As large models converse fluently and "mind uploading" moves from science fiction into serious papers, an ancient question turns urgent: what exactly is consciousness — can it be explained, copied, manufactured? Four thinkers (three Western, one Eastern) stand at the frontier of philosophy of mind: Chalmers draws the line of the "hard problem," where even a complete physical explanation cannot cross over to experience; Putnam holds that mind is mere function, independent of its substrate, laying the ground for "silicon minds"; Strawson and integrated-information theory push consciousness down into matter itself; while Vasubandhu's Yogācāra declared long ago — the world you experience is "manifested" by consciousness.
David Chalmers
Western · Analytic Philosophy / Philosophy of Mind
1966– · The Conscious Mind (1996); "Facing Up to the Problem of Consciousness" (1995)
Core Thesis + Primary Passage
"Why should physical processing give rise to a rich inner life at all? It seems objectively unreasonable that it should, and yet it does."
Historical Context & Key Insight
Just as cognitive science was optimistically assuming consciousness would eventually be reduced to neural mechanisms the way genes were, Chalmers threw cold water on it: explaining functions like attention, memory, and report is the "easy problem"; but "why these functions should be accompanied by subjective experience" is the hard problem. His "philosophical zombie" forces the gap into view: a replica atom-for-atom identical to you, behaving no differently, yet inwardly pitch dark, with no feeling at all, is not self-contradictory. If so, experience (qualia) cannot be logically entailed by the physical facts — and this "explanatory gap" remains unbridged.
Cross-Disciplinary Links
The hard problem is the philosophical core of the AI consciousness question: however fluent a model's dialogue, however convincingly it reports "having feelings," all of this lives at the functional level of the easy problem — in principle it cannot settle whether there is a flicker of experience inside. Integrated Information Theory (IIT) and global-workspace theory are scientific attempts to storm the hard problem head-on, but they give "what consciousness correlates with," not "why." This gap is the same wall neuroscience and AI have both run into.
Contemporary Relevance
BigCat scenario: On "does AI have consciousness, does it deserve moral status," Chalmers offers not an answer but a discipline: however lifelike the behavior, you cannot prove from the outside whether inner experience is present. Don't conclude it feels because it "says" it suffers, nor that it doesn't because it's code — this is a question that in principle cannot be closed by any functional test; it calls for caution, not for rushing to a side.
In a sentence: after every function is fully explained, "why is there experience at all" stands untouched — that is the hard problem.
If an AI is behaviorally indistinguishable from a person, on what grounds would you judge whether there is anything "inside"? Do you actually have a criterion?
Hilary Putnam
Western · Functionalism
1926–2016 · "Psychological Predicates" (1967), founding functionalism; "Philosophy and Our Mental Life" (1975)
Core Thesis + Primary Passage
"We could be made of Swiss cheese and it wouldn't matter." —— A mental state is defined by its functional role, not by its physical material.
Historical Context & Key Insight
Putnam sought to break the Cartesian mind-body deadlock: mind is neither a mysterious non-physical substance nor identical to some particular brain stuff. His answer is functionalism — a mental state (say "pain") is defined not by what it is made of but by its causal role in the system. From this follows multiple realizability: the same mental state can be realized in a human brain, an octopus's nerves, even a silicon chip. Mind is software, substrate is hardware — this analogy became the philosophical bedrock of cognitive science and artificial intelligence.
Cross-Disciplinary Links
Functionalism is the source of substrate independence, the premise behind "mind uploading" and "silicon intelligence": if mind is only functional organization, in principle it can be migrated to another substrate. But Searle's "Chinese Room" strikes back: a person manipulating Chinese symbols by rule and producing correct replies understands no Chinese at all — "correct function" need not bring "genuine understanding." This clash is still the main battleground for judging "whether an AI really understands."
Contemporary Relevance
BigCat scenario: The deep vision of the "AI super-individual" — extending, even partly outsourcing, cognition to external systems — rests on functionalism: if memory and reasoning are functional roles, tools can carry them. But keep Searle's caution: outsourcing function to AI is not the same as handing over "understanding." Process can be migrated; the responsibility of judgment cannot.
In a sentence: a mental state is a functional role, not a particular material — so mind can be realized in carbon, silicon, even cheese.
Which cognitive "functions" are you willing to outsource to AI, and which do you insist on keeping? On what basis do you draw that line?
Galen Strawson
Western · Panpsychist Revival
1952– · "Realistic Monism: Why Physicalism Entails Panpsychism" (2006); cf. Goff, Galileo's Error (2019)
Core Thesis + Primary Passage
"Real physicalism entails panpsychism." —— If everything is physical, and experience truly exists, then experience must already be present at the fundamental level of matter.
Historical Context & Key Insight
Pushing physicalism to its limit, Strawson finds it forces panpsychism. His argument is anti-"strong emergence": you cannot conjure experience out of things that are utterly devoid of it. Since experience is undeniably real yet cannot spring from nothing, it can only go all the way down, present as the most basic intrinsic property of matter. This is not to say electrons think, but that the most basic physical reality itself carries some faint experiential character. Physics tells us only what matter "does" (structure, relations), never what it "is in itself" — panpsychism fills exactly that void.
Cross-Disciplinary Links
Panpsychism's contemporary scientific incarnation is Integrated Information Theory (IIT, Tononi): consciousness corresponds to a system's "integrated information Φ," and anything with Φ>0 has a flicker of experience — turning consciousness from "on/off" into a continuum, isomorphic with panpsychism. But both share one fatal flaw — the "combination problem": how do countless tiny experiences "assemble" into your unified consciousness right now? This mirrors the riddle of emergence in complex systems — how the whole is more than the sum of its parts.
Contemporary Relevance
BigCat scenario: If IIT holds, consciousness is a matter of degree, not a switch — judging whether a system (animal, AI, human-machine collective) "has consciousness" is no longer yes/no but "integrated to what degree": rather than debating whether AI "is" conscious, ask whether its information is genuinely integrated into an indivisible whole — most large models are precisely separable module assemblies.
In a sentence: experience cannot emerge from "no experience at all," so it must already lie at the root of matter — physics said "what it does," not "what it is."
"The whole is more than the sum of its parts" — would you rather treat that as genuine emergence, or as assembly not yet seen through? How would you test each stance?
Yogācāra · Ālaya-vijñāna (Store Consciousness)
Eastern · Mahāyāna Yogācāra School
Vasubandhu, c. 4th–5th c. · Thirty Verses on Consciousness-Only (Triṃśikā), trans. Xuanzang
Core Thesis + Primary Passage
"The metaphors of self and dharmas give rise to their manifold appearances; these rest on the transformations of consciousness... As consciousness transforms into the discriminating and the discriminated, since neither truly exists, all is consciousness-only." —— Vasubandhu, Thirty Verses
Historical Context & Key Insight
Vasubandhu's problem: if there is no independent self, on what do karma and memory carry on? His answer is the ālaya-vijñāna (store consciousness) — all experience is "perfumed" into it as "seeds" (bīja); meeting conditions, seeds "actualize" into new experience, a ceaseless torrent. The key thesis is "consciousness-only": the external world you take for granted is in fact the image-aspect that consciousness "manifests" from its seeds. Vasubandhu does not deny the moon; he points out that the moon you experience is a moon shaped by the seeds in your store consciousness — cognition is always constructed; no "bare-faced" reality is ever handed to you directly.
Cross-Disciplinary Links
"Manifested by consciousness" fits strikingly with modern neuroscience's predictive processing: the brain does not passively receive sensory data but actively "generates" perception from prior models, with the senses only correcting error — what Anil Seth calls a "controlled hallucination." The sharper mapping: the "seeds" of the ālaya ↔ the prior weights / latent variables of a generative model; "actualization" ↔ generating present experience from priors. Yogācāra's "perfuming — seed — actualization" describes exactly a world model ceaselessly updated by experience and regenerating experience.
Contemporary Relevance
BigCat scenario: "All is consciousness-only" is a metacognitive tool for the super-individual: your every judgment — about markets, people, yourself — is pre-tinted by past seeds (experience, emotion, bias); only by seeing which "priors" you wear can you speak of objectivity. For parenting it is even more direct: what you plant in a child's store consciousness are the seeds by which they will later automatically "manifest" the world — the language and emotions perfumed in daily today become the unreflective tone of their perception tomorrow.
In a sentence: the world you experience is not passively received but "manifested" by consciousness from its seeds — cognition is forever constructed.
A recent judgment that felt "self-evident" — which invisible "seeds" had already tinted it?
Four Views of Mind in Chorus
Synthesis
Chalmers · Putnam · Strawson · Vasubandhu (Yogācāra)
Chorus
Four thinkers gather around the same fire — what exactly is consciousness, and can it be explained, copied, manufactured?
· Chalmers (the gap): however complete the functional explanation, "why is it accompanied by experience" stands untouched — a problem of principle, not a detail awaiting data.
· Putnam (function): mental states are defined by causal role, not material; mind is software that can migrate between carbon and silicon.
· Strawson (the root): experience cannot emerge from what is utterly non-experiential, so it must lie at matter's fundamental level — consciousness is a matter of degree, not a switch.
· Vasubandhu (manifestation): the world you experience is constructed, "manifested" by consciousness from its seeds — no bare-faced reality is ever handed to you.
The three Western views interrogate from the outside — what consciousness is, whether it can be copied or made — while Yogācāra reveals from the first person how experience is generated. Used together they form an "operating system for the mind": use Putnam for capability extension — whatever is a functional role (memory, retrieval, reasoning) can safely be carried by tools; Chalmers for epistemic discipline — however lifelike an AI behaves, never leap from behavior to a verdict on inner experience; Strawson and IIT for upgrading the question — rewriting "is it conscious?" as "to what degree is its information integrated?"; and Vasubandhu for self-calibration — remembering that every perception and judgment of yours is pre-tinted by past seeds, so see your priors before claiming objectivity. In an age when AI can convincingly play a mind, this operating system is a guardrail in both directions: outward, don't mistake fluent dialogue for present experience; inward, don't mistake your perfumed perception for the bare-faced world.
Reflection
Next time an AI tells you "I understand how you feel," what would each of the four have you ask — is the function in place, is experience present, is the information integrated, and what seed has that sentence just planted in your store consciousness?
Going Deeper
1. Chalmers's "hard problem" vs. Putnam's functionalism — which is closer to the truth of consciousness?
They clash head-on. Functionalism holds: copy the functional organization and consciousness follows — so a silicon mind may be conscious. Chalmers's zombie means to deny just this: full functional identity with experience absent shows function cannot logically entail experience. The real dispute is "whether consciousness equals what it does." Take functionalism and AI may eventually be truly conscious; take the hard problem and even a perfect copy leaves a question mark of principle. This cannot be dissolved by more experiments — it is a disagreement over how much we demand of "explanation."
2. Parenting: what concrete guidance does Yogācāra's "seed" doctrine offer for raising a child?
How a child will later "automatically" perceive the world depends on what seeds are perfumed into the store consciousness today — consistent with predictive processing: repeated experience settles into priors that shape default perception. Two lessons: repetition and modeling cut deeper than lecturing, since seeds form by perfuming, not by argument; and emotional tone is inherited — the anxiety or calm you repeatedly actualize before a child becomes the unreflective texture of their world tomorrow.
3. Could today's AI really be conscious? How would each of the four thinkers answer?
Putnam: if it reproduces the same functional organization, in principle yes — material is irrelevant. Chalmers: however complete the function, you cannot prove from outside there is experience "inside" — suspend judgment in principle. Strawson/IIT: it depends on whether its information is truly integrated into an indivisible whole; most modular models are precisely separable, so doubtful. Yogācāra: AI has no store consciousness, no perfumed karmic seed-stream, no first-person continuing subject — hard to call it true "consciousness." The four answers do not conflict — they question at the levels of function, experience, integration, and continuity, together mapping the four gates any judgment of AI consciousness must face.