dominant culture. I might also have some hardwired notion akin to the idea of sex, possibly (but not provably) along the same axis.

But they are not the only way I construct an image of “gender” in my brain. There are so many clues of various form that concur to this image. Think all the little signals conveyed through apparel, style, anatomy, sensorial experience, demeanor, social role, relationships, behavior, gait, voice, language, facial expressivity, cultural frame of reference, and so on.

Most of these signals are not distributed along the man-woman axis, although correlations exist. And taken together they are probably unique enough that they can be used to fingerprint anyone down to the individual level.

The observed person might explicitly state their gender identity, of course, but we are not going to pay attention to that for a second. Rude. But useful for two reasons. The first is that we are making an attempt to understand the perception of gender as its own experience, without words to describe it. We are looking at the colors themselves, regardless of their name. The second reason is that no one is making a statement when you are both the observer and the observed.

Just like color is not directly a property of light, but a phenomenon that arises in the eye (and visual cortex) of the beholder, gender might not be directly a property of a person, but a perceptual phenomenon that arises in close connection to other brain functions. Maybe a linguistic need to categorize in labels; or a social need to find people with similar traits; or a sexual need that drives us close to potential partners. Something like that.

One’s experience of their own gender might or might not walk along the same lines. Experiencing your own gender is a completely different matter than looking at someone else’s. When it comes to communicating one’s own experience to others, though, you definitely need some shared language. It’s not only about names: the whole idea of gender expression is about all the signifiers that outwardly communicate gender. And culture impacts self-perception too, not only communication. When words like “genderqueer” or “non-binary” went mainstream, millions of people found them useful to take a more nuanced look at themselves.

Now, is gender a thing at all? It’s a legitimate question. Maybe all those different dimensions and nuances and stuff should be called something different. I don’t care much about the name myself. But the fact is that, ever since the day someone birthed the notion that humans can be classified in two sex buckets, people have been finding out ways the classification was unable to accurately describe themselves. Draw a graph starting from that erroneous binary classification, and any associated concept goes under the “gender” umbrella. Additionally, the need to classify is still felt strongly by tons of folks, cis and trans, so we can’t just do away with it. Matters not that it’s a heterogeneous collection of different traits, as long as we know we are talking about the same stuff.

One more question is: if gender is a perceptual phenomenon, does it reflect a more complex physical reality? What is the gender equivalent of electromagnetic radiation over a range of wavelengths? The whole depth of the question is well beyond me, but let me at least attempt to say that there is necessarily a physical reality to any perception. And it’s a safe guess to imagine our perception offers a low-resolution, compressed view of that physical reality.

Given we are speaking of a concept with cultural ramifications, we are also expanding the perception, not only compressing, by enriching it with attached semantics. Could that mean that the whole thing about gender is a just a fancy way to talk about cultural identities of some other sort. Well, first, that might be and it’s an explanation as good as any. After all, the bucketization in men and women is exactly that. I prefer to consider more dimensions than the social one, though. The other thing is that attaching semantics does not negate the possibility that we are compressing information. Saying that yellow is annoying or red means danger doesn’t bring back any of the dimensionality of the electromagnetic spectrum.

Perceptual is not not physical

For one last time, let’s jump across the two sides of the gender-color analogy. When I mentioned the hue wheel, I focused on the perceptual reality: colors have a distinct qualitative aspect that allows us to say that green is not the same as yellow or magenta, but it’s more similar to the former than to the latter. But the wheel itself is organized in a way that reflects physical reality.

Let’s unroll the wheel. The hues from red (0°) to blue (240°), through green, approximate the spectral colors arranged on a line of decreasing wavelength (from 700 nm to 460 nm). The remaining third of the circle is closed with the line of purples, that is, all the blends from spectral blue to spectral red. If we had better technology we could go further down to 380 nm to show violet, but we still would need the line of purples to perceptually close the circle back to longest wavelength, red.

Technicalities aside, the point is that, while I insisted that colors are a thing of our perception, our perception still maintains some structure from the physical world. If you organize hues based on perceptual similarity, you get the same rainbow you get when splitting sunlight. Spectral colors are the weird reality that make the structure apparent, while purples are the reminder that we are making this up in our brains.

(Marginally, you gotta appreciate the—accidental, as far as I know—symbolism of the bi pride flag being the missing piece to close the circle of the rainbow gay pride flag. Oh, and the trans flag adding unsaturated colors!)

If you went along with the mental exercise to imagine “gender hues” as more colorful than the dichromatic boy-girl scheme, you might have also tried to give some structure to those distinct colors. And if you were frustrated (or elated, not to assume your agenda) not to be able to find a structure, you were at least in good company. I find it very hard to imagine that there’s only one way to describe the transitory space between a supposed masc-fem continuum. Even harder when it comes to the gender-hues of any two given people.

But, if I assume that gender has both a physical and a perceptual reality, I can finally assemble all these contradictory and ineffable clues into a slightly more manageable conceptualization. On the physical level, gender might be a highly-dimensional mixture of things, maybe resident somewhere between neural connections, neurotransmitters and other hormones, chromosomes, epigenetics, microbiome, and who knows what. On the perceptual level it can manifest in billions of ways (some specifically reserved to the person experiencing gender on their own skin, some observed by people on the receiving end of gender expression, some emerging exclusively through interactions between folks).

Physical gender might include things that are highly correlated to some man-woman duality, and these things might be reflected in some compressed form in perceptual gender, to the confusion of every person who ever experienced gender dysphoria (hey, that might include you even if you are a cis person reading this). And it definitely includes things that are not, just to make things easier (sarcasm) but also more colorful.

Gender metamerism

This is it, that’s the idea. Gender may not be a spectrum, but gender lies on a spectrum. Each point of the spectrum, like a different wavelength, contributes to gender experience in a different way, through a continuum of varying perceptual quality. Plus a few other non-continuous dimensions, I guess. When perceived, like color, it’s dimensionally reduced in a way manageable for our brain’s information-processing capability.

The implication would be that even the most refined perception won’t be able to distinguish a gender made of pure 600 nm light from a mixture of reddish and greenish gender-light, as we are physically wired to see either as an orange gender. It would be the implication if we attributed the perception of gender purely to a physical capability. But! We know that a major source of information loss related to gender perception is cultural, that is, we use learned information to make sense of the perception. This means that we can increase our ability to perceive gender by learning about it.

Since this is a delicate topic, and I don’t know who the audience of this post might be, I feel obligated to add a serious note: please don’t rely on your perception to attribute gender to other people. No matter how well-educated you are, you don’t have the same type and amount of information that a person has about themselves. And misgendering has a cultural significance such that it can have extreme consequences. There is a serious reason why it’s commonly advised to avoid the risk, and “don’t assume gender” is a simple and pragmatic policy that many (myself included) recommend to follow at all times.

On the other hand, an enhanced skill to savor (see, I can do other analogies!) the multidimensionality of gender is extremely powerful when applied on oneself. I should be extremely suspicious if I look at myself and at someone else I know, and we use the same terminology to describe ourselves. Sure, we both might be perceived as green to an untrained eye, but is it the same shade of green? Is it constant under different lighting conditions? Does it still look the same when I apply different filters? Generally, the answer is no.

Just to be clear. This is not about the difference between one’s own gender, and the same gender as understood by others. From the plane we are looking from, these are both colors, only one is special because self-observation is a unique vantage point. If I were to find an analogous in color perception, it could be that the color you see when you close your eyelids is still a color. It might be silly to put self-perception and other-perception in the same basket, but it would be harder to split them, and preserve the useful property that gender is something people can talk about. The point is more that whatever gender we perceive, there is a complex set of characteristics that determine our perception which are not visible to the naked eye, and we cannot trust similarity alone to tell the whole story.

At the end of the tunnel

Like I said, both color and gender are extremely prone to misconceptions. Perhaps it’s their nature as perceptual phenomena that makes them very subjective. Or, perhaps, it’s the fact that both are very generalized experiences that receive only cursory introduction at a very early age, and are rarely considered in their complexity.

This post might have furthered the confusion, in suggesting it might have a point to make. It’s more of a lighthearted excuse to be a nerd around two fascinating topics. The idea that genders have a physical reality akin to that of light is suggestive fiction. The complexity of gender lies more in the socio-political frameworks that construct (and deconstruct) it. Still, the metaphor has been a valuable reminder that most of the complexity is unseen, and that two similar gender expressions can hide two very different personal worlds.

Bonus stuff

• “Autism is a Spectrum” Doesn’t Mean What You Think addresses some of the same misunderstandings about the notion of a spectrum when used to describe people.
• The Hitchhiker’s Guide to Digital Colour does a great job at demystifying colors for anyone using them in digital media. It includes a thorough discussion of perceptual color.
• This community gender spectrum page (Nonbinary wiki) has some historical diagrams reminding ideas also seen on this post.
• Another gender spectrum page (Gender Wiki), with another familiar diagram.
• Micheal Bach’s collection of Visual Phenomena & Optical Illusions includes a section about color and does a great job at demonstrating (and documenting) how color vision is way more complicated, and sometimes surprising, than just sensing light.
• If the word “hue” is confusing, there are good explanations out there, but I can’t suggest any particular one. In case you missed it, though, I made a sorta lightweight intro to colors and hues.

Go to Top