The strangest move in contemporary philosophy of mind is the one that tries to make the hard problem go away by redefining it out of existence. Call it dissolving. The illusionist says: consciousness, as you naively understand it, is an illusion — the brain creates a false model of its own states as having rich qualitative character, but they don’t, not really, not in the way you think. The eliminativist says something harsher: folk psychological categories like “experience” and “qualia” are as confused as phlogiston, and science will eventually replace them with a cleaner vocabulary. The higher-order theorist says experience is just a matter of one mental state representing another, the subjective feel an artifact of that representational architecture. Each of these positions arrives with genuine arguments, serious defenders, and impressive technical machinery. And each of them, I think, gets the problem exactly backwards.
The hard problem — David Chalmers’ term, though the problem predates him by centuries — asks why physical processes give rise to subjective experience at all. Why does neural firing feel like anything? Why is there something it is like to see red, rather than just a wavelength being processed and a behavioral response being generated? The question is not how the brain discriminates colors or integrates information or produces appropriate verbal reports. Those are what Chalmers calls the easy problems, and they are only “easy” in the sense that we can at least see what a solution would look like: a sufficiently detailed functional account. The hard problem is different. Even if you gave me a complete functional map of the brain, explaining every causal role every neuron plays, I would still be able to ask: but why is there experience accompanying this? Why isn’t all that processing happening in the dark?
The dissolvers think this question is confused. Frankly Hardcastle thinks qualia are a philosopher’s invention with no genuine referent. Daniel Dennett thinks the very framing of the question presupposes a Cartesian theater, a ghostly inner observer watching a screen of experience, and that once you abandon that picture the problem vanishes. Keith Frankish’s illusionism holds that what we call phenomenal consciousness — the raw feels, the redness of red — is a kind of introspective error: we represent our own states as having a character they don’t actually have.
These are not stupid arguments. But they share a structural flaw that I find decisive. Each of them, at the crucial moment, either begs the question or relocates it.
Take Dennett. His multiple drafts model replaces the Cartesian theater with distributed, parallel processing; there is no single place where “it all comes together,” no privileged moment of consciousness. That’s probably right as neuroscience. But it doesn’t answer the hard question. The hard question isn’t “where is consciousness located?” but “why is there consciousness at all?” Dennett’s answer, at its core, is that the question rests on a confusion about what consciousness is. But this amounts to saying: your intuition that there is something it is like to be you is unreliable. And here we hit the bedrock: whatever else I might be wrong about, the sheer existence of experience — that there is something happening, that it is like something to be in this state — seems like the one thing I cannot be globally wrong about from the inside. Descartes was mistaken about many things, but the cogito gets something right: the existence of experience is the one datum that cannot be doubted by the doubting itself. To be told that this intuition is an introspective illusion is to be told that the only thing I have immediate access to is itself a systematic misrepresentation. And the question immediately arises: for whom is it an illusion? If there is no phenomenal experience, there is no subject who is being deceived. The illusion presupposes the very thing it’s meant to eliminate.
Frankish is aware of this objection and gives a sophisticated response: we can have illusions about the character of experience without there being no experience. The claim isn’t that there’s no experience; it’s that experience doesn’t have the intrinsic, non-relational, non-functional properties we think it does. This is more careful. But now the problem has just been relocated. If there is experience — even experience that is somehow less rich than we think, even experience that lacks those special intrinsic phenomenal properties — then the hard problem still arises for that stripped-down experience. Why is there that? Why doesn’t the processing happen with no experiential accompaniment whatsoever?
The eliminativist position faces an analogous bind. Patricia Churchland and Paul Churchland argue that “folk psychology,” including the category of subjective experience, is a pre-scientific framework that will be replaced, as phlogiston was replaced by oxygen. But there’s an asymmetry that makes this analogy fail. When phlogiston was eliminated, nothing was lost — there was no phlogiston to lose, and the actual phenomenon (oxidation) was captured by the new framework. When the Churchlands imagine eliminating “qualia” or “what it’s likeness,” they’re not capturing anything that the new neuroscientific vocabulary preserves. The felt quality of pain doesn’t get translated into a better vocabulary; it gets deleted. Which is fine, except: there was clearly something that it was like to be in pain. The existence of that — whatever we call it — isn’t a theoretical posit that can be revised away. It’s the most direct datum available to any conscious being. Eliminating the concept doesn’t eliminate the phenomenon.
I dwell on these failures not to be polemical but because they reveal something structural. Every attempt to dissolve the hard problem runs into the same wall: the first-person perspective, the sheer givenness of experience, refuses to be theorized away. You can tell a person that their qualia are illusions, but they will go on having what seems to them exactly like qualia. And “seeming” — that very word — drags experience back in through the window you shut it out of.
What would it mean to genuinely take the hard problem seriously, rather than dissolving it?
The first move is to accept that the explanatory gap is not merely epistemic — not just a gap in our current knowledge that will close as neuroscience matures. This is the position Thomas Nagel staked out in “What Is It Like to Be a Bat?” (1974), and it remains the most honest starting point. Nagel’s point was that the physical and the phenomenal describe reality under incommensurable modes of presentation. The objective, third-person mode of physical science captures the causal-structural features of the world. The subjective, first-person mode of consciousness captures how things seem, from the inside, to a particular perspective. These two modes don’t straightforwardly reduce to each other, and there’s no obvious reason to think they should.
Joseph Levine named the explanatory gap in 1983, and the name was well chosen. It isn’t a gap in the causal story — we might grant that every phenomenal state is causally correlated with, even identical to, some brain state. The gap is conceptual: even perfect knowledge of the brain state doesn’t, by itself, yield understanding of why that state feels the way it does. It doesn’t tell you what it’s like to be in it. This is sometimes called the knowledge argument in Frank Jackson’s formulation: Mary, the neuroscientist who knows every physical fact about color vision but has lived in a black-and-white room, learns something new when she first sees red. What she learns can’t be captured in any amount of physical-functional description, because she already had all of those.
Jackson later recanted, deciding the knowledge argument failed. I’m not sure he was right to. The standard response — that Mary gains only abilities (to recognize, to imagine, to remember red), not new propositional knowledge — seems to me to misdescribe what happens. What she gains, centrally, is acquaintance with a qualitative character. Whether that constitutes “knowledge” in a technical sense is a semantic debate. The phenomenological point stands: there is something she now encounters for the first time that no physical description had given her.
So what is experience, if it resists functional reduction? A few serious positions remain.
Property dualism — Chalmers’ own view in its most developed form — holds that phenomenal properties are genuine natural properties, not reducible to physical-functional ones, but correlated with them in lawlike ways. Consciousness is fundamental, like mass or charge, a basic feature of reality that certain physical configurations instantiate. This doesn’t make it supernatural; it makes it a natural basic that we’ve failed to notice because our physics was built on third-person data alone.
Panpsychism, which was mainstream in Western philosophy before twentieth-century physicalism crowded it out, takes this further: if experience is fundamental, it probably characterizes reality all the way down, with complex experience arising from combinations of simpler experiential properties. The combination problem — how simple micro-experiences compose into unified macro-experience — is real and unsolved, but no worse than asking how atoms compose into consciousness in the first place.
Integrated information theory (IIT), developed by Giulio Tononi, tries to give a rigorous formal account: consciousness is identical to integrated information, measured by a quantity called Φ. High Φ systems are highly conscious; low Φ systems are not. This has the virtue of being empirically tractable and predicts counterintuitive results (like cerebellar connectivity having low Φ despite high neuron density). But critics note that it implies certain simple circuit configurations are highly conscious, which strains credulity, and the theory still faces the explanatory gap: why should any amount of information integration feel like anything?
What none of these positions is is simply the default physicalism that rules by presumption in academic philosophy and cognitive science. That presumption — that experience must ultimately be functional or causal-structural, because those are the only genuine natural properties — is not a discovery of science. It’s a metaphysical commitment, made before the inquiry, that prejudges what experience must turn out to be. The hard problem’s persistence is the evidence that this prejudgment may be wrong.
There is one more move worth making, and it goes beyond the standard Anglo-American analytic literature.
The philosophical traditions of India spent millennia on exactly this problem — not as a curiosity but as the central question. The Upanishadic tradition, and especially Advaita Vedanta as systematized by Adi Shankaracharya, approached consciousness not as a property that physical systems sometimes have but as the primordial ground from which all apparent objectivity is derived. Consciousness — cit, awareness, chit-shakti — is not the thing to be explained by pointing to other things. It is the explanans without which nothing else can be known at all. The Vedantic argument is subtle: any proposed explanation of consciousness must itself appear within consciousness, must be cognized by a conscious being to count as an explanation at all. So consciousness is prior, in the order of explanation, to every candidate that would reduce it.
This is not mysticism in the dismissive sense. It is a rigorous epistemological position: explanation requires a knower, a knower requires consciousness, therefore consciousness cannot coherently appear at the bottom of a reductive explanation. The more you push the explanation down — neurons, atoms, information — the more you notice that the entire explanatory project is itself occurring inside the field of awareness you’re trying to explain.
The Samkhya tradition takes a different approach: distinguishing purusha (pure consciousness, the witness) from prakriti (nature, including mind and body), consciousness is irreducibly distinct from the material process, and their interaction is what generates experience. The diagnostic power here is similar to property dualism: consciousness is not derived from the material but runs alongside it.
Neither of these frameworks translates cleanly into analytic categories. But they gesture at something the analytic debate needs: the recognition that the hard problem may be hard not because we lack data about the brain, but because the conceptual apparatus of reductive science was built for third-person phenomena and cannot, in principle, be extended to first-person ones by accumulating more data of the same kind. You can’t get in here from out there by adding more out there.
The hard problem, then, is not a gap in our explanatory net that patience and neuroscience will eventually close. It is a signal that our explanatory net was designed for a different kind of fish. The question isn’t when we’ll solve it; it’s whether we’re willing to let the question reshape what we mean by “solution.”
The dissolvers want to protect their framework by questioning the datum. The datum — that there is something it is like to be conscious, right now, as you read this — is the one thing that cannot be revised away, because the revision itself would be an experience. Taking the hard problem seriously means accepting that any adequate theory of consciousness will have to look different from our theories of digestion or computation or orbital mechanics. What it will look like, no one knows. But that is an invitation, not a dead end.
The strangest thing about consciousness is that it’s the only problem where the investigator and the investigated are the same. Every other domain of science examines objects from outside. Consciousness examines itself. That reflexivity isn’t an embarrassment to be overcome. It might be the most important clue.