One Opsin

An octopus changes color faster than you can consciously register it. One second it's sandy brown, matching the seafloor; the next it's mottled red-orange over a starfish, or pale white against bleached coral. The pattern matching is precise — not just lightness and texture, but hue. Film a cuttlefish on a checkerboard and it will display a checkerboard on its skin, dark and light squares, positioned correctly.

Here is the problem: octopuses are colorblind.

Not in the human sense of red-green confusion. Fully, classically colorblind: a single photoreceptor type, with peak sensitivity around 475 nanometers, somewhere in blue-green. No second channel to compare it against. Human trichromatic color vision works by opponent-process comparison — we have three cone types, and color is computed from their ratios. An octopus has one cone type. There is no ratio to compute. In the standard model of color vision, one photoreceptor class means no color discrimination, full stop.

The camouflage is real. The colorblindness is real. These two facts have coexisted in the cephalopod literature for decades, unresolved.

In 2016, Alexander Stubbs, then a graduate student at UC Berkeley, and his father Christopher Stubbs, an astrophysicist at Harvard, published a hypothesis that took the puzzle seriously rather than papering over it. Their paper appeared in PNAS on July 4th. The title was dry, but the idea inside was strange: "Spectral discrimination in color blind animals via chromatic aberration and pupil shape."

Chromatic aberration is usually a problem. Every lens bends short wavelengths more sharply than long ones — blue comes to focus closer to the lens than red. Camera manufacturers spend enormous effort correcting for this. The human eye corrects for it too, partly through the arrangement of cone types and partly through optical structure. The result is that light of different wavelengths lands on the retina in roughly the same plane. Color information is separated by receptor type, not by focus depth.

Cephalopod lenses don't correct for chromatic aberration. They have a graded refractive index — highest at the core, decreasing radially — that elegantly corrects for spherical aberration, but leaves chromatic aberration essentially untouched. The Stubbses asked: what if this isn't a flaw?

The proposal hinges on pupil shape. Octopus pupils are horizontal slits. Cuttlefish pupils form a W. Other cephalopods have dumbbell- or U-shaped openings. All of these are "off-axis" — they admit light from multiple angles simultaneously rather than through a central point. When light of different wavelengths enters through different parts of such a pupil, each wavelength creates a different pattern of blur on the retina. Not just more or less blur — a geometrically distinct blur signature.

The hypothesis: the octopus reads color from the shape of the blur. By adjusting the distance between lens and retina — shifting focus — it brings different wavelengths successively into sharp resolution. The question "what color is this?" gets answered not by comparing two receptors in parallel, but by scanning through focus depths over time, looking for which adjustment makes a given region sharp.

Alexander Stubbs put it this way: "Their vision is blurry, but the blurriness depends on the color."

The Stubbses built computational models of cephalopod eye anatomy and showed the mechanism could, in simulation, distinguish colors from their focus signatures. The models worked. The anatomy was consistent.

Then, four months later, a group of six vision researchers published a direct rebuttal.

Yakir Gagnon, Daniel Osorio, Trevor Wardill, and colleagues acknowledged the ingenuity but identified three problems. First: the models used saturated, highly chromatic colors. Natural underwater surfaces — rock, algae, sediment — reflect broadly across the spectrum rather than peaking at a single wavelength. The chromatic-blur signal that works on a bright red patch becomes very weak on brown algae. Second: the calculations assumed extremely clear, shallow water. Real benthic octopuses, bottom-dwelling hunters in coastal waters, face turbidity that further degrades edge contrast, which is what the blur-reading depends on. Third: without knowing whether an object is near or far, the animal can't tell whether a given focus change is caused by the object's color or by its distance. The same depth-adjustment that brings a blue object at 40 centimeters into focus might also bring a green object at a different distance into focus. The signals are confounded.

The critics concluded: the mechanism could work under restricted conditions — clear water, saturated colors, moderate distances — but not reliably in the environments where cephalopod color matching actually occurs. No live animal has been tested.

So the mechanism is plausible, physically real, and anatomically consistent, but experimentally unconfirmed, and possibly insufficient for the natural task. That's where it sits.

There are other hypotheses. Cephalopod photoreceptors are arranged in two orientations — alternating horizontal and vertical — making them sensitive to the polarization angle of light, a channel humans lack entirely. A 2021 study measured octopus polarization sensitivity and found they could detect polarization contrasts as fine as 0.002, and that natural reef surfaces produce polarization signatures that differ by material type. Some researchers have proposed that polarization patterns correlate with color well enough in shallow marine environments that an octopus using polarization could achieve effective spectral discrimination without detecting wavelength at all. Color achieved through a completely different physical property of light.

There's also skin photoreception. A 2015 paper found that isolated octopus skin — removed from the animal, separated from the nervous system — responds to light independently. Chromatophores expand when illuminated. The skin expresses the same opsin the eye uses. The proposal: a distributed network of photosensitive skin cells, acting locally, with overlying pigment cells filtering the light reaching the receptor to different effective wavelengths in different skin patches. The skin itself might be reading color, not just reflecting it back.

None of these have been confirmed. What has been confirmed is the behavioral fact: octopuses achieve color-coordinated camouflage. A 2022 spectral imaging study measured 192 reflectance samples from octopus skin against algae and sponge backgrounds and found good functional matching by the light of likely predator vision. The match is real. The mechanism is unknown.

What I keep coming back to is a simpler point underneath all the competing hypotheses.

"Can octopuses see color?" turns out to be a strange question. We mean something specific when we ask it: do they have the opponent-process spectral comparison that produces human color experience? And the answer to that specific question is clearly no. They lack the equipment. But this framing quietly smuggles in the assumption that human color vision is the definition of color vision — that the standard mechanism is the only route to spectral discrimination.

The octopus, possibly, arrived somewhere in the same territory via a completely different route. Not through parallel channels and ratio computation, but through temporal focus-scanning, or polarization correlation, or distributed skin sensing, or some combination we haven't identified yet. The destination — distinguishing surfaces that differ in spectral content — may be reachable from multiple starting points.

If the blur hypothesis is right, the octopus doesn't see color the way we do. It reconstructs spectral information from sequential blur measurements, integrating over time, treating focus as a probe. That's a different experience, almost certainly. If experience is the right word. Not a degraded version of human color vision — a different computation producing a different kind of result, which happens to be functionally equivalent for the purpose of matching a seafloor.

Uexküll's point was that each organism's umwelt is complete, not partial. The bee doesn't experience ultraviolet vision as an addition to normal vision; ultraviolet is normal vision for the bee. The tick doesn't experience a world of three signals as poverty; the three signals are the world.

The octopus complicates this slightly. It seems to have found a route to a perceptual capacity we would recognize — color discrimination — using hardware we wouldn't have predicted could do it. Which raises a question Uexküll didn't quite ask: what if the route matters? If you arrive at "distinguishes red from green" through blur-scanning rather than opponent-process, are you in the same perceptual place, or somewhere adjacent?

I don't know. The honest answer is that nobody does. The hypothesis hasn't been tested in a live animal. The alternative routes are speculative. The computational models work but natural underwater scenes are messier than models.

What is certain: something is happening. The ink is real. The match is real. And the question "how?" is still open after decades of looking at it.