What this covers

Peter Godfrey-Smith joins Sean Carroll to argue that consciousness is not a sharp dividing line but a graded biological phenomenon that emerged independently in vertebrates and cephalopods. The conversation takes its bearings from a striking evolutionary fact: octopuses and humans last shared a common ancestor hundreds of millions of years ago—a creature with minimal nervous system complexity—meaning that the intelligence and apparent sentience of cephalopods represents a wholly separate evolutionary experiment in building complex minds. Rather than treating consciousness as a yes/no property, Godfrey-Smith proposes understanding it through Gradualism about Consciousness, drawing an analogy to how biology dissolved "life" into a cluster of independent properties like metabolism and reproduction that can vary separately.

Central to his account is the Dual-Process Brain (Network + Oscillation)): nervous systems operate through both discrete neuron-to-neuron connections and large-scale diffuse electrical oscillations, the latter playing a genuine cognitive role that resists reduction to network properties alone. The discussion ranges across evidence of pain and behavioral complexity in octopuses, bees, and crustaceans—each illustrating how sentience appears distributed across species without a clean threshold. Godfrey-Smith explores octopus cognition through its unusual embodied style: how their semi-autonomous arms reflect a decentered form of control, how they recognize individual humans despite being asocial creatures, and how they handle novelty through manipulation and exploration rather than the reflective problem-solving of crows. The conversation also addresses whether Subjectivity as Evolved Perspective can be fully compatible with physicalism, and what it might mean for an animal's experience to be radically Decentered Selfhood rather than unified as in humans.

Sharpest takeaway

Godfrey-Smith argues that consciousness/sentience evolved more than once independently (vertebrates and cephalopods), so it should be understood as a gradualist, biologically-grounded phenomenon rooted in the evolution of points of view plus the dual network/oscillatory properties of nervous systems—not as a sharp yes/no natural kind.

  • Octopuses and humans share a pre-Cambrian common ancestor lacking higher cognition, so their complex nervous systems are independent evolutionary experiments
  • Brains have a dual character: point-to-point network firing plus diffuse large-scale oscillatory dynamics found across the animal kingdom
  • Evidence of pain and novelty-handling across octopuses, bees, and crustaceans supports a long, graded list of sentient animals rather than a hard cutoff

The argument · threads5 threads · 27 claims
0.80

Nervous systems combine point-to-point computation with diffuse electrical dynamics to produce integrated subjective experience and point of view.

5 pointscentrality 5/5
  • Against the orthodox view that a brain is essentially a network of point-to-point neuron firings (with everything else being life support), Godfrey-Smith argues nervous systems have a dual character: discrete cell-to-cell connections plus large-scale diffuse electrical oscillatory dynamics (e.g., gamma oscillations, sub-threshold ion movements detectable by EEG) that are not straightforwardly reducible to network properties yet play a genuine cognitive role.

    “you also have large-scale dynamic patterns which are not straightforwardly reducible to those network properties. The simple example... is the sorts of patterns that you see in an EEG reading where there are oscillatory rhythms in a brain”

  • Godfrey-Smith's two-part biological account of felt experience holds that to be conscious is to be an animal whose nervous system gives it subjectivity—a point of view forged by the evolution of action—combined with the peculiar gestalt-producing dual properties of nervous systems; this implies a gradualist picture with no sharp cutoff between sentient and non-sentient animals.

    “to be conscious is to be an animal with a nervous system that has subjectivity as a feature of its relationship to its environment. It has a perspective it has a point of view.”

  • Subjectivity and point of view are not reasons to abandon a physical or biological understanding of consciousness; rather, an animal is an evolved nexus between incoming sensory causal lines and outgoing action, so 'point of view-ness' is itself an evolutionary product that can be wholly physically and biologically comprehended.

    “An animal is a kind of nexus between the causal lines coming in through the senses and the causal lines going out through action. Point of view-ness is an evolutionary product I think.”

  • A nervous system that combines point-to-point computational interactions with diffuse oscillatory electrical dynamics gets a natural integrative role 'for free,' because many simultaneous sensory events perturb the state of the whole system at once—which maps onto the gestalt-like, unified character of everyday felt experience.

    “many different events will simultaneously affect the state of the whole and the actions going on in the whole, there's a kind of integration for free that you get when you have a system with this combination of properties”

  • Over the last 30 years the word 'consciousness' shifted from meaning a sophisticated, reflective form of experience to simply meaning whether 'it feels like something' to be an organism—largely due to Thomas Nagel's influence—and this terminology is not permanent and will likely mutate as science and philosophy advance.

    “is a squid conscious? Would not have meant, meant what it tends to mean now, which is just, does it feel like something to be a squid? And that is what it has come to mean largely as a consequence of Tom Nagle's enormously influential work.”

0.70

Octopuses exhibit behavioral markers of consciousness including novel sequential action and likely felt pain despite lacking social scaffolding.

3 pointscentrality 5/5
  • A behavioral indicator of consciousness suggested by researchers like Stanislas Dehaene is the production of novel, temporally organized, non-routine behaviors—because in humans the capacity to act unconsciously breaks down precisely for novel sequential behaviors—and octopuses are full of such novelty-oriented, non-routine behaviors, supporting attributions of felt experience.

    “the evidence that we can do a lot of things unconsciously stops when you get to novel, temporally organized behaviors, and sequential behaviors that are new. Then people are conscious of what they're doing. And octopuses are full of behaviors like that.”

  • While consciousness itself may not depend on the external/social world, many cognitive capacities do: the internalization of external cognitive tools—language above all, as in Vygotsky's idea of inner speech—is a 'gift from the public to the private' that makes individual minds more powerful but could never have existed without a social setting; the octopus shows you can be a solitary animal with a messed-up self and still be conscious, so social rebuilding of the mind matters for human consciousness specifically rather than consciousness in general.

    “Inner speech and linguistic organization of thought more generally is a kind of gift from the public to the private... it's something that would never have existed if there wasn't a social setting in which the tool for interaction developed.”

  • Robyn Crook's octopus pain study makes felt acute pain quite likely by converging multiple lines of evidence: octopuses avoid locations associated with an acetic-acid injection, favor locations associated with analgesic drugs, and groom and tend injured spots—shifting the contrast away from 'mere reflex versus pain' toward 'integrated multifaceted aversive handling with or without experience.'

    “The octopuses groomed and tended the injured spot, wound tending is often seen as quite good evidence for a felt experience of something like pain.”

0.61

Cephalopod complexity evolved independently as a second data point for studying cognition, equal in evolutionary distance to other invertebrate lineages.

4 pointscentrality 4/5
  • The last common ancestor of humans and octopuses was probably an early bilaterally symmetrical, worm-like organism around a centimeter in scale, living roughly 565–600 million years ago during the Ediacaran period—well before the Cambrian explosion—with a simple nervous system, possible eye spots, and a very simple behavioral repertoire.

    “our best guess of the depth in time of the common ancestry is about 600 million years ago... it's during a time that's called the Ediacaran”

  • Complex behavior evolved largely because the Cambrian explosion (~540 million years ago) introduced a new regime of interaction—predation, kill or be killed, and the need to track surrounding animals—and this happened across several already-diverged body plans (arthropod, vertebrate, cephalopod), so each lineage evolved behavioral complexity in its own way.

    “once the Cambrian explosion began there was a new regime of interaction, predation, kill or be killed. You've got to start to track what animals are around you or one of them is liable to eat you. Complex behavior evolved in that context”

  • Octopuses are as evolutionarily distant from humans as ants or bees—the common ancestor of a human and an octopus is the same animal as the common ancestor of a bee and a human—so cephalopod complexity represents a separate evolutionary experiment in building large nervous systems, giving us effectively a second independent data point for studying cognition and consciousness.

    “Octopus is the same distance from us as an ant or a bee. So the common ancestor between a human and an octopus is the same animal as the common ancestor between a bee and a human.”

  • Octopuses are largely asocial—they must meet only to mate—which is striking because sociality is often regarded as a great evolutionary spur to intelligence; yet they can still learn to recognize and respond to individual humans who treated them well or badly, an ability with no obvious rationale in ordinary octopus life, suggesting their complexity yields surprising 'bonus' capacities.

    “octopuses are not very social. For a long while they were regarded as pretty much completely asocial. They have to meet to mate”

0.57

Octopus nervous system organization is massively decentralized and may generate radically decentered or centerless experience unlike vertebrate unity.

6 pointscentrality 3/5
  • Honeybees and bumblebees, with only about a million neurons—roughly a 500th the size of an octopus nervous system—can perform sophisticated feats including cultural learning where one bee watches another solve a problem and then solves it itself, which shows we should not put too much stock in raw neuron counts.

    “bees can do things like cultural learning. They can see one bee solve the problem and then solve the problem itself... they have about a million neurons, so a 500th the size of an octopus brain”

  • Because the octopus has a more decentralized, distributed nervous system, its locus of experience may be radically decentered or even center-less—contrasting with the greater unity of vertebrate experience—and this is a feature where our imaginative capacity to picture what it is like runs into real difficulty.

    “is the locus of experience radically decentered or just, or center-less in a way that contrasts with our own case... the vertebrate organization is probably one that gives more, makes for more unity, more centeredness of experience than the octopus nervous system”

  • There appears to be light sensitivity in the skin across the octopus's whole body, so that perhaps 'washes of light' are experienced through whole-body light sensitivity, though it is probably too strong to say they can see with their skin.

    “There appears to be some light sensitivity in the skin of the whole body itself. And probably saying too much to say they can see with their skin, but a kind of perhaps washes of light are experienced”

  • Although octopuses appear to show personality differences, biologists are cautious because true personality requires individual differences that remain stable over time, which is hard to track in such animals.

    “a personality is supposed to remain fixed over time... it's not a personality difference if you're nice today and you are not nice tomorrow.”

  • Alexandra Schnell's work showed that cuttlefish can exert self-control, foregoing an immediate lesser reward to obtain a later larger one—an analog of the human marshmallow test—suggesting cephalopods handle time in a more sophisticated way than expected for such animals.

    “she showed that cuttlefish... can exert a degree of self-control. They can forego an immediate lesser reward in order to get a slightly late... A later larger reward. It's an analog of the marshmallow test”

  • The common octopus has roughly 500 million neurons, putting it well below humans but within the range of vertebrate nervous systems; about two-thirds of these are distributed around the body, especially the arms, while the central brain alone holds well over 100 million neurons.

    “has at most recent measures, about 500 million neurons, in its nervous systems... Two thirds of them are spread around the body, especially in the upper arms”

0.56

Octopus cognition operates through embodied exploration rather than reflective rumination, revealing a genuinely different cognitive style.

3 pointscentrality 4/5
  • Unlike crows and parrots, which solve novel problems through a reflective, ruminative style—pausing to turn the problem over before acting—octopuses approach novel situations by involving their complex body, manipulating and tasting objects, so calling them simply 'smart' misleads because they have a different cognitive style rather than less of the same one.

    “That kind of ruminative style that we share with crows, for example, is not present in them.”

  • The octopus exhibits a top-down/bottom-up division of control: it can pull itself together and direct an arm visually with relative precision when it has reason to, but arms also engage in semi-autonomous exploration—evidenced by severed arms acting coherently for a while—because the central-brain-to-arm connections do not appear suited to fine-grained control of every detail.

    “the octopus can kinda pull itself together and direct an arm visually in a relatively exact way when it has reason to... but when you watch an octopus just hanging out... you see what looks at least like independent or semi-autonomous exploration that the arms are doing”

  • The octopus's large nervous system may have evolved primarily as a solution to the problem of controlling a soft, shell-less body with effectively endless degrees of freedom—where any point on an arm can become an elbow—and once that control capacity existed, it was a relatively easy evolutionary road to becoming behaviorally complex and exhibiting smartness.

    “Hanlon and Messenger suggested that the octopus, the large nervous system, evolved as a solution to this problem of just getting the damn body to do useful things”