How Octopuses Solve Problems Without Bones, Without Centralized Brains, and Sometimes Without Their Owners' Knowledge

The octopus is the most cognitively sophisticated invertebrate that exists. It is also a creature that, in evolutionary terms, has almost nothing in common with the animals we usually compare to humans. The last common ancestor of an octopus and a human was a small wormlike thing that lived around 600 million years ago and had nothing that we would recognize as a brain. Every cognitive structure octopuses have, they evolved independently. They are, in the literal sense, the closest thing we have to an alien intelligence on this planet, and the way their cognition is built differs from vertebrate cognition in ways that are still being worked out.

The basic facts are striking enough on their own. An octopus has about 500 million neurons (roughly the count of a dog or a cat). About two-thirds of those neurons are not in the central brain but are distributed in two large nerve cords running down each arm and in clusters of ganglia throughout the body. Each arm has, in effect, its own peripheral computing system that can sense, reach, grasp, and manipulate objects with substantial autonomy. The central brain coordinates the arms but does not micromanage them.

This architecture is not a quirk; it is a deeply different solution to the problem of being a body. The vertebrate solution centralizes computation in a single brain that issues commands to peripheral muscles via long nerves. The octopus solution distributes computation across the body, with the central brain serving more as an arbitrator and overall planner than as a master controller. The biological reason is structural: an octopus has no bones, so its body cannot be controlled by a central system that issues commands to a fixed skeletal geometry. Instead, the body is a constantly changing soft structure, and the only way to control it efficiently is to let local segments handle local problems.

The famous escapes

The folklore around octopuses is full of escape stories, and most of them are true. Octopuses routinely escape aquariums by squeezing through small holes (their bodies have no rigid components except the beak, so they can compress themselves to the size of the smallest opening their beak will fit through). They climb out of tanks at night and drop to the floor. They figure out how to push the lids off their enclosures. They open jars containing food, both from the inside (when placed inside the jar) and from the outside (a more difficult task that requires holding the jar steady while turning the lid).

The most-studied escape behaviors come from the New Zealand octopus Inky, who in 2016 climbed out of his aquarium tank, crossed the floor, found a drainpipe leading to the ocean, and disappeared into it. His enclosure had a small gap at the top of the lid; once Inky discovered the gap, he was capable of squeezing through it and locating the drain. The escape required, in sequence: noticing that the lid had a gap, planning a route to the drain (which Inky had not used before), executing the climb, executing the squeeze, navigating the floor, and entering the drain. Whether this is "tool use" or "planning" depends on which philosopher you ask, but it is clearly something more than reflexive behavior.

Octopuses in laboratory settings have been observed performing similar feats. They unscrew the lids of food jars (some species) or pry them off (others). They push aside the false bottoms of puzzle boxes. They learn, by observing other octopuses, how to solve mechanical puzzles. The observational learning is particularly striking because octopuses are not social animals in the wild; they live solitary lives and meet only for mating. Yet in captivity, an octopus that watches another octopus solve a puzzle will, on average, solve the puzzle faster on its own first attempt than an octopus that has not watched.

Color, camouflage, and a strange fact about vision

An octopus changes color. The skin of an octopus contains thousands of cells called chromatophores, each containing a pigment sac and surrounded by muscle fibers; when the muscles contract, the pigment expands and the cell becomes more visible. Different chromatophores contain different pigments (typically red, orange, yellow, brown, and black). By coordinating the chromatophores across an area of skin, an octopus can change the apparent color of that area in milliseconds.

The strange fact is that the octopus is colorblind. Behavioral and electrophysiological tests have repeatedly shown that octopuses have only one type of photoreceptor, which means they cannot distinguish colors in the conventional sense. And yet they match the colors of their surroundings precisely, including in environments where they cannot see their own bodies (because they are camouflaging the side facing the predator while looking the other direction).

The leading hypothesis, formally proposed by the Hanlon and Mäthger labs in 2010 and refined since, is that octopus skin contains its own photoreceptive proteins (opsins) and is itself sensitive to light, including some wavelength discrimination. The mechanism is not visual in the usual sense (the skin does not "see" in the way an eye does) but it can detect the spectral composition of incoming light and adjust the chromatophores accordingly. The 2015 Ramirez and Oakley paper found opsin expression in the skin of cephalopods, which gave the hypothesis a molecular foundation. The full story is still being worked out, but the rough picture is that an octopus is, in some sense, seeing with its skin.

Recognition and personality

Octopuses recognize individual humans. The most thorough study of this was done at the Seattle Aquarium in 2010 by Roland Anderson and colleagues, who exposed two octopuses (Octopus rubescens) repeatedly to two different humans: one who fed them and one who poked them with a bristly stick. After about two weeks, the octopuses behaved differently toward the two humans on first sight: they approached the feeder, retreated from or sprayed water at the poker. The behavior held even when the humans wore identical clothes and entered the room from the same direction.

The recognition is presumably visual; octopuses have excellent vision in monochrome and can resolve fine detail. The relevant point is that the octopus is forming, retaining, and acting on long-term individual associations with specific humans, which requires both individual recognition and long-term memory. Both of these capacities have been studied extensively in octopuses and they are present, although the details (how long the memories last, how they degrade, how they are encoded) are still active research questions.

The personality literature is similarly fascinating. Individual octopuses, even of the same species and from the same clutch, show consistent behavioral differences across time and contexts. Some are bolder, some are shyer. Some are aggressive in food competition, some are not. Some are curious about novel objects, some retreat. The differences are stable enough that aquarists can identify individuals after months of separation, and they have measurable consequences (bold octopuses live shorter lives in the wild but acquire more food).

The sleep question

The most surprising recent finding is that octopuses sleep, and they have something that looks remarkably like REM sleep. The 2021 Iglesias paper documented two distinct sleep states in Octopus insularis: a quiet sleep state with motionless body and pale skin, and an active sleep state with rapid color and shape changes accompanied by twitching of the arms and eye movements. The active state lasted about forty seconds and recurred every thirty minutes or so. The pattern is structurally similar to vertebrate REM sleep, in which most dreaming occurs.

Whether octopuses dream in any meaningful sense is unanswerable on current evidence. The active sleep state is consistent with what dreaming would look like behaviorally, but consistency is not proof. What the finding does establish is that two cognitive architectures (vertebrate and cephalopod) that diverged 600 million years ago and developed completely independently have both arrived at a sleep state that involves rapid neural activity, behavioral disconnection from the environment, and what appears to be internally generated experience. This is striking convergent evolution, and it suggests that whatever cognitive function REM-like sleep performs is fundamental enough to be worth re-evolving.

What we still do not know

The honest answer is most of it. We do not know how memory works in a distributed nervous system. We do not know what the central brain does when an arm is solving a local problem. We do not know whether the seeing-with-skin hypothesis is correct or just one piece of a larger puzzle. We do not know how octopuses learn from observation given that they are not social. We do not know whether the dream-like sleep state involves anything we would recognize as experience.

The reason most of it is unknown is that octopuses are difficult research subjects. They live short lives (most species live one to three years; the longest-lived, the giant Pacific octopus, lives about five). They are solitary and cannot be kept together. They are demanding to keep in captivity (water quality, enrichment, escape-proofing). Many of the relevant cognitive tests were designed for vertebrate subjects and translate poorly. And the research community working on cephalopod cognition is small relative to the depth of the questions.

What is becoming clear is that the octopus is not a less-organized version of vertebrate intelligence; it is a different organization entirely, and one that has solved many of the same cognitive problems by different means. It is the closest thing we have to an experiment showing that the kind of intelligence vertebrates have is not the only kind that can evolve. That alone is worth more attention than it gets.