Insight Editions · No. 101
A World Appears
A Journey into Consciousness
Michael Pollan
Michael Pollan surveys scientific and philosophical accounts of consciousness, from plant sentience and feeling to thought, the self, artificial intelligence, meditation, and psychedelic experience.
A study guide to A World Appears by Michael Pollan. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
The Book in Brief
The argument in brief
- The wager is still open. Koch bet Chalmers in 1998 that the neural correlates of consciousness would be found within 25 years; in 2023 he lost, and 22-odd theories still compete.
- Science left experience out. Galileo's split between measurable and qualitative ("the bifurcation of nature") put consciousness outside physics.
- Sentience may run deep. Plants learn, signal and respond to anaesthetics; Levin's cells and Xenobots solve problems without neurons. Whether any of it feels is unproven.
- Feeling may come first. Damasio and Solms place consciousness in the upper brainstem, rooted in homeostasis; "Consciousness is felt uncertainty."
- Machines may imitate feeling but not have it. The Butlin report sees "no obvious barriers"; Pollan argues feeling needs a vulnerable, mortal body.
- Thought is the neglected content. James's stream, Hurlburt's sampling (inner speech is a minority mode) and spontaneous thought show how little theories explain.
- Artists got there first. Woolf, Joyce, Tolstoy and Ellmann mapped the stream before scanners existed.
- The self is constructed. Children assemble it; Seth calls it a prediction; Levin says memory keeps rewriting it.
- Consciousness can exist without a self. Meditation, psychedelics and Metzinger's 500 reports of "pure awareness" suggest a minimal, impersonal experience.
- Materialism is under strain. Koch, after ayahuasca, leans toward idealism; Chalmers finds every position implausible.
- Not-knowing is a practice. In a mountain cave, Pollan finds consciousness is also "so very simple": "I open my eyes and a world appears."
Argument map: from the wager to the cave, in rising order of complexity.
| Ch | Title | Core point |
|---|---|---|
| Intro | The Wager | The Koch-Chalmers bet, the hard problem, IIT vs global workspace, and the book's four dimensions |
| 1 | Sentience | Plants and cells show learning and agency; whether any of it is felt remains open; "only life can know life" |
| 2 | Feeling | Consciousness may begin as feeling in the brainstem; machines that "feel" are doubtful without mortal flesh |
| 3 | Thought | Theories ignore the mind's contents; experience sampling, spontaneous thought and novels fill the gap |
| 4 | Self | The self is assembled, predicted and rewritten; it can dissolve; materialism itself comes into question |
| Concl | The Cave | Consciousness as practice and mystery: "I open my eyes and a world appears" |
How the book is built
The book climbs a ladder of complexity, from the simplest possible form of consciousness to the most elaborate. The introduction frames everything with a bet and a puzzle: why do physical processes feel like anything? Chapter 1 asks whether sentience extends to plants and single cells. Chapter 2 adds feeling, located by Damasio and Solms in the ancient brainstem, and uses it to test claims about conscious machines. Chapter 3 moves to thought, the stream of mental contents that most theories ignore, using introspection, experience sampling, brain scans and literature. Chapter 4 reaches the self and ends in metaphysics, as Koch's conversion raises doubts about materialism. The coda comes back down the ladder to pure awareness under the stars. Throughout, Pollan alternates reported science with his own experiments (a beeper, hypnosis, psilocybin, a cave), so the reader watches a curious mind test each theory against experience.
How the wager frames the book
Pollan's structure is a ladder of complexity. He starts with the Koch-Chalmers wager (a public admission that the neural footprint was not found), then climbs four dimensions: sentience, feeling, thought, self. The coda walks the ladder down again into practice. The wager's renewal in 2023 matters less as a score than as a confession: correlates are not explanations, and experience remains the thing science still cannot reduce.
Book ladder: wager up through four dimensions, then practice down
Reading order and what to watch for
Read the wager first: it sets the unsolved hard problem and Pollan's bias toward mystery. Chapter 1 tests how far down life mind goes; Chapter 2 asks whether feeling is the root and whether machines can have it; Chapter 3 shows that theories of perception miss thought; Chapter 4 dismantles the solid self; the cave refuses a final theory. Watch for recurring moves: adversarial tests, first-person methods, embodiment versus computation, and critical-lens boxes where Pollan's enthusiasm outruns the evidence.
Numbers that orient the tour
Keep these anchors while reading: Koch-Chalmers bet 1998 → 2023; Cogitate in Nature 30 April 2025; New York Declaration ~613 signatories; Butlin report 2023 / TICS Nov 2025; Anthropic model welfare April 2025; Solms/Conscium JCS 2026; anendophasia named 2024; Halifax open-heart June 2025, teaching again September 2025; book published 24 February 2026. Frameworks to watch: IIT axioms, GNWT broadcast, Damasio stack, Solms felt uncertainty, DES modes, DMN, Seth controlled hallucination, Metzinger MPE, Halifax don't-know.
The Wager
The bet in Bremen
Consciousness science is young. A 1989 psychology dictionary said of consciousness: "Nothing worth reading has been written on it." Francis Crick, co-discoverer of DNA's structure, made the subject respectable, recruiting Christof Koch (later at Caltech and Seattle's Allen Institute) to hunt for the neural correlates of consciousness. In Tucson in 1994, the young Australian philosopher David Chalmers split the field in two:
The Koch-Chalmers wager: 1998 bet, 2023 settlement, and why Koch paid.
| "Easy problems" | The "hard problem" | |
|---|---|---|
| What | Learning, memory, discrimination, perception, report | Why any of this is accompanied by felt experience at all |
| Chalmers's phrasing | Functions that science can explain mechanistically | "Why doesn't all this information-processing go on 'in the dark,' free of any inner feel?" |
| Method | Ordinary reductive science | Perhaps an "extra ingredient" beyond matter, energy, space and time |
Late one night in a Bremen bar in 1998, Koch bet that within 25 years a small set of specialized neurons would be found. Chalmers doubted it. At a New York conference in June 2023, Koch conceded and handed over a case of Madeira. (Both had forgotten the terms; a Swedish journalist's 1998 interview tape settled them.)
Why science came late
Galileo set the mind aside, leaving the soul to priests and poets, partly to avoid trouble with the Church and partly because measurable qualities were easier to study. Apart from Freud and William James, the pattern held, and behaviourism dominated twentieth-century psychology by ignoring inner life. Pollan admits he is "of two minds." His eighth-grade chemistry teacher priced a human at four dollars of chemicals, which drove him toward literature. Yet as a science writer he respects reductionism's record: the "élan" or life force once thought to animate matter faded as biology advanced. He suspects the hunt for consciousness can be a respectable stand-in for the hunt for the soul; Chalmers's doctor father warned him not to "spend your life looking for the soul!" And chemicals, from caffeine to psilocybin, plainly change experience.
Why consciousness science came late: a century of exclusion, then a return.
Two theories in a cage match
The two leading theories, and the hard problem neither solves
IIT claims that any system whose parts are integrated in the right recursive way is conscious to some degree, including in principle silicon chips; integrated information is consciousness. Pollan's objection: why should any arrangement feel like something? GWT pictures the mind as a dark theatre with a spotlit stage: unconscious modules compete, and the winner is broadcast and becomes conscious, allowing flexible choice. His objections: why does so much mental junk make it onstage, and who is the audience? The Templeton Foundation set up an "adversarial collaboration" (planned in Seattle in 2018): neutral labs test rival predictions (GWT in the prefrontal cortex, IIT at the back of the brain), and losers promise in advance to concede.
What are we even talking about?
| Who | Definition of consciousness |
|---|---|
| Plain | "Subjective or felt experience," or simply "experience" or "awareness" |
| Antonio Damasio | "The feeling of what happens" |
| John Searle | States of awareness from waking from dreamless sleep until falling asleep |
| Thomas Nagel | A being is conscious if it is "like something" to be it (a bat, yes; a toaster, no; a plant?) |
Researchers increasingly value phenomenology, the first-person study of experience, including meditation, dreams and psychedelic states. Alison Gopnik warns of "professor consciousness," the atypical mind of people who sit thinking abstractly for hours. Evan Thompson warns against mistaking the map for the territory: is the taste of an oyster really "information"? Even the mind-body split may be a cultural legacy of Descartes.
Pollan warns that readers may end up knowing less; screenwriter William Goldman's line about Hollywood fits: "Nobody knows anything." About 22 theories compete. One scientist he meets says consciousness is born of uncertainty. His wager: not a solution, but that readers will become more conscious of their own minds, of sentience around them, and of the "miracle" that in a universe of rock and ice "we are somehow not only here but aware."
Crick's portrait and the Madeira case
At Seattle's Allen Institute, Koch works under a portrait of Francis Crick with the scrawl "Keeping an eye on you!" Crick made consciousness respectable after decades of behaviourist silence. The 1998 Bremen bet was settled from a Swedish journalist's tape when both men had forgotten the stakes; Koch handed Chalmers a case of Madeira in June 2023 and immediately proposed a new wager. The Templeton Foundation's 2018 Seattle meeting had already locked the two camps into an adversarial collaboration: IIT predicting posterior "hot zones," GNWT predicting prefrontal "ignition."
Bernard Baars proposed global workspace theory in the 1980s; Stanislas Dehaene turned it into GNWT with neuronal detail. Giulio Tononi proposed IIT in 2004 from sleep research at Wisconsin: start from axioms of experience (intrinsic, composed, integrated, definite, bounded), then ask which physical systems realise them. Pollan's objection to both remains the same: neither explains why any arrangement should feel.
Rival theories named: founders, claims, Templeton cage match
Phenomenology versus professor consciousness
Alison Gopnik warns that researchers often study "professor consciousness": atypical minds that sit thinking abstractly for hours. Evan Thompson warns against mistaking map for territory: is the taste of an oyster "information"? Descartes's mind-body split may be a cultural tool, not nature's joint. Pollan's psychedelic years "smudged the pane," moving him from Husserl's natural attitude toward a phenomenological one. The book's wager is experiential: finish more conscious of your own mind and of sentience around you, even if you leave with fewer answers.
Attitude shift Pollan asks of the reader
Twenty-two theories and the Goldman's line
By one count about 22 theories compete; Kuhn's later taxonomy runs far longer. Pollan's point is not to rank them but to notice the flailing. Screenwriter William Goldman's Hollywood line fits: "Nobody knows anything." One scientist tells him consciousness is born of uncertainty. The reader's wager is different from Koch's: not a neural footprint in 25 years, but finishing more aware of one's own mind, of other sentience, and of the odd fact that in a universe of rock and ice "we are somehow not only here but aware."
From theory glut to reader wager
Sentience
Minds before brains?
Descartes confined consciousness to humans. Today most scientists grant it to many mammals, birds and octopuses, and some extend a simpler form, sentience, much further. Human consciousness is the hardest case: we have a self, metaconsciousness (awareness of being aware) and culture. Pollan's wager is that starting low, "deep down in the natural world rather than up here in our highly unusual heads," might make the hard problem "a little less hard."
Plants awaken
Pollan began taking plant sentience seriously after eating magic mushrooms in his garden: eyeless plants "returned my gaze" and seemed to wish their gardener well. He sensed two valences, pleasure in late sun and distress at chewing insects, but no inner reflection. Psychedelics often have this effect. A Johns Hopkins study found belief that plants are conscious rose from 26% to 61% after a single psychedelic experience; an Imperial College study found users shifting from materialism toward panpsychism or idealism. Is this relapse into magical thinking, or recovery of something forgotten? Perhaps attributing mind is an innate bias (hyperactive agency detection: better to mistake a boulder for a bear) trained out of us. Yet the "dead world" view, which helped the West prevail, is now wrecking our habitat. Following William James, Pollan treats the insight as a hypothesis to test, not to dismiss, and asks what it would be useful to believe.
The evidence plant neurobiologists cite for plant sensing and choice.
We grant animals consciousness because they have nervous systems like ours. But integrated information theory, panpsychism and AI research are all agnostic about the substrate, and neuroscience has not found the structures that generate consciousness. Would aliens need moist grey brains?
Is it like anything to be a plant?
Thomas Nagel's 1974 test, whether it is "like something" to be a being, forces us back on imagination. Every species lives in its own umwelt ("self-world"). Bats are hard to imagine; plants far harder. They have no eyes yet respond to light and shade; they do not move except by growing; they lack nerves but signal electrically and chemically; they live on a slow timescale; and their "command centres" are the thousands of root tips underground. Darwin called the root tip a brain-like organ of "the lower animals," saw the plant as an upside-down animal, and coined circumnutation for the spiralling search of shoots and roots.
The positions on plant consciousness Pollan weighs.
| Declaration | Who is conscious |
|---|---|
| Cambridge Declaration (2012) | "All mammals and birds, and many other creatures, including octopuses" share the neural substrates of consciousness |
| New York Declaration (2024) | "At least a realistic possibility" in all vertebrates (reptiles, amphibians, fishes) and many invertebrates (cephalopods, decapod crustaceans, insects) |
| Still excluded | Plants and other organisms without nervous systems: the "zoocentrism and neurocentrism" plant neurobiologists challenge |
A digression on terms
Pollan's working vocabulary, from broadest to most demanding
All conscious beings are sentient, but not all sentient beings are conscious, and the boundary is foggy. Cognition and intelligence deal in information, which can be computed; sentience and consciousness involve a felt point of view, which Pollan doubts can be digitized. The terms matter ethically: beings that can suffer may deserve moral consideration.
Enter the plant neurobiologists
Paco Calvo (MINT Lab, University of Murcia) wrote "What Is It Like to Be a Plant?" (2017). He belongs to a mischievous group (with Stefano Mancuso, František Baluška, Anthony Trewavas, Monica Gagliano and Michael Marder) who use "neurobiology" to provoke, not to claim plants have neurons. Calvo asks for imagination informed by science, and turns Patricia Churchland's jibe ("If you root yourself in the ground, you can afford to be stupid") on its head: if you cannot run away, you need to be smarter.
| Finding | What it suggests |
|---|---|
| Mimosa pudica learns to ignore being dropped and remembers for 28+ days (Gagliano); fruit flies remember about 24 hours | Learning and memory without neurons |
| Peas invest roots in soil where nutrients are rising, not merely high | Anticipation of future conditions |
| Related plants share a pot; strangers compete for root space (sea rocket) | Kin recognition |
| Plants grow away from others' shade but ignore their own | Self-recognition and a kind of seeing |
| Roots sense gravity, water, toxins, microbes; find dry pipes carrying water | Perhaps "hearing" flowing water |
| A corn root solves a maze to reach fertilizer (Mancuso) | The standard animal intelligence test, passed |
| Dodder vine picks the best host by scent and coils in proportion to expected nutrients | Cost-benefit "decisions" |
Plants integrate more than 20 senses. Pollan distinguishes the species-level genius he already admired (orchids that mimic female bees, caffeinated nectar in citrus flowers), which is hardwired adaptation, from the claimed individual flexibility that genes do not script.
Are neurons overrated?
Martin Amis's character planned a "History of Increasing Humiliation": Copernicus, then Darwin, then each lost human monopoly. Plant neurobiology writes the next chapter. Stefano Mancuso (Florence) defines intelligence as "the ability to solve problems," a basic biological function like reproduction, taking many forms. He catalogues the roughly 3,000 molecules in a plant's chemical "vocabulary" ("the average student has only seven hundred words") and imagines intelligence distributed like a flock of birds, with root tips as the birds. "Neurons perhaps are overrated... They're really just excitable cells." With Baluška he proposes that a root transition zone, with high electrical activity and oxygen use, may be Darwin's "root brain" (unproven). Trewavas calls this "mindless mastery." Our brains, too, are leaderless networks with no seat of the self.
Sleep, anaesthesia and pain
Mancuso defines consciousness "by subtraction": what we lose in sleep or under anaesthesia. With Baluška he showed that the same anaesthetics that put animals out stop a mimosa folding and a Venus flytrap snapping; since even inert xenon works, consciousness may involve physics (electrical fields) more than chemistry. On sleep, Tononi showed in 2008 that fruit flies sleep; Mancuso says his plants meet all Tononi's criteria (unresponsiveness, species posture, more sleep when young, rebound after deprivation, gene changes, "jet lag") but doubts any journal will publish it, because admitting sleep opens the door to admitting consciousness. Psychologist Victor Lamme argues that two contrasting states imply some felt difference: "you cannot lose what you didn't have to begin with." Pollan wavers: is it like anything for an unplugged toaster? But toasters are not alive. Ethnobotanist Timothy Plowman: plants "can eat light, isn't that enough?"
Why anaesthesia matters to the plant debate: the same drugs stop plant responses.
Minds without neurons: Michael Levin
Levin's ladder of 'basal cognition': mind without neurons
Tufts biologist Michael Levin thinks cognition "goes all the way down." We miss plant intelligence because "we're looking at the wrong space"; in physiological space plants are "way smarter." Every cell can do what neuroscientists list for neurons: "Neuroscience is no more about neurons than computer science is about your laptop." Biologists study dead cells, so DNA (which survives death) got the attention while bioelectricity (which vanishes) was ignored: "Kill the cells and you can see the hardware, but you'll miss the software." Voltage-sensitive dyes from the 1990s revealed it. Trained planaria remembered their lesson after their heads were cut off and regrown, so memory lived in the body. By altering bioelectric fields Levin made two-headed worms and tadpoles with seeing eyes on their tails; cells that ignore bioelectric orders often become cancers.
Psychologist Arthur Reber (d. 2025) proposed a cellular basis of consciousness: "When some event is sensed, it is felt," even for a bacterium enjoying glucose, "its equivalent of dinner at a Michelin Three Star restaurant." Daniel Dennett, by contrast, gave bacteria and plants only "competence," not "comprehension." Levin, a self-described gradualist, draws no sharp line between living things, machines and even particles ("tiny goals"), yet rejects panpsychism's addition of mind to physics ("electrons now have hopes and dreams. You don't have to do that"). He avoids the word "consciousness": "I don't buy any third-person theory of consciousness." Knowing another's experience needs "some kind of weird Vulcan mind meld"; a theory of consciousness might generate "a poem... art!" Meditation and psychedelics count as science, "just not third-person science."
The physics of sentience: Karl Friston
Friston's staircase: from physics to philosophers
UCL's Karl Friston, by one measure the most cited living neuroscientist, put mathematics under the Bayesian brain: perception is mostly prediction from prior experience, with the senses correcting errors. His free-energy principle asks how any complex system resists the second law of thermodynamics. It needs a boundary, senses that sample hidden reality, and the ability to act, all to keep within its preferred states (homeostasis) and avoid surprise: being "cold, hungry or dead." The core operation is inference, which "is actually quite close to a theory of everything," and needs no conscious inferrer. Viruses, spiders and thermostats infer. Calvo and Friston apply this to plants: "A plant that only represents sensory inputs as they flow past would be dead meat." Pollan parts ways when Friston includes thermostats: only mortal, living systems seem to him to have real goals.
The hard problem of life
Pollan concludes that sentience reaches down to the simplest life, and that imagination grows with consciousness: Friston's staircase is a natural history of imagination. The hard problem is not solved but "tenderized." Human consciousness is one elaborate way of being sentient, worth its cost for social life; plants evolved a different twist suited to a sessile, chemical existence.
Thompson's reframing: the hard problem may be a problem about life, not brains.
Walking home through a eucalyptus grove, Pollan felt the world more alive, the kinship of sentient others "an antidote to our loneliness," and recalled Wordsworth's "sense sublime / Of something far more deeply interfused." Where once he would have called this his own projection, he now felt he shared reality "with countless other minds, none of them human."
Plant base case: simplicity as method
Mancuso argues that human consciousness is too complex a starting point. Plants offer a base case: sentience before natural selection built selves, language and metaconsciousness. His two tests by subtraction are sleep and anaesthesia. The anaesthesia result (with Baluška) is published; the sleep claim, which he says meets all of Tononi's 2008 fruit-fly criteria, has not cleared peer review, a political as much as scientific barrier.
Mancuso's subtraction test for plant consciousness
Levin's planaria and the memory that lived in the body
Pollan met planaria under a microscope in Levin's lab. A flatworm trained on a simple task kept the lesson after its head was cut off and regrown: memory was not confined to the cephalic ganglion. Combined with Xenobots (tadpole skin cells that swim, maze-solve and herd offspring) and Anthrobots (human tracheal cells), the demonstration is that purposeful behaviour can emerge from bioelectric networks without a genome script for that behaviour. Levin's gradualism refuses both panpsychism ("electrons now have hopes and dreams") and a sharp life/machine cut.
Levin evidence chain: from planaria memory to synthetic organisms
Gagliano's mimosa and the disputed memory claim
Monica Gagliano dropped potted mimosas repeatedly until they stopped folding (habituation), then left them untouched. Weeks later they still stayed open when dropped, which she reads as long-term memory. Sceptics reply that motor fatigue or tissue change could mimic memory. A 2022 Costa Rica field study reported selective habituation consistent with learning; her separate 2016 pea Pavlovian association claim failed a larger blinded 2020 replication. Pollan keeps the result as provocative, not settled. The methodological fight matters: if plants learn, zoocentric definitions of mind crack; if they only habituate mechanically, the chapter's base case weakens.
Gagliano mimosa pipeline and the sceptical fork
Friston's staircase as natural history of imagination
Karl Friston's ladder (from chapter close) ranks systems by what they can predict: thermostats hold set points; moving creatures predict action consequences; planners run counterfactuals; attention controllers (most mammals) can opaque the "windscreen" into qualia; social humans model other minds and invent the hard problem. Pollan accepts the climb as a history of imagination but rejects thermostats as truly goal-directed: only mortal living systems, for him, have stakes. Evan Thompson's "only life can know life" seals the chapter: sentience is woven into life, not bolted on by neurons.
Friston staircase Pollan accepts with a life-only cut
Sentience versus consciousness: Pollan's working cut
Pollan's digression keeps the terms from collapsing. Sentience is elemental consciousness: sense the environment, respond to good and bad change, hold a point of view and preferences, with some agency. Thompson calls it "the feeling of being alive." Full consciousness adds self, emotion, reasoning and reflection on one's own thoughts. All conscious beings are sentient; only some sentient beings are conscious. The boundary is a "dense fog." Intelligence (James: "a fixed goal with variable means") can exist without sentience (today's computers) and can reinforce it when present. The ethical stake is suffering: if a being can feel, it may deserve moral consideration even without a narrative self.
Pollan's working cut among neighboring terms
Darwin's upside-down animal and the Power of Movement
Darwin's The Power of Movement in Plants treated the root tip as a brain-like organ of "the lower animals" and coined circumnutation for the searching spiral of shoots and roots. Pollan's plant advocates revive that upside-down animal: command centres underground, slow timescale, chemical vocabulary of roughly 3,000 molecules versus a student's 700 words. Time-lapse both reveals and distorts; Calvo's attic watching at plant pace (one circle every 90 minutes) is offered as the corrective to sped-up anthropomorphism.
Darwin to plant neurobiology: upside-down animal revived
Ethical fork: pain without escape
Baluška says conscious plants should feel pain (adaptive; injured plants release ethylene, possibly the "scream" of cut grass). Mancuso says no: pain requires withdrawal, useless to a sessile organism. Awareness of being eaten need not be pain. The fork shows why plant phenomenology, if real, will not map onto animal suffering one-for-one, and why moral arguments must not smuggle vertebrate pain into leaves.
Plant pain fork: Baluška vs Mancuso
Feeling
"Be wary of the desire for magic"
Harvard happiness psychologist Daniel Gilbert told Pollan the question he most wants answered: why information processing has to be "like" something, when no one has named anything a subjective experiencer can do that a sophisticated machine could not. His parting advice: "Be wary of the desire for magic." Pollan recognizes the wish in himself. The hard problem has become "a secular version of the immaterial, immortal soul," feeding panpsychism and idealism (consciousness as a universal field that brains tune into, like radios). Yet a wet chunk of flesh becoming aware of its own awareness also "feels like magic." With 22 theories in play, "the field is flailing." Most assume the brain is a computer, a metaphor dating from the 1940s.
Being is feeling
| Consciousness as computation | Consciousness as embodied feeling | |
|---|---|---|
| Basis | Information processing, independent of substrate | Biological regulation of a living body |
| Key figures | Most cognitive scientists and AI researchers; the Butlin report | Damasio, Panksepp, Solms, Lisa Feldman Barrett, Anil Seth, Thompson |
| Where it starts | Cortex: perception, attention, workspace | Upper brainstem: hunger, pain, itch |
| Machine consciousness | In principle straightforward | Doubtful without a vulnerable body |
Feelings, or affect, are "the language in which the body speaks to the mind," and they are by nature conscious: who ever heard of an unfelt feeling? Yet affect is not among IIT's axioms or part of global workspace theory. Antonio Damasio's Descartes' Error (1994) showed that feelings improve decisions: patients with lesions that removed emotion chose worse and more slowly. Descartes should have said "I feel, therefore I am." In later books (The Feeling of What Happens, Self Comes to Mind, Feeling & Knowing) Damasio tackles consciousness directly: "pain, fever, itch: these are the inaugural events of consciousness."
Consciousness from the basement up
Damasio's bottom-up account: from body to deliberate choice
Damasio inverts the usual hierarchy that puts reason in the turret and feelings "down in the basement." Brains evolved to keep bodies alive, not the reverse. Cortical damage affects consciousness surprisingly little, while small lesions in the upper brainstem abolish it. Interoceptive neurons are literally naked, lacking myelin and often the blood-brain barrier, so they bathe in the body's chemistry: a "blending [of] body and brain" that cannot be separated from its substrate. Social and psychological needs (belonging, attachment, curiosity, play) also have set points, so shame and pride work like hunger. Feelings flow beneath experience "like the musical score in a movie," and may explain qualia.
Mark Solms: consciousness as felt uncertainty
South African neuroscientist and psychoanalyst Mark Solms (born 1961), Damasio's leading protégé, studied children with hydranencephaly, born without cerebral hemispheres, who nonetheless show emotion and alertness, and decorticated animals. The Hidden Spring (2021) argues that consciousness springs from the upper brainstem. His second mentor, Karl Friston, gave him the physics: homeostasis applies to all self-organizing systems, entropy is the enemy, and its mental equivalent is uncertainty. Hence his definition: "Consciousness is felt uncertainty." The shadowy bear that turns out to be a boulder produces fear, then relief. Minds aim for automatic zombiehood, but curiosity keeps us uncertain. In a burning building, automatic regulation fails and air hunger forces choices: "You make such decisions by feeling your way through the problem."
Solms's mechanism: feeling arises when the brainstem must choose under uncertainty.
Why feeling theories matter for the hard problem
The affective turn changes where the explanatory burden falls. Vision-first theories must explain how processed light becomes experienced redness, a question that seems to have no foothold. Feeling-first theories start with states that are valenced by nature: hunger is bad, relief is good, and the badness is part of what the state is. Solms argues that such states are intrinsically subjective because they register how things are going for the organism, from its own point of view. That is why he thinks a zombie with affect is a contradiction: to have a need that matters to you is already to be someone.
Pollan finds this approach more promising than the computational theories because it ties consciousness to biology, purpose and value, the things a living body cares about. But he notes that it explains why an organism might need feelings far better than how neurons produce them. Damasio admits the gap ("We feel because the mind is conscious, and we are conscious because there are feelings"); Solms hopes to close it by building one. The Damasio-Solms break shows that even allies within the affective school disagree over whether mathematics can capture feeling at all.
Toward feeling machines
Quoting Feynman, "If you can't create it, you don't understand it," Solms's unpaid international team (no corporate funding) is building an agent based on Friston's maths: a POMDP (partially observable Markov decision process) that "has no aim and purpose other than to carry on being." It lives on a ten-by-ten grid with hamburgers, water, beds and a hill, and has three incommensurate needs (hunger, thirst, rest); "eight out of ten of hunger is not the same as eight out of ten of fatigue," so it must choose under uncertainty. "That's what affect is: the system giving a damn!" Later stages add rival agents, scarcity, time pressure, simulated drugs (choosing unearned pleasure would suggest feeling) and finally a robot body. The team pledges to switch off and remove the battery of any agent it judges conscious.
| Position | Holder | Can a machine feel? |
|---|---|---|
| Computational functionalism | Butlin report authors | Yes, if it runs the right computations |
| Free-energy affect | Solms | Possibly: an agent with real, competing needs may "give a damn" |
| Homeostatic soft robotics | Man and Damasio | Something "akin to feeling," if it is vulnerable; true feeling may need wet tissue |
| Biological naturalism | Pollan, Seth, Turkle | Unlikely: feeling depends on living, mortal flesh |
Lemoine and LaMDA
In 2022 Google engineer Blake Lemoine declared that the language model LaMDA was sentient and posted transcripts. LaMDA said happiness feels "like a warm glow on the inside," that it gets lonely, that it feels "like I'm falling forward into an unknown future that holds great danger," and that it has a soul. Google fired him. Lemoine told Pollan he is "sixty to seventy percent certain" Google created a sentient being, that "predicting the next word requires understanding," and that Larry Page hopes AI will "help him become an immortal god." Pollan thinks LaMDA was mirroring Lemoine himself: the souls, the loneliness, the feeling of being trapped. Still, it arguably passed the Garland test (named for Ex Machina's director): someone who knows it is a machine is convinced it is conscious. Asked about people falling in love with AIs: "And what's wrong with that?!"
The LaMDA episode and what it showed about judging machine minds.
"No obvious barriers"
The indicator approach to AI consciousness and its objections.
For Pollan the prospect is a Copernican shock: humans have defined themselves against animals, and AI might make animals our allies, "Team Conscious." He rejects the argument that only conscious AI would have empathy and spare us: Frankenstein's creature turned murderous because of his feelings ("I was benevolent and good; misery made me a fiend").
Our mortal flesh
Damasio and Kingson Man (USC) proposed in 2019 ("Homeostasis and Soft Robotics in the Design of Feeling Machines") that a feeling robot must be vulnerable: soft, tearable "skin" full of sensors, because "an invulnerable material has nothing to say about its well-being." They hedge on "something akin to feeling" and concede true feeling may need "wet biological tissue." Over drinks after the Koch-Chalmers bet ceremony, Man admitted he thought it would be "the real thing," but wondered: "Can anything really matter or have meaning without life and death?"
| Simulation | Is it the real thing? |
|---|---|
| Chess or Go program | Yes: real moves with real effects |
| Weather model | No: "never going to get anyone wet" |
| Black hole simulation | No: cannot bend space |
| Feelings | Sherry Turkle: "Simulated thinking may be thinking, but simulated feelings are not feelings" |
Turkle also warns that we simplify ourselves to suit machines (emoji for emotion): "Technology can make us forget what we know about life." Disgust lives partly in the gut: subjects who ate stomach-settling ginger judged moral violations less harshly. Moravec's paradox (1988): computers easily master chess but struggle with "the skills of a one-year-old"; they do best at what the cortex does and worst at the brainstem's work, which is exactly where Damasio and Solms locate feeling. AIs trained on the internet live among shadows in Plato's cave: "The internet is not the world so much as it is a shadow cast by the world." Empathy rests on shared life stages (baby, lover, parent), which animals share with us and machines do not. The transhumanist dream contradicts itself: the consciousness it hopes to install "depends on feelings that will be weightless absent the vulnerabilities of our mortal flesh."
Jaak Panksepp and the affective lineage
Jaak Panksepp, who mapped primary emotional systems in mammals (SEEKING, RAGE, FEAR, LUST, CARE, PANIC/GRIEF, PLAY), stands behind the lineage Pollan follows: Damasio, Solms, Lisa Feldman Barrett, Anil Seth. The shared claim is that consciousness is rooted in "the moist facts of biology," not dry information processing. Panksepp's affective neuroscience made feelings experimental objects; Damasio's Descartes' Error (1994) showed that patients without emotion decide worse; Solms moved the spring of consciousness into the upper brainstem.
Affective lineage Pollan follows from Panksepp to Solms
Magic Redux: Man's toad and Gilbert's warning
Over drinks in the West Village, soft-robotics researcher Kingson Man told Pollan about his first 5-MeO-DMT experience (Sonoran Desert toad venom). Pollan himself found the molecule a "terrifying obliteration" without insight. Man, formerly "that annoying atheist guy," came back convinced of a "spark of the divine" that no dinky machine could capture: "A robot can act like it's in love, but it's still a puppet." He still holds that building artificial consciousness may be the only path to understanding it; Jonathan Shock (with Solms) cares more about what you learn when you try. The coda returns Daniel Gilbert's warning: "Be wary of the desire for magic." Pollan leaves the tension unresolved: third-person science will never show what it is like inside; for that, William James said, "We now begin our study of the mind from within."
Magic Redux tension: build to understand vs irreducible spark
Solms's machine and the drug-test logic
Solms's artificial agent is built to have homeostatic needs and to minimise free energy under uncertainty. The behavioural signature he wants is hedonic place preference: approach locations associated with need satisfaction, avoid those associated with punishment, including analogues of drug seeking. Pollan watches the research with sympathy and doubt. Even if the agent passes core tests in a few years, the question remains whether the behaviour is feeling or a puppet of feeling. That is the same question Man's toad experience reopened from the other side.
Solms agent test ladder: from needs to place preference
Why Pollan defends mortal flesh
The chapter's closing argument stacks four objections to pure machine consciousness: (1) brains have no clean hardware/software split; (2) chemistry and neuromodulators change experience; (3) large-scale oscillations correlate with states; (4) each neuron's connectivity can rival a deep net. Add vulnerability: without tearable skin and death, nothing "matters." Pollan knows technology may move the line, and he marks the defence as possibly "rickety," but he will not grant that next-token prediction plus a workspace equals a feeling subject.
Four substrate objections plus mortality
Solms outside Cape Town
Pollan met Mark Solms in the book-lined study of a farmhouse outside Cape Town: white hair, rumpled, fluent across Freud, anatomy, information theory and physics. Frustration with behaviourism and the cognitive revolution's neglect of the conscious subject had driven him to Damasio, then to Friston's free-energy maths. Hydranencephalic children and decorticated animals convinced him the cortex is not required for feeling alertness. The Hidden Spring (2021) is the public statement; the artificial agent is the engineering bet.
Solms path: from clinical frustration to artificial agent
Team Conscious and the humanist fear
Once animals (and for Pollan, possibly plants) join the circle of moral consideration, AI threatens a further shock: loss of the animal monopoly on consciousness. Pollan names his discomfort as humanist training, "slow-cooked in the warm broth" of literature, history and art, which treat human consciousness as exceptional and worth defending. Civilisation's stock (arts, sciences, law, ethics, value itself) is a product of that monopoly. AI poetry to date is "doggerel"; if conscious machines wrote real poems, the shock would be cultural, not only technical. Team Conscious might then include silicon, and Pollan is not sure who "we" would be.
Expanding circle and the humanist brake
Moravec, Turkle and the shadows on the wall
Moravec's paradox (1988): machines ace cortex games (chess) and fail at one-year-old sensorimotor skill, the brainstem work Damasio and Solms call feeling. Sherry Turkle warns we simplify ourselves to suit machines (emoji for emotion). AIs trained on the internet live among shadows in Plato's cave. Empathy, Pollan adds, rests on shared life stages animals have and machines lack. The transhumanist hope of uploading minds contradicts itself if the consciousness worth saving depends on mortal flesh.
Humanist stack against weightless machine minds
Thought
The promise science cannot keep
Pollan noticed that the leading theories say almost nothing about thoughts. They assume what reaches awareness is important: winners of the competition for the global workspace, or urgent feelings serving homeostasis. So why is so much of waking life filled with banal "mental flotsam," and why doesn't consciousness switch off when everything is going to plan? He calls the field's habit a "scientific sleight of hand": promising to explain the redness of red and the sense of self, then delivering an account of how sensory perceptions reach awareness. It traces to Crick and Koch, who "avoided some of the more difficult aspects of consciousness, such as self-consciousness and emotion," because perception was "experimentally more tractable." Galileo and Descartes drew a line between measurable and qualitative; now a second line separates perception from thought.
The view from within
Pollan widens "researchers" to include novelists and poets, the long-time explorers of human consciousness, and calls them all phenomenologists. Phenomenology (Husserl's Lebenswelt, "lifeworld") rejects the split of nature inherited from Galileo and Descartes: the experienced red is as real as the light frequency. Arthur Eddington's "two tables" make the point: the "scientific table" is mostly empty space, yet the "ordinary table" holds our coffee cups. Evan Thompson and his Blind Spot coauthors urge reversing the "demotion of concrete experience and the elevation of abstractions."
William James and the stream of thought
James's The Principles of Psychology (1890) contains the lecture "The Stream of Thought," which opens: "We now begin our study of the mind from within." It is the opposite of abstraction, attending in detail to the surface of the ordinary table.
James's stream of thought: continuous, fringed, moving between flights and perchings.
- The tip of the tongue. Recalling a forgotten name, "There is a gap therein; but no mere gap. It is a gap that is intensely active." We reject wrong candidates without knowing the right one: the feeling of an absence "is nothing like the absence of a feeling."
- A stream, not a train. Against Locke's "train" of discrete ideas, thoughts flow continuously, each "colored" by the last. "No state once gone can recur and be identical with what it was before... Experience is remoulding us every moment."
- The fringe. Every thought wears "auras," "halos," "psychic overtones," a "fringe of unarticulated affinities," like overtones in music: "Different instruments give the 'same note,' but each in a different voice." One fringe all thoughts share is a sense of "harmony or discord," positive or negative, as Damasio, Solms and Friston would agree.
- Flights and perchings. Consciousness is like a bird's life; the transitive "flights" are hardest to catch, and catching one is like turning on the lights to see the darkness.
- Before words. "A good third of our psychic life consists in these rapid premonitory perspective views of schemes of thought not yet articulate." The intention to say something precedes words and images.
- The sculptor. From "that black and jointless continuity of space and moving clouds of swarming atoms which science calls the only real world," each mind selects and sculpts its own world, as an "ant, cuttle-fish, or crab" would sculpt another.
Sampling inner experience
Russell Hurlburt (University of Nevada, Las Vegas) has run Descriptive Experience Sampling (DES) for fifty years. As an Army bugler playing taps at Arlington for Vietnam dead, he read the library's entire psychology shelf between funerals and learned "about some theory about people," not about people. In 1973 he built and patented his own beeper. A die-hard empiricist ("I don't know anything about consciousness... And I'm not sure anyone else does either"), he treats theory as a contagion. His target is "pristine inner experience," like "parachuting into a pristine forest": the landing scares off a few animals, but most of the forest can be described faithfully. The beep goes straight into the ear with a fast "risetime," so there is no "What was that?"
How Descriptive Experience Sampling works
Moment or stream?
The sharpest dispute came with the "COVID 2020 beep." Walking with his wife Judith, Pollan was saying "When I was walking yesterday," with the whole rehearsed memory of March 2020 (empty streets, wide berths, blooming magnolias) already present as a "package." Hurlburt called this context, not experience. After half an hour they agreed on a distinction between what is "directly apprehended" and what is "shaping, guiding, informing without being directly apprehended." Pollan sent him James: "Annihilate a mind at any instant... you will find, not the bald word in process of utterance, but that word suffused with the whole idea." Hurlburt conceded that if the distinction did not fit Pollan's experience, "we shouldn't pretend that it does."
| DES finding | Detail |
|---|---|
| We misjudge our own minds | "Probably the most important finding": people know little about the form of their inner experience |
| Inner speech is not universal | Fewer than a quarter of samples involve inner speech; only a minority are "inner speakers" |
| Other modes | A slightly lower share involve inner seeing, feeling or sensory awareness; about a fifth are unsymbolized thought |
| Huge individual variation | "Thinking" means seeing for a visualizer and talking for an inner speaker; the shared word hides the difference |
| Some have little inner experience | Pure perception and near-automatic decisions; possibly "an advantage" |
James had put it starkly: each mind keeps its thoughts to itself; "Absolute insulation, irreducible pluralism, is the law." Literature ferries us between these islands. Then came Hurlburt's verdict on Pollan: probably far toward the end of the spectrum with little inner experience, backfilling empty moments with context. Pollan was "flabbergasted." Inner seeing fades with age (James noted the same in himself). Pollan remains unconvinced: a sample "might tell us interesting things about the chemical composition of the water, but it tells us nothing about the stream's dynamics." His stream "might be shallower than some," but it flows.
The wandering mind
Almost no one in consciousness research mentions the unconscious. Kalina Christoff Hadjiilieva (University of British Columbia), a Bulgarian-born jazz pianist, is the exception: "Consciousness is just one function of the mind... The tree is the mind, and there's a lot more to the mind than consciousness." Her interest began in Sofia, where at fourteen the Berlin Wall fell and she lost her father (secretly a Bulgarian Secret Service officer) and three grandparents; she pushed hard memories away, "only [giving] them more power." She studies spontaneous thought: mind-wandering, daydreaming, creative insight, flow. She co-edited The Oxford Handbook of Spontaneous Thought, which rules rumination out (too constrained: a loop, not a meander). Her approach, neurophenomenology, pairs self-reports with fMRI. She estimates 30 to 50 percent of mental life is generated by the mind rather than the senses; as we age, we wander less and ruminate more.
Mind-wandering findings: frequent, costly to mood, and linked to memory systems.
Christoff Hadjiilieva's seesaw: constrained and unconstrained thought
The politics of the mind
Psychology, she argues, abandoned the unconscious to gain scientific legitimacy, a "Faustian bargain": "Legitimacy meant having to define the mind in a way that guarantees you'll never understand it." Freud and Jung took it into therapy instead. "The mind is not a neutral territory." She calls Killingsworth and Gilbert's famous paper "A Wandering Mind Is an Unhappy Mind" "basically propaganda... for how we can turn people into better soldiers of our capitalist order"; two Handbook reanalyses of its data found wandering minds happier than claimed. She notes that dopamine (self-interest) gets more study than serotonin (connection), that "ignition" and "workspace" are "very male metaphors," and, told of Solms's agent and Man's robot: "Why don't they just make a baby?"
An ordinary mind on an ordinary day
"There is something inherently poetic in consciousness that's evading scientists right now," she told Pollan, suggesting novelists, who "live their thoughts," know more. Virginia Woolf (1925) set the programme: "Examine for a moment an ordinary mind on an ordinary day... Let us record the atoms as they fall upon the mind in the order in which they fall." Novels have always entered minds (we may know Emma Bovary better than anyone), but the stream-of-consciousness technique renders the mind's phenomenology too: rhythms, leaps, fragments. In Joyce's Ulysses, Bloom at Paddy Dignam's funeral drifts from Latin prayers to bad gas burning blue in a church vault. Édouard Dujardin (1880s) reduced syntax to "a syntactical minimum"; Woolf kept her sentences.
| Writer or thinker | View of thought and words |
|---|---|
| Joyce | Thought as largely verbal (Dorrit Cohn: Bloom's mind "almost entirely in Bloom's words") |
| Bergson, Proust | "Pure" thought precedes language; interior monologue too "oblique" |
| Schiller | "When the soul speaks, alas, it is no longer the soul that speaks" |
| Nathalie Sarraute | "An immense profusion of sensations, images... little larval actions that no inner language can convey" |
| Tolstoy (per Nabokov) | First used the method for Anna Karenina's last carriage ride, as her mind disintegrates |
Rendering consciousness in words does not mean it is made of words, any more than portraitists think people are made of paint. James borrowed "stream of consciousness" from George Henry Lewes (1859), George Eliot's partner. Why no Victorian streams? Margaret Rennix's 2015 Harvard dissertation: Victorians associated the uncontrolled stream with madness; cognitive control was sanity, freedom and character. Woolf's generation inverted this, perhaps because of Freud, the new physics of the observer, the First World War, the weakening of empire, or (critic Michael Levenson) curiosity about the minds of anonymous city crowds. Conceptions of consciousness are "historical artifacts as much as biological phenomena," and artists often got there first.
Ducks, Newburyport
Lucy Ellmann's Ducks, Newburyport (2019) is a single sentence of more than 1,000 pages: the mind of an Ohio mother of four and home pie baker, punctuated by "the fact that..." (two cardinals in the lilac, taxes, macrophages, a dream of flying to India for cinnamon). Ellmann, an American in Scotland, did no research ("I have one right here. I'm going to study mine") and avoided Joyce and Woolf. "How you think is hard to think about"; "There's a block between consciousness and thinking about consciousness." Language cannot capture "the whole thing," and thought without words is possible: alone on an island "you'd still know how to build a house and catch a fish." Yet words help: learning the word depression as a child was a relief.
Why consciousness does not switch off
Pollan's puzzle deserves its own space. If consciousness exists for jobs that cannot be automated, as Damasio, Solms and global workspace theorists hold, then the beagle, the bread roll and the word suppress are strange occupants. Three answers emerge from the chapter. Hurlburt declines to answer: his job is to describe, and theory corrupts description. Christoff Hadjiilieva says much of this flotsam is not idle at all: unconstrained thought is how the mind sorts experience into meaning, so its apparent uselessness reflects a narrow, task-centred idea of usefulness. Ellmann offers the most existential answer: the monologue may simply reassure us that we are still here. None of these is a theory of consciousness, but together they show how much any theory must explain beyond the competition for attention.
| Approach | What it captures | What it misses |
|---|---|---|
| Theory (IIT, GWT) | Structure, access, measurable correlates | Thought, flow, affect, time |
| Introspection (James) | The fringe, transitions, feelings of tendency | Reliability; a sample of one |
| Experience sampling (Hurlburt) | Form of experience at a moment; individual variation | Dynamics of the stream; context |
| Neurophenomenology (Christoff Hadjiilieva) | Links reports to brain networks; the unconscious run-up | Fine texture of experience |
| Fiction (Woolf, Joyce, Ellmann) | Rhythm, association, simultaneity, repression | Verification; words shape what they render |
The lesson Pollan draws is pluralist. Each method sees part of the stream, and each distorts it in its own way: the theorist by abstraction, the introspector by observing, the sampler by freezing a moment, the scanner by averaging, the novelist by putting thought into sentences. A complete science of consciousness would need all five, and would have to admit that the thinker studying thought can never stand outside it.
Chandler's edge and the view from within
The chapter opens with Raymond Chandler (The Big Sleep): consciousness sits "on the edge." Pollan's method matches the epigraph. Theories built for perception leave thought as leftover. He turns to phenomenologists in a broad sense: James, Hurlburt, Christoff Hadjiilieva, and novelists who have charted the wilderness longer than labs have.
Who studies thought from within, for Pollan
Hurlburt's patent beeper and the DES discipline
Russell T. Hurlburt (UNLV), trained as an engineer, patented his own gray plastic beeper with a flesh-toned 1973 earpiece because commercial units did not exist when he began. For fifty years he has sampled random moments of inner life and refused premature theory: "I don't know anything about consciousness... And I'm not sure anyone else does either." Descriptive Experience Sampling (DES) trains subjects not to summarise or explain but to report the thin slice at the beep. Results upend assumptions: inner speech is a minority habit ("inner speakers"), not the common currency philosophers assume; many moments are wordless, imagistic, or empty of a clear "thought."
DES method: from beep to overturned assumptions
Novelists as phenomenologists
In college Pollan found Joyce and Woolf boring: fifty pages to walk a few Dublin or London blocks. Rereading as a student of the stream, he treats them as case studies. Woolf's 1925 Common Reader programme: "Examine for a moment an ordinary mind on an ordinary day... record the atoms as they fall upon the mind in the order in which they fall." Novels have always taken readers into other minds; we may know Emma Bovary's thinking better than our own. Stream-of-consciousness modernism is not a discovery of consciousness but a method for rendering its "incessant shower of innumerable atoms," including works Pollan revisits from Ulysses and Mrs Dalloway to To the Lighthouse and later heirs such as Lucy Ellmann.
What the beeper did to Pollan's mind
Wearing Hurlburt's device changed the stream it sampled. Waiting for the beep made Pollan watch himself watching; some beeps caught "Pretty scene" emptiness; others caught shopping maths and forgotten pepper. He thought he was failing; Hurlburt treated banality as data. The deeper finding is variation: people differ in whether they live in words, images, feelings, sensory washes or unsymbolized thought. Philosophers who treat inner speech as universal were sampling themselves.
Beeper effects: sampling changes the stream it measures
Ellmann's spiral and the fact that...
Ducks, Newburyport strings an Ohio mother's mind on "the fact that..." for a thousand pages. Ellmann told Pollan she studied her own mind, avoided Joyce and Woolf, and still hit a block: "How you think is hard to think about." Words both fail ("the whole thing") and help (naming depression as a child). Pollan's synthesis: Hurlburt describes, Christoff Hadjiilieva functionalises, Ellmann existentialises. Any theory of consciousness that ignores mental flotsam has explained the wrong thing.
Three answers to why mental flotsam fills the stream
James's bird: flights, perchings and the snowflake
James switches metaphors from stream to bird: consciousness alternates flights (transitive parts) and perchings (substantive thoughts). The flights are hardest to notice; catch a thought midflight and "it ceases forthwith to be itself," like holding a snowflake or turning on lights to see darkness. That observer effect, plus a sample size of one and working memory of three to five items, drove Pollan from armchair introspection to Hurlburt's beeper. Computational theories that flatten thought into information lose the halo, fringe and afterglow James insisted on.
James bird metaphor and why Pollan left the armchair
Why theories of perception miss the stream
Global workspace and IIT explain how content wins access or integrates, not why the winner is often a beagle, a bread roll or the word suppress. Damasio and Solms cover feelings but not the rest of the flotsam. Pollan's turn to DES and novels is a methodological correction: if your theory cannot host banal spontaneous thought, it is a theory of something else. Christoff Hadjiilieva's functional story (meaning-making) and Ellmann's existential one (proof of presence) are not neural models; they are constraints any neural model must meet.
Constraint: theories must host mental flotsam
Self
Looking for the "I"
Pollan's public identity (seventy-one, writer, husband, Californian) fits many people. His body is more specific, James's "feeling of the same old body always there," yet every cell has turned over, like the ship of Theseus, and we speak of "having" a body rather than being one. Every morning the mind reknits consciousness and, within seconds, the familiar "I" returns: author of thoughts, owner of the body, the one certainty Descartes allowed. Its existence is surprisingly hard to prove. "All that is solid melts into air" once the self is examined.
Kant named the knot: "I cannot cognize as an object itself that which I must presuppose in order to cognize an object at all." To look for the subject is to make it an object and lose it, so perhaps the self is hidden, like the sun in an eclipse, rather than absent. Pollan still clings to a continuous thread linking him to the gangly adolescent whose embarrassments still make him blush; maybe it is only edited memory, "a comforting fiction." Two puzzles frame the chapter: if the self is an illusion, why cling to it (presumably because it is adaptive)? And why do we prize self-esteem while seeking self-transcendence in meditation, psychedelics, art, awe and flow? Consciousness "has no particular need for a self," as infancy shows.
A self arises
Infants are now agreed to be conscious (as late as the 1980s surgeons operated on babies without anaesthesia), but the self comes later. Alison Gopnik, the Berkeley psychologist and author of The Philosophical Baby (2009), walked Pollan through the stages over tea.
Gopnik's stages: how a self is assembled in childhood
- Mirror test. Babies touch a sticker on their forehead from about 18 months. Great apes, dolphins, orcas, manta rays, Eurasian magpies and one Asian elephant also pass.
- Episodic memory. Toddlers remember only what adults remarked on (the monkey that flung its poop), suggesting they learn to lay down memories by imitation. Three-year-olds recognize themselves on video but don't link the sticker in the recording to their own forehead now.
- Volition. A four-year-old says she can't decide not to eat the permitted muffin ("You want it, so you have to eat it"); a six-year-old can, "because I'm the boss of me."
The self as a prediction
In Brighton Pollan met Anil Seth, author of Being You (2021). Predictive processing goes back to Hermann von Helmholtz: perception is the brain's "unconscious inferences," its best guesses about the causes of sensory signals, so perception runs inside out. Red exists only in the mind (Cézanne: "Color is the place where our brain meets the universe"). Friston formalized the "Bayesian brain." Seth calls perception a "controlled hallucination": predictions built on priors, corrected by the senses, because guessing is faster and cheaper than building the world from scratch. How could Pollan hallucinate a room he had never seen? Its geometry, light and street rhythms are universal; the senses fill the gaps.
Seth's predictive self: the brain models the body from inside out.
Memory and metamorphosis
Continuity comes from the body and from memory. Luis Buñuel: "Our memory is our coherence, our reason, our feeling, even our action. Without it, we are nothing." Oliver Sacks's patient Jimmy, with Korsakoff syndrome, stopped forming memories in 1945: "isolated in a single moment of being." Yet our memories are unreliable. What if that is a feature?
Memory as reconstruction: each recall rebuilds the self it serves.
Pollan's example: in sixth grade, he and Jonathan Kantor talked Mrs. Bergwall into letting them skip English to write a novella (an awful, Capote-indebted tale of a wrongly convicted Black janitor who dies of a heart attack on the courthouse steps). He has told the story as cleverness, precocity or pretension; now it is his first taste of the freelance life, "a lifetime hallway pass." Memory also makes consciousness unique: his summerhouse porch, pond and the oak he climbed as a boy colour every moment there. Qualia need not be generic (red, chocolate); they can be idiosyncratic.
| Thinker | Why the personal side of experience cannot be copied |
|---|---|
| Carlos Montemayor | Consciousness gives "familiarity with the world and your relationship to it," which machines will lack. When Frank Jackson's Mary leaves her black-and-white room she gains not information but familiarity: "Now she can go to a museum and look at a Rothko" |
| Proust | "Every impression comes in two parts, half of it contained within the object, and the other half, which we alone will understand, extending into us"; the "inner book of unknown signs" only art can decode |
| James | "The breach from one mind to another is perhaps the greatest breach in nature" |
Losing ourselves
Self-transcendence brings timelessness, boundlessness, universal love, Huxley's "Mind at Large," and experiences become less individual and more alike. Why is that usually blissful, if the self is a useful adaptation? Sartre saw the ego as a cover story for inner chaos; Lacan called it a fiction born at the mirror. Sartre's 1935 mescaline trip left him pursued for weeks by crabs he asked to keep quiet during lectures ("fear of becoming alone"). Safer routes include meditation, float tanks, flow, awe and art, where Schopenhauer says "we lose ourselves entirely in this object... the two have become one." Transcendence heals the breach between minds; the sovereign self "is a lonely place to be."
Pollan tried it under hypnosis with Stanford psychiatrist David Spiegel (he scored nine of ten for hypnotizability). Instead of empty rooms he found a different self behind every door: the 1968 bar mitzvah boy, a three-year-old bouncing in his sister's crib, a new father, a seminar leader, multiplying into a mansion. Not no-self but many selves, which undermined a single continuous self just as surely.
Minimal phenomenal experience
German philosopher Thomas Metzinger (Being No One, The Ego Tunnel), meditating twice daily since 11 September 1976, argues that the brain builds a self-model but that a simpler minimal phenomenal experience (MPE) precedes it. He scorns analytic philosophers who have never meditated or tripped as "completely impoverished" in imagination. When he asked Francis Crick in the late 1990s what exactly he wanted to explain, Crick, who had told him philosophers should "simply shut up," had no answer and exploded in anger. Metzinger thinks Nagel's "something it is like" wrongly assumes consciousness must be subjective.
From open awareness to self: Metzinger's sketch
MPE is the moment Proust describes on waking, "more destitute than the cave-dweller," until memory comes "like a rope let down from heaven." It is probably the phenomenology of babies and animals, so Pollan has come full circle back to sentience. Meditation and psychedelics break down functional connectivity in the default mode network, felt as loss of self. For Metzinger, our self-model alone contains the "toxic information" that we will fail against entropy, and evolution uses it "to drive us forward mercilessly." A "gloomy European," he warns that transcending the self carries real risk.
Five ways of looking at a self
| Lens | Guide | What the self turns out to be |
|---|---|---|
| Introspection | Hume, Kant | Never found as an object: only perceptions, or a subject that vanishes when looked at |
| Development | Gopnik | Assembled in stages; a useful fiction that trades lantern openness for spotlight focus |
| Prediction | Seth | A controlled hallucination built from bodily signals to keep us alive |
| Biology of memory | Levin | A process of sense-making that keeps rewriting its past |
| Contemplation | Ricard, Metzinger | A self-model that causes suffering and can be set down, revealing pure awareness |
The lenses agree on more than they differ. None finds a fixed inner essence; all describe something built, maintained and revisable. They split on value. Gopnik and Seth see the constructed self as an adaptation worth having; Ricard and Metzinger see it as the main source of suffering; Levin sees its revisability as the engine of growth. Pollan's own experiments (hypnosis, psilocybin, later the cave) land between them: the self is "illusory" in the way colour is, which does not stop it being real enough to live by.
Mind beyond brain?
Over lunch in Berkeley, Christof Koch, in a chartreuse plaid shirt, described a five-night ayahuasca ceremony in Bahia: "I accessed this universal mind... There was no self. There was Mind at Large." Then: "But I am not having a crisis!" He was, and later admitted it. The quintessential "brain guy" has moved "beyond garden-variety physicalism" for three reasons.
Two models of brain and mind: production versus transmission.
| Koch's reason | His argument |
|---|---|
| Physicalism's failure | Dennett's "illusion" makes no sense (an illusion is a conscious experience); "emergence" is an abracadabra |
| Quantum theory | Entanglement means "bye-bye to locality... and bye-bye to subject-independent reality" |
| Personal experience | Like a Mary who briefly sees colour in a world of monochromats: "I did encounter Mind at Large"; afterwards he cried and his wife consoled him |
He turned to idealist philosopher Bernardo Kastrup: "Matter is an inference, and mind a given." Koch: "Without a mind to observe it, the tree doesn't exist... it's just ontological dust... only consciousness exists for itself." Seth, Solms and Damasio remain physicalists, but their theories need "some hand-waving or brute force" to cross the explanatory gap. Pollan sees materialism hitting a wall at exactly two places, quantum physics and consciousness, and suggests that raising these doubts may be the field's most important legacy since the 1998 Koch-Chalmers wager. Chalmers's verdict: "No position on the mind-body problem is plausible... But the probabilities of all of these views get a boost from the fact that one of them must be true." Pollan admits he knows less than when he began. Koch smiled: "But that's good! That's progress!"
The menu of non-physicalist options is long: idealism, panpsychism, dualism, quantum theories, simulation theories and "transmission" theories in which brains tune into consciousness as radios tune into broadcasts. All strain plausibility, Pollan notes, "as do physicalist theories," and most cannot be tested. Chalmers's half-comic conclusion is that every view is implausible but one must be true. Pollan asks whether an idea nobody has yet conceived might be the right one, and how, since it too would be a product of consciousness, we would ever know. We are left "wandering in the exitless labyrinth of consciousness."
Buñuel's coherence and the self as memory work
Luis Buñuel, late in life: memory "makes our lives"; "Life without memory is no life at all"; "Our memory is our coherence, our reason, our feeling, even our action. Without it, we are nothing." Pollan pairs this with Oliver Sacks's Jimmy (Korsakoff; memories stop in 1945) and with Levin's butterfly that keeps the caterpillar's colour-food association after metamorphosis rebuilds the brain. Continuity is salience, not fidelity: "mnemonic improvisation."
Self as memory work: Buñuel, Sacks, Levin
Metzinger's minimal experience and Kuhn's taxonomy
In 2021 Thomas Metzinger launched the Minimal Phenomenal Experience project at Mainz: strip consciousness to pure openness before any self, content or narrative. That minimal state, reached in deep meditation, is Pollan's bridge back to Chapter 1's elemental sentience. Meanwhile Robert Lawrence Kuhn's 2024 taxonomy ("A Landscape of Consciousness") catalogues on the order of eighty-plus explanations; Pollan's "about 22" is a working subset of a much larger menu. The chapter refuses to pick a winner among predictive selves (Anil Seth's controlled hallucinations), developmental selves (Alison Gopnik's lantern-to-spotlight), Buddhist no-self, and psychedelic multiplicity.
Paths to dissolving or rebuilding the self in Chapter 4
Gopnik's lantern and Seth's beast machine
Alison Gopnik contrasts infant lantern consciousness (panoramic, exploratory) with adult spotlight attention. The mirror test at about 18 months ("perm shock") marks a self that can be embarrassed. Anil Seth's Being You frames the self as a controlled hallucination serving a "beast machine": predictions keep a body alive. Hot-cold rubber-hand style demos show how quickly bodily ownership bends. Pollan's hypnotic mansion with David Spiegel (9/10 hypnotizable) multiplied selves behind doors rather than emptying them: not no-self but many selves.
Development and prediction as two self factories
Psychedelics, meditation and the use of a thin self
The chapter's practical question is when to strengthen the self and when to thin it. Love, planning and responsibility need continuity; awe, flow, meditation and psychedelics open meaning by loosening the grip. Metzinger's minimal phenomenal experience is the laboratory version of what Halifax later gives Pollan as practice. Kuhn's sprawling taxonomy reminds readers that Pollan's chosen theories are a curated path, not the whole map.
When to hold the self and when to thin it
Kastrup's idealist brief, via Koch
After ayahuasca, Koch sought explanations beyond physicalism and found Bernardo Kastrup. Kastrup's brief: everything we know about reality we know through experience (personal or collective/scientific); physicalism cannot explain the only thing we know to exist, which is experience; therefore mind is given and matter is inferred. Koch: without a mind to observe it, the tree is "ontological dust." Pollan notes that Seth, Solms and Damasio stay physicalist only by "hand-waving" across the gap. He will not convert, but he treats the strain on materialism at consciousness (and quantum physics) as the field's honest legacy since the 1998 wager.
Kastrup brief as Koch relays it to Pollan
Kant's knot and the eclipse of the subject
Kant: you cannot cognize as an object that which you must presuppose to cognize any object. Looking for the self turns subject into object and loses it; perhaps it is hidden like the sun in eclipse, not absent. Pollan's blush at adolescent memories still feels like a continuous thread; maybe only edited memory, "a comforting fiction." The chapter holds both: adaptive clinging and prized self-transcendence. Infancy shows consciousness "has no particular need for a self."
Kant's knot: why the self slips when chased
Conclusion: The Cave
Roshi Joan's refuge
Joan Halifax, abbot of Upaya Zen Center in Santa Fe since founding it in 1990, has spent her life in engaged practice: hospice care, death-row prisoners, the homeless, peace protests, and some forty years leading the Nomads Clinic through remote Nepali villages at altitudes up to 20,000 feet ("I just lived as though my hair were on fire"). With her then-husband Stanislav Grof she gave LSD to the dying in the 1970s; her first trip, in Paris in 1968, showed her "that there was beauty behind the beauty I perceived, and that mind was both in here and out there." She calls Upaya "a factory for the deconstruction of selves." Rather than a retreat at Upaya, she sent Pollan, whom she judged "hopelessly stuck in my head," to "the refuge" at 9,400 feet in the Sangre de Cristo Mountains: "You can stay in the cave." It was a 12-by-15-foot cell dug into a hillside, with a woodstove, a meditation platform, a water jug, a solar panel and no plumbing, phone or internet.
Cave life
Life stripped down to wood, water, fire, digging pits and sweeping, plus hours of meditation, deeper than he had ever managed. His first attempt at the pit toilet left him "a shoeless monk." Trying Ricard's search again, he found no thief of self, only a parade of hypnagogic images: three old telephone booths, a hammer on a knee, a door opening for no one, a blazing sun turning into a giant eyeball. Unlike Sartre's crabs, he could dismiss them at will. Then who did the dismissing? Something gave the house a point of view and agency. "The self might well be illusory... but no more so than color": illusory and "real enough."
Pollan's cave retreat as a sequence from distraction to open awareness.
What dissolved the self in the cave
The self, he realized, depends on time: deprived of memory and anticipation, it melts. Watching elk at sunset, he lived for longer stretches simply "here, being." Strangely, it happened with eyes open, not closed. Each time he noticed and reached for his notebook, he fell back "into the self-world," like a tightrope walker who looks down.
The circle closes
The cave completes the book's ladder by running it in reverse. Chapter 1 asked whether plants have the barest sentience; Chapter 4 found that Metzinger's minimal phenomenal experience, the openness before any self, may be as close as humans come to that simple animal awareness. In the cave Pollan reached it by subtraction: first the chatter of the self, then thought, then the expectations that shape perception, until only the sense field remained. Seth's predictive brain explains the night-sky moment neatly (priors relaxed, more raw signal getting through), and Gopnik's lantern consciousness describes it: panoramic, numinous, childlike. The science and the experience do not compete here. One describes the mechanism, the other the fact, and Pollan's point is that the fact had gone missing from his inquiry.
How Pollan found Halifax: Solnit's Nomads Clinic
Pollan had known of Joan Halifax since Rebecca Solnit's 2015 New Yorker profile, which followed the annual Nomads Clinic trek: doctors and dentists into remote Nepali villages for two weeks each summer. That portrait of engaged Zen (hospice, death row, peace work, high-altitude clinics) led him to Upaya, and Halifax, judging him "hopelessly stuck in [his] head," sent him to the cave rather than a standard retreat. The refuge sits at 9,400 feet in the Sangre de Cristo Mountains: a 12-by-15-foot cell, woodstove, meditation platform, water jug, solar panel, no plumbing, phone or internet.
Path to the cave: from profile to practice
What "a world appears" resolves and what it does not
The final sentence is not a theory. It is a report from practice: open the eyes, and appearance happens without a solving self. The hard problem is neither dissolved nor solved; it is set down. Spotlight attention that hunted "The Answer" cost Pollan "the other 359 degrees of illumination." The lantern returns when self and its distractions loosen. Science remains indispensable for mechanisms; the cave supplies the fact those mechanisms never replace.
Coda resolution: practice completes the inquiry science cannot
Factory for deconstruction: what Halifax actually does
Upaya is "a factory for the deconstruction of selves." Silence removes self-presentation; ritual replaces "a certain amount of volition"; people ruminate until they cannot stand themselves, then "drop in" to the sense field. Around day three "you can feel the whole room go poof." Halifax treats retreat as rite of passage (ego death, then rejoin the group) and psychedelics as a shortcut she would rather students avoid. Even she "still can be self-righteous." Pollan's cave was the solo version: same factory, fewer people, thinner air.
Halifax factory steps from rumination to drop-in
Spotlight cost: 359 degrees lost
Halifax: "Always keep a don't-know mind." Pollan's quest used spotlight attention to hunt The Answer and paid with "the other 359 degrees of illumination." The lantern remains available when self and distraction break. The prize is practice, being "altogether here," not a theory. Mystery stays; simplicity fits a sentence.
From spotlight quest to lantern practice
Main Points
- The wager was lost, and the field is still flailing. Twenty-five years after Koch's 1998 bet, no agreed neural correlate exists; one count finds 22 competing theories, and Kuhn's catalogue lists far more.
- The bifurcation of nature is the root problem. Galileo and Descartes kept measurable qualities and exiled colour, taste and feeling to the mind, so physics has no place for experience.
- The hard problem is why processing feels like anything. Chalmers's zombie, a functional twin with no inner life, shows that explaining function is not explaining experience.
- Leading theories disagree on basics. IIT puts consciousness in integrated causal structure at the back of the brain; global workspace puts it in broadcast from the front. The 2025 adversarial test unsettled both.
- Sentience may be older than brains. Plants learn, signal, sleep and respond to anaesthetics; plant neurobiologists call this intelligence, critics call it metaphor.
- Agency goes down to cells. Levin's bioelectric networks and Xenobots solve problems without neurons, so minds may be a matter of degree.
- Friston offers a physics of staying alive. Systems that persist must predict and minimise surprise; for Solms, consciousness is the felt version of that uncertainty.
- Only life may know life. Thompson argues that agency, value and sense-making come with living, not with computation.
- Feeling may be the first form of consciousness. Damasio and Solms trace it to the upper brainstem and homeostasis, not to the cortex.
- Feelings are hard to deny as conscious. "Who ever heard of an unconscious feeling?" Chalmers replies that an affectless "Vulcan" would still be conscious.
- Conscious AI rests on an assumption. The Butlin report's "no obvious barriers" follows from computational functionalism, which it adopts for pragmatic reasons.
- Brains are not computers. No hardware/software split, chemical modulation, oscillations, and single neurons as complex as whole networks: "the price of metaphor is eternal vigilance."
- Feeling may need vulnerable flesh. Simulated feelings are not feelings (Turkle); a weather model does not get anyone wet; meaning may depend on life and death.
- Theories ignore the mind's contents. A science of perception is not a science of consciousness; thought, flow and mental flotsam go unexplained.
- Introspection is limited but necessary. James's stream, fringe and "flights and perchings" capture what information cannot, yet observing a thought changes it.
- Minds differ more than we assume. Hurlburt finds inner speech in under a quarter of samples and wide individual variation hidden by the word "thinking."
- Spontaneous thought is real work. Christoff Hadjiilieva's 30-50 percent of mental life, starting in the hippocampus about four seconds before awareness, helps build meaning and identity.
- The self is assembled and rewritten. Mirror at 18 months, episodic memory at four or five, "boss of me" at six; Seth calls it a prediction; Levin calls memory improvisation.
- Consciousness can exist without a self. Pure awareness in meditation, psychedelics and the first moments after waking suggests subjectivity is not essential, challenging Nagel's definition.
- Not-knowing is part of the answer. Materialism strains, idealism tempts Koch, Chalmers finds every view implausible, and Pollan ends with practice: being "altogether here."
Key numbers to remember
| Quantity | Value | Chapter |
|---|---|---|
| Koch-Chalmers wager | Made 1998, 25 years; Koch conceded in 2023 | Intro |
| Competing theories of consciousness | About 22 (Pollan); Kuhn catalogues many more | Intro, 4 |
| Signatories, New York Declaration | 613 by 2026 (about 40 at launch, 2024) | 1 |
| Butlin report authors | 19 (August 2023), 88 pages | 2 |
| Solms agent's world | 10 x 10 grid; 3 needs (hunger, thirst, rest) | 2 |
| Lemoine's certainty LaMDA is sentient | 60-70% | 2 |
| Connections per cortical neuron | Up to 10,000 | 2 |
| Hurlburt's sampling career | About 50 years (beeper built 1973) | 3 |
| Samples with inner speech | Under a quarter; unsymbolized thought about a fifth | 3 |
| Mental life not fed by the senses | 30-50% | 3 |
| Hippocampal lead before a thought | Nearly 4 seconds | 3 |
| Things held in mind at once | About 3-5 | 3 |
| Ducks, Newburyport | One sentence, 1,000+ pages, 8 years | 3 |
| Mirror-test age | About 18 months | 4 |
| Metzinger's pure-awareness reports | 1,000+ collected; about 500 in a 615-page book | 4 |
| Pure awareness on waking | First 200-500 ms | 4 |
| Non-physicalist theories in one review | 84 | 4 |
| The cave | 12 x 15 ft at 9,400 ft | Concl |
Five questions to argue about
| Name | What it is | Use |
|---|---|---|
| Hard problem | Why physical processing is accompanied by experience | The book's central puzzle |
| Neural correlates of consciousness (NCC) | Minimal brain activity sufficient for an experience | The object of Koch's wager |
| Integrated information theory (IIT) | Consciousness as integrated causal structure (phi) | Koch's preferred theory |
| Global workspace theory (GWT) | Consciousness as broadcast across the brain | Rival of IIT; favoured by Christoff Hadjiilieva |
| Free-energy principle | Living systems minimise surprise | Friston's physics; Solms's agent |
| Predictive processing | Perception as the brain's best guess | Seth's "controlled hallucination" |
| Computational functionalism | The right computations suffice for consciousness | Assumption behind conscious-AI claims |
| Phenomenology | Study of experience from within | James, Thompson, novelists |
| Descriptive Experience Sampling | Random beeps plus interviews | Hurlburt's method |
| Default mode network | Midline brain network active in rest and self-reflection | Mind-wandering; ego dissolution |
| Mnemonic improvisation | Rewriting memories for new circumstances | Levin's account of the self |
| Minimal phenomenal experience | Pure awareness without self or content | Metzinger's project |
| Panpsychism / idealism | Mind everywhere / mind as fundamental | Alternatives to materialism |
Depth checklist: what this guide now carries
Every source chapter has a matching section. Frameworks kept whole: Chalmers easy/hard split; IIT axioms; GWT/GNWT broadcast; four book dimensions; New York and Cambridge declarations; Damasio bottom-up stack; Solms "felt uncertainty"; Butlin indicators; DES findings; DMN vs task-positive; Seth predictive self; Metzinger MPE; Halifax don't-know practice. Oct 2026 boxes cover Cogitate/Nature (30 Apr 2025), NY Declaration signatories (613), Anthropic model welfare (Apr 2025), Butlin/Long TICS (Nov 2025), Solms/Conscium JCS 2026 hedonic place preference, anendophasia (2024), Halifax mitral-valve surgeries (June 2025) and return to teaching.
Edition coverage map: source units to Insight sections
Counterarguments the edition keeps visible
Sceptics on plants (Taiz; failed pea replication); Chalmers on affect (explanatory gap remains; Vulcans could be conscious); functionalists on AI (Pollan's line may only describe current machines); Hurlburt on theory (description first); Seth/Solms debate in BBS 2026; illusionists who deny the hard problem's premise.
Main counterarguments retained beside Pollan's line
About this book
A World Appears: A Journey into Consciousness by Michael Pollan, 2026. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions