The philosopher and historian of science Peter Godfrey-Smith, Professor at the University of Sydney, has studied cephalopods, particularly octopuses. In conversation with Curt Jaimungal for the format "Theories of Everything," he explained his views on mind, consciousness, and selfhood. He also develops a speculative theory according to which rhythmic electrical activity of the brain is essential for consciousness. Godfrey-Smith addresses these topics in his book Other Minds: The Octopus and the Deep Origins of Consciousness.

The Octopus as an Encounter with the Alien

Godfrey-Smith describes encountering an octopus as "probably the closest we will come to meeting an alien" that humans can experience. Octopuses, cuttlefish, and squid form a group of closely related animals. They possess large nervous systems and display complex behavior. According to Godfrey-Smith, there are good reasons to assume that octopuses are conscious. Less is known about squid, but in his assessment, their similarities to octopuses also point in that direction.

The evolutionary distance is large. All living things are connected through a common family tree. According to Godfrey-Smith, the last common ancestor of humans and octopuses lived about 600 million years ago. More recent estimates put it at roughly 570 million years. For most animals that are considered intelligent and probably conscious, by contrast, the common ancestor lies much closer. That ancestor was likely a worm about one centimeter long or less. It had a nervous system and lived in the sea.

Mind, Consciousness and Intelligence

Conceptual Distinctions

Godfrey-Smith understands the mind as a broad category. It encompasses two groups of properties:

• Intelligence-related properties: information processing, perception, memory and action selection
• Experience-related properties: subjective experiencing

In his view, both groups can be separated from each other. Intelligent systems can exist without sentience. Animals with little intelligence, on the other hand, can show signs of experience.

Selfhood and Information Processing

Godfrey-Smith sees mind at the intersection of two dimensions:

1. Selfhood: the property of being a bounded, self-maintaining whole
2. Information processing (cognition): perception, memory and action selection

In his assessment, a bacterium possesses a fairly pronounced selfhood, since it maintains itself and sustains its organization against the tendency toward dissolution. However, it does not have enough information processing to possess a mind. He therefore answers the question of whether bacteria have a mind in the negative.

Counterexamples show the weight of both dimensions:

• Sports teams, clubs, bands and companies have a great deal of information processing. Their selfhood, however, is low. It lies predominantly with the individual persons.
• Computers have no selfhood. They also lack some important features of information processing.
• Plants possess both dimensions only to a small degree. An oak tree resembles a collective more than a clearly bounded organism. Nevertheless, it has extensive signaling and information processing.

According to Godfrey-Smith, selfhood is a matter of degree. An organism such as a human possesses a great deal of selfhood. Humans have clear boundaries, can be counted, and maintain themselves both consciously and unconsciously. A group of children playing a game such as "Hot Potato" also forms a whole. Their unity, however, is minimal, since the children can separate at any time.

Terminological Dispute and Stipulation

Godfrey-Smith considers it unavoidable that terms such as mind be stipulatively defined. The concept is historically burdened with ethical, religious and everyday notions. Some researchers automatically associate mind with consciousness or responsibility. Others use a scientific term with different connotations.

As an example, he names Susan Schneider. She does not use the term mind for systems without any spark of consciousness. Godfrey-Smith himself, by contrast, sees consciousness as a special, refined property of some minds. In his view, mentality can be unconscious. He does not consider a necessary connection between mentality and consciousness to be given.

As a comparison, he refers to the physicist Wayne Myrvold, who explained in an episode of the podcast that entropy is not a single term but encompasses several concepts with similar character. Depending on the context, a different concept comes to the fore each time.

Gradation Instead of a Yes-No Distinction

Godfrey-Smith assumes that a scientific concept of mind will be graded. The question of whether an animal has a mind is then not answered with yes or no. Instead, one describes the extent to which certain features are present. For an earthworm, an answer in 50 years might state that it possesses a lot of one aspect of mentality and very little of another.

Anand Vaidya, with whom Jaimungal had previously discussed the topic, held the view that consciousness can be graded and that every thing possesses some of it. Jaimungal objected that the question then becomes whether there are things with zero consciousness. He compared this to a charged particle, which already counts as charged at a minimal but nonzero charge.

Godfrey-Smith considers it very likely that there are numerous objects without any mind at all. In his assessment, most things in his surroundings possess no mind whatsoever. He regards a distinction between a tiny but real portion and true zero as a conceptual remnant that a future specialist terminology will probably not adopt.

Analogy to Life

Godfrey-Smith compares the way we approach the mind with the way biology approaches life. Biology, he argues, has in part explained and in part demystified the concept of life.

From Vitalism to the Dissolution of the Concept

At the end of the 19th century, life was regarded as a special property that required its own explanation. Vitalism assumed that living things possess a distinct life force. Biology did not answer this question but dissolved the question itself. Today, one no longer asks where life begins or ends. Instead, one describes the emergence of self-maintaining chemical patterns.

According to Godfrey-Smith, biology describes life through two feature clusters:

Prions are misfolded proteins that, upon contact with proteins of the same kind, trigger the same abnormal folding pattern. They cause diseases in animals and humans.

Borderline Cases

Viruses possess many features of replication but have no metabolism of their own. They are complete metabolic parasites and use the metabolism of the infected cell. Viroids and simple systems from hypothetical origin-of-life scenarios, by contrast, possess a small share of metabolic features but no reproduction. Godfrey-Smith considers the question of whether such systems are "alive" unproductive. The term is best used without forcing borderline cases into a yes-or-no scheme.

In his assessment, the development of the concept of mind proceeds in a similar way. Here too, the question of a sharp boundary will lose its significance.

Schlaf, Traum und Evolution

Godfrey-Smith berichtet über Hypothesen, nach denen Schlaf der evolutionäre Ausgangszustand sei und Wachheit später entstanden sei. Er nennt Rodolfo Llinás, Neurowissenschaftler an der NYU, und Yuri Busacki, dessen Arbeiten er eingehend gelesen habe. Michael Walker, ein Schlafforscher, vertritt nach Jaimungals Angaben eine ähnliche spekulative Theorie. Godfrey-Smith hält diese Vorstellung für schwer nachvollziehbar. Ein Tier müsse Handlungen wählen und auf die Umwelt reagieren. Es könne nicht einfach nur ruhen.

Er verweist auf folgende Befunde:

• Bei Organismen ohne Nervensystem ist kein Schlaf oder Traum belegt.
• Bei Quallen, die kein Gehirn, aber ein Nervensystem besitzen, gibt es Hinweise auf Schlaf.
• Aktiver Schlaf, beim Menschen als REM-Schlaf bezeichnet, ist bei vielen Tieren verbreitet. Analoga finden sich bei Fliegen, Bienen und Oktopussen.
• Bei Einzellern wie dem Pantoffeltierchen (Paramecium) gibt es Hinweise auf Reaktionen auf Narkosemittel, die denen von Tieren ähneln.

Aktiver Schlaf geht nach Godfrey-Smith mit einem traumähnlichen Zustand einher. Die Verbindung sei jedoch nicht einfach. Menschen träumen auch außerhalb des REM-Schlafs, etwa im Non-REM-Schlaf. Dennoch bestehe eine enge Verbindung zwischen REM-Schlaf und Träumen. Daraus folge, dass ein traumähnliches Analogon bei vielen Tieren vorkommen könnte.

Bei Oktopussen lassen sich Hinweise auf Schlafzustände über Farb- und Musteränderungen der Haut beobachten. Die Haut wirkt dabei wie ein Fenster zum Gehirn. Ob Oktopusse träumen, ist nach Godfrey-Smith nicht bekannt. Dies sei bei Tieren schwer zu untersuchen. Er selbst hielt früher für sehr unwahrscheinlich, dass der gemeinsame Vorfahr von Mensch und Oktopus schlief oder träumte. Heute halte er diese Annahme für weniger sicher, da Schlaf offenbar weit verbreitet ist.

The Octopus Brain

The octopus brain differs greatly from the human brain. It has the shape of a ring that encircles the esophagus. Behind each eye lies an optic lobe. An unpaired region, the so-called vertical lobe, is considered the seat of more abstract cognitive functions. A cortex or a cerebellum is absent. Typical regions of the human brain have no clearly defined equivalents in the octopus.

About two-thirds of the neurons lie in the peripheral nervous system, that is, in the arms. There are strong horizontal connections between the arms. Some authors describe this as a second system. Others speak of a model consisting of a central brain and eight peripheral systems. Godfrey-Smith considers it unclear which description is accurate. What is decisive, he says, is that more than half of the processing power does not lie in the central brain.

The esophagus passes through the brain. Sharp pieces of food can therefore cause damage in the brain.

Autonomy of the Arms

The arms display behavior that appears autonomous. A relaxed octopus explores its surroundings with its arms without the animal as a whole acting in a targeted way. Godfrey-Smith assumes that there is a central self of the octopus that coordinates the arms and occasionally grants them autonomy. However, the arms do not possess enough selfhood to have a mind of their own. He considers it too far-reaching to assume that the arms have their own experiences. He cannot, however, rule this out.

Severed arms can, according to reports from experiments, still show a certain capacity for exploration. They then retain a remnant of their own ability to touch and probe.

Jaimungal drew a comparison with human limbs. When cycling, the arms and legs perform movements that are not consciously controlled. Godfrey-Smith considers such actions to be brain-controlled, even though they proceed unconsciously. He emphasizes that the distinction between conscious and unconscious is a point he views critically.

An Experiential Profile Instead of a Single Consciousness

Godfrey-Smith refers to David Armstrong, an Australian philosopher who described a thought experiment in the late 1960s or early 1970s. A driver covers a long distance without consciously registering the journey. He listens to the radio or thinks about the next day. At his destination, he realizes that he has driven about 100 miles.

Many philosophers and psychologists hold that conscious thought at any given time comprises only one or a few things. Stanislas Dehaene, a neuroscientist and psychologist, holds this position. According to Godfrey-Smith, he repeated it in a commentary on a paper by Anthropic about a global workspace structure in the AI system Claude. The concept goes back to Bernard Baars. Dehaene developed it further in his research, including together with Nakachi and others.

Godfrey-Smith does not consider this view compelling. He proposes an experiential profile. This profile consists of facets, not separate building blocks. It includes:

• the sensory-attentive facet, such as the screen in front of you
• mood
• energy level
• temperature
• a sense of the position of one's own body
• hunger and the time elapsed since the last meal

In his view, these facets contribute to what it feels like to be a particular person. Many people would instead assume that there is a rapid switching between contents and that only one thing is ever conscious at a time. Godfrey-Smith considers this an illusion that does not match his experience.

He also suspects that numerous things make small contributions to the experiential profile. One example is a text being written while attention is focused on a conversation. Some would classify such processes as entirely outside consciousness. Godfrey-Smith considers it more likely that they make a minor contribution. He concedes that this is difficult to test empirically.

The Question of Unity

Jaimungal asked how many units a conscious experience comprises. A color impression, such as a brown-black kettle, could count as one unit. A sentence or a mixture of senses could also be grouped together. Godfrey-Smith considers multisensory experience a sufficient criterion for it not being a single unit. His experiential profile encompasses several modalities, such as seeing, hearing, and bodily feeling.

Researchers in consciousness science speak of the visual "gist" of a scene. According to this, the entire scene can function as one component. Godfrey-Smith considers this possible when only vision is considered. However, when many modalities are involved at once, a description as a single unit is no longer possible.

In octopuses, according to Godfrey-Smith, it remains open whether there are multiple subjects and one mind, or the reverse. He suspects a central subject of the octopus that steers the animal as a whole. The arms could at times develop their own dynamics. He considers it possible that language will in the future develop terms for such intermediate stages.

Artificial Systems and the Simulation Hypothesis

Consciousness in Artificial Systems

Godfrey-Smith does not consider it well founded that today's computer systems are conscious. He finds this assumption unlikely, both because of the way they process information and because of their hardware. In his view, there is a hardware restriction on consciousness. He considers it possible that consciousness is tied to certain physical properties.

Regarding AI development, he said that he had long paid it little attention. Large language models, however, have greatly changed teaching at his university. He emphasizes that, as a traditional university lecturer, he values essays as a form of examination, and that his relationship with these models is burdensome for him. He now considers it necessary to engage with them more intensively. However, he is not yet able to answer the question of whether the formation of a global workspace is surprising. This concerns in particular the question of whether language models map human cognitive structures, since they are trained on human language.

The Simulation Hypothesis

Godfrey-Smith does not consider the simulation hypothesis to be seriously justified. It holds that humans are artifacts in a computer-based simulation. He regards the probability that some people ascribe to this hypothesis as too high.

He engages with a probabilistic argument advanced by Nick Bostrom and David Chalmers. According to this reasoning, future societies would, if they were able to, create numerous simulated agents. The number of simulated minds would then be greater than the number of biological minds. Godfrey-Smith considers the premise too far-reaching. He finds it surprising that many people agree with this premise.

According to Godfrey-Smith, Chalmers considers it possible, in his book Reality Plus, that there is a probability of about 25 percent that we live in a simulation. Godfrey-Smith also considers this value too high. He cites several reasons:

1. Technical difficulty: Virtual reality is visually impressive. Everyday experiences such as hot showers, falls, overeating, or waves while swimming, however, are physically diverse. In his assessment, they would be significantly harder to simulate.
2. Motivation: The more difficult it is to simulate realistic experiences, the less likely it is that future societies would create billions of such systems.
3. Hardware restriction: If consciousness is tied to biological hardware, a purely software-based simulation becomes less likely.

In this context, Godfrey-Smith distinguishes two scenarios. A scenario with biological brains in a matrix, in which cells and chemicals are present and receive only unusual inputs, is in his assessment far more elaborate than a purely software solution with computers and solar installations.

Rhythms and Consciousness

The Central Thesis

Godfrey-Smith develops a hypothesis according to which consciousness rests on two sides of brain activity:

• Networked point-to-point interaction: One cell fires and triggers the next to fire.
• Large-scale rhythmic dynamics: electrical oscillations, such as those recorded in the EEG.

In his view, both sides are closely connected but not identical. The oscillations are not simply a by-product. Many researchers previously regarded them as mere background noise. Godfrey-Smith now considers it more likely that these patterns play a unifying or organizing role.

Findings

Godfrey-Smith cites several connections between rhythmic activity and states of consciousness:

• Characteristic patterns occur in sleep and in the waking state.
• Under anesthesia, different patterns appear.
• During focused visual attention with eyes open, the patterns differ from those of an inattentive waking state with eyes closed.

According to his account, these connections do not exist only in humans. The research group of Bruno van Swinderen at the University of Queensland studies fruit flies (Drosophila). Their nervous system comprises several hundred thousand neurons. The flies show sleep, including active sleep, and attention. When they direct their attention to one of several objects in their visual field, a connection appears with a large-scale rhythmic pattern. It is a beta rhythm of about 20 cycles per second. The pattern resembles the one in humans, but is not identical to it.

Godfrey-Smith sees this as an argument for a functional role. The common ancestor of humans and flies lies far back in time. A similar resource is used in similar ways in both lineages. Van Swinderen and his colleagues write in one of their papers that this does not look like a coincidence. Godfrey-Smith considers this argument strong, but not one that can be pushed too far.

The Role of Rhythms

According to Godfrey-Smith, synchronization can serve a central function. The oscillations help to coordinate events within the brain in time. Without them, synchronization would be difficult to achieve. If synchronization is essential for consciousness, the oscillations would be a prerequisite without themselves being consciousness.

According to Godfrey-Smith, computers possess the networked side, at least if one understands neural networks as genuine networks. But they lack the second side. Jaimungal summarized this by saying that parts of brain activity have so far been regarded as noise that does not belong to consciousness. Godfrey-Smith agrees with this summary, but emphasizes that he is not simply proposing an addition to the functional model of consciousness. He considers the physical dimension important.

Artificial Neurons and Brain Replication

Thought Experiment

Jaimungal proposed a thought experiment: a single neuron is replaced by a silicon chip that behaves functionally the same. Godfrey-Smith considers this experiment the point at which he would drop out. A neuron has membranes, ion currents, and chemical sensitivities. These properties influence its excitability and its role in messenger substances such as nitric oxide. An artificial component without these features could show similar properties in the network, but not exactly the same ones.

In his assessment, a single artificial neuron would have only a very small effect. With thousands, hundreds of thousands, or millions of exchanges, however, the system would change. It would think differently and speak differently. Godfrey-Smith considers the idea of a behaviorally identical replacement a fantasy. He expects the change would become noticeable. He would probably feel strange if the synchronization through the oscillations were impaired.

Mind Uploading

Jaimungal asked whether mind uploading could be derived from this. Godfrey-Smith does not consider this idea realistic. If the brain were copied as biological tissue, the result would be a copy that quickly diverges. According to his account, biological systems are noisy. The result would be an independent entity with its own deviations.

He does, however, consider it possible that future hardware will emerge that resembles neither an ordinary computer nor a living brain, but combines the properties relevant to nervous systems. He describes this as an open question. He has no well-founded ideas about it.

Functional Side Effects

Jaimungal pointed to examples such as "junk DNA" and the microbiome, which were long considered insignificant and are today considered important. Godfrey-Smith agrees that this is a good example. He opposes the idea that organs have only a single function. A heart, for instance, is more than a pump. It contains fibroblasts, immune components, extracellular matrix, and macrophages. A pacemaker can take over the pumping function, but cannot replace the entire role of the heart. Whether this shows up in lifespan, he considers an unsuitable test.

Cephalopods: Cuttlefish, Squid and Giant Squid

Differences Between the Groups

Complex cephalopods form three groups: octopuses, cuttlefish and squid. Cuttlefish and squid are more closely related to each other than either of these groups is to octopuses. According to Godfrey-Smith, the split occurred far back, well over 100 million years ago.

According to him, cuttlefish display high behavioral complexity. The work of Alexandra Schnell, an Australian researcher, describes a test similar to the marshmallow test. In it, subjects choose between a lower-value option available immediately and a higher-value option available after a waiting period. The cuttlefish forwent the immediately available, lower-value food when they could wait for the higher-value food. They performed better on the test than many other animals and better than Godfrey-Smith had expected.

According to him, cuttlefish brains are roughly as large as those of octopuses. Cuttlefish display less dramatic behaviors. However, the Australian giant cuttlefish is curious and shows interest in humans.

Giant Squid

Godfrey-Smith explained that giant squid are difficult to study. Jaimungal reported on researcher Edie Widder, who developed about 13 years ago a light system called E-Jelly. During deep-sea dives, researchers found no giant squid. However, calculations derived from finds in the stomachs of sperm whales suggested a larger population. Widder suspected that light keeps squid away. With the light system, which resembles a certain jellyfish, a giant squid was reportedly lured within 30 seconds to one minute. It was the first live-captured giant squid from the deep sea.

Research Focus

Godfrey-Smith describes the question of the nature of mind as the central problem of his work. He pursues an approach that approaches mind through its components. These include selfhood and information processing. Explanation arises from an understanding of these aspects. He describes this as an analogy to life. He is working on a more precise elaboration of this approach.

For his own position, he uses the term "neural dynamics of subjectivity." The term emphasizes the importance of the neural, of dynamics, and of selfhood. He regards subjectivity as closely linked to selfhood, viewpoint, and perspective.

On the problem of consciousness, Godfrey-Smith has expressed the following. David Chalmers and Thomas Nagel had increased the pressure on materialist explanations. Godfrey-Smith shares this pressure, but rejects dualist and panpsychist answers, since he does not consider them substantiated. He refers to the concept of the explanatory gap, which Joe Levine coined. This gap exists between the physical and the experiential.

Currently, Godfrey-Smith is concerned with pain and pain-like experiences in various animals, particularly insects. In recent years, numerous works have appeared that make pain in insects more likely than before. This changes the ethical assessment. He emphasizes that he understands this as a development of research.