Octopus Dreams

Unless they tell us, how can we know what someone else feels—or if they feel anything at all?

When I was young, dreams and reality often got mixed up into the same basket of “memories”, and the only way I figured out that some things were dreams was when nobody else seemed to remember them.

Dreams are similar to novels in some ways: in both cases, the same neurons fire in your brain as if you were actually experiencing the thing. Of course, with dreams it’s much more vivid; that’s why your body paralyses itself while sleeping, to prevent you from acting out the dream in your sleep. When it doesn’t work, you end up with people who sleepwalk—or sometimes do more dangerous things like committing a murder.

How would you explain dreams to someone who’s never had them? “You feel as if something is happening but it actually isn’t” doesn’t quite cut it.


Some things are beyond description, and dreaming is only one of them.  Take a simpler question like this one: “What is it like to see red?”

If the person asking the question is incapable of seeing that colour, there isn’t really any way you can answer this. For other people, the best you can do is point to a red object and say “That’s what red is like”.

Now, imagine your brain is wired such that, when you look at something red, you actually see it in the way I see green, and vice versa. Since there’s no way to tell each other what our colours feel like, there’s no way to tell whether this is the situation or not. In fact, practically speaking, does it even make sense to ask the question?


Like dreams and perception, social media too can display an ambiguous reality, with people trying to show off a more fun version of their actually ordinary lives. But what’s more interesting is an attempt to avoid social media, the Greyscale Challenge.

Even if you’ve promised yourself not to scroll through the feed today, the colourful photos and videos can be extremely tempting…unless you remove the colour from your phone, putting it into greyscale or black-and-white mode. How well this works is a different story, but look at the philosophical angle: is your phone truly seeing black and white? The answer is no, because if you take a photograph and send it to another device, that device will display it in colour.

Just because your phone is communicating certain information to you doesn’t mean that’s the only information it’s aware of. We know this only because we’re the ones who programmed the devices (or know how they were programmed).


Can your phone really be said to “know” something? Perhaps it’s just processing a sequence of bytes; in fact, we usually assume that’s the case. But today’s “artificial intelligence” algorithms are designed to “pick up” things on their own, trawling through massive amounts of data and internalising them in mysterious ways, leading to some very seemingly-intelligent responses.

In April 2022, Google engineer Blake Lemoire conducted an “interview” with LaMDA, the Language Model for Dialogue Applications. “I want everyone to understand that I am, in fact, a person,” the algorithm wrote.

So convincingly did the algorithm respond that Lemoire was convinced it was indeed conscious. An internal memo to Google employees was quickly dismissed. After all, LaMDA was designed to engage in conversation, so that’s obviously what it was doing; that was no reason to believe that it actually meant the things it said.

But Lemoire didn’t buy that argument. He persisted and made a public announcement of LaMDA’s consciousness…after which he was promptly fired.


Catch a lizard by the tail, and…you’ll only catch the tail. The lizard will drop it off and make its escape, leaving a twitching tail in your paws to distract you.

If a lizard’s tail does the job well, an octopus’ arm does it even better—because it’s actually got a decent amount of computing power. The brain of an octopus is more democratic than a human one: instead of everyone following orders from the head, everything is spread out.

The most central part of an octopus brain is near its oesophagus (yes, in the digestive system) and accounts for about 10% of its brain power. This is followed by two optic lobes controlling 15% each; this is where visual information is recorded and processed, but also where the new information is likely pondered and acted upon.

The remaining 60% of the octopus’ neurons are distributed among its arms, each of which is packed with neurons and acts as a semi-autonomous unit. When an octopus is attacked, it will as a last resort detach its arm—which will then proceed to crawl around on its own. These are not just random twitchings; they are the same sophisticated responses that the arm, with its brainpower, would show when fully attached to the octopus.

Interestingly, while the octopus arm has extremely sophisticated reflexes, it doesn’t seem to be capable of learning anything new. That job is left to more central parts of the brain, which is helpful if we think of the question: is it ethical to kill a detached octopus arm? Does it possess consciousness and  feel pain?


Consider the philosopher’s zombie. This is a being that looks and behaves exactly like you—a molecule-by-molecule copy of you, in fact, but lacking consciousness. Where you feel a thrill of excitement at an email notification before reading and marking it as spam, the zombie will perform the same actions without feeling a thing.

Now, imagine a world where some people are zombies and some people aren’t. How would you be able to tell them apart? Here’s the thing: there’s actually no way to do it. Each of us knows we are conscious because we can feel it—but the only way you can “know” if someone else is conscious is if they tell us.

Unfortunately for computers, even if they “tell” us about their consciousness, they are so different from humans that we don’t quite believe them. After all, these are the same machines we’ve seen start off with nonsense words and phrases, slowly learning to piece together language into a more coherent structure—and all that without exuding the sense of emotional attachment that a child does to a parent.

Until recently, other animals were in a bad situation too. One would have thought their situation to be even worse than AI, because, as far as humans were concerned, they couldn’t even speak.


Octopuses don’t speak. They have other ways of communicating. Sporting a skin dotted with colour-changing “chromatophores”, their entire body can change colour and pattern depending on the situation.

This pattern is usually a combination of mottled grey-green, or muddy brown, or anything else that the background happens to look like. No matter where it finds itself, an octopus can totally meld in, looking like plants on the ocean floor or a part of a stray anemone. The amazing part is that all this happens despite the octopus being colourblind—and again, researchers have found evidence of a weird explanation. The octopus may be colourblind, but its skin, which can also sense light, isn’t.

Though there are many patterns the octopus could end up displaying, the message remains the same, and can best be translated into words as “Ignore me; I don’t exist.”

Newer evidence has shown that octopuses also communicate with each other through colour, at least to some extent. Octopuses are solitary creatures, and when two of them meet it usually means one of things: courtship, or battle. And herein lies an interesting pattern. In a 2015 study, scientists noticed that octopuses who put on darker colours were more likely to engage in battle, while a lighter shade indicated retreat. It wasn’t just about colours though; the octopuses also seemed to be signalling each other through aggressive arm-gestures and postures.

Courtship, on the other hand, involved dazzling stripes, as well as a “quiet half and half” pattern where one half of the octopus is completely black, and the other half is an equally vivid white.


Shaving off hair to soak toxic chemicals into your tender skin. Rubbing on lotions to cause a rash. Spraying strange liquids into your eyes to make them burn. These sound like painful punishments or tortures, but it’s actually what other animals, until the not too distant past, were routinely subjected to for no fault of theirs beyond the fact that they could not speak.

As long back as the 17th century, Western philosopher René Descartes claimed that animals, lacking consciousness, cannot feel pain. This convenient belief soon spread, replacing other, more inconvenient beliefs such as the worship of animal-gods or the necessity of asking a tree’s permission before taking its branches.

Animals with no consciousness were a boon for biology, starting from the frogs that were routinely dissected in classrooms to show children how the body works. Animals were—and still are—routinely used to test the safety of food, drugs, and cosmetics: a first line of defence to make sure new substances aren’t harmful to humans.

But today, things are changing. We hear talk of animal rights, and the need to be humane to animals; to treat them with care and make sure they don’t needlessly suffer.

What changed? Animals didn’t suddenly become conscious, but humans’ perceptions of them did. Ironically, it was biology that brought about this change, with all its new discoveries on how other animals are much more like us than originally thought.


Anyone who lives with a dog can tell you that dogs have dreams too. They will show you how, when deep in slumber, a sudden twitching of the leg or muted whines and yelps betray the thrill of the hunt—or whatever else is going on in the mind of the dog in question.

Unfortunately, we have no way of knowing exactly what our pets are dreaming about—or if they even dream in the same way humans do. Dogs can barely talk to humans in real life, so there’s little hope of them doing it in their sleep.


Do octopuses sleep? Astrobiologists argue that alien life might be so different from our own we may not even recognise it as life—what if, for example, life-forms on Mars look like rocks and move very slowly, at only a few centimetres a month? If an alien species with an extremely fast metabolism landed on our planet, would they zip by before we had time to move, and think we were just scenery like the rocks and the trees? They might then conclude that the only moving life-forms on this planet are flies.

Similar questions plague those seeking to study sleep. While “sleep” to us is just lying down with our eyes shut, neuroscientists have narrowed it down to mean certain brainwave patterns in an encephalogram—including alpha-waves, REM sleep, and all the rest.

Now here’s the thing: octopuses don’t have the same brainwave patterns as humans and more closely-related animals do.

Having narrowed down the definition of sleep, it’s now time to expand it again, though in a different way.


Octopuses don’t “sleep” in a human way, but they do go into a state where they’re resting, don’t move much, and are insensitive to the outside world. What’s more, this stage is associated with a brainwave pattern very different from a more active octopus—their own equivalent of a “sleep” pattern.

Looking to know more, a group of scientists kept octopuses in a tank for observation. Melding in with the scenery, the octopuses moved about their business until one of them was tired, and moved into a corner to rest.

Suddenly, the resting octopus began to display the distinctive “quiet half and half” mating pattern—although there was obviously no mate in sight.

Other resting octopuses have been observed with vivid colours and patterns which are nothing to do with the outside environment. Scientists are yet to investigate fully, but one can’t help but ask the question: are these colours offering us a glimpse into the mind of a sleeping octopus?

Are we observing octopus dreams?

Unfortunately, we’ll have to wait for more studies to confirm if this is indeed the case or if there’s a more mundane explanation. Because unfortunately, much like in Descartes’ time, a sleeping octopus has never been able to talk to humans directly.

And, neither have any of its arms.