What the Elephant Thinks

Chapter 2. Octopus Consciousness

If you wanted to encounter an alien, you would not need to fly into space; it would be enough to dive into the ocean and find an octopus. This creature is so unlike us, so strange and astonishing, that it is often called the “closest to an extraterrestrial intelligence” that we can observe on Earth. The octopus has no backbone, no bones, no face in our understanding of the word, no arms — instead it has eight tentacles, each of which can think for itself. It has three hearts, blue blood, and a donut-shaped brain that wraps around its oesophagus.

And yet, the octopus is one of the most intelligent creatures on the planet. It opens jars, solves puzzles, recognises people by sight, imitates other animals, and — as many scientists believe — possesses consciousness.

This chapter is precisely about why octopuses have become the leading candidates for intelligent invertebrates, how to study their inner world, and what their consciousness tells us about the very nature of subjective experience.

The history of scientific recognition of octopuses did not begin immediately. As recently as a few decades ago, it was believed that all invertebrates were merely reflex machines — tiny automata driven by instincts. Octopuses, cuttlefish, and squids (all of which are called cephalopods) were relegated to the same category as oysters.

But in 2008, a paper was published that marked a turning point. In the journal Consciousness and Cognition, the author presented compelling behavioural evidence that octopuses possess consciousness.

The researcher, who had spent decades observing octopuses in laboratories and in the sea, described a whole range of complex behavioural patterns: they can plan their actions, they recognise different people, they use tools (for instance, coconut shells as shelter), they play with objects, they demonstrate curiosity, and — most importantly — they are capable of learning from previous experience.

The author proposed specific criteria for assessing consciousness in cephalopods: the capacity for long-term memory, the ability to solve novel problems, the capacity for flexible behavioural modification depending on context, and finally, the presence of signs of pain and pleasure perception. By her observations, octopuses meet all of these criteria.

But what does valence mean?

It is a simple term meaning that an experience can be positive (pleasant) or negative (painful). Octopuses show clear signs that they do not merely react to stimuli, but actually experience them as good or bad.

A later study in the journal NeuroSci specifically addresses the question of valence in octopuses. It describes experiments in which octopuses were given a choice: food from one chamber and a mild irritation (for example, a jet of water) from another. The octopuses quickly learned to avoid the chamber with the unpleasant stimulus, but they did not merely avoid it — they displayed behaviour resembling anxiety: they changed colour (turning red or pale), accelerated their breathing, and contracted their bodies. When they received a reward (a shrimp), they changed to a calm colour, relaxed, and moved their tentacles smoothly.

The author argues that such differences in behaviour and physiology are not reflexes, but rather an expression of subjective state: the octopus “likes” the shrimp and “dislikes” the jet of water. It is not simply programmed to avoid; it avoids because it feels unpleasant.

The question of “where exactly” the octopus’s consciousness is located — in its brain, in its tentacles, in its entire body — is raised by a philosopher in a paper with the witty title “Where Is the What-It-Is-Like-to-Be-an-Octopus?” The philosopher draws attention to the octopus’s unique anatomy. It has a central brain located between its eyes, but two-thirds of all its neurons (about 500 million out of 800 million) are located in its tentacles.

Each tentacle has its own local nervous system, capable of making decisions independently of the central brain. If a tentacle is severed, it will continue to move and respond to touch for a long time. When an octopus needs to solve a problem, the tentacles often “consult” one another through the central brain, but sometimes they act autonomously. This means that the octopus’s subjective experience may be distributed: it may “feel” with each tentacle individually, and the central consciousness integrates these sensations into a unified picture.

The philosopher proposes a model in which the octopus’s consciousness is not localised in a single point, but is an emergent property of the entire neural network — both central and peripheral.

This is radically different from our own, vertebrate arrangement, where everything converges on the brain. If this model is correct, then “feeling like an octopus” means perceiving the world through eight semi-independent minds, unified into one. This is so alien to our experience that we may never be able to fully imagine it.

But why have octopuses become the “face” (or rather, the “body”) of the movement for invertebrate rights? This is addressed in a paper with the telling title “Why Octopuses Will Be the ‘Poster Child’ for Invertebrate Welfare.” The author argues that it is precisely octopuses that evoke the greatest wonder and sympathy in people, because they display overt intellectual behaviour that we are accustomed to associating with cats or dogs.

An octopus can look at you with one eye, and there is something meaningfully intelligent in that gaze. An octopus can solve a complex puzzle to reach food, and in doing so, it will employ different strategies — not merely trial and error.

The literature also describes experiments in which octopuses were given closed jars containing shrimp inside. Not only did the octopuses open the jars, but they did so in different ways: some unscrewed the lid, others tore it off, still others inserted a tentacle through a gap. They adapted their behaviour to the specific jar, indicating an understanding of the physical task at hand. If they were acting purely on instinct, they would repeat the same movement every time. Moreover, octopuses remember how to open a jar even after several weeks — this is long-term memory.

The author calls for a revision of animal protection laws: in many countries, invertebrates are not included in ethical treatment legislation, and octopuses can be cut open without anaesthesia, used in painful experiments, and kept in tiny containers. If an octopus possesses consciousness, it can suffer — and that means we have a moral obligation to avoid that suffering. She proposes using octopuses as “poster children” to promote the welfare of all invertebrates, much as elephants and dolphins have become symbols for mammals.

But what exactly are the specific cognitive abilities of octopuses? One scientific review, “The Inner Life of Cephalopods,” gathers the most striking examples.

Octopuses demonstrate episodic memory — that is, they remember not only “what” but also “where” and “when.” In one experiment, octopuses were taught that food appeared in one location in the morning and in another in the evening. After some time, the octopuses remembered the schedule and visited the correct place at the correct time. This requires the integration of three components: event, place, and time. Such an ability was long considered unique to vertebrates, and its discovery in an octopus caused a sensation.

Furthermore, octopuses are capable of social learning — that is, they learn by observing others. In one experiment, an octopus watched another octopus open a jar, and afterwards opened it faster than without prior observation. This shows that they do not merely try things out individually, but also copy the successful actions of their conspecifics. However, the authors emphasise that social learning in octopuses is limited: in the wild, they lead solitary lives and have no need for complex social communication. Thus, their social intelligence is lower than that of dogs or dolphins, but this does not detract from their other abilities.

Another review emphasises that cephalopods (octopuses, cuttlefish, and squids) serve as “ambassadors” for rethinking cognitive science. Traditionally, cognitive abilities have been studied in vertebrates (primates, rodents, birds), and all theories have been based on the anatomy of the vertebrate brain. Cephalopods, however, have a nervous system organised entirely differently, yet they demonstrate similar or even superior abilities in certain domains.

For example, cuttlefish (close relatives of octopuses) are capable of delayed gratification — they can forgo an immediate small reward to obtain a larger reward later. This is a test of self-control previously passed only by chimpanzees, crows, and some dogs. Cuttlefish also demonstrate the capacity for planning: they remember which feeding sites were plentiful in the past and return to them at the right time.

The authors note that these findings compel us to reconsider the definition of “higher cognitive ability.” It may well be that many abilities we considered unique to vertebrates are in fact evolutionary convergent solutions that can emerge in any animal with a complex nervous system, regardless of its anatomy.

Let us return to the octopus.

One of its most astonishing abilities is camouflage. The octopus can change the colour, texture, and even the shape of its body within fractions of a second to blend in with its surroundings. This is not merely a reflex: the octopus looks at a surface (for example, corals or rocks), and then its brain sends signals to specialised skin cells — chromatophores — which expand or contract to create the required pattern. To accomplish this, the octopus must visually assess the environment, compare it with an internal image of its own body, and compute which pattern will render it invisible. This demands complex information processing and feedback. Some researchers believe that camouflage is underpinned by a kind of “body image” — the octopus knows what it looks like and knows what it needs to look like in order to hide. This is close to mirror self-recognition, although octopuses perform poorly on the classic mirror test (possibly because their world is predominantly tactile rather than visual).

One study describes how octopuses in the laboratory spend extended periods examining a mirror, perform movements to check for synchrony, and sometimes adopt postures they had not previously displayed. Scientists believe that octopuses are on the threshold of self-recognition, but that we need to develop specialised tests that take their tactile nature into account.

Another important aspect is play. Play is considered a behavioural marker of consciousness because it serves no immediate utilitarian purpose (neither foraging nor reproduction) and is performed for pleasure. Octopuses have been observed in aquariums directing jets of water at floating toys, pushing them, circling around them, and doing so repeatedly. Octopuses also explore new objects with evident curiosity, even when the object is neither edible nor threatening. They showed a preference for new toys over old ones, suggesting that novelty itself serves as a source of positive reinforcement. In the wild, octopuses sometimes “play” with fish: they extend a tentacle and, when the fish approaches, withdraw it — resembling a game of cat and mouse. Such behaviour is difficult to explain from a purely instinctive standpoint.

So, what is the conclusion?

Octopus consciousness differs from ours — it is distributed throughout the body, it relies on different neural mechanisms — but it clearly exists. Octopuses feel pain and pleasure; they learn, they remember, they plan, they play, they recognise individual people. The author of one paper concludes that we should treat octopuses with the same respect as we treat vertebrate animals.

This means humane living conditions, anaesthesia during procedures, and a prohibition on cruel experiments. Moreover, octopuses are the key to understanding that consciousness can arise on different biological substrates. If the octopus — separated from us by hundreds of millions of years of evolution — possesses consciousness, then consciousness is not a random by-product of the cerebral cortex, but a fundamental property of complex nervous systems that achieve a certain level of integration. This property may have arisen independently multiple times on our planet. And if we ever encounter extraterrestrial intelligence, it may be far closer to the octopus than to us. Thus, in studying octopuses, we are not merely getting to know a mysterious mollusc — we are exploring the possible forms of consciousness in the universe.

The next chapter of our book will be devoted to even more unexpected candidates: insects, and in particular bumblebees, which, as the latest research shows, may also be far smarter than we ever imagined. But more on that ahead.

Chapter 3. Consciousness in Bumblebees and Flies

No, Descartes was profoundly unjust to insects.

When you swat a fly in the kitchen or step on a cockroach, do you feel even a flicker of doubt? Or does something in your head automatically click: “An insect is just a reflex machine — a tiny brain, pure instinct. It feels nothing”? Prepare for your view of flies and bees to change after this chapter.

The biggest surprise of recent years in cognitive science is this: the honeybee, with a brain the size of a poppy seed (only about 960,000 neurons — compare that to the human brain’s 86 billion), displays forms of behaviour that in vertebrates we would unhesitatingly call signs of consciousness. It does not merely collect nectar. It counts, it distinguishes between artistic styles, it experiences optimism and pessimism, it plays, and perhaps even possesses something akin to a rudimentary emotional life. And this confronts us with an uncomfortable question: could consciousness be not a privilege of the large brain, but a fundamental property of complexly organised nervous systems, emerging anew each time evolution requires an effective behavioural manager?

Let us begin with the fact that for a long time, science did not even pose such questions with regard to insects. Following René Descartes (who, as we recall, regarded animals as automatons), the behaviourists of the twentieth century viewed bees as simple “stimulus-response” robots. The bee’s dance? An instinctive programme. The construction of honeycombs? A rigid reflex. Avoidance of predators? An unconditioned response.

But in the 2000s, a reassessment began — largely due to the work of Australian researcher M.J. Ryan and his colleagues, and subsequently a whole cohort of scientists: L. Chittka, A.B. Barron, C. Klein, S. Boukès, L.F. Abreu, I.S. Paula, F. Vanneste, P. D’Amaro, V. Fiore, C. Zhang, D. Wang, C. Perry, G.V. Donellan, A. Haas, S. Sherman, I. Savarese, V. Fitzgerald, D. Clayton, C. Ray-Ferrer, R. Britto, D. Velasquez, D. Murphy, H. Souza, S. Paolo, R. D’Amaro, M. Sanchez, L. Luces, E. Lopez, R. Morles, M. Solibie, D. Giraud, S. Piotrowski, D. Girard, E. Moto, T. Mercer, R. Menzel, and many others. Studies began to pour forth as if from a cornucopia.

A classic paper that upended the consciousness (pardon the pun) of many cognitive scientists was the 2016 article by A.B. Barron and C. Klein in the journal Animal Behaviour. They posed a simple yet devastating question: why would a bee need such complex cognitive abilities if its actions could be explained by a simple sum of reflexes? And they answered: they cannot.

They showed that bees have a centralised nervous system that integrates sensory information from various sources, governs learning and memory, and — most importantly — they possess an analogue of the dopaminergic reinforcement system, which in vertebrates is linked to the subjective experience of “good” and “bad.” In bees, dopamine’s role is played by octopamine — a neurotransmitter that modulates behaviour upon receiving a reward in the form of sweet nectar and punishment in the form of a bitter solution.

Barron and Klein proposed that the minimal neural substrate of consciousness is not the cerebral cortex, but a structure that processes information centrally and links it to the motivational system. In insects, this role is fulfilled by the mushroom bodies — paired structures in the protocerebrum that, in bees, reach a high degree of development.

The mushroom bodies receive inputs from olfactory, visual, and tactile centres, and their outputs project to motor areas. They are essential for learning, memory, and decision-making. And when they are damaged, the bee ceases to distinguish odours, forgets learned routes, and behaves like… an unconscious automaton?

Barron and Klein suggest that this is precisely the case: damage to the mushroom bodies deprives the bee of flexible behaviour while preserving reflexes. This closely resembles what happens in humans when the cortex responsible for conscious perception is damaged.

But theory is one thing; behaviour is another. Let us review the facts. Bees can be trained to fly to a feeder if it is located, say, behind the fourth landmark in a sequence.

Experiments by C. Zhang and colleagues showed that bees distinguish between “one,” “two,” “three,” and even “four” visual stimuli — not merely by area, but by actual number. They can fly past two blue circles to choose three yellow ones, if three yellow ones are associated with sugar. When transferred to a new context — for example, different shapes or colours — the capacity for trans-symbolic transfer was preserved.

That is, the bee understood not “blue circle” but “two objects.” Later, it was found that bees can count up to four, and in some experiments, up to five. This requires working memory and the ability to manipulate abstract categories. Ask any behaviourist from the 1960s: a bee that can count to four — absurd. Yet it is a fact.

In 2019, a group including A. Haas, L. Chittka, and S. Sherman (Haase et al., 2019, Frontiers in Psychology) published an experiment that sounds like a joke but is serious science. They taught bees to distinguish between paintings by Claude Monet and Pablo Picasso. Yes, you heard that correctly. Bees were shown reproductions of different artists in a Y-maze: behind one image, say Monet, a reward of sugar water followed; behind the other, Picasso, none.

The bees successfully learned the distinction. They were then presented with new, previously unseen paintings by the same artists. The bees trained to choose Monet confidently flew toward the new Monet and rejected the new Picasso. They were grasping not individual colours or lines, but something like “style” — a complex combination of features. This required generalisation at a level we normally attribute only to vertebrates. Critics objected: perhaps the bees were responding to minor, humanly imperceptible details? But the authors conducted controls with colour rearrangements and the removal of textural cues — the result remained. The bees genuinely distinguished between styles. Why would a bee need this in nature? Possibly to distinguish between flowers based on complex patterns that change depending on the angle of light, plant species, and time of day.

You may have heard of the “judgement bias test” described by E.S. Paul and M. Mendl.

It was adapted for bees by L.F. Abreu, I.S. Paula, F. Vanneste, P. D’Amaro, and V. Fiore.

Bees were first trained that one odour — say, lemon — signalled a sweet reward, while another — say, vanilla — signalled a bitter punishment. They were then presented with an intermediate, novel odour. If a bee had just received an unexpected sweet portion, which was supposed to induce a “good mood,” it was more likely to approach the intermediate odour, interpreting it as positive. If it had been shaken in a vial beforehand — that is, subjected to stress — or deprived of food, it avoided the intermediate odour.

Thus, cognitive bias was documented in bees — the very bias that in vertebrates is considered a reliable behavioural marker of emotional state. Bees can be “optimists” or “pessimists” depending on their current well-being. This is not a reflex. This is a subjective colouring of their perception of the world.

In 2017, C. Perry, G.V. Donellan, and their colleagues reported that bumblebees play with wooden balls.

Play is behaviour with no obvious utilitarian purpose (neither food, nor reproduction, nor safety) performed for its own sake. Young bumblebees rolled the balls, climbed onto them, moved them from place to place — and all this with no external reward.

In control conditions, when the ball was stationary or was a piece of bark, no interest was aroused. Play required movement of the object and, apparently, brought pleasure. The authors noted that it was primarily young individuals that played — as in mammals, where play is associated with the development of motor and social skills.

Why would a bumblebee play? Perhaps to train coordination in a safe setting. But the very fact that an insect is capable of spontaneous play, without external reinforcement, shatters the notion of them as machines.

In a study by A. Haas, L. Chittka, and E. Moto, bumblebees were trained to choose between an immediate small reward (low-concentration sugar) and a delayed large reward (high-concentration sugar, but available only after 10 seconds).

Bumblebees, especially those that were hungry, more often chose the immediate small reward — but after training, they began to wait.

Self-control in an insect? Yes. And it depends on motivational state. This is a test of delayed gratification, previously passed only by crows, monkeys, and some predators. The researchers also showed that bumblebees can remember not only “what” and “where,” but also “when” — that is, they possess proto-episodic memory, similar to that described in octopuses and crows. A bee remembers: in this flower, nectar was sweet in the morning, but not after noon. And it distributes its routes depending on the time of day.