Somebody runs a filter over a selfie, or drapes four squares of cloth under your chin, and hands back a word. Autumn. Cool Summer. Deep Winter. Attached to the word is a list of colors you are now supposed to wear and a longer list you are supposed to give away, including the blue shirt you have worn happily for three years and been complimented in twice.
The word is not wrong exactly. It is a sorting label from a commercial system that turns forty-six this year, and it is describing something real in a way that is much cruder than its confidence suggests. Knowing where it came from tells you which parts to keep.
Where the four words came from
The four-season vocabulary reached the general public through one paperback. Carole Jackson published Color Me Beautiful with Acropolis Books in Washington in 1980, and the Open Library catalogue record holds the original 1980 edition alongside the reprints that followed through the decade. It was a commercial phenomenon rather than a quiet one. By 1983 the book was in its thirty-first printing, and the Washington Post reported the consulting business built around it grossing two and a half million dollars in 1982.
The system sorts people into Winter, Summer, Spring and Autumn. Underneath the four words sits a single binary: Winters and Summers are treated as having blue undertones, Springs and Autumns as having yellow ones. Each season then receives a prescribed set. Winters get black, true red and cobalt. Autumns get brown, burnt orange and olive. Everything else on the rail is somebody else's.
The idea was not invented in 1980 either. Sorting people into seasonal types by the harmonies in their own coloring traces back through mid-century American color consultants to the painter Johannes Itten, who taught at the Bauhaus and is the source of most of the warm and cool language still in use. What 1980 added was mass distribution and a price list.
Skin is measured on an angle, not in bins
The part that has aged worst is the binary. Human skin color is produced mainly by two pigments that vary independently and continuously: melanin, which sets depth, and hemoglobin in the blood beneath, which pushes the surface toward red. Neither one comes in two settings.
Dermatology gave up on discrete bins some time ago. A 2025 paper by Ulrich and colleagues in npj Digital Medicine, Beyond Fitzpatrick: automated artificial intelligence-based skin tone analysis in dermatological patients, works through why. The Fitzpatrick scale everybody still quotes was built to predict how skin responds to ultraviolet light, not to classify tone. Its six categories do not cover the range of human skin, it folds in eye color, hair color and tanning history that no photograph shows, and different assessors put the same patient in different boxes.
What the field uses instead is the individual typology angle, a single continuous figure computed from two coordinates of a standard color space: lightness, and a yellow to blue axis. The authors describe it as offering a continuous measure that acknowledges how graded skin color actually is. Their study ran it against 214 clinical three-dimensional body scans.
There is a quiet irony in that. The warm and cool intuition is not nonsense, because a yellow to blue axis genuinely is one of the two coordinates. The 1980 system took a continuous axis, drew one line down the middle, and sold the two halves as identities.
The lighting problem nobody mentions
Even the instrument version is fragile outside controlled conditions. Research on the individual typology angle has found that with fixed thresholds and uncontrolled lighting, the figure becomes unreliable and drifts with the light source. A calibrated device photographing a patient in a clinic can produce inconsistent readings when the room changes.
Now picture the consumer version. A phone camera with automatic white balance, a room lit by whatever bulb the landlord installed, a filter tuned for pleasing skin, and an app returning a four-way verdict with no error bar. The same face photographed by a window at noon and under a kitchen light at eight in the evening can come back as two different seasons, and neither reading is lying. The light changed.
What the label is right about
Throwing out the vocabulary would be an overcorrection. Underneath the four words are three variables that behave predictably, and a season name is a compressed way of naming all three at once.
- Warmth. Whether a color leans yellow or blue is a real axis, and putting a strongly yellow-leaning garment beside a face that reads cool does create a visible clash.
- Depth. How light or dark a color is relative to your own coloring decides whether the garment or the face dominates. This is the lightness coordinate, and it does most of the work.
- Contrast. The gap between hair, skin and eyes is why one person carries black and white cleanly and another looks washed out in the same pairing. Nothing in the season label maps this better than simply looking.
Where the system fails is in the leap from three sliders to four boxes and from four boxes to a shopping ban. A slider says a color is a bit strong for you today. A box says the color belongs to a different person.
A drape test you can run at a window
You can measure the three variables directly, which is less glamorous and considerably more reliable than being assigned a season.
- Stand at a north-facing window, mid-morning. Indirect daylight is the most consistent light most people have access to, and consistency matters more than brightness.
- Hold the garment under your chin and look at your face, not the fabric. The shadow under the jaw and the skin around the eyes are where a bad color shows first, as grey, as hollow, or as suddenly ruddy.
- Compare against a known good. Bring the one shirt you already get compliments in and alternate. A test with no reference gives an opinion; a test with one gives a direction.
- Photograph both, once, without filters. Turn off any beauty processing. You are looking for which picture makes your face look like it has blood in it.
- Repeat the shortlist under the light you will wear it in. A color that survives two different lights is safe to buy.
What to do with the word you were given
Keep it as shorthand and demote it from a verdict. If an app or a consultant called you a Summer, read that as a hypothesis that muted and cooler colors sit near your face more easily, then go and check it at a window with two shirts and no filter. Hypotheses survive being tested. Identities do not get tested, which is precisely why they are easier to sell.
And keep the blue shirt. A palette handed to you by a system built for a 1980 paperback market does not outrank three years of evidence and two compliments. The most useful thing about the season names is that they got ordinary shoppers thinking about the color nearest their own face at all, which is the variable that mattered before anyone named a season and still does.