Intelligence beneath the surface: learning, remembering and solving problems

In a marine biology laboratory in Naples, a live crab moves inside a glass jar. A common octopus, Octopus vulgaris, faces the jar, which is sealed with a plastic stopper. The crab’s movements can draw the octopus towards the glass. The octopus approaches and runs its arms over the jar, its suckers gripping the surface.

The crab stays out of reach until the stopper comes free. Each of the octopus’s eight flexible arms carries suckers that can grip and taste what they touch. Together they give the animal fine control over an object. That still leaves the problem of what to do with the stopper. The octopus has to find a way to pull it out and then get hold of the crab.

Over successive trials, failed attacks became less frequent. The octopuses got quicker at removing the stopper and seizing the crab, while the time they spent exploring stayed much the same. The task stayed the same from one trial to the next. What changed was how the octopus approached it. The improvement could come from trial and error or from learning to associate a movement with its outcome. How the octopus understands the jar remains an open question.

The researchers could see the improvement in the time it took to get the crab. What happened on one attempt helped shape what the animal did on the next. Another experiment asked what would happen if an octopus watched one of its own kind make a choice.

Researchers give an octopus two objects that differ in colour. Choosing one earns a reward. A second octopus, new to the task, watches as the trained animal makes its choices. Then the demonstration ends.

The observer gets the same choice on its own. In the experiment published in 1992, observers tended to choose the object they had seen the demonstrator select. The preference could persist for several days without further training. Watching the other octopus made a difference. How it did so remains debated. The demonstration may simply have drawn attention to one object.

For an animal that generally lives alone, another octopus can still be a source of information. The result comes from the laboratory. Whether octopuses learn from one another this way in the wild is a separate question. An octopus’s nervous system is organised very differently from a vertebrate’s. Many neurons lie in the arms and their nerve networks, where they help process sensations and control movement.

Those circuits let each arm do some of the work of exploring a surface or handling an object locally, as part of a coordinated animal. Touch and movement are closely linked as a sucker grips a surface. Watching another animal is one way to pick up information in a lab. On the seabed off Indonesia, a different octopus shows what it can do with an object it finds.

A moray can use the same refuges for a long time. A reef fish can return to an area it knows. An octopus can find a shelter. These behaviours seem familiar. We too find our home, recognise a path or remember a place where we found food. Returning to a familiar spot could reflect several kinds of memory.

Smell, shape or a nearby landmark might guide an animal back to a cavity. It might also repeat a familiar route without relying on a map of the area. Researchers can test those possibilities by changing the conditions. They might remove a landmark, move a reward or block the usual route.

Does the animal still find its destination? Can it pick up a new cue or learn another route?

The change makes it possible to see what information the animal uses. Finding a place is one of several things a fish can remember. A fish can learn a sequence of movements, link a signal to food or adjust how it responds to another fish of the same species. These abilities do not all require the same nervous mechanisms.

In vertebrates, several brain regions are involved in the processing of spatial, sensory and social information. In octopuses, the parts of the nervous system involved in learning and memory are organised differently. Both kinds of nervous system can support complex behaviour. The differences matter when we compare what animals can do.

An octopus can explore an object with its arms. A cleaner wrasse can reach a marked spot by scraping against a surface. A moray explores cracks that the grouper cannot reach. Each faces problems shaped by its anatomy, senses and way of life.

A jar, a mirror and a crevice put different abilities to the test. To see what an animal has learnt, look at how its response changes with experience. The change can be small. It may repeat a familiar movement in a different way.

It may take another route to the prey. Or it may respond to a signal at the right moment.

The crab is still behind the glass. For an octopus, a sealed jar is an unfamiliar obstacle. Its arms and suckers are already capable of opening it. The suckers can grip the stopper while the arms move it.

At first, much of that effort goes into movements that leave the stopper in place. The octopus needs a movement that will pull it free. Across trials, failed attacks became less frequent and the stopper came out faster. The prey becomes accessible.

The physical task stayed the same. Experience changed the way the octopus tackled it.

What happened on earlier attempts helped the octopus reach the crab on later ones.

How the octopus understands the stopper is still open. What the experiment measures is the improvement with practice. The crab was visible from the beginning. It took practice to get to it.