Start with the strength of the opposing case, because it is considerable and any honest version of this question has to survive it.
Drink four glasses of wine and your consciousness changes in ways you can predict from the chemistry. Take a small quantity of propofol into a vein and it stops, completely, for as long as the drug is present, and then resumes. Damage the fusiform gyrus and a person loses the ability to recognize faces while keeping everything else. Damage Broca’s area and language production goes while comprehension stays. Wilder Penfield, operating on conscious epilepsy patients in the 1950s, touched an electrode to the temporal lobe and people reported hearing music that was not playing. Sever the corpus callosum and you get a person whose left hand will button a shirt while the right unbuttons it, and who will confabulate a reason for doing so.
Nothing in that record is ambiguous. Whatever consciousness is, the brain is intimately, specifically, and causally involved at a resolution of millimeters.
So the question in the headline is not asking whether the brain matters. It is asking something narrower and stranger, which is whether the brain being necessary means consciousness is located there, in the way a fire is located in a fireplace.
The state of the science, which is worse than you think.
In April 2025 the Cogitate Consortium published in Nature the most rigorous test the field has managed. Twenty-plus laboratories, 256 participants, functional MRI and magnetoencephalography and intracranial EEG recording directly from electrodes inside people’s heads. The design was adversarial: proponents of the two leading theories, integrated information theory and global neuronal workspace theory, agreed in advance on which findings would falsify their own positions, and preregistered them.
Neither theory came through with its core predictions intact. IIT’s requirement of sustained posterior synchronization during conscious experience failed. Global workspace predictions about prefrontal involvement and broadcasting also failed. Christof Koch, one of IIT’s architects and among the paper’s authors, described the challenge afterward as finding a way to explain the null result.
Around the same period, more than a hundred researchers signed a letter arguing that IIT is unfalsifiable in practice and should be classed as pseudoscience. The dispute is still running.
So the honest summary is that after thirty years of serious work, the two dominant scientific theories of consciousness are both in trouble, and the field’s most expensive attempt to adjudicate between them adjudicated nothing. This is the ground on which anyone asking where consciousness is has to stand.
Answer one: nowhere, because it is not that kind of thing.
The philosopher Alva Noë argues that we have been asking about consciousness the way we would ask about a substance, when it behaves more like an activity. His example is dancing. A dancer requires legs and a nervous system and a floor, and no amount of study of the legs will locate the dancing inside them, because dancing is something a whole person does in an environment rather than something a body part secretes.
On this reading, the brain is necessary to consciousness the way legs are necessary to walking. Ask where consciousness is and you have made the same mistake as asking where the walking is.
This has real explanatory appeal and one substantial weakness, which is that it does not obviously touch the hard problem. Dancing has no felt quality that needs accounting for. Experience does. Reclassifying consciousness as an activity leaves you needing to explain why this activity, alone among the activities in the universe, is like something from the inside.
Answer two: wider than the skull.
In 1998 Andy Clark and David Chalmers proposed that the mind extends into the tools it uses. Their example was a man with memory impairment who keeps a notebook and consults it exactly as another person consults recollection. If the notebook plays the functional role memory plays, they argued, the notebook is part of his mind, and the boundary at the skin is a convention rather than a discovery.
Related work in embodied and enactive cognition, going back to Varela, Thompson, and Rosch, treats mind as a loop running through body and environment rather than a computation happening in a box.
This is the most defensible answer on the list, and it earns real ground on cognition. On consciousness specifically it earns less, because most versions extend the machinery of thought without extending the felt quality of experience, and the felt quality is what the question was about.
Answer three: everywhere.
Panpsychism holds that experience is a fundamental feature of matter, present in some minimal form throughout, and that brains do not create consciousness so much as organize it into the complex kind we have. It is a serious position with serious contemporary defenders, Philip Goff and Galen Strawson among them, and it appeals to physicists and philosophers for the same reason: it dissolves the problem of getting experience out of non-experience by declining to start with non-experience.
Its problem has a name, the combination problem, and it has resisted solution for a century. If every particle has a speck of experience, how do trillions of specks combine into one unified point of view that sees a single room? Nobody has an account. Adding up small experiences does not obviously produce a large one any more than adding up small squares produces a circle.
Answer four: elsewhere, with the brain as receiver.
This is the answer the question is usually reaching for, and it has a distinguished pedigree. William James proposed in 1898 that the brain’s function might be transmissive rather than productive, filtering or focusing a consciousness that exists independently, the way a prism does not manufacture light. Bergson argued something similar. Aldous Huxley picked it up. It is the view underneath most contemporary claims that consciousness survives death.
The evidence usually offered for it deserves careful handling, because it does not say what it is said to say.
The hydrocephalus cases are the strongest card and they are weaker than advertised. John Lorber, a Sheffield pediatrician, reported in the late 1970s on a large series of patients whose ventricles had expanded enormously, including a mathematics student with an IQ of 126 and a cortical mantle Lorber described as around a millimeter thick. In 2007 French neurologists published a case in the Lancet of a civil servant with massive ventricular enlargement, married, employed, two children, tested at an IQ of 75.
Both are astonishing. Neither shows a person without a brain. They show brain tissue compressed and displaced rather than absent, in people whose condition developed slowly, usually from infancy, which is precisely the circumstance in which neural reorganization is most capable. The French patient was diagnosed at six months and shunted. Lorber’s series, it should be said, was never published in a peer-reviewed journal.
What the cases actually demonstrate is that the brain is far more plastic and far less dependent on gross anatomy than mid-century neurology assumed. That is a large finding. It is a finding about brains.
The deeper problem with the receiver hypothesis is not evidential. It is that the theory predicts nothing the productive theory does not also predict. Damage the receiver, the signal degrades. Damage the producer, the output degrades. Both accounts explain every observation equally well, which means no observation can distinguish them, which means the choice between them is being made on grounds other than evidence. That is worth admitting rather than obscuring.
The word doing the damage.
Notice that all four answers are attempts to supply a location, and that the question demanded one before any of them started.
Location is a relation between objects in space. Asking where a thing is presupposes it is the kind of thing that occupies a position, and there are perfectly respectable items that are not. The number seven is nowhere. Wednesday is nowhere. The key of D minor is nowhere. None of these is thereby mysterious or supernatural, and nobody demands a laboratory that will find them.
It is possible that consciousness belongs to that category. If so, the question in the headline is malformed in the way that asking what lies north of the North Pole is malformed, and no amount of searching resolves it because there is no direction left to look.
That move deserves suspicion, though, because “you have committed a category error” is also the thing people say when they have no answer and would prefer to sound as though the question were beneath them. The felt quality of experience is not obviously like the number seven. Seven does not hurt. A person in pain is not confused about whether there is something happening, and if you tell them the location of their pain is a malformed question they will reasonably conclude that you have changed the subject.
Where this leaves us.
A reader wrote to me after the last of these essays and made a point I have been carrying since. He suggested that the identity we imagine losing may be less solid than we assume, and that when that becomes experiential rather than theoretical, the fear attached to it changes character. He was also careful to say that whatever can be reported from inside a living body cannot license conclusions about what follows the death of one, and that carefulness is why the point landed.
The same discipline applies here. Everything we know about consciousness we know from inside one, using instruments built by others, correlated against brains that were all, without exception, working at the time.
We have no theory. We have two leading candidates that just failed the most demanding test anyone has designed, an active dispute over whether one of them is science at all, and four philosophical alternatives that each solve part of the problem by relocating the rest of it somewhere the solver does not have to look.
What I am fairly confident of is that anyone who tells you where consciousness is, in a tone of settled knowledge, is telling you about their temperament rather than about the world. The materialist and the mystic are both working from the same data, which is almost none, and the confidence is coming from somewhere other than the evidence in both directions.
That is an unsatisfying place to end an essay. It is also, as far as I can tell, the accurate one, and the accurate one is worth more than the satisfying one, particularly here, where the temptation to settle it is so strong and the cost of a false settlement is a person who has stopped asking.